Leisure-Time Physical Activity and Ischemic Stroke Risk

The Northern Manhattan Stroke Study

Ralph L. Sacco, MD, MS; Robert Gan, MD; Bernadette Boden-Albala, MPH; I-Feng Lin, MS;
Douglas E. Kargman, MS, MD; W. Allen Hauser, MD; Steven Shea, MD, MS; Myunghee C. Paik, PhD
Background and PurposePhysical activity reduces the risk of premature death and cardiovascular disease, but the relationship
to stroke is less well studied. The objective of this study was to investigate the association between leisure-time physical
activity and ischemic stroke in an urban, elderly, multiethnic population.
MethodsThe Northern Manhattan Stroke Study is a population-based incidence and case-control study. Case subjects had
first ischemic stroke, and control subjects were derived through random-digit dialing with 1:2 matching for age, sex,
and race/ethnicity. Physical activity was recorded through a standardized in-person interview regarding the frequency and
duration of 14 activities over the 2 prior weeks. Conditional logistic regression was used to calculate odds ratios (OR)
and 95% confidence intervals after adjustment for medical and socioeconomic confounders.
ResultsOver 30 months, 369 case subjects and 678 control subjects were enrolled. Mean age was 69.9612 years; 57% were
women, 18% whites, 30% blacks, and 52% Hispanics. Leisure-time physical activity was significantly protective for stroke
after adjustment for cardiac disease, peripheral vascular disease, hypertension, diabetes, smoking, alcohol use, obesity,
medical reasons for limited activity, education, and season of enrollment (OR50.37; 95% confidence interval50.25 to
0.55). The protective effect of physical activity was detected in both younger and older groups, in men and women, and
in whites, blacks, and Hispanics. A dose-response relationship was shown for both intensity (light-moderate activity
OR50.39; heavy OR50.23) and duration (,2 h/wk OR50.42; 2 to ,5 h/wk OR50.35; $5 h/wk OR50.31) of
physical activity.
ConclusionsLeisure-time physical activity was related to a decreased occurrence of ischemic stroke in our elderly,
multiethnic, urban subjects. More emphasis on physical activity in stroke prevention campaigns is needed among the
elderly. (Stroke. 1998;29:380-387.)
Key Words: cerebrovascular disorders n elderly n epidemiology n ethnic groups n racial differences n risk factors

egular exercise has well-established benefits for reducing

the risk of premature death and many diseases, including
cardiovascular disease. In Healthy People 2000, the US Department of Health and Human Services targeted physical activity
in health objectives 1.3 to 1.7 for health promotion and disease
prevention. The aim by the year 2000 is to increase the
proportion of people who engage in regular physical activity
and reduce the proportion of those who engage in no
leisure-time physical activity, particularly among people aged
65 years and older.1 Current guidelines endorsed by the
Centers for Disease Control and Prevention and the National
Institutes of Health recommend that Americans should exercise for at least 30 minutes of moderately intense physical
activity on most, and preferably all, days of the week.2,3
The relationship between physical activity and stroke is less
well studied than for other cardiovascular diseases. Only a few
previous studies have evaluated the association between physical activity and the risk of stroke.4 12 The beneficial effects

have been predominately described among white populations,

have been more apparent for men than women, and generally
have been described for younger rather than older adults. The
aim of this population-based case-control study was to investigate the relationship between leisure-time/recreational physical activity and ischemic stroke in an elderly, urban, multiethnic population.

Subjects and Methods

The Northern Manhattan Stroke Study (NOMASS) is an ongoing,
prospective, population-based incidence and case-control study designed to determine stroke incidence, risk factors, and prognosis in a
multiethnic, urban population. Northern Manhattan consists of the
area north of 145th Street and south of 218th Street, bordered on the
west by the Hudson River, and separated from the Bronx on the east
by the Harlem River. In 1990 nearly 260 000 people lived in the
region, with 40% older than 39 years. The racial/ethnic distribution in
this community is approximately 20% black, 63% Hispanic, and 15%
white residents.

Received September 17, 1997; final revision received November 6, 1997; accepted November 21, 1997.
From the Department of Neurology (R.L.S., R.G., B.B.-A., D.E.K., W.A.H., I-F.L.); Sergievsky Center (R.L.S., W.A.H.); Division of Epidemiology
(R.L.S., W.A.H., S.S.); Department of Medicine (Division of General Medicine) (S.S.); and Division of Biostatistics (I-F.L., M.C.P.), Columbia University
College of Physicians and Surgeons and School of Public Health, New York, NY.
Presented in part in abstract form at the 48th Annual Meeting of the American Academy of Neurology, San Francisco, Calif, March 2330, 1996.
Correspondence to Ralph L. Sacco, MD, MS, Neurological Institute, 710 W 168 St, New York, NY 10032. E-mail RLS1@Columbia.edu
1998 American Heart Association, Inc.

380
Downloaded from http://stroke.ahajournals.org/
by guest on March 5, 2016

Sacco et al

381

Index Evaluation of Case and Control Subjects

ADL
BMI
CI
NOMASS
OR
QWB
TIA

Selected Abbreviations and Acronyms

5 Activities of Daily Living
5 body mass index
5 confidence interval
5 Northern Manhattan Stroke Study
5 odds ratio
5 Quality of Well-being
5 transient ischemic attack

Selection of Case Subjects

Case subjects eligible for the case-control study were prospectively
enrolled if they met the following criteria: (1) diagnosed with first
cerebral infarction after July 1, 1993; (2) older than 39 years at onset
of the stroke; and (3) resided in the Northern Manhattan community
in a household with a telephone. Even patients who died from their
initial stroke were enrolled by relying on surrogate data. Patients with
TIA were excluded, ie, neurological deficits lasting less than 24 hours
and no ischemic infarct found on brain imaging. Prospective case
surveillance consisted of daily screening of all admissions, discharges,
and head CT scan logs at the Presbyterian Hospital in the City of New
York, the only hospital in the community where approximately 80%
of all patients in Northern Manhattan with cerebral infarction are
hospitalized.13 To ensure complete incident stroke enumeration in the
region, cases with International Classification of Diseases, 9th revision
codes 430 to 438, 446, and 447 were also identified through discharge
lists from 14 hospitals outside the immediate region. Extensive
ongoing community-based surveillance for nonhospitalized stroke was
done through random household telephone surveys (Audit and
Survey, Inc) and frequent interval contacts with community physicians, senior citizen centers, visiting nurse services, and other social
and cultural community agencies. Direct community outreach strategies and the local media were also utilized to encourage self-referral
of patients with stroke.
Patients diagnosed with stroke as well as a variety of other
neurological syndromes (eg, aphasia, hemiparesis, weakness, coma,
syncope) were screened by the research assistants, and the case was
discussed with a study neurologist to confirm eligibility. After permission was received from the attending physician, written informed
consent was obtained from the patient or the family. The study was
approved by the institutional review boards at Columbia-Presbyterian
Medical Center and the other primary hospitals.

Selection of Control Subjects

Community control subjects were eligible if they (1) had never been
diagnosed with a stroke, (2) were older than 39 years, and (3) resided
in Northern Manhattan for at least 3 months in a household with a
telephone. Stroke-free subjects were identified by random-digit dialing with dual-frame sampling to identify both published and unpublished telephone numbers.14,15 When a household was contacted, the
research objectives were explained, and a resident older than 39 years
was interviewed briefly to record age, sex, race/ethnicity, and biobehavioral and vascular disease risk factors. The telephone interviews
were conducted by Audits and Surveys, Inc, a survey research firm
that used trained interviewers bilingual in English and Spanish.
The interview data from control-eligible subjects were downloaded
to the NOMASS computer system and sorted by sex and race/
ethnicity. These subjects were recontacted by the NOMASS staff and
invited to participate in the study. Appointments were made for
in-person evaluations at the hospital or home for those who could not
come in person (7% were done at home). Two concurrent control
subjects were randomly selected from among the participants within
the appropriate stratum and matched to each stroke case subject by age
(within 5 years in either direction), sex, and race/ethnicity. Our
telephone survey response rate was 94%, and our in-person participation rate was 70% among the telephone responders. At the time of
the in-person visit, written informed consent was obtained.

Data were collected through interview of the case and control subjects
by trained research assistants, review of the medical records, physical
and neurological examination by the study physicians, in-person
measurements, and fasting blood specimens for lipid and glucose
measurements. Data were obtained directly from study subjects by
means of standardized data collection instruments. When the subject
was unable to answer questions because of death, aphasia, coma,
dementia, or other conditions, a proxy who was knowledgeable about
the patients history was interviewed (data were obtained from the
study subject in 74% of case subjects). Stroke-free control subjects
were interviewed in person and evaluated in the same manner as case
subjects (data were obtained from the study subject in 99% of control
subjects). Case subjects were interviewed as soon as possible after their
stroke, within a median time of 4 days from stroke onset.
All assessments were conducted in English or Spanish depending on
the primary language of the participant. Race/ethnicity was based on
self-identification through a series of interview questions modeled
after the US census. All participants responding affirmatively to being
of Spanish origin or identifying themselves as Hispanic were classified
as such. All participants classifying themselves as white without any
Hispanic origin or black without any Hispanic origin were classified as
white, non-Hispanic, or black, non-Hispanic, respectively.
Subjects were interviewed regarding sociodemographic characteristics, stroke risk factors, and other medical conditions. Standardized
questions were adapted from the Behavioral Risk Factor Surveillance
System by the Centers for Disease Control and Prevention regarding
the following conditions: hypertension, diabetes, hypercholesterolemia, peripheral vascular disease, transient ischemic attack, cigarette
smoking, alcohol use, and cardiac conditions such as myocardial
infarction, coronary artery disease, angina, congestive heart failure,
atrial fibrillation, other arrhythmias, and valvular heart disease.16
Subjects also completed a comprehensive functional status battery,
which included the Barthel ADL, the QWB , and the Geriatric Social
Readjustment Rating scales.1719
Blood pressure was measured with the use of a calibrated standard
aneroid sphygmomanometer (Omron). After the subject had 5 minutes of relative immobility in a sitting position, two blood pressure
measurements separated by 15 minutes were recorded. In subjects
with blood pressure recordings discrepant by .10 mm Hg, a third
measurement was obtained by the study physician and entered into the
database along with the closer blood pressure reading by the research
assistant. Anthropometric measurements of height and weight were
determined by the use of calibrated scales. Blood samples were sent for
complete blood count on admission or enrollment. Fasting glucose
was measured with a Hitachi 747 automated spectrometer (Boehringer). Fasting lipid panels (including total cholesterol, LDL, HDL, and
triglyceride) were measured with a Hitachi 705 automated spectrometer (Boehringer).
Hypertension was defined as a systolic blood pressure recording
$160 mm Hg or a diastolic blood pressure recording $95 mm Hg
(based on the average of the two blood pressure measurements) or the
patients self-report of a history of hypertension or antihypertensive
use. Diabetes mellitus was defined by a fasting glucose .140 mg/dL
(7.7 mmol/L), the patients self-report of such a history, or insulin or
hypoglycemic use. For this analysis, smoking was defined as currently
smoking cigarettes, and heavy alcohol use was defined as current
drinking of $14 drinks per week. BMI was calculated as weight
(kilograms) divided by height (meters) squared, and obesity was
defined as BMI $27.8 for men and $27.3 for women.1

Physical Activity Assessment

A questionnaire adapted from the National Health Interview Survey
of the National Center for Health Statistics was used to measure recent
leisure-time/recreational physical activities.20 This survey form has
been found to be reliable in evaluating elderly subjects.21 The
questionnaire records the frequency and duration of 14 different
recreational activities during the 2-week period before the interview.
In light of the US Surgeon Generals report on physical activity and
health,2,22 activities surveyed in our elderly population were classified
as light-moderate (walking, calisthenics, dancing, golf, bowling,

Downloaded from http://stroke.ahajournals.org/ by guest on March 5, 2016

Physical Activity and Stroke

382

horseback riding, and gardening) and heavy (hiking, tennis, swimming, bicycle riding, jogging, aerobic dancing, and handball or
racquetball or squash). A series of yes/no responses were recorded for
each of the questions, posed as In the last 2 weeks, have you engaged
in
for physical activity? For stroke case subjects,
the question was phrased as In the last 2 weeks prior to your
stroke . . . . Each affirmative response was followed by two other
questions: On average, how many times did you perform this activity
over the last 2 weeks? and On average, how many minutes each
time? From these responses the frequency and duration of each
activity were computed.

TABLE 1.

Demographics and Vascular Risk Factors

Case Subjects
(n5369)

Control Subjects
(n5678)

White

63 (17%)

121 (18%)

Black

106 (29%)

204 (30%)

Hispanic

196 (53%)

349 (51%)

206 (56%)

394 (58%)

Mean6SD

70.0612.4

69.8611.8

Age $65

254 (69%)

472 (70%)

116 (32%)

307 (45%)

295 (80%)

584 (86%)

Hypertension

264 (72%)

393 (58%)

Diabetes

140 (38%)

131 (19%)

Demographics
Race/ethnicity

Sex

Statistical Analyses

Female

The crude frequency and mean duration of any physical activities, as

well as the individual physical activities and their subgroups (lightmoderate and heavy), were examined in the case and control subjects.
Multivariate conditional logistic regression for matched data was used
to calculate the ORs and 95% CIs for physical activity and stroke, with
adjustment for potential stroke risk factors such as hypertension,
diabetes, cardiac disease, smoking, heavy alcohol use, and obesity,
education, and season (winter versus other seasons). The latter was
done to control for any seasonal mismatch between the case and
control subjects, since subjects enrolled in winter may be less physically active. Adjusted analyses were performed overall and stratified by
age (,65 and $65 years), sex, and race/ethnicity. To examine for a
dose-response relationship, physical activity was categorized by level
of intensity into three groups (none, light-moderate, and heavy
activities) and by duration of physical activity (none, ,2 h/wk, 2 to
,5 h/wk, and $5 h/wk). Subjects performing both light-moderate
and heavy activities were classified as heavy. We assessed for a linear
trend in duration as a continuous variable by testing whether the
coefficient of the quadratic term was zero. Interactions between any
physical activity and other stroke risk factors were assessed in the
multivariate model.
Subgroup analyses were performed to evaluate for any confounding
due to recent infection or TIA preceding the stroke. Proxy reliability
of the physical activity questionnaire was assessed among 16 stroke
case subjects and their proxies. After historical information from the
patient was obtained in the usual manner, a surrogate familiar with the
patients medical history and habits, but not aware of his/her responses
to the questionnaire, was interviewed with the same instrument. To
evaluate whether using the proxy variable led to any bias in estimates
of the ORs, a subgroup analysis was done restricting the data to only
nonproxy physical activity data.

Results
From July 1, 1993, to December 31, 1995, 369 case subjects
with first cerebral infarction and 678 age-, sex-, and race/
ethnicity-matched control subjects were enrolled in the casecontrol study. Among the case subjects, 7% died within 30 days
of the stroke, and both fatal and nonfatal strokes were included
in the analyses. Overall, there were 18% whites, 30% blacks,
and 52% Hispanics. Mean age was 69.9612 years, and women
constituted 57%. Hypertension, diabetes, peripheral vascular
disease, smoking, and cardiac disease were more common
among the case subjects, while obesity was more prevalent
among the control subjects. The control subjects were slightly
more educated and were enrolled slightly more frequently
during nonwinter months than the case subjects (Table 1).
Only 16% of the case and control subjects were working .10
h/wk at the time of enrollment. Based on the recruitment
telephone-interview data, our nonresponse analysis showed
that control subjects who participated in person had a similar
frequency of hypertension, diabetes, and cardiac disease but
were more likely to be overweight and smoke cigarettes than
subjects who did not participate.

Age, y

Education
$HS graduate
Season of enrollment
Nonwinter*
Risk factors

PVD

85 (23%)

89 (13%)

Smoking

81 (23%)

117 (18%)

Cardiac disease\

141 (39%)

158 (23%)

Obesity

128 (37%)

286 (42%)

33 (9%)

43 (7%)

Heavy alcohol intake#

HS indicates high school; PVD, peripheral vascular diseases.

*March 21 to December 20.
History of hypertension or systolic blood pressure $160 mm Hg or diastolic
blood pressure $95 mm Hg.
History of diabetes or glucose $140 mg/dL (7.7 mmol/L).
Currently smoking.
\History of myocardial infarction, coronary artery disease, congestive heart
failure, atrial fibrillation, or valvular heart disease.
BMI $27.8 for men and $27.3 for women.
#.2 drinks per day.

Construct validity of our physical activity assessment was

demonstrated within our cohort. Physical activity duration
measurement correlated inversely with BMI among the control subjects (Pearsons r52.125; P,.001) and directly with
ADL scores (Pearsons r5.157; P,.0001) and other activity
questions on the QWB scale (Pearsons r5.182; P,.0001). A
seasonal variation in the performance of physical activity was
observed. Our control subjects were more active during
nonwinter (March 21 to December 20) than winter months
(December 21 to March 20) (P,.0005). The physical activity
assessment was reasonably reliable when obtained from proxies, with a crude concordance rate of 0.69 for engaging in any
physical activity in our proxy reliability substudy.
During the 2 weeks before study enrollment, 49% of case
subjects and 25% of control subjects reported no leisure-time
recreational physical activity (Table 2). Among those who
performed any physical activity, walking was the most frequently reported form of exercise in both case and control
subjects, followed by calisthenics. The mean duration for any
activity was 4.064.4 h/wk for case subjects and 4.764.8 h/wk

Downloaded from http://stroke.ahajournals.org/ by guest on March 5, 2016

Sacco et al
TABLE 2.

383

Intensity and Duration of Physical Activities

Case Subjects

Control Subjects

Hours/Week

Hours/Week

Never

179 (49%)

174 (25%)

Any activities

190 (51%)

4.064.4

504 (75%)

4.764.8

Light

186 (50%)

3.864.4

489 (72%)

4.564.6

174 (47%)

3.663.5

468 (69%)

4.264.4

20 (5%)

1.561.5

106 (16%)

1.561.4
1.761.7

Walking
Calisthenics
Dancing

3 (1%)

5.368.4

13 (2.0%)

Golf

2 (1%)

18.5623.3

4 (1%)

1069.7

Gardening

1 (0.3%)

6 (1%)

2.463.7

Bowling

0 (0%)

2 (0.3%)

1.860.4

Horseback riding

3.360
0

0 (0%)

0 (0%)

16 (4%)

2.261.8

64 (9%)

2.463.4

Bicycle

6 (2%)

2.261.5

17 (3%)

2.563.7

Swimming

0 (0%)

13 (2%)

1.661.3

Hiking

0 (0%)

3 (0.4%)

1.861.0

Tennis

0 (0%)

2 (0.3%)

0.7561.7

Aerobic dance

8 (2%)

2.361.1

Jogging

4 (1%)

0.860.6

Handball

0 (0%)

Moderate/heavy
Moderate

Heavy

for control subjects. As expected in this elderly sample, the

mean duration of the heavy types of physical activity was less
than the light-moderate forms of activity.
Engaging in any physical activity (light-moderate or heavy)
was independently related to a reduced risk of stroke
(OR50.37; 95% CI50.25 to 0.55) after adjustment for
hypertension, diabetes, cardiac disease, peripheral vascular
disease, smoking, obesity, heavy alcohol use, medical reasons
for limited activity, education, and season of enrollment. The
protective effect of exercise for stroke was demonstrated in all
subgroups after stratification by age, sex, and race/ethnicity
(Fig 1). Our smaller white subgroup had the widest CI in the
adjusted multivariate analyses; however, the point estimate of
the OR for physical activity was still in the protective
direction.
To assess for other confounding factors that may have
affected the level of physical activity in our study subjects, we
performed several analyses. The effect of physical activity was
similar in our subgroup analysis restricted to nonproxy data
(OR5.40). After exclusion of case subjects who experienced
TIA during the preceding 30 days, the adjusted OR for
physical activity was comparable to our original model
(OR50.42; 95% CI50.28 to 0.63). Symptoms of infection
were reported in only 9% of our case subjects and 6% of our
control subjects and were not statistically different, and the
association between physical activity and stroke remained even
after adjustment for admission white blood cell count
(OR50.36; 95% CI50.22 to 0.60). The median prestroke or
preenrollment Barthel ADL scores were 100 for both our case
and control subjects, and the mean QWB scores were similar
between case subjects (0.760.14) and control subjects

28 (4%)
9 (1.3%)
0 (0%)

2.163.9
2.262.75
0

(0.760.16) (Table 1). With the use of the Geriatric Social

Readjustment Rating Scale to survey recent changes in the
subjects life events that may affect physical activity, the case
and control groups were identical when we compared stressful physical illness (painful arthritis, eyesight failure, hearing
failure, or feeling of slowing down) and impactful acute loss
events (death of spouse, death of close friend, death of
relative, or loss of job).
When we used the no-exercise group as a reference, heavy
activity (OR50.23; CI50.10 to 0.54) appeared slightly more
protective than light-moderate activity (OR50.39; CI50.26

Figure 1. Association between physical activity and stroke.

Adjusted ORs and 95% CIs from multivariate conditional logistic
regression analysis of the association between ischemic stroke
and physical activity are shown overall and stratified by age,
sex, and race/ethnicity. No physical activity is the reference category within each stratum.

Downloaded from http://stroke.ahajournals.org/ by guest on March 5, 2016

Physical Activity and Stroke

384

Figure 2. Dose-response relationship between physical activity

and stroke. ORs and 95% CIs from multivariate conditional
logistic regression analysis of the association between ischemic
stroke and physical activity are shown stratified by intensity of
activity and duration in hours per week. Mod/Hvy indicates
moderate-heavy activity.

to 0.58), suggesting a dose-response relationship between the

intensity of physical activity and stroke. The same doseresponse relationship was seen for duration of physical activity,
with the greatest protection seen in those performing $5 h/wk
of exercise (Fig 2). Test for linear trend of the dose response for
physical activity duration was significant (P5.006), indicating
that every unit increment in duration of physical activity was
associated with a significant decrement in stroke risk. The
dose-response relationships were also observed for each of the
age, sex, and race/ethnicity strata.
The protective effect of physical activity in our subjects was
not fully explained by elevation in HDL levels. Mean plasma
HDL levels were not related to leisure-time physical activity,
and the independent effect of physical activity remained after
HDL was added to our model in an analysis among the 954
subjects (91% of study sample) with HDL data. Moreover, the
effect of physical activity was not altered by other lipid levels,
including fasting cholesterol, triglyceride, or LDL. Although
the mean diastolic blood pressure level was slightly lower
among physically active compared with inactive control subjects, the protective effects of physical activity were also
independent of hypertension and other stroke risk factors.
Furthermore, the protective effect did not differ by risk factor
strata for hypertension, diabetes, any cardiac disease, or smoking status, and no significant interactions were detected between physical activity and the covariates in the multivariate
analyses.

Discussion
The cardiovascular benefits of physical activity have been
emphasized by numerous organizations, including the Centers
for Disease Control, National Institutes of Health, and the
American Heart Association. Recent guidelines recommended
that individuals perform moderate exercise for approximately
3.5 h/wk. This suggestion arose from accumulating data
regarding the beneficial effects of physical activity in reducing
the incidence of heart disease. The benefits for stroke, however, have not always been consistent across age and sex
subgroups. Furthermore, little is known about physical activity

in different nonwhite race/ethnic groups, in whom the stroke

burden has been reported to be greater.23,24
In our study we detected a beneficial relationship between
leisure-time physical activity and stroke. The protective effects
were found after adjustment for medical and socioeconomic
confounders, season of recruitment, and smoking and alcohol
use. Similar benefits were observed for those younger and
older than 65 years, both men and women, and among whites,
blacks, and Hispanics. The mean duration of leisure-time
physical activity in our population approximated that recommended by the Centers for Disease Control and National
Institutes of Health, with much less participation in heavy
forms of activity. In our relatively older cohort, walking was
the most common form of recreational physical activity reported, which is consistent with observations in other comparably aged populations.21,25 The finding of an overall protective
effect for leisure-time physical activity suggests that even
light-moderate forms of activity may confer some benefits for
the elderly. Our results are particularly relevant in the more
sedentary and older population who have a greater risk of
stroke and more prevalent cardiovascular risk factors and
whose physical activity pattern may comprise mainly light to
moderate activity of shorter duration.
The protective effects of physical activity for stroke risk
noted in other studies were limited to certain age, sex, and risk
factor subgroups. The Honolulu Heart Program, which investigated older middle-aged men of Japanese ancestry, showed a
protective effect of habitual physical activity from thromboembolic stroke only among their nonsmoking group.5 The
Framingham Study demonstrated the benefits of combined
leisure and work physical activities for men but not for
women.6 In the Oslo Study of men aged 40 to 49 years,
increased leisure physical activity was related to a reduced
stroke incidence.7 For women aged 40 to 65 years, the Nurses
Health Study showed an inverse association between level of
physical activity and the incidence of any stroke.8 For white,
lower and middle class, urban women participating in the
Copenhagen City Heart Study, lack of physical activity had an
effect similar to that of cigarette smoking, with a relative risk of
stroke of 1.4.9 In the National Health and Nutrition Examination Survey I follow-up study, low level of recreational or
nonrecreational activities was slightly associated with an increased risk of stroke for both men and women and among
blacks.10
The dose-response relationship between level and duration
of leisure physical activity and protective effects from stroke
was consistent in our study population and among each of the
age, sex, and race/ethnic subgroups. Although performance of
leisure-time physical activities conferred protection, more
vigorous (heavy) forms of physical activities provided additional benefits compared with light-moderate activities. The
same was true for duration of physical activity. Although the
odds of stroke were reduced in the group performing ,2
h/wk of physical activity, additional protection was observed
with increasing duration of exercise. The beneficial effects of
even small amounts of leisure-time physical activity in the
older population have not been reported previously, and the
dose-response relationship is encouraging for those individuals
who can safely increase their level of physical activity.

Downloaded from http://stroke.ahajournals.org/ by guest on March 5, 2016

Sacco et al
The dose-response relationship between increasing amounts
of physical activity and the reduction in the risk of stroke has
not always been demonstrated. Wannamethee and Shaper11
found an inverse relationship between physical activity and risk
of stroke, but the benefit of vigorous physical activity for stroke
was offset by an increased risk of heart attack. In Framingham
among the older cohort, the strongest protective effect was
detected in the medium tertile physical activity subgroup, with
no additional benefit gained from higher levels of physical
activity.6 Among subjects in a case-control study in West
Birmingham, UK, who were free of cardiac disease, peripheral
vascular disease, and poor health, recent vigorous exercise was
no more protective than walking.12
The protective effect of physical activity may be partly
mediated through its role in controlling various known risk
factors for stroke. Exercise has been shown to lower blood
pressure in certain groups.26 28 It is also associated with a lower
incidence of cardiovascular disease,29 32 improved diabetes
control,33,34 better dietary habits, and lower body weight.35,36
Cigarette smokers are less likely to participate in exercise
programs.37 Furthermore, people with a higher level of education participate in more leisure-time activity.38 The finding
of an independent effect of physical activity after adjustment
for these factors suggests that mechanisms other than the
control of risk factors may be responsible for the protective
effect of physical activity for stroke.
Other biological mechanisms are also associated with physical activity, including reductions in plasma fibrinogen and
platelet activity and elevations in plasma tissue plasminogen
activator activity and HDL concentrations.39 42 In a preliminary analysis from our multiethnic population, we found HDL
concentration to be inversely related to the risk of stroke.43
However, we found no significant differences in plasma HDL
levels among our inactive, light-moderately active, and heavily
active groups in either our case or control subjects, and the
independent effect of physical activity remained after HDL and
other lipid levels were added to our model in a subgroup
analysis.
There are some limitations to our study. Our design was a
case-control study and not a prospective cohort study. However, our subjects were population based, and selection bias is
less likely than in hospital-based case-control studies. The OR
more closely approximates a relative risk in population-based
case-control studies, since the case and control subjects are
derived from the same community and provide a close
approximation of the prevalence of exposure in those with and
without incident disease.44 Recall bias, a frequent source of bias
in case-control studies, seems less likely since the case and
control subjects were not aware of the hypotheses being tested.
In addition, recall physical activity surveys like the instrument
we used are less likely to influence a subjects typical behavior
than activity diaries or logs.22
There may have been some unmeasured confounders that
led to a reduction in recent physical activity among the stroke
case subjects or influenced the selection of control subjects
who were more physically active. This could have resulted in
an overestimation of the OR. However, we made a concerted
effort to adjust for such confounders in our analyses and
performed nonresponse analyses which showed that the con-

385

trol subjects who participated in the study had risk factor

profiles similar to those who did not participate in person,
except for greater BMI and smoking among the participants.
Both of these conditions would lead to bias of the OR toward
the null value, since such conditions would decrease the
likelihood of engaging in physical activity. Some have argued
that recent infections are more frequent in stroke patients,45,46
which could also confound the physical activity association.
We found no direct association between clinical symptoms of
infection and physical activity among our case subjects, and our
findings were the same even after we controlled for admission
white blood cell count. There was likewise no relationship
between occurrence of TIA within 30 days before stroke and
physical activity to suggest that TIA differentially reduced the
level of physical activity in our case subjects. Furthermore, the
prestroke or preenrollment median ADL scores, mean QWB
scores, and Geriatric Social Readjustment Scale assessments
were similar among our case and control subjects.
Physical activity questionnaires have differed in the time
frame of assessment, and some have included occupational as
well as leisure or recreational activities.4755 The majority of our
subjects were elderly, with only 16% working at the time of
enrollment. Among those who were working, the job types
infrequently involved strenuous physical activities. Our physical activity assessments were not designed to evaluate lifelong
exercise practices, exercise patterns at younger ages, physical
conditioning, or quantitative estimations of energy expenditure. Our instrument measured leisure physical activity during
the preceding 2 weeks. The 2-week period of activity recall
was deemed reliable, short enough to allow reasonably accurate recall in an elderly sample, yet long enough to represent
the usual activity patterns of most people.55 The inverse
correlation of physical activity with BMI in our study could
argue for our measurements being correlated with more
sustained patterns of physical activity. It is not clear from our
study whether the pattern of physical activity in the elderly
correlated with a life-long pattern of physical activity that
conferred the protective effect or whether recent physical
activity, regardless of life-long patterns, can confer some
benefits.
Our study demonstrates the importance of leisure-time
physical activity in the prevention of stroke among the elderly,
in men and women, and in all race/ethnic groups. While
currently available methods of physical activity measurement
may not be ideal, they have provided a means of demonstrating
important benefits of physical activity.22 The benefits in stroke
are apparent even for light-moderate activities, such as walking, and the data support additional benefits to be gained from
increasing the level and duration of ones recreational activity.
The potential undesirable consequences of extreme exercise,
such as alterations in hormonal levels in women, musculoskeletal injuries, loss of bone mineral content, and risk of acute
myocardial infarction,56 62 should be considered when advising
individual patients, particularly the elderly, to increase their
physical activity. Nevertheless, physical activity is a modifiable
behavior that requires greater emphasis in stroke prevention
campaigns. As long as the adverse effects of high-intensity
physical activity are minimized, leisure-time physical activity
could translate into a cost-effective means of decreasing the

Downloaded from http://stroke.ahajournals.org/ by guest on March 5, 2016

Physical Activity and Stroke

386

public health burden of stroke and other cardiovascular diseases

among our rapidly aging population.

21.

Acknowledgments
This study was supported by grants from the National Institute of
Neurological Disorders and Stroke (R01 NS 27517, R01 NS 29993,
and T32 NS 07153) and the General Clinical Research Center (2 M01
RR00645). We acknowledge the support of Dr J.P. Mohr, Director
of Cerebrovascular Research; the statistical assistance of Qiong Gu;
the processing of the lipids by Drs Henry Ginsberg and Lars F.
Berglund in the Division of Preventive Medicine and Nutrition and
the Irving Center for Clinical Research; and the help provided by Drs
J. Kirk Roberts and Mitchell S. Elkind, fellows in stroke
and neuroepidemiology.

References
1. Department of Health and Human Services. Healthy People 2000: National
Health Promotion and Disease Prevention Objectives. Washington, DC:
Department of Health and Human Services; 1991. US Department of
Health and Human Services publication PHS 9150213.
2. Pate RR, Pratt M, Blair SN, Haskell WL, Macera CA, Bouchard C,
Buchner D, Ettinger W, Heath GW, King AC, Kriska A, Leon AS, Marcus
BH, Morris J, Paffenbarger RS, Patrick K, Pollock ML, Rippe JM, Sallis J,
Wilmore JH. Physical activity and public health: a recommendation from
the Centers for Disease Control and Prevention and the American College
of Sports Medicine. JAMA. 1995;273:402 407.
3. NIH Consensus Development Panel on Physical Activity and Cardiovascular Health. Physical activity and cardiovascular health. JAMA. 1996;276:
241246.
4. Fletcher GF. Exercise in the prevention of stroke. Health Reports. 1994;6:
106 110.
5. Abbott RD, Rodriguez BL, Burchfiel CM, Curb JD. Physical activity in
older middle-aged men and reduced risk of stroke: the Honolulu Heart
Program. Am J Epidemiol. 1994;139:881 893.
6. Kiely DK, Wolf PA, Cupples LA, Beiser AS, Kannel WB. Physical activity
and stroke risk: the Framingham Study. Am J Epidemiol. 1994;140:
608 620.
7. Haheim LL, Holme I, Hjermann I, Leren P. Risk factors of stroke
incidence and mortality: a 12-year follow-up of the Oslo Study. Stroke.
1993;24:1484 1489.
8. Manson JE, Stampfer MJ, Willett WC, Colditz GA, Speizer FE,
Hennekens CH. Physical activity and incidence of coronary heart disease
and stroke in women. Circulation. 1995;91(suppl):5.
9. Lindenstrom E, Boysen G, Nyboe J. Lifestyle factors and risk of cerebrovascular disease in women: the Copenhagen City Heart Study. Stroke.
1993;24:1468 1472.
10. Gillum RF, Mussolino ME, Ingram DD. Physical activity and stroke
incidence in women and men: the NHANES I Epidemiologic Follow-up
Study. Am J Epidemiol. 1996;143:860 869.
11. Wannamethee G, Shaper AG. Physical activity and stroke in British middle
aged men. BMJ. 1992;304:597 601.
12. Shinton R, Sagar G. Lifelong exercise and stroke. BMJ. 1993;307:231234.
13. Statewide Planning and Research Cooperative System. New York State
Department of Health; 1990.
14. Olson SH, Kelsey JL, Pearson TA, Levin B. Evaluation of random digit
dialing as a method of control selection in case-control studies. Am J
Epidemiol. 1992;135:210 222.
15. Thornberry OT, Massey JT. Coverage and response in random digit
dialing national surveys. In: Proceedings of the Section on Survey Research
Methods, American Statistical Association. Alexandria, Va: American Statistical
Association; 1983:654 659.
16. Gentry EM, Kalsbeek WD, Hooelin GC, Jones JT, Gaines KL, Forman
MR, Marks JS, Trowbridge FL. The Behavioral Risk Factor Surveys, II:
design, methods and estimates from combined state data. Am J Prev Med.
1985;1:9 14.
17. Mahoney FI, Barthel DW. Functional evaluation: the Barthel Index. Md
State Med J. 1965;14:61 65.
18. Kaplan RM, Bush JW. Health-Related Quality Of Life Measurement For
Evaluation Research, and Policy Analysis. Health Psychol. 1982;1:61 80.
19. Amster LE, Krauss HH. The relationship between life crises and mental
deterioration in old age. Int J Aging Hum Dev. 1974;5:5155.
20. Moss AJ, Parsons VL. Current estimates from the National Health
Interview Survey, United States, 1985. Washington, DC: National Center

22.

23.
24.

25.

26.

27.
28.

29.

30.

31.

32.
33.

34.

35.
36.

37.

38.

39.

40.

41.

42.

43.

for Health Statistics. US Department of Health and Human Services publication PHS 86 1588. Vital Health Stat 10 (160); 1986.
McPhillips JB, Pellettera KM, Barrett-Connor E, Wingard DL, Criqui
MH. Exercise patterns in a population of older adults. Am J Prev Med.
1989;2:6572.
Department of Health and Human Services. Physical Activity and Health: A
Report of the Surgeon General. Atlanta, Ga: Department of Health and
Human Services, Centers for Disease Control and Prevention, National
Center for Chronic Disease Prevention and Health Promotion; 1996.
Sacco RL, Hauser WA, Mohr JP. Hospitalized stroke in blacks and Hispanics in northern Manhattan. Stroke. 1991;22:14911496.
Sacco RL, Kargman DE, Gu Q, Zamanillo MC. Race-ethnicity and
determinants of intracranial atherosclerotic cerebral infarction: the
Northern Manhattan Stroke Study. Stroke. 1995;26:14 20.
Frandin K, Grimby G, Mellstrom D, Svanborg A. Walking habits and
health-related factors in a 70-year-old population. Gerontology. 1991;37:
281288.
Kokkinos PF, Narayan P, Colleran JA, Pittaras A, Notargiacomo A, Reda
D, Papademetriou V. Effects of regular exercise on blood pressure and left
ventricular hypertrophy in African-American men with severe hypertension. N Engl J Med. 1995;333:14621467.
Martin JE, Dubbert PM, Cushman WC. Controlled trial of aerobic
exercise in hypertension. Circulation. 1990;81:1560 1567.
Hagberg JM, Montain SJ, Martin WH III, Ehsani M. Effects of exercise
training in 60- to 69-year-old persons with essential hypertension. Am J
Cardiol. 1989;64:348 353.
Blair SN, Kampert JB, Kohl HW III, Barlow CE, Macera CA, Pafferbarger
RS Jr, Gibbons LW. Influences of cardiorespiratory fitness and other
precursors on cardiovascular disease and all-cause mortality in men and
women. JAMA. 1996;276:205210.
DAvanzo B, Santoro L, La Vecchia C, Maggioni A, Nobili A, Iacuitti G,
Franceschi S, for the GISSI-EFRIM Investigators. Physical activity and the
risk of acute myocardial infarction. Ann Epidemiol. 1993;3:645 651.
Lakka TA, Venalainen JM, Rauramaa R, Salonen R, Tuomilehto J,
Salonen JT. Relation of leisure-time physical activity and cardiorespiratory
fitness to the risk of acute myocardial infarction in men. N Engl J Med.
1994;330:1549 1554.
Berlin JA, Colditz GA. A meta-analysis of physical activity in the prevention of coronary heart disease. Am J Epidemiol. 1990;132:612 628.
Manson JE, Rimm EB, Stampfer MJ, Colditz GA, Willett WC, Krolewski
AS, Rosner B, Hennekens CH, Speizer FE. A prospective study of physical
activity and incidence of non-insulin-dependent diabetes mellitus in
women. Lancet. 1991;338:774 778.
Kemmer FW, Berger M. Exercise and diabetes mellitus: physical activity as
part of daily life and its role in the treatment of diabetic patients. Int J Sports
Med. 1983;4:77 88.
Garrow JS. Effect of exercise on obesity. Acta Med Scand. 1987;711(suppl):
6773.
Pavlou KN, Krey S, Steffee WP. Exercise as an adjunct to weight loss and
maintenance in moderately obese subjects. Am J Clin Nutr. 1989;49:
11151123.
Dishman RK, Sallis JF. Determinants and interventions for physical activity
and exercise. In: Bouchard C, Shephard RJ, Stephens T, eds. Physical
Activity, Fitness, and Health. Champaign, Ill: Human Kinetics Publishers;
1994:214 238.
Caspersen CJ, Christenson GM, Pollard RA. The status of the 1990
Physical Fitness Objectives: evidence from NHIS 85. Public Health Rep.
1986;101:587592.
Lakka TA, Salonen JT. Moderate to high intensity conditioning leisure
time physical activity and high cardiorespiratory fitness are associated with
reduced plasma fibrinogen in eastern Finnish men. J Clin Epidemiol. 1993;
46:1119 1127.
Wang JS, Jen CJ, Chen HI. Effects of exercise training and deconditioning
on platelet function in men. Arterioscler Thromb Vasc Biol. 1995;15:
1668 1674.
Rangemark C, Hedner JA, Carlson JT, Gleerup G, Winther K. Platelet
function and fibrinolytic activity in hypertensive and normotensive sleep
apnea patients. Sleep. 1995;18:188 194.
Williams PT. High-density lipoprotein cholesterol and other risk factors for
coronary heart disease in female runners. N Engl J Med. 1996;334:
1298 1303.
Kargman DE, Berglund LF, Gu Q, Boden-Albala B, Gan R, Roberts K,
Hauser WA, Shea SC, Paik M, Ginsberg H, Sacco RL. High density
lipoprotein: a potentially modifiable stroke risk factor: the Northern Manhattan Stroke Study. Neuroepidemiology. 1996;15:282. Abstract.

Downloaded from http://stroke.ahajournals.org/ by guest on March 5, 2016

Sacco et al
44. Rothman KJ. Modern Epidemiology. Boston, Mass: Little, Brown & Co;
1986:63 64.
45. Grau AJ, Buggle F, Heindl S, Steichen-Wiehn C, Banerjee T, Maiwald M,
Rohlfs M, Suhr H, Fiehn W, Becher H. Recent infection as a risk factor
for cerebrovascular ischemia. Stroke. 1995;26:373379.
46. Bova I, Bornstein NM. Acute infection as an independent risk factor for
ischemic stroke. Stroke. 1996;27:187. Abstract.
47. Washburn RA, Montoye HJ. The assessment of physical activity by questionnaire. Am J Epidemiol. 1986;123:563576.
48. Osler M, de Groot LC, Enzi G, for the Euronut SENECA Investigators.
Life-style: physical activities and activities of daily living. Eur J Clin Nutr.
1991;45(suppl 3):139 151.
49. Chasan-Taber S, Rimm EB, Stampfer MJ, Spiegelman D, Colditz GA,
Giovannucci E, Ascherio A, Willett WC. Reproducibility and validity of
a self-administered physical activity questionnaire for male health professionals. Epidemiology. 1996;7:81 86.
50. Wolf AM, Hunter DJ, Colditz GA, Manson JE, Stampfer MJ, Corsano KA,
Rosner R, Kriska A, Willet WC. Reproducibility and validity of a selfadministered physical activity questionnaire. Int J Epidemiol. 1994;23:
991999.
51. Baecke JAH, Burema J, Frijters ER. A short questionnaire for the measurement of habitual physical activity in epidemiological studies. Am J Clin
Nutr. 1982;36:936 942.
52. Paffenbarger RS Jr, Blair SN, Lee IM, Hyde RT. Measurement of physical
activity to assess health effects in free-living populations. Med Sci Sports
Exerc. 1993;25:60 70.
53. Ainsworth BE, Leon AS, Richardson MT, Jacobs DR, Paffenbarger RS Jr.
Accuracy of the College Alumnus Physical Activity Questionnaire. J Clin
Epidemiol. 1993;46:14031411.

387

54. Paffenbarger RS Jr, Laughlin ME, Gima AS, Black RA. Work activity of
longshoremen as related to death from coronary heart disease and stroke.
N Engl J Med. 1970;282:1109 1114.
55. Lamb KL, Brodie DA. Leisure-time physical activity as an estimate of
physical fitness: a validation study. J Clin Epidemiol. 1991;44:4152.
56. Shangold MM. Exercise and the adult female: hormonal and endocrine
effects. Exerc Sport Sci Rev. 1984;12:553579.
57. Klesges RC, Ward KD, Shelton ML, Applegate WB, Cantler ED, Palmieri
GM, Harmon K, Davis J. Changes in bone mineral content in male
athletes: mechanisms of action and intervention effects. JAMA. 1996;276:
226 230.
58. Rencken ML, Chesnut CH III, Drinkwater BL. Bone density at multiple
skeletal sites in amenorrheic athletes. JAMA. 1996;276:238 240.
59. Fiatarone MA, ONeill EF, Ryan ND, Clements KM, Solares GR, Nelson
ME, Roberts SB, Kehayias JJ, Lipsitz LA, Evans WJ. Exercise training and
nutritional supplementation for physical frailty in very elderly people.
N Engl J Med. 1994;330:1769 1775.
60. Maron BJ, Shirani J, Poliac LC, Mathenge R, Roberts WC, Mueller FO.
Sudden death in young competitive athletes: clinical, demographic, and
pathological profiles. JAMA. 1996;276:199 204.
61. Requa RK. The scope of the problem: the impact of sports-related injuries.
In: Proceedings of the Conference on Sports Injuries in Youth: Surveillance
Strategies. National Institutes of Health, The National Institute of Arthritis
and Musculoskeletal and Skin Diseases. Bethesda, Md: National Institutes
of Health; 1992. US National Institutes of Health publication 933444.
62. Mittleman MA, Maclure M, Tofler GH, Sherwood JB, Goldberg RJ,
Muller JE. Triggering of acute myocardial infarction by heavy physical
exertion. N Engl J Med. 1993;329:16771683.

Downloaded from http://stroke.ahajournals.org/ by guest on March 5, 2016

Leisure-Time Physical Activity and Ischemic Stroke Risk: The Northern Manhattan
Stroke Study
Ralph L. Sacco, Robert Gan, Bernadette Boden-Albala, I-Feng Lin, Douglas E. Kargman, W.
Allen Hauser, Steven Shea and Myunghee C. Paik
Stroke. 1998;29:380-387
doi: 10.1161/01.STR.29.2.380
Stroke is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright 1998 American Heart Association, Inc. All rights reserved.
Print ISSN: 0039-2499. Online ISSN: 1524-4628

The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://stroke.ahajournals.org/content/29/2/380

Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Stroke can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office.
Once the online version of the published article for which permission is being requested is located, click
Request Permissions in the middle column of the Web page under Services. Further information about this
process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Stroke is online at:
http://stroke.ahajournals.org//subscriptions/

Downloaded from http://stroke.ahajournals.org/ by guest on March 5, 2016