American Journal of Epidemiology
Copyright ª 2005 by the Johns Hopkins Bloomberg School of Public Health
All rights reserved; printed in U.S.A.
Vol. 163, No. 2
DOI: 10.1093/aje/kwj019
Advance Access publication November 17, 2005
Original Contribution
Cardiorespiratory Fitness Is an Independent Predictor of Hypertension Incidence
among Initially Normotensive Healthy Women
Carolyn E. Barlow, Michael J. LaMonte, Shannon J. FitzGerald, James B. Kampert, Joe L. Perrin,
and Steven N. Blair
Received for publication November 18, 2004; accepted for publication August 30, 2005.
The authors examined the association between cardiorespiratory fitness and incident hypertension in women
who were normotensive and free of cardiovascular disease at baseline in the Aerobics Center Longitudinal Study
(Dallas, Texas), 1970–1998. A total of 4,884 women performed a maximal treadmill exercise test and completed
a follow-up health survey. During an average follow-up of 5 years, 157 incident cases of self-reported, physiciandiagnosed hypertension were identified from the health surveys. The cumulative incidence of hypertension was
3.2%. Compared with the rates of low-fit women, crude hypertension rates were 60% and 79% lower among
women in the moderate and high fitness categories, respectively (p < 0.001). After adjustment for several potential
confounders, the odds ratios for hypertension were 1.0, 0.61 (95% confidence interval (CI): 0.30, 1.21), and 0.35
(95% CI: 0.17, 0.73) in low-, moderately, and highly fit women, respectively (ptrend < 0.01). Each 1-metabolic
equivalent increment in treadmill performance was, on average, associated with a 19% (95% CI: 10, 27; p < 0.001)
lower odds of incident hypertension. The pattern and strength of association between fitness and hypertension
persisted in analyses stratified on body mass index, age, and the presence of prehypertension at baseline. An
active lifestyle should be promoted for the primary prevention of hypertension in women.
cohort studies; hypertension; motor activity; physical fitness; women
Abbreviations: CI, confidence interval; MET, metabolic equivalent; PAR, population attributable risk.
In May 2003, the National High Blood Pressure Education Program released the seventh report of the Joint National Committee on the Prevention, Detection, Evaluation,
and Treatment of High Blood Pressure (1). Attention was
called to the escalating economic and public health burden
of hypertension and related health conditions. Hypertension
afflicts nearly 29 percent of the US adult population or
approximately 58 million individuals (2). Age-adjusted hypertension prevalence is currently higher among women
than men, and a significant increase in the prevalence of
hypertension was observed during the period from 1988 to
2000 in women but not in men (2). A woman who is hypertension free at age 55 years has a 91 percent risk of de-
veloping future hypertension (3). Among middle-aged and
older women, elevations in both systolic and diastolic blood
pressure are associated with strong, continuous increases in
risk of stroke, coronary heart disease, congestive heart failure, peripheral vascular disease, and renal disease across
a very broad range of blood pressure levels (4–7). Even mild
elevations in blood pressure increase the risk of hypertension development (8) and vascular mortality (5, 6, 9). In
women with established cardiovascular disease, elevated
systolic blood pressure is among the strongest predictors
of secondary cardiovascular disease events (10). For this
reason, leading health organizations have intensified efforts
aimed at the primary prevention of hypertension, with a
Correspondence to Carolyn E. Barlow, Centers for Integrated Health Research, The Cooper Institute, 12230 Preston Road, Dallas, TX 75230
(e-mail: [email protected]).
142
Am J Epidemiol 2006;163:142–150
Downloaded from http://aje.oxfordjournals.org/ at Pennsylvania State University on February 27, 2014
From The Cooper Institute, Dallas, TX.
Cardiorespiratory Fitness and Hypertension among Women
resting blood pressure of 140 mmHg systolic or 90
mmHg diastolic (n ¼ 138); had an abnormal resting or
exercise electrocardiogram (n ¼ 403); or achieved less than
85 percent of their age-predicted maximal heart rate during
graded treadmill exercise testing (n ¼ 211). These criteria
resulted in 4,884 women aged 20–79 years for analysis.
Clinical examination
The baseline examination was conducted following a 12hour fast and has been described in detail elsewhere (31,
34). Briefly, body mass index (kg/m2) was computed from
measured height and weight. Waist girth (cm) was measured
level with the umbilicus following a normal exhalation.
Resting blood pressure was auscultated as the first and fifth
Korotkoff sounds according to a standard sphygmomanometer protocol (35). Serum samples were analyzed for lipids
and glucose by use of automated bioassays in accord with
the Centers for Disease Control and Prevention Lipid Standardization Program.
Cardiorespiratory fitness was quantified as the duration
(minutes) of a maximal graded treadmill exercise test using
a modified Balke protocol as previously described (31, 34).
Exercise duration from this protocol correlates highly (r ¼
0.94) with measured maximal oxygen uptake in women (36)
and explains a large proportion of the variation (R2 ¼ 0.60–
0.80) in physical activity habits in our population sample
(37, 38). Study participants were placed into age-specific
fitness categories defined by the distribution of maximal
exercise duration for all women in the Aerobics Center Longitudinal Study (31, 34) according to the following age
groups: 20–39, 40–49, 50–59, and 60 or more years. The
least-fit 20 percent of women in each age-specific distribution of maximal exercise duration were classified as having
low fitness, the next 40 percent as having moderate fitness,
and the remaining 40 percent as having high fitness. Low
fitness by this definition is consistently an independent predictor of morbid and mortal events in the Aerobics Center
Longitudinal Study population (39–41), and it is analogous
to conventional definitions of hypertension or obesity in epidemiologic work. To standardize the estimate of cardiorespiratory fitness, we computed maximal metabolic equivalents
(METs) (1 MET ¼ 3.5 ml of oxygen uptake per kg of body
mass per minute) from the final treadmill speed and grade
(42). Table 1 shows the maximal exercise duration and estimated METs according to the age-specific fitness categories
in women from the Aerobics Center Longitudinal Study.
Morbidity surveillance
MATERIALS AND METHODS
Women in the current study received a preventive medical
examination at the Cooper Clinic in Dallas, Texas, during
1970–1998 and responded to at least one follow-up mail
survey conducted in 1982, 1986, 1990, 1995, or 1999 as
part of the Aerobics Center Longitudinal Study. Participants
were excluded if at baseline they reported a personal history
of hypertension (n ¼ 666), cardiovascular disease (n ¼ 44),
diabetes (n ¼ 89), or cancer (n ¼ 215); had a measured
Am J Epidemiol 2006;163:142–150
The incidence of hypertension was ascertained from the
1982, 1986, 1990, 1995, and 1999 mail-back health surveys.
A case-finding question for physician-diagnosed illness was
used to identify cases of hypertension. Women were asked if
a physician had ever told them they had hypertension. If yes,
respondents were asked to report the year of diagnosis. For
women who completed multiple surveys, the first survey in
which hypertension was reported was used in analyses. This
method of case ascertainment has been used in other large,
Downloaded from http://aje.oxfordjournals.org/ at Pennsylvania State University on February 27, 2014
major emphasis on modifying adverse lifestyle behaviors
such as physical inactivity (11–13).
It is well established that physical activity is associated
with lower cardiovascular morbidity and mortality (14, 15).
Accumulating evidence indicates that regular moderate-tovigorous physical activity significantly lowers resting blood
pressure in normotensive and hypertensive adults (16–18).
Prospective observational studies consistently show that
higher levels of self-reported physical activity are associated
with lower hypertension incidence in men (19–24). Data in
women, however, are equivocal (21, 23–26), which may be
partly explained by imprecise assessment of physical activity in women when using activity recall instruments developed and validated among men (27, 28).
Cardiorespiratory fitness, assessed with maximal exercise
testing, is stronger than self-reported physical activity as
a predictor of cardiovascular outcomes, most likely because
fitness assessment provides a more objective and reliable
index of recent activity habits than do self-reported data
(29). Fitness may also be better than recalled activity as
a predictor of health outcomes, because factors other than
activity may influence both fitness levels and health status
through shared biologic pathways (30). Few prospective
studies have related a baseline fitness exposure to the development of hypertension in initially normotensive individuals (31–33). Reporting on 3,350 middle-aged Japanese
men, Sawada et al. (32) observed a 90 percent higher 5-year
age- and covariate-adjusted incidence of hypertension for
men in the lowest compared with the highest quintile of fitness estimated from submaximal cycle ergometry. Carnethon
et al. (33) reported that, after adjustment for age, sex, and
other confounding factors, each 1-minute decrease in maximal treadmill performance was associated with a 19 percent
higher 15-year risk of incident hypertension among men and
women. An early analysis of 6,039 men and women from
our cohort showed a 52 percent higher 4-year age-, sex-, and
covariate-adjusted incidence of hypertension for individuals
with low versus high maximal treadmill exercise tolerance
(31). A relatively small number of women precluded sexspecific analysis in our previous report.
Given the recent attention to the growing health burden of
hypertension (1), the recent increases in hypertension
among women (2), and the current emphasis on increasing
population levels of physical activity in the primary prevention of hypertension (12), we set forth to examine the association between cardiorespiratory fitness and incident
hypertension in a large cohort of women who were normotensive and healthy at baseline.
143
144 Barlow et al.
TABLE 1. Age-specific cardiorespiratory fitness
classifications among women, Aerobics Center Longitudinal
Study, Dallas, Texas, 1970–1998
Age
(years)
20–39
40–49
50–59
Maximal treadmill
time (minutes)*
Maximal
METsy
Low
<10.3
<8.2
Moderate
10.3–15.0
8.2–10.4
High
>15.0
>10.4
Low
<8.9
<7.6
Moderate
8.9–11.0
7.6–9.4
High
>11.0
>9.4
Low
<6.9
<6.7
Moderate
6.9–10.7
6.7–8.5
High
>10.7
>8.5
Low
<5.5
<5.8
Moderate
5.5–9.0
5.8–7.6
High
>9.0
>7.6
* Modified Balke treadmill protocol: 3.3 mph (5.311 km/hour), 0%
for first minute, 2% for second minute, plus 1% for each additional
minute to 25%, and then plus 0.2 mph (0.3219 km/hour) for each
additional minute until exhaustion.
y METs, metabolic equivalents, estimated from the final treadmill
speed and grade (42).
population-based epidemiologic studies of hypertension (20,
31, 43). We previously verified the accuracy of self-reported,
physician-diagnosed hypertension in this cohort and observed 98 percent sensitivity and 99 percent specificity (31).
Statistical analysis
The primary exposure variable in the current study
was cardiorespiratory fitness, defined categorically as low,
moderate, and high. Once a case was identified, additional
follow-up on subsequent surveys for this individual was not
undertaken. To account for differences in survey response
patterns among study participants, we created indicator variables (yes/no) for each survey period. To standardize for
surveillance and period and length of follow-up, we entered
these variables, as well as the year of the baseline examination, into our analyses as covariables. Cumulative incidence
rates were computed by dividing the number of hypertension cases by the population at risk for the total sample and
within categories of fitness. Because the data included in
this analysis were from periodic surveillance with unknown
or uncertain event times, logistic regression was used to
compute odds ratios and 95 percent confidence intervals
for incident hypertension according to categories of cardiorespiratory fitness. Adjusted models controlled for the potential confounding effects of baseline age (years), smoking
(yes/no), alcohol intake (drinks/week), systolic and diastolic
blood pressure (mmHg), family history of hypertension
(yes/no), waist girth (cm), glucose (mg/dl), and triglycerides
(mg/dl). Tests for linear trend of the fitness-hypertension
association were conducted by entering the three-category
fitness variable into the regression model as an ordinal term.
RESULTS
Study participants were mostly Caucasian (>95 percent)
and from middle to upper socioeconomic strata. Women
who developed hypertension were somewhat older and less
physically fit and had a higher prevalence of being 55 or
more years of age and prehypertensive at baseline than
women who remained normotensive through follow-up
(table 2). Significant (p < 0.01) Spearman correlations were
observed with systolic and diastolic blood pressure, respectively, for age (r ¼ 0.27 and r ¼ 0.24), body mass index (r ¼
0.23 and r ¼ 0.21), waist girth (r ¼ 0.21 and r ¼ 0.19),
triglycerides (r ¼ 0.19 and r ¼ 0.16), glucose (r ¼ 0.19 and
r ¼ 0.16), and maximal METs (r ¼ 0.18 and r ¼ 0.17).
A total of 157 women reported developing hypertension
during an average of 5 years’ follow-up, which resulted in
a 3.2 percent cumulative incidence of hypertension. A steep
inverse gradient (p < 0.001) was observed across the rates
of hypertension and cardiorespiratory fitness (figure 1).
Crude incidence rates were significantly lower among
women with moderate and high fitness than among women
with low fitness (p < 0.05 each).
Odds ratios were used to quantify the strength of association between cardiorespiratory fitness and incident hypertension (table 3). Women in the low fitness category were
the referent group. Unadjusted analyses demonstrated an
inverse dose-response relation (p < 0.0001) between fitness
and hypertension. Women in the moderate and high fitness
categories, respectively, had a 61 percent (95 percent confidence interval (CI): 43, 73) and 81 percent (95 percent CI:
71, 87) lower relative odds of developing hypertension when
compared with their low-fit counterparts. The strength and
pattern of association persisted with increasing control of
potential confounding influences. In the fully adjusted
model, each 1-MET increment in maximal exercise performance was, on average, associated with a 19 percent
(95 percent CI: 10, 27) lower odds of incident hypertension
(p < 0.001).
We examined the influence of fitness on hypertension incidence within strata of known hypertension antecedents
(table 4). Higher levels of fitness were associated with lower
incidence of hypertension in women who were normal
weight (p < 0.01) and overweight (p ¼ 0.02). When stratified according to age, an inverse gradient for incident
Am J Epidemiol 2006;163:142–150
Downloaded from http://aje.oxfordjournals.org/ at Pennsylvania State University on February 27, 2014
60
Fitness
category
To examine potential modifying effects of select variables
on the fitness-hypertension association, we conducted logistic regression analyses according to strata of body mass index (<25 vs. 25 kg/m2), age (<55 vs. 55 years), and
baseline prehypertension (present/absent). The population
attributable risk (PAR) of hypertension was estimated for
fitness and selected risk factors to quantify the influence that
changing these risk factors might have on hypertension incidence in our population sample. PAR was computed as
Pc(1 1/ORadj), where Pc is the prevalence of a risk factor
among hypertension cases, and ORadj is the multivariable
adjusted odds ratio for hypertension associated with the
specified risk factor (44). All p values are two sided, with
an alpha level of 0.05.
Cardiorespiratory Fitness and Hypertension among Women
12
TABLE 2. Baseline characteristics by hypertension status at
follow-up among women, Aerobics Center Longitudinal Study,
Dallas, Texas, 1970–1998
Age (years)
42.8 (9.9)
Hypertensive
(n ¼ 157)
(mean (SD) or %)
46.9 (9.5)
Body mass index (kg/m2)
22.3 (3.5)
23.5 (3.7)
Waist girth (cm)
65.2 (21.6)
70.3 (25.0)
Treadmill time (minutes)
13.8 (4.6)
10.9 (4.0)
Maximal METs*
9.9 (2.3)
8.2 (2.0)
Total cholesterol (mg/dl)
198.6 (43.1)
210.3 (40.4)
61.6 (17.5)
60.6 (13.7)
Triglyceride (mg/dl)
84.9 (84.5)
97.2 (56.5)
Glucose (mg/dl)
93.2 (9.4)
95.8 (11.6)
Systolic blood pressure
(mmHg)
109.5 (11.1)
119.1 (11.6)
Diastolic blood pressure
(mmHg)
73.4 (7.8)
79.1 (7.6)
Alcoholic drinks per weeky
3.3 (7.2)
4.4 (11.3)
Follow-up interval (years)
4.9 (3.4)
7.9 (5.7)
55 years of age (%)
13.7
26.1
Prehypertensionz (%)
35.8
63.1
Current smoker (%)
9.5
10.8
Family history of
hypertension (%)
43.9
49.7
Fitness level (%)
Low
11.8
39.5
Moderate
36.7
34.2
High
51.5
26.3
* SD, standard deviation; METs, metabolic equivalents; HDL
cholesterol, high density lipoprotein cholesterol.
y One unit of alcohol is defined as 12 ounces (3.41 dl) of beer,
5 ounces (1.421 dl) of wine, or 1.5 ounces (0.4262 dl) of hard liquor.
z Resting systolic blood pressure of 120–139 mmHg or diastolic
blood pressure of 80–89 mmHg.
hypertension across levels of fitness was seen in women
aged less than 55 years (p < 0.0001). A similar pattern of
association was observed in women who were aged 55 or
more years; however, the p value was nonsignificant (p ¼
0.22) in this age stratum. Prehypertension was present at
baseline in 63.1 percent of women who became cases, compared with 35.8 percent (v21df ¼ 46:3; p < 0.0001) of women
who remained hypertension free. The age-adjusted odds
ratio of incident hypertension was 2.8 (95 percent CI: 1.9,
3.9) in women with prehypertension at baseline. However,
an inverse association between fitness and hypertension was
observed in women both with and without prehypertension
at baseline (p < 0.01).
The relative importance of low fitness and other hypertension risk factors is shown in table 5. After adjustment for
each risk factor shown in the table, low cardiorespiratory
fitness was a significant predictor of incident hypertension
on the same or higher order of magnitude as having prehypertension, being overweight, and being 55 or more years of
Am J Epidemiol 2006;163:142–150
10
8
Linear trend p < 0.001
6
62
4
44
2
0
Low
Moderate
High
Cardiorespiratory Fitness Level
FIGURE 1. Cumulative incidence of hypertension by level of cardiorespiratory fitness among women, Aerobics Center Longitudinal
Study, Dallas, Texas, 1970–1998. Bars are 95% confidence intervals.
The number of cases is shown above the bars.
age. To place the risk of hypertension for each exposure
category in the context of population-disease burden, we
computed PAR estimates based on the prevalence and
strength of association with hypertension for each exposure.
Accordingly, if all unfit women in our population sample
became fit, a reduction in new-onset hypertension of 22
percent might be expected.
DISCUSSION
The data reported here show an inverse dose-response
association between cardiorespiratory fitness and 5-year incidence of hypertension in initially normotensive and
healthy middle-aged women. The association persisted after
extensive adjustment for potential confounding factors and
when stratified on the presence of known hypertension precursors, including baseline prehypertension. Strengths of
the study are use of maximal exercise testing to quantify
cardiorespiratory fitness, women free of hypertension and
other forms of cardiovascular disease at baseline, and a large
sample with extensive follow-up for incident cases. We believe this is the first study of cardiorespiratory fitness and
hypertension in women that meets all of these criteria.
There are three principal findings from the current study.
First, the inverse gradient of hypertension rates across fitness categories persisted through extensive control for confounding factors including age, smoking status, baseline
blood pressure, family history of hypertension, waist girth,
and metabolic risk factors known to synergize in the development of hypertension (45, 46). The general pattern
of association was for substantially lower relative odds of
incident hypertension in women with moderate levels of
Downloaded from http://aje.oxfordjournals.org/ at Pennsylvania State University on February 27, 2014
HDL cholesterol* (mg/dl)
51
Cumulative Incidence Rate (%)
Normotensive
(n ¼ 4,727)
(mean (SD*) or %)
145
146 Barlow et al.
TABLE 3. Odds ratios and 95% confidence intervals for incident hypertension by level of
cardiorespiratory fitness among women, Aerobics Center Longitudinal Study, Dallas, Texas, 1970–1998
Fitness level
Moderate (n ¼ 1,795)
Low (n ¼ 609)
High (n ¼ 2,480)
Odds
ratio
95%
confidence interval
Odds
ratio
95%
confidence interval
0.13, 0.29*
Unadjusted
1.0
0.39
0.27, 0.57
0.19
Multivariabley
1.0
0.45
0.30, 0.68
0.27
0.17, 0.43*
Multivariablez
1.0
0.47
0.31, 0.72
0.31
0.19, 0.50*
Multivariable§
1.0
0.61
0.30, 1.21
0.35
0.17, 0.73**
cardiorespiratory fitness. Adjusted odds ratios showed a 39–
55 percent lower incidence of hypertension in moderately fit
women compared with women in the low fitness category.
Women with higher levels of fitness category had even lower
relative odds of hypertension than did moderately fit participants. The observation that moderate and higher levels of
cardiorespiratory fitness were associated with lower incidence of hypertension in our cohort of initially normotensive
and healthy women has clinical and public health importance. The national prevalence of hypertension has risen
significantly in women (2), and hypertension is a potent risk
factor for new and recurrent cardiovascular disease events in
women (7, 10). Experimental studies suggest that exercise
training-induced reductions in blood pressure are most pronounced in individuals with established hypertension (13,
16). The prospective data reported here indicate that higher
cardiorespiratory fitness may be an important homeostatic
factor that contributes to the regulation and maintenance of
healthy blood pressure levels. Our findings underscore the
importance of promoting increases in physical activity and
higher levels of fitness as a primary method of preventing
hypertension in women (11–13).
TABLE 4. Odds ratios and 95% confidence intervals for incident hypertension according to body mass
index level, age, and prehypertension status by cardiorespiratory fitness level among women, Aerobics
Center Longitudinal Study, Dallas, Texas, 1970–1998
Fitness level
Moderate (n ¼ 1,795)
Low (n ¼ 609)
High (n ¼ 2,480)
Odds
ratio
95%
confidence interval
Odds
ratio
95%
confidence interval
Body mass index
(kg/m2)y
<25
1.0
0.59
0.33, 1.05
0.39
0.19, 0.72*
25
1.0
0.41
0.20, 0.79
0.40
0.18, 0.94**
<55
1.0
0.53
0.33, 0.86
0.29
0.17, 0.51***
55
1.0
0.46
0.17, 1.2
0.45
0.17, 1.3****
Present
1.0
0.29
0.14, 0.63
0.24
0.11, 0.52***
Absent
1.0
0.71
0.40, 1.3
0.42
0.22, 0.79***
Age (years)z
Prehypertension§
* ptrend ¼ 0.004; **ptrend ¼ 0.02; ***ptrend < 0.01; ****ptrend ¼ 0.22.
y Adjusted for age, examination year, survey response pattern, current smoking, family history of hypertension,
and baseline blood pressure.
z Adjusted for examination year, survey response pattern, current smoking, family history of hypertension, body
mass index, and baseline blood pressure.
§ Adjusted for age, examination year, survey response pattern, current smoking, family history of hypertension,
and body mass index.
Am J Epidemiol 2006;163:142–150
Downloaded from http://aje.oxfordjournals.org/ at Pennsylvania State University on February 27, 2014
* ptrend < 0.0001; **ptrend < 0.01.
y Adjusted for age, examination year, and survey response pattern.
z Adjusted for the above variables and smoking, alcohol intake, systolic and diastolic blood pressure, and family
history of hypertension.
§ Adjusted for the above variables and waist girth, glucose, and triglyceride concentration. Note: the results of the
third multivariable model were materially unchanged when body mass index was substituted for waist girth.
Cardiorespiratory Fitness and Hypertension among Women
147
TABLE 5. Odds ratios and population attributable risk of incident hypertension for low cardiorespiratory
fitness and selected hypertension risk factors among women, Aerobics Center Longitudinal Study, Dallas,
Texas, 1970–1998
Prevalence
(%)*
Odds ratio
Unadjusted
Adjustedy
95%
confidence intervaly
Population
attributable risk (%)z
15.5
3.5
2.3
1.5, 3.4
22.3
Family history of hypertension
44.2
1.3
1.4
0.99, 1.9
14.1
Current smoker
9.5
1.2
0.95
0.54, 1.7
0.6
Prehypertension
36.7
3.1
2.6
1.8, 3.8
38.7
Body mass index 25 kg/m2
16.7
3.1
1.8
1.2, 2.7
16.4
Glucose 110 mg/dl
4.3
2.1
0.95
0.48, 1.9
0.3
Triglyceride 150 mg/dl
7.5
2.6
1.4
0.86, 2.4
2.0
14.1
2.2
1.9
1.3, 3.1
12.4
Age 55 years
* Prevalence among the entire population sample.
y Adjusted for examination year, survey response pattern, and the other variables in the table.
z Computed as Pc(1 1/ORadj), where Pc is the prevalence among hypertension cases, and ORadj is the
multivariable adjusted odds ratio (44). Pc (ordered as listed in the table) was 39.5, 49.7, 10.8, 63.1, 36.9, 6.1, 7.1,
and 26.1.
A second major observation from our analyses was the
consistent inverse gradient between hypertension and cardiorespiratory fitness within various strata of other hypertension predictors (table 4). Body mass index is a predictor
of hypertension in women (23, 25, 26, 31, 43, 45). Whether
classified as normal weight or overweight on the basis of
their body mass index, women with moderate and high cardiorespiratory fitness in our study had relative odds of hypertension 41–61 percent lower compared with their low-fit
peers. Our finding is similar to the pattern of association
between recalled physical activity and incident hypertension
reported across the strata of body mass index in a cohort of
Finnish women (23) and previously reported in men (19, 20,
23). Body fat per se may not be the principal vector of hypertension risk; rather, overweight in the presence of physical
inactivity and low fitness may represent the high-risk phenotype. Although overweight and obesity are significant population health concerns associated with hypertension (47),
the results from our study and those reported by others (19,
20, 23) emphasize the importance of physical activity and
fitness in preventing hypertension among individuals with
normal weight as well as individuals who are overweight or
mildly obese. Additional studies in more diverse population
samples of women and men with objective measures of
physical activity, fitness, and body fat are needed to better
understand this complex issue.
Data from the Framingham Heart Study indicate that
women who are normotensive at age 55 years have greater
than a 90 percent risk of developing hypertension during
their remaining years of life (3). The older segment of the
US population is rapidly increasing (48). It is, therefore, conceivable that most adults living beyond their sixth decade
would require pharmacotherapy for hypertension, which is
costly and potentially hazardous (1). Our analyses show a
significant inverse association between fitness and hypertension incidence in women aged less than 55 years and a similar, albeit nonsignificant (p ¼ 0.22), pattern of association
Am J Epidemiol 2006;163:142–150
in women aged 55 or more years, which may be attributed to
the small number of cases (n ¼ 9) in the high-fitness category. It appears as though moderate and higher levels of fitness may prevent the development of hypertension in young
and older women alike.
It has become apparent that even high normal blood pressure increases the risk of developing hypertension (8) and
cardiovascular events (5). Accordingly, national blood pressure guidelines now include prehypertension (systolic blood
pressure of 120–139 mmHg or diastolic blood pressure of
80–89 mmHg) as the first category of abnormal blood pressure to be targeted with therapeutic intervention (1). Prehypertension was present at baseline in 63 percent of women
who became hypertensive during follow-up. The relative
odds of developing hypertension were threefold higher in
women with prehypertension at baseline. However, even in
women with prehypertension at baseline, the incidence of
hypertension was 29 percent and 58 percent lower among
moderately and highly fit women, respectively, compared
with low-fit women. Current guidelines recommend intensive lifestyle modifications, including increased physical
activity, as the primary therapy for prehypertension (1,
12). Our prospective data indicate that higher levels of fitness confer substantial protection against clinically manifest
hypertension in women who are at risk because of the presence of prehypertension.
The third major finding from our study was the substantial
strength and independence of low cardiorespiratory fitness as
a predictor of incident hypertension (table 5). The 2.3-fold
increase in the multivariable odds of incident hypertension
associated with low fitness was slightly lower than that observed for prehypertension and higher than the adjusted odds
ratios associated with a body mass index of greater than 25,
age of 55 or more years, family history of hypertension,
current smoking, and high levels of metabolic antecedents
(45, 46) for hypertension. After extensive adjustment for
potential confounders, the strength of association between
Downloaded from http://aje.oxfordjournals.org/ at Pennsylvania State University on February 27, 2014
Low fitness
148 Barlow et al.
sion risk in our population sample of women. However,
because the women included in this report were relatively
young and apparently healthy at baseline, it is not likely
that menopausal issues or medication use explained a large
proportion of the observed associations reported herein.
Nonetheless, these issues should be addressed in future prospective analyses on fitness and hypertension risk.
Although cardiorespiratory fitness was predicted from
maximal exercise test performance rather than measured
with indirect calorimetry, correlations between predicted
and measured fitness levels in women are large (r ¼ 0.94)
(36). Recent evidence of only a modest genetic contribution
(25–45 percent) to cardiorespiratory fitness (52), a high correlation (r ¼ 0.70–0.90) between cardiorespiratory fitness
and detailed records of usual physical activity habits (37,
38), reported training gains of up to 30 percent (42), and
rapid detraining-related losses (53) illustrate the plasticity of
cardiorespiratory fitness and should temper arguments of
genetic transmission as its principal determinant.
The presence of subclinical disease at baseline could have
affected treadmill performance or increased the incidence of
hypertension during follow-up. We attempted to account for
this by excluding women with abnormal resting or exercise
electrocardiograms, women who did not achieve at least 85
percent of the age-predicted maximal heart rate during the
treadmill exercise test, and women with known or suspected
cardiovascular disease and cancer at baseline. Baseline risk
factor profiles (table 2) were similar between women who
remained hypertension free and women who developed hypertension in follow up. This adds confidence to our assertion that undetected subclinical disease is not the underlying
difference in hypertension incidence between fit and unfit
women. The independent dose-response gradient between
fitness and hypertension also supports a causal inference.
In conclusion, our prospective findings in a large cohort
of initially normotensive and healthy middle-aged women
show that cardiorespiratory fitness is an independent predictor of incident hypertension in women. Even small reductions
in blood pressure associated with modest increases in physical activity and fitness may have a substantial public health impact on hypertension prevalence and related health sequellae.
We believe physicians should counsel sedentary patients to
become physically active and to improve their cardiorespiratory fitness for the primary prevention of hypertension.
ACKNOWLEDGMENTS
Supported by grants AG06945 and HL62508 from the
National Institutes of Health. Additional support has been
provided by the Communities Foundation of Texas on
recommendation of Nancy Ann and Ray L. Hunt.
The authors thank Dr. Kenneth H. Cooper for establishing
the Aerobics Center Longitudinal Study, the Cooper Clinic
physicians and technicians for collecting the baseline data,
Melba Morrow for editorial assistance, the Center for Data
Management for conducting the mail surveys, and the data
entry staff for processing data on patients and the survey.
Conflict of interest: none declared.
Am J Epidemiol 2006;163:142–150
Downloaded from http://aje.oxfordjournals.org/ at Pennsylvania State University on February 27, 2014
low fitness and incident hypertension in our cohort of
women was higher than reported among Japanese men (relative risk 1.8) (32), higher than the sex- and covariateadjusted association previously reported in our cohort of
healthy women and men (relative risk ¼ 1.5) (31), and similar to the sex- and covariate-adjusted association (hazard
ratio 2.4–3.0) reported in men and women from the
Coronary Artery Risk Development in Young Adults Study
(33). Clearly, low levels of cardiorespiratory fitness should
be seen as a potent risk factor for hypertension development.
Increasing population levels of fitness through greater physical activity should be a cornerstone element in the primary
prevention of hypertension.
We computed PAR values to estimate the burden of hypertension attributable to low fitness and other risk predictors. We observed that 22 percent of incident hypertension
cases might be averted or delayed if all unfit women in our
population sample became fit. Our PAR for low fitness and
hypertension is consistent with the 21 percent recently reported for men and women combined in the Coronary Artery Risk Development in Young Adults Study (33). In our
study, the PAR associated with low fitness was second in
magnitude only to prehypertension (PAR ¼ 39 percent);
however, our data also indicate that moderate and higher
levels of fitness protect against hypertension development
even in women with prehypertension. Experimental evidence indicates that most women can achieve moderate levels of cardiorespiratory fitness (e.g., 6–9 METs) through
regular participation in moderate intensity activities like
brisk walking or bicycling (49, 50). Although genetic susceptibility is an important contributor, other mechanisms
underlie hypertension such as architectural changes in the
vascular wall, altered endothelial cell function, increased
activity of the sympathetic nervous system, hyperinsulinemia, hypertriglyceridemia, and increased sodium retention
(51), each of which may be favorably modified by regular
physical activity (13). Unlike antihypertensive drugs, which
are costly, largely limited to blood pressure control, and
often have adverse effects, physical activity is a relatively
safe and inexpensive therapy with favorable effects on a
broad spectrum of cardiovascular disease antecedents and
outcomes, including hypertension (15).
Similar to other large, prospective, epidemiologic studies,
the Aerobics Center Longitudinal Study has limitations that
should be considered. The external validity of our findings is
specific to women with demographics similar to those enrolled in the Aerobics Center Longitudinal Study. The homogeneity of our cohort enhances the internal validity of our
findings because it lowers the likelihood of confounding by
these characteristics. Further, our study participants are very
similar on key clinical variables such as lipids, glucose, and
blood pressure to participants in other large epidemiologic
studies in the United States (34, 43). While we urge caution
against overgeneralizing the results reported herein, we are
not aware of published data suggesting that the health benefits of enhanced fitness would be any less among women of
other races/ethnicities or socioeconomic disposition.
Lack of sufficient data on nutrient intake, menopausal
status, and medication use precluded examination of their
influences on the association between fitness and hyperten-
Cardiorespiratory Fitness and Hypertension among Women
REFERENCES
Am J Epidemiol 2006;163:142–150
20. Paffenbarger RS Jr, Wing AL, Hyde RT, et al. Physical activity
and incidence of hypertension in college alumni. Am J Epidemiol 1983;117:245–57.
21. Pereira MA, Folsom AR, McGovern PG, et al. Physical activity and incident hypertension in black and white adults: the
Atherosclerosis Risk in Communities Study. Prev Med 1999;
28:304–12.
22. Hayashi T, Tsumura K, Suematsu C, et al. Walking to work
and the risk for hypertension in men: the Osaka Health Survey.
Ann Intern Med 1999;131:21–6.
23. Hu G, Barengo NC, Tuomilehto J, et al. Relationship of physical activity and body mass index to the risk of hypertension:
a prospective study in Finland. Hypertension 2004;43:25–30.
24. Haapanen N, Miilunpalo S, Vuori I, et al. Association of leisure time physical activity with the risk of coronary heart
disease, hypertension and diabetes in middle-aged men and
women. Int J Epidemiol 1997;26:739–47.
25. Folsom AR, Prineas RJ, Kaye SA, et al. Incidence of hypertension and stroke in relation to body fat distribution and other
risk factors in older women. Stroke 1990;21:701–6.
26. Dyer AR, Liu K, Walsh M, et al. Ten-year incidence of elevated blood pressure and its predictors: the CARDIA Study.
Coronary Artery Risk Development in (Young) Adults. J Hum
Hypertens 1999;13:13–21.
27. Ainsworth BE. Issues in the assessment of physical activity in
women. Res Q Exerc Sport 2000;71(suppl):S37–42.
28. LaMonte MJ, Ainsworth BE. Quantifying energy expenditure and physical activity in the context of dose response.
Med Sci Sports Exerc 2001;33(suppl):S370–8.
29. Blair SN, Cheng Y, Holder JS. Is physical activity or physical
fitness more important in defining health benefits? Med Sci
Sports Exerc 2001;33(suppl):S379–99.
30. Haskell WL. Physical activity and health: need to define the
required stimulus. Am J Cardiol 1985;55:4D–9D.
31. Blair SN, Goodyear NN, Gibbons LW, et al. Physical fitness
and incidence of hypertension in healthy normotensive men
and women. JAMA 1984;252:487–90.
32. Sawada S, Tanaka H, Funakoshi M, et al. Five year prospective study on blood pressure and maximal oxygen uptake.
Clin Exp Pharmacol Physiol 1993;20:483–7.
33. Carnethon MR, Gidding SS, Nehgme R, et al. Cardiorespiratory fitness in young adulthood and the development of cardiovascular disease risk factors. JAMA 2003;290:3092–100.
34. Blair SN, Kohl HW 3rd, Paffenbarger RS Jr, et al. Physical
fitness and all-cause mortality: a prospective study of healthy
men and women. JAMA 1989;262:2395–401.
35. Perloff D, Grim C, Flack J, et al. Human blood pressure determination by sphygmomanometry. Circulation 1993;88:2460–70.
36. Pollock ML, Foster C, Schmidt D, et al. Comparative analysis
of physiologic responses to three different maximal graded
exercise test protocols in healthy women. Am Heart J 1982;
103:363–73.
37. Blair SN, Mulder RT, Kohl HW. Reaction to ‘‘secular trends in
adult physical activity: exercise boom or bust?’’. Res Q Exerc
Sport 1987;58:106–10.
38. Paffenbarger RS Jr, Blair SN, Lee IM, et al. Measurement of
physical activity to assess health effects in free-living populations. Med Sci Sports Exerc 1993;25:60–70.
39. Blair SN, Kampert JB, Kohl HW, et al. Influences of cardiorespiratory fitness and other precursors on cardiovascular
disease and all-cause mortality in men and women. JAMA
1996;276:205–10.
40. Kampert JB, Blair SN, Barlow CE, et al. Physical activity,
physical fitness, and all-cause and cancer mortality: a prospective study of men and women. Ann Epidemiol 1996;6:452–7.
Downloaded from http://aje.oxfordjournals.org/ at Pennsylvania State University on February 27, 2014
1. Chobanian AV, Bakris GL, Black HR, et al. The Seventh Report of the Joint National Committee on Prevention, Detection,
Evaluation, and Treatment of High Blood Pressure: the JNC 7
report. JAMA 2003;289:2560–72.
2. Hajjar I, Kotchen TA. Trends in prevalence, awareness, treatment, and control of hypertension in the United States, 1988–
2000. JAMA 2003;290:199–206.
3. Vasan RS, Beiser A, Seshadri S, et al. Residual lifetime
risk for developing hypertension in middle-aged women
and men: the Framingham Heart Study. JAMA 2002;287:
1003–10.
4. Stamler J, Stamler R, Neaton JD. Blood pressure, systolic and
diastolic, and cardiovascular risks. US population data. Arch
Intern Med 1993;153:598–615.
5. Vasan RS, Larson MG, Leip EP, et al. Impact of high-normal
blood pressure on the risk of cardiovascular disease. N Engl
J Med 2001;345:1291–7.
6. Benetos A, Thomas F, Safar ME, et al. Should diastolic and
systolic blood pressure be considered for cardiovascular risk
evaluation: a study in middle-aged men and women. J Am Coll
Cardiol 2001;37:163–8.
7. Vasan RS, Massaro JM, Wilson PW, et al. Antecedent blood
pressure and risk of cardiovascular disease: the Framingham
Heart Study. Circulation 2002;105:48–53.
8. Vasan RS, Larson MG, Leip EP, et al. Assessment of frequency
of progression to hypertension in non-hypertensive participants in the Framingham Heart Study: a cohort study. Lancet
2001;358:1682–6.
9. Lewington S, Clarke R, Qizilbash N, et al. Age-specific
relevance of usual blood pressure to vascular mortality: a
meta-analysis of individual data for one million adults in 61
prospective studies. Lancet 2002;360:1903–13.
10. Mason PJ, Manson JE, Sesso HD, et al. Blood pressure
and risk of secondary cardiovascular events in women: the
Women’s Antioxidant Cardiovascular Study (WACS). Circulation 2004;109:1623–9.
11. 1999 World Health Organization–International Society of
Hypertension guidelines for the management of hypertension.
Guidelines Subcommittee. J Hypertens 1999;17:151–83.
12. Whelton PK, He J, Appel LJ, et al. Primary prevention of
hypertension: clinical and public health advisory from the
National High Blood Pressure Education Program. JAMA
2002;288:1882–8.
13. Pescatello LS, Franklin BA, Fagard R, et al. American College
of Sports Medicine position stand. Exercise and hypertension.
Med Sci Sports Exerc 2004;36:533–53.
14. Kohl HW 3rd. Physical activity and cardiovascular disease:
evidence for a dose response. Med Sci Sports Exerc 2001;33
(suppl):S472–83.
15. NIH Consensus Development Panel on Physical Activity and
Cardiovascular Health. Physical activity and cardiovascular
health. JAMA 1996;276:241–6.
16. Fagard RH. Exercise characteristics and the blood pressure
response to dynamic physical training. Med Sci Sports Exerc
2001;33(suppl):S484–92.
17. Ishikawa-Takata K, Ohta T, Tanaka H. How much exercise is
required to reduce blood pressure in essential hypertensives:
a dose-response study. Am J Hypertens 2003;16:629–33.
18. Ketelhut RG, Franz IW, Scholze J. Regular exercise as an
effective approach in antihypertensive therapy. Med Sci Sports
Exerc 2004;36:4–8.
19. Paffenbarger RS Jr, Lee IM. Intensity of physical activity
related to incidence of hypertension and all-cause mortality:
an epidemiological view. Blood Press Monit 1997;2:115–23.
149
150 Barlow et al.
47. Mokdad AH, Ford ES, Bowman BA, et al. Prevalence of
obesity, diabetes, and obesity-related health risk factors, 2001.
JAMA 2003;289:76–9.
48. Yanowitz FG, LaMonte MJ. Physical activity and health in the
elderly. Curr Sports Med Rep 2002;1:354–61.
49. Duncan JJ, Gordon NF, Scott CB, et al. Women walking for
health and fitness: how much is enough? JAMA 1991;266:
3295–9.
50. Branch JD, Pate RR, Bourque SP. Moderate intensity exercise training improves cardiorespiratory fitness in women.
J Womens Health Gend Based Med 2000;9:65–73.
51. Kaplan NM. Clinical hypertension. Philadelphia, PA:
Williams & Wilkins, 1998.
52. Bouchard C, Malina RM, Perusse L, eds. Genetics of fitness and
physical performance. Champaign, IL: Human Kinetics, 1997.
53. Coyle EF, Martin WH 3rd, Sinacore DR, et al. Time course
of loss of adaptations after stopping prolonged intense endurance training. J Appl Physiol 1984;57:1857–64.
Am J Epidemiol 2006;163:142–150
Downloaded from http://aje.oxfordjournals.org/ at Pennsylvania State University on February 27, 2014
41. Wei M, Gibbons LW, Mitchell TL, et al. The association between cardiorespiratory fitness and impaired fasting glucose
and type 2 diabetes mellitus in men. Ann Intern Med 1999;
130:89–96.
42. American College of Sports Medicine. ACSM’s guidelines
for exercise testing and prescription. Philadelphia, PA:
Lippincott Williams & Wilkins, 2000.
43. Huang Z, Willett WC, Manson JE, et al. Body weight,
weight change, and risk for hypertension in women. Ann
Intern Med 1998;128:81–8.
44. Rothman KJ, Greenland S. Modern epidemiology. Philadelphia, PA: Lippincott-Raven Publishers, 1998.
45. Haffner SM, Miettinen H, Gaskill SP, et al. Metabolic precursors of hypertension. The San Antonio Heart Study. Arch
Intern Med 1996;156:1994–2001.
46. Hunt SC, Stephenson SH, Hopkins PN, et al. Predictors of an
increased risk of future hypertension in Utah. A screening
analysis. Hypertension 1991;17:969–76.