American Journal of Epidemiology
Copyright © 1997 by The Johns Hopkins University School of Hygiene and Public Health
All rights reserved
Vol. 146, No. 4
Printed in U.S.A
Physical Activity and Cardiovascular Disease Risk Profile in Women
Margreet A. Pols,1 Petra H. M. Peeters,1 Jos W. R. Twisk,2 Han C. G. Kemper,2 and Diederick E. Grobbee1
In a population of 4,576 Dutch women aged 49-70 years who participated in the European Prospective
Investigation into Cancer and Nutrition (EPIC) in 1993-1995, the relation between physical activity and the
presence of cardiovascular disease risk indicators was assessed cross-sectionally. Physical activity was
determined from a self-administered questionnaire, while blood pressure, heart rate, body mass index,
waist/hip ratio, and waist circumference were measured at the study center. Mean risk indicator levels were
calculated for different activity categories. Blood pressure was most clearly associated with time spent in
sports (mean systolic blood pressure, adjusted for age, level of education, and smoking, 128.9 mmHg in the
highest sports tertile, and 132.1 mmHg in the lowest sports tertile; mean diastolic blood pressure, 77.8 mmHg
and 79.0 mmHg, respectively). Body mass index, waist/hip ratio, and waist circumference showed an inverse
relation with cycling, gardening, do-it-yourself-activities, and sports. In this population, leisure-time activity
was inversely related to cardiovascular disease risk indicators, but work activity and housework were not. The
authors conclude that if investigators wish to measure physical activity in women over age 50 years with the
aim to identify high- and low-risk groups for cardiovascular disease, they should consider not only housework
activity, but also leisure-time activities such as cycling, sports, and do-it-yourself activities. Am J Epidemiol
1997;146:322-8.
aged; blood pressure; cardiovascular diseases; exercise; heart rate; obesity; risk factors; women
Numerous studies have been conducted to investigate the relation between physical activity and cardiovascular disease. Powell et al. (1) extensively reviewed most studies performed up to 1985, and they
concluded that, especially in the better designed studies, physical activity showed an inverse relation with
the incidence of coronary heart disease. In a metaanalysis conducted by Berlin et al. (2) of the cohort
studies described by Powell et al. and of more recent
publications, the same pattern was found, with the
better designed studies yielding higher relative risks
for less active subjects compared with more active
subjects.
Physical activity may lower the risk of coronary
heart disease through mediation of cardiovascular disease risk factors. Indeed, in the literature, physical
activity has been associated with favorable levels of
cardiovascular disease risk factors (3-5). Most of
these studies, however, were performed in men (4, 5).
For women, the relations could be different, because
sex-specific factors such as hormonal changes influence the risk of cardiovascular disease for women
importantly (6). The aim of this study was to investigate whether an association between physical activity
and cardiovascular disease risk factors is also present
in a population of women aged 49-70 years, and
whether different types of activity are differently related to systolic and diastolic blood pressure (7), heart
rate, body mass index (8), waist/hip ratio (9), and
waist circumference (10).
MATERIALS AND METHODS
The European Prospective Investigation into Cancer
and Nutrition (EPIC) is a large international cohort
study investigating the relation of nutrition and other
life-style factors with the development of chronic diseases (11). Apart from cancer, cardiovascular disease
is one of the endpoints (EPIC-Heart Study). This study
is being carried out in nine European countries (Denmark, France, Germany, Great Britain, Greece, Italy,
the Netherlands, Spain, and Sweden). By the end of
1998, data from 400,000 subjects will have been collected. Half of the Dutch cohort consists of women
aged 49-70 years and is known as the EPIC-Prospect
Study. A dietary and a general questionnaire are completed, blood samples are taken, and anthropometric
measurements are carried out. In the coming decades,
Received for publication August 29, 1996, and accepted for
publication April 7, 1997.
Abbreviation: EPIC, European Prospective Investigation into
Cancer and Nutrition.
1
Department of Epidemiology, Faculty of Medicine, Utrecht University, Utrecht, the Netherlands.
2
EMGO-lnstitute, Faculty of Medicine, Vrije Universiteit, Amsterdam, the Netherlands.
322
Physical Activity and Cardiovascular Disease Risk in Women
endpoints will be collected and risk estimates can be
made.
In the Netherlands, all women are invited for breast
cancer screening on a 2-yearly basis, starting in the
calendar year they turn age 50 years, and repeated
until they reach age 70 years. From November 1993 to
April 1995, all women who were invited for the national breast cancer screening program in Utrecht and
surrounding areas received an invitation to participate
in the EPIC-Prospect Study. A total of 5,948 women
aged 49-70 years (±35 percent of the invited women)
agreed to take part.
Participating women received a dietary questionnaire and a general questionnaire. The general questionnaire comprised sections on socioeconomic factors, reproductive history, menstruation and hormonal
therapy (contraceptive or other), past and current
physical activity, past and current use of tobacco and
past use of alcohol, medical history, and family history. Subjects completed the questionnaires at home,
and brought them along when they visited the study
center. The questionnaires were checked by trained
interviewers for completeness and inconsistencies.
Women who reported that they suffered a myocardial infarction or a stroke were excluded from the
analyses (n - 227). Data on current antihypertensive
medication were not available. However, the subjects
were asked if they had ever been treated for hypertension and, if so, in what year the treatment started.
Because subjects who were ever treated and were now
normotensive might be so with or without treatment,
only subjects who reported no treatment for hypertension were included in the analyses. Table 1 shows the
general characteristics of the 4,576 remaining subjects.
Physical activity
The EPIC physical activity questions concerned
work and leisure-time activities. Subjects were asked
to classify their paid or voluntary job into sedentary
work, standing work, manual work, or heavy manual
work. In the case of retirees, the question referred to
the occupation longest held by the subject. Therefore,
for retired women, reported work activity represented
past occupational activity. For leisure-time activities,
women estimated the amount of time spent in walking,
cycling, gardening, do-it-yourself activities, sports,
and housework in a normal week during the summer
and winter of the past year. In the analyses, the mean
number of hours of summer and winter was used.
Time spent in total leisure-time activity was computed
as the sum of time spent in walking, cycling, gardening, do-it-yourself activities, and sports. Furthermore,
subjects were asked whether they perspired while carrying out the above-mentioned activities in a normal
Am J Epidemiol
Vol. 146, No. 4, 1997
323
TABLE 1. Characteristics of 4,576 Dutch women aged 49-70
years who participated in the European Prospective
Investigation into Cancer and Nutrition, 1993-1995
Mean (SD*)
or%
Variable
Age (years)
Height (m)
Body weight (kg)
BMI* (kg/m2)
Waist circumference (m)
Hip circumference (m)
Waist/hip ratio
Mean systolic blood pressure (mmHg)
Mean diastolic blood pressure (mmHg)
Mean heart rate (beats/minute)
Smoking (%)
Yes
No
Level of education (%)
Primary
Lower vocational
General secondary
Senior secondary vocational or senior
general secondary >3 years
Higher vocational or university >3 years
:
57.1 (5.8)
1.64(0.06)
69.3(10.7)
25.7 (3.8)
0.83(0.1)
1.05(0.08)
0.79 (0.06)
131 (19)
78 (10)
74(11)
23.4
76.6
23.1
25.8
24.5
13.8
12.8
SD, standard deviation; BMI, body mass index.
week (yes/no). Reproducibility and validity of the
physical activity questions were estimated in a pilot
study preceding the start of the EPIC-Prospect Study,
and have been published separately (12). In brief,
reproducibility was assessed with 5 and 11 months
test-retest intervals, and Spearman correlation coefficients (13) ranged from 0.47 to 0.89 in men and from
0.49 to 0.81 in women for specific activity categories.
Relative validity, using an activity diary as a reference
method, ranged from 0.32 to 0.81 in men and from
0.28 to 0.72 in women (12).
Blood pressure and heart rate
Blood pressure and heart rate were measured in sitting position with a BOSO Oscillomat (Bosch+Sohn
GmbH u. Co., Jungingen, Germany). Two measurements were taken with a 5-10 minutes interval, and
the mean was used in the analyses.
Anthropometric measurements
Anthropometric measurements (height, weight,
waist circumference, and hip circumference) were
taken with the subjects wearing indoor clothes and no
shoes. Body mass index was calculated as weight
(kg)/height (m)2. Waist/hip ratio was computed as
waist circumference/hip circumference. Both being
overweight and having a predominantly abdominal fat
distribution add to the risk of developing cardiovascular disease. A body mass index of &25 kg/m2 is
324
Pols et al.
commonly regarded as overweight (14) and has been
found to be associated with an increased risk of coronary heart disease in women (8). A high waist/hip
ratio (>0.80) corresponds with a pattern of abdominal
fat distribution that increases the risk of cardiovascular
disease (9). A recent study (10) suggests that a waist
circumference of ^ 8 0 cm in women could be an even
stronger cardiovascular disease risk factor than the
waist/hip ratio.
TABLE 2. Physical activity variables (% of population for
categorical variables; means (SD*) for continuous variables)
for 4,576 Dutch women aged 49-70 years who participated in
the European Prospective Investigation into Cancer and
Nutrition, 1993-1995
%or
mean (SD)
Variable
Work activity (%)
Sedentary
Standing
Manual
Heavy manual
32.5
32.7
19.5
15.4
Perspiring (%)
Yes
No
67.0
33.0
Type of activity (hours/week), mean (SD)
Housework
Walking
Cycling
Gardening
Do-it-yourself
Sports
Total leisure-time activity
19.3(11.3)
8.0(9.1)
3.9 (4.7)
1.6(2.4)
1.0(2.9)
1.3(2.1)
15.8(12.5)
Data analysis
Categorical activity variables were treated as such
(work activity: categories 1-4, and perspiring: yes/
no). Continuous variables were divided into tertiles. In
case more than 33 percent of subjects reported that
they did not engage in a specific activity at all, they
were all classified in the lowest tertile for that activity
(e.g., gardening, 35.4 percent; sports, 47.1 percent;
and do-it-yourself-activities, 72.8 percent, resulting in
two groups).
A multiple linear regression model (15) was used to
calculate mean levels of the risk indicators for categories or tertiles of the activity variables, adjusted for
age, level of education, and current smoking status.
Alcohol consumption was not included as a covariate,
because data on alcohol consumption was only available for about 60 percent of the population and alcohol
use in this population was very low (74 percent of the
women consumed 0-1 alcoholic drinks per day).
To assess whether less active women more frequently have multiple risk indicators, the cardiovascular disease risk indicators were dichotomized using
boundaries derived from the literature. The mean number of positive risk indicators (above the cut-off point)
was calculated for different levels of activity. For
systolic and diastolic blood pressure, we used the
criteria proposed in the fifth report of the Joint National Committee on Detection, Evaluation, and Treatment of High Blood Pressure (JNC V) (7), i.e., 140
mmHg for systolic blood pressure and 90 mmHg for
diastolic blood pressure. A high resting heart rate is
considered as a sign of a poor physical condition,
although there is no clearly defined threshold level.
Therefore an arbitrary value of 75 beats/minute was
chosen. The other cut-off points were: body mass
index, 25 kg/m2; waist/hip ratio, 0.80; and waist circumference, 80 cm.
RESULTS
Measures of activity are summarized in table 2. In
the study population, 50.3 percent still had a paid or
voluntary job and 44.4 percent used to have a paid or
voluntary job but did not work any more. Of the
* SD, standard deviation.
subjects who ever had a job, 32.5 percent had a sedentary job and 32.7 percent had a standing job.
Housework was the activity most frequently reported in the EPIC hours/week question. Of the leisure-time activities, walking was performed the most
frequently, followed by cycling. Perspiring while doing one of these activities was reported by 67 percent
of the subjects.
Mean levels of the cardiovascular disease risk indicators are presented in table 1. Twenty-six percent of
the population had systolic hypertension ( a 140
mmHg) and 12 percent had diastolic hypertension
(>90 mmHg). A heart rate of S75 beats/minute was
measured in 43 percent of the women. Obesity (body
mass index ^25 kg/m2) was found in 53 percent, and
a waist/hip ratio of ^0.8 in 38 percent. Finally, a waist
circumference of ^ 8 0 cm was measured in 59 percent
of the population.
Relations between inactivity and cardiovascular disease risk are shown in table 3. In this population, the
cardiovascular disease risk indicators seemed to be
positively related to work activity, and inversely related to leisure-time activities.
Both systolic and diastolic blood pressure were
highest in the women who performed heavy manual
work. There was a similar relation between systolic
blood pressure and housework: women in the lowest
tertile of the time spent engaged in housework had
lower blood pressures. Blood pressure was inversely
related to the time spent in cycling or sports, and in all
leisure-time activities together.
Am J Epidemiol
Vol. 146, No. 4, 1997
o_
Am J Epidemi
CD
CO
p
z
a>
_ i
Vol
78.4 ± 0.25
78.3 ± 0.27
78.4 ± 0.25
78.7 ± 0.25
78.5 ± 0.26
78.0 ± 0.26
131.1 ±0.45
129.9 ± 0.50
130.6 ±0.45
131.1 ±0.46
131.2 ±0.47
129.4 ± 0.46*
± 0.30
± 0.28
± 0.37
± 0.44
75.3 ± 0.29
74.1 ±0.28
73.2 ± 0.27*
74.6 ± 0.29
74.0 ± 0.20
78.9 ± 0.27
78.1 ±0.24
78.2 ± 0.25*
78.0 ± 0.25
78.6 ±0.18
131.4 ±0.49
130.3 ± 0.44
130.0 ± 0.46*
129.5 ± 0.46
131.1 ±0.33*
25.4 ± 0.09
25.9 ± 0.07*
25.9 ±0.10
25.6 ± 0.09
25.7 ± 0.09
25.9 ± 0.09
25.7 ±0.12
25.6 ± 0.09*
75.1 ±0.24
74.4 ± 0.34
72.7 ± 0.28*
79.0 ± 0.22
78.0 ± 0.32
77.8 ± 0.25*
26.0 ±0.10
25.7 ± 0.09
25.5 ± 0.09*
26.0 ±0.10
25.7 ± 0.09
25.6 ± 0.09*
25.8 ± 0.09
25.8 ±0.10
25.6 ± 0.09
25.8 ± 0.07
25.6 ±0.10
132.1 ±0.41
129.9 ± 0.59
128.9 ± 0.44*
±0.10
±0.10
±0.12
±0.15*
25.7 ± 0.09
25.8 ± 0.09
25.7 ±0.10
25.5
25.8
25.5
26.5
Mean ± SE
(kg/m2)
Body mass index
74.4 ±0.19
73.7 ± 0.30
78.6 ±0.18
77.9 ± 0.27
74.5 ± 0.28
73.9 ± 0.27
74.2 ± 0.30
75.2 ± 0.28
74.2 ± 0.28
73.1 ±0.28*
74.3 ± 0.27
74.4 ± 0.30
73.9 ± 0.27
74.0 ± 0.27
74.3 ± 0.29
74.3 ± 0.29
74.2
74.1
73.6
74.5
Mean ± SE
Heart rate
(beats/minute)
130.9 ± 0.32
129.9 ± 0.48
77.9 ± 0.25
78.7 ± 0.25
78.6 ± 0.26*
78.0 ± 0.25
79.0 ± 0.26
78.2 ± 0.26
129.2 ± 0.44
131.4 ±0.48
131.1 ±0.47*
129.7 ± 0.47
131.2 ±0.45
130.9 ± 0.47
78.3 ± 0.27
78.2 ± 0.26
77.9 ± 0.34
79.2 ± 0.40
Mean ± SE
Mean ± SE
130.4 ±0.48
129.8 ± 0.47
130.3 ± 0.63
132.9 ±0.76*
Dlastolic
±0.0015
±0.0014
±0.0019
± 0.0022*
0.783 ±0.0014
0.786 ±0.0010
0.787 ±0.0014
0.785 ±0.0014
0.784 ±0.0014
0.788 ±0.0012
0.785 ±0.0017
0.783 ±0.0014*
0.787 ±0.0010
0.782 ±0.0015*
0.788 ±0.0014
0.783 ±0.0013
0.784 ±0.0014*
0.790 ± 0.0014
0.782 ±0.0014
0.784 ±0.0014*
0.785 ±0.0013
0.785 ±0.0014
0.786 ±0.0014
0.786 ±0.0014
0.786 ±0.0014
0.783 ±0.0014
0.782
0.787
0.784
0.793
Mean ± SE
Waist/hip
ratio
±0.24
± 0.24
± 0.31
± 0.38*
81.9 ±0.23
82.9 ±0.17*
83.2 ± 0.25
82.3 ± 0.23
82.2 ± 0.22*
83.1 ±0.21
82.3 ± 0.28
81.9 ±0.22*
82.7 ±0.16
82.1 ±0.25*
83.2 ± 0.25
82.2 ± 0.22
82.2 ± 0.23*
83.5 ± 0.24
82.0 ± 0.23
82.0 ± 0.23*
82.8 ± 0.22
82.7 ± 0.25
82.2 ± 0.23
82.6 ± 0.23
82.7 ± 0.24
82.4 ± 0.23
81.9
82.6
82.0
84.3
Mean ± SE
Waist
(cm)
* Significant trend (three or four categories) or significant difference (two categories), after adjustment for age, smokinc), and level of education (p < 0.05).
t Means are adjusted for age, smoking, and level of education.
Do-it-yourself (hours/week)
0
£!4
Sports (hours/week)
0
14-1
£114
Total leisure-time activity (hours/week)
£814
9-1614
£17
Perspiring
No
Yes
• & .
14-1 Mt
0
£414
Gardening (hours/week)
2-4
*7VS>
Cycling (hours/week)
Work activity
Sedentary
Standing
Manual
Heavy manual
Housework (hours/week)
21314
14-2116
2:22
Walking (hours/week)
£3
314-7
Activity and level
Systolic
Blood pressure (mmHg)
± 0.043
± 0.042
± 0.056
± 0.063*
2.21 ±0.041
2.36 ± 0.029*
2.44 ± 0.042
2.27 ± 0.040
2.22 ±0.041*
2.42 ± 0.036
2.30 ± 0.051
2.16 ±0.040*
2.35 ± 0.028
2.22 ± 0.045*
2.33 ± 0.041
2.33 ± 0.040
2.26 ± 0.042
2.44 ± 0.040
2.27 ± 0.041
2.20 ± 0.042*
2.34 ± 0.039
2.33 ± 0.043
2.26 ± 0.041
2.25 ± 0.040
2.35 ± 0.041
2.32 ± 0.042
2.24
2.31
2.23
2.55
Mean ± SE
No. of
risk factors
TABLE 3. Mean levelst * standard errors (SE) of the cardiovascular disease risk factors for categories or tertiles of physical activity among 4,576 Dutch women aged
49-70 years who participated in the European Prospective Investigation into Cancer and Nutrition, 1993-1995
a
>
CO
o
X
c/>"
CD
2
$
o
c
CO
QJ
§
o
3
vity c
ten
325
326
Pols et al.
Heart rate showed no association with work or
housework. On the other hand, cycling, sports, and
total leisure-time activities were all inversely associated with heart rate.
The relations between work activity and body mass
index, waist/hip ratio, and waist circumference all
showed a similar pattern, with the women who reported that they did heavy manual work having the
highest levels of the risk indicators. No relation was
found with housework. The three obesity parameters
were in general inversely associated with the leisuretime activities, especially with cycling, gardening, and
sports.
The general pattern present for individual risk indicators was also found if we considered clustering of
risk indicators (occurrence of multiple risk factors),
i.e., we observed a positive relation with work and
housework, and an inverse one with the leisure-time
activities. Again, the strongest associations were seen
for cycling, do-it-yourself activities, sports, and time
spent in all leisure-time activities combined.
Relations between reported perspiring and risk indicators were very weak, and, if seen at all, were not
in the expected direction. However, after we separated
the "perspirers" (n = 1,503) from the "non-perspirers"
(n — 3,044), we found that perspirers spent significantly more time in leisure-time activities (16.9 vs.
13.7 hours/week, p < 0.001) than did the non-perspirers. Among the non-perspirers, a relation between total
leisure-time activity and the risk indicators was observed only for heart rate, whereas among the perspirers, a relation was found for every risk indicator except body mass index.
DISCUSSION
In our study among 4,576 women aged 49-70 years,
the relations between the activity parameters and cardiovascular disease risk indicators indicate that in general a higher activity level of work or housework is
associated with a higher level of the risk indicators,
which was not expected. On the other hand, the inverse relation between leisure-time activities and risk
profile was in agreement with the hypothesis that
physical activity lowers cardiovascular disease risk.
The differences in risk indicator levels between the
activity groups were rather small and their clinical
relevance may be questioned. The fact that several
differences still reached statistical significance reflects
the large number of subjects included in the study.
Because the positive association between work activity and the risk indicators and the inverse relation
between leisure-time activity and the risk indicators
was fairly consistent for most activities and most risk
indicators, it is not likely that these results are a chance
finding. In a population of healthy women in a rather
narrow age range, in which the variation in physical
activity probably is quite small, one would not expect
to find huge contrasts in risk indicator levels. The
effects of activity pattern on cardiovascular disease
found in this study can therefore be considered real
and relevant, although they are admittedly of small
magnitude.
Participants in the study were women who agreed to
complete two rather extensive questionnaires, to spend
an hour extra at the breast cancer screening unit to
have the questionnaires checked and the measurements taken, and to provide a blood sample. Probably
these women are interested in health-related topics and
recognize the importance of scientific investigations.
Furthermore, women who have a busy job might be
less inclined to participate. Therefore, the women in
the study could have been a selected population. Data
on non-responders are scarce. Van Leer et al. (16)
studied blood pressure in a monitoring project on
cardiovascular disease risk factors in the Netherlands
between 1987 and 1991, and they reported a mean
systolic blood pressure of 125 mmHg and a mean
diastolic blood pressure of 79 mmHg in 5,278 women
aged 50-59 years. Body mass index, waist/hip ratio,
and waist circumference did not differ from those
found by Den Tonkelaar et al. (17) in women who
participated between 1984 and 1986 in the same breast
cancer screening project from which the women in our
study were recruited. From these data, it seems that
selection does not play an important role in our study.
The analyses presented in this paper are
cross-sectional. This implies that no conclusion can be
drawn regarding cause and effect. It might be possible
that women who are more active in their leisure time
as a result have lower blood pressure and are leaner,
but on the other hand such women could be more
active because they feel more fit.
We recognize that one study limitation is that we did
not have data on current use of antihypertensive medication, but only on "ever" use of such medication.
The exclusion of subjects who were ever treated for
hypertension resulted in a lower mean blood pressure
for the whole population (systolic blood pressure,
130.6 vs. 133.6 mmHg; diastolic blood pressure, 78.4
vs. 79.7 mmHg). The relations with the activity variables, however, were similar to those in the whole
population. Furthermore, no data on angina pectoris
were available and the possible inclusion of women
with angina may have confounded the results to some
extent, because both level of physical activity and
level of the risk indicators can be affected by the
presence of coronary heart disease.
Am J Epidemiol
Vol. 146, No. 4, 1997
Physical Activity and Cardiovascular Disease Risk in Women
Few data have been reported on the relation between
occupational activity and risk indicators in female
populations. Albanes (18) compared body mass index
for low, moderate, and high levels of non-recreational
activity, and found inverse relations for both men and
women in the second National Health and Nutrition
Examination Survey (NHANES II). Stender et al. (5)
reported a higher body mass index in men with higher
levels of job activity compared with that in inactive
men. Among 628 Shanghai women aged 35-64 years,
Hong et al. (19) found inverse relations between
weekly frequency of periods of activity that caused
shortness of breath, increase in pulse rate, and perspiration on the one hand, and systolic and diastolic blood
pressure, body mass index, and heart rate on the other
hand. Hong et al. did not, however, discriminate between occupational and leisure-time activity.
Few studies have been published that have compared different types of leisure-time activity with levels of cardiovascular disease risk indicators in female
populations. Among 7,722 German females aged
25-69 years (20), leisure-time physical activity was
inversely associated with systolic and diastolic blood
pressure, and with body mass index. The differences in
risk indicator levels between activity groups were of
the same order of magnitude as those observed in our
study. Bijnen et al. (4) studied the relation between
various types of activity and cardiovascular disease
risk indicators in a population of 1,402 men aged
69-90 years from Finland, Italy, and the Netherlands,
and they reported no relation between physical activity
and body mass index or blood pressure, but generally
favorable associations between total weekly physical
activity on the one hand and resting heart rate and high
density lipoprotein (HDL) cholesterol on the other
hand. These results point in the same direction as our
results regarding leisure-time activity.
Work activity and time spent on housework showed
a similar pattern in the relation with cardiovascular
disease risk profile: both were either positively or not
at all related with the risk indicators. The positive
relation between work and housework activity and risk
indicators could not be explained by the theory that
women who were involved in heavy jobs were less
active during leisure time. On the contrary, a positive
linear association was seen in total leisure-time activity for increasing levels of occupational activity.
The women who performed heavy work or who
spent a lot of time on housework might be a selected
group. For example, the results could reflect a remaining effect of socioeconomic status, although the phenomenon remained after adjustment for level of education. In this population, the amount of time spent on
housework was large, but it obviously is not suitable to
Am J Epidemiol
Vol. 146, No. 4, 1997
327
identify women with a lower risk for cardiovascular
disease.
The risk indicators were inversely related to most
leisure-time activities, as has been reported previously
for men (4). For blood pressure, the relation was most
prominent for sports and total leisure-time activity.
The exclusion of subjects who were ever treated for
hypertension might have weakened the associations
because more women with a relatively high blood
pressure were excluded. Time spent in cycling, sports,
and all leisure-time activities did show the strongest
influence on heart rate, probably because these activities are of higher intensity than, for example, walking
or gardening. The anthropometric parameters (body
mass index, waist/hip ratio, and waist circumference)
were inversely related to all leisure-time activities
except walking.
In this population of women aged 49-70 years, a
higher level of participation in leisure-time activities
was generally associated with a more favorable level
of cardiovascular disease risk indicators, especially
indicators of being overweight. However, by contrast,
work activity and housework were positively or not at
all associated with the level of most risk indicators.
We therefore recommend that investigators who study
physical activity in women aged over 50 years in order
to identify high- and low-risk groups for cardiovascular disease should consider all aspects of activity—
work and housework activity, as well as such leisuretime activities as cycling, sports, and do-it-yourself
activities.
ACKNOWLEDGMENTS
This study was supported by the Commission of European Communities contract no. Soc 95 200500 05F02.
REFERENCES
1. Powell KE, Thompson PD, Caspersen CJ, et al. Physical
activity and the incidence of coronary heart disease. Annu Rev
Public Health 1987;8:253-87.
2. Berlin JA, Colditz GA. A meta-analysis of physical activity in
the prevention of coronary heart disease. Am J Epidemiol
1990; 132:612-28.
3. Dannenberg AL, Keller JB, Wilson PWF, et al. Leisure time
physical activity in the Framingham Offspring Study: description, seasonal variation, and risk factor correlates. Am J Epidemiol 1989; 129:76-88.
4. Bijnen FCH, Feskens EJM, Caspersen CJ, et al. Physical
activity and cardiovascular risk factors among elderly men in
Finland, Italy, and The Netherlands. Am J Epidemiol 1996;
143:553-61.
5. Stender M, Hense HW, Doring A, et al. Physical activity at
work and cardiovascular disease risk: results from the
MONICA Augsburg study. Int J Epidemiol 1993;22:644-50.
6. Schouw YT van der, Graaf Y van der, Steyerberg EW, et al.
328
7.
8.
9.
10.
11.
12.
13.
Pols et al.
Age at menopause as a risk factor for cardiovascular mortality.
Lancet 1996;347:714-18.
The Joint National Committee on Detection, Evaluation, and
Treatment of High Blood Pressure. The fifth report of the Joint
National Committee on Detection, Evaluation, and Treatment
of High Blood Pressure (JNC V). Arch Intern Med 1993;153:
154-83.
Manson JE, Colditz GA, Stampfer MJ, et al. A prospective
study of obesity and risk of coronary heart disease in women.
N Engl J Med 1990;322:882-9.
Lapidus L, Bengtsson C, Larsson B, et al. 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. BMJ 1984,289:1257-61.
Han TS, Leer EM van, Seidell JC, et al. Waist circumference
action levels in the identification of cardiovascular risk
factors: prevalence study in a random sample. BMJ 1995;311:
1401-5.
Riboli E. Nutrition and cancer: background and rationale of
the European Prospective Investigation into Cancer and Nutrition (EPIC). Ann Oncol 1993;3:783-91.
Pols MA, Peeters PHM, Ock6 MC, et al. Estimation of reproducibility and relative validity of the questions included in the
EPIC physical activity questionnaire. Int J Epidemiol 1997;
26(Suppl 1):S181-189.
Kendall M, Gibbons JD. Rank correlation methods. 5th ed.
London: Edward Arnold, 1991: chapter 1.
14. World
Health
Organization.
Measuring
obesity—
classification and description of anthropometric data.
Copenhagen: Regional Office for Europe, 1989.
15. Kleinbaum DG, Kupper LL, Muller KE. Applied regression
analysis and other multivariable methods. 2nd ed. Belmont,
CA: Wadsworth Publishing Co, 1988: chapter 8.
16. Leer EM van, Seidell JC, Kromhout D. Levels and trends in
blood pressure and prevalence, and treatment of hypertension
in the Netherlands, 1987-1991. Am J Prev Med 1994;10:
194-9.
17. Tonkelaar I den, Seidell JC, Noord PA van, et al. Fat distribution in relation to age, degree of obesity, smoking habits,
parity and estrogen use: a cross-sectional study in 11,825
Dutch women participating in the DOM-project. Int J Obes
1990; 14:753-61.
18. Albanes D. Potential for confounding of physical activity risk
assessment by body weight and fatness. Am J Epidemiol
1987;125:745-6.
19. Hong Y, Bots ML, Pan X, et al. Physical activity and cardiovascular risk factors in rural Shanghai, China. Int J Epidemiol
1994;23:1154-8.
20. Helmert U, Herman B, Shea S. Moderate and vigorous leisuretime physical activity and cardiovascular disease risk factors
in West Germany, 1984-1991. Int J Epidemiol 1994;23:
285-92.
Am J Epidemiol
Vol. 146, No. 4, 1997