Self-Reported Stress and Risk of Stroke
The Copenhagen City Heart Study
Thomas Truelsen, MD, PhD; Naja Nielsen, BMsc; Gudrun Boysen, MD, DMSc; Morten Grønbæk, MD, DMSc
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
Background and Purpose—Lay people often mention stress as one of the most important risk factors for stroke. Stress
might trigger a cerebrovascular event directly or could be associated with higher levels of blood pressure or an
unfavorable lifestyle. To examine these possibilities, we analyzed the association between self-reported stress frequency
and intensity and risk of stroke.
Methods—Data from the second examination, 1981 to 1983, of participants in the Copenhagen City Heart Study were
analyzed with Cox regression after a mean of 13 years of follow-up. A total of 5604 men and 6970 women were
included, and 929 first-ever strokes occurred, of which 207 (22%) were fatal within 28 days after onset of symptoms.
The stress frequency categories were never/hardly ever, monthly, weekly, or daily. The stress intensity categories were
never/hardly ever, light, moderate, or high.
Results—Subjects with high stress intensity had almost a doubled risk of fatal stroke compared with subjects who were
not stressed (relative risk [RR], 1.89; 95% CI, 1.11 to 3.21). Weekly stress was associated with an RR of 1.49 (95% CI,
1.00 to 2.23). There was no significant effect of stress in analyses of nonfatal strokes. Subjects who reported to be
stressed often were more likely to have an adverse risk factor profile.
Conclusions—Self-reported high stress intensity and weekly stress were associated with a higher risk of fatal stroke
compared with no stress. However, there were no significant trends, and the present data do not provide strong evidence
that self-reported stress is an independent risk factor for stroke. (Stroke. 2003;34:856-862.)
Key Words: Denmark 䡲 epidemiology 䡲 stress 䡲 stroke
I
the effect of stress reduction indicated that it could decrease
carotid atherosclerosis in hypertensive blacks.17 These results
suggest that stress might be a potential modifiable risk factor
for stroke.
The examination of the relationship between stress and risk
of stroke is often hampered by relatively small sample sizes
and the lack of consensus on how stress should be defined or
measured,18 and some of the studies have presented results
only for middle-aged men.6,8 In the present study, data from
men and women participating in the Copenhagen City Heart
Study (CCHS) on the intensity and frequency of self-reported
stress have been analyzed for all strokes, nonfatal stroke, and
fatal stroke. As part of the analyses, we have compared the
risk factor profiles of the different stress groups. Furthermore,
SES is a well-established risk factor for stroke.19 –22 Previous
studies have indicated that low SES may enhance the effect of
stress16,23; therefore, we have also examined the possibility of
an interaction between SES and stress level.
n population surveys on the knowledge of stroke risk
factors, stress is one of the most frequently mentioned
factors, often before smoking and hypertension.1–5 The existing literature on the association between stress and risk of
stroke is inconclusive, with some studies showing that there is
a relation6 – 8 and others indicating that there is none.9 There
are different possible biological mechanisms by which psychological stress may increase risk of stroke. One is related to
behavior in which active coping has been related to increased
catecholamine release and sympathetic activation, which may
either directly or indirectly affect the vascular system, eg, via
an effect on the blood pressure. Previous studies have
addressed the association between stress and changes in
blood pressure levels.10 –13 Another possibility for an association between stress and risk of stroke might be related to
behavioral differences such as smoking, physical activity, and
alcohol consumption or to socioeconomic status (SES).14
Stressed subjects may be more likely than nonstressed subjects to have an adverse risk factor profile, which could
explain an increased risk of stroke. Psychological stress has
also been associated with the progression of atherosclerotic
changes of the carotid artery,15,16 and preliminary findings on
Subjects and Methods
Population
The CCHS is a prospective observational study that was initiated in
1976 when 19 698 subjects living in Østerbro and Nørrebro in
Received August 9, 2002; final revision received September 16, 2002; accepted October 22, 2002.
From the Institute of Preventive Medicine, Kommune Hospitalet (T.T.); Centre for Alcohol Research, National Institute of Public Health (N.N., M.G.);
and Department of Neurology, Bispebjerg Hospital (G.B.), Copenhagen, Denmark.
Correspondence to Thomas Truelsen, Institute of Preventive Medicine, Kommunehospitalet, Øster Farimagsgade 5, 1399 Copenhagen K, Denmark.
E-mail [email protected]
© 2003 American Heart Association, Inc.
Stroke is available at http://www.strokeaha.org
DOI: 10.1161/01.STR.0000062345.80774.40
856
Truelsen et al
Copenhagen were invited to participate in the first study examination. In 1981 to 1983 and 1991 to 1994, additional study examinations were done. The response rates at the 3 examinations were 70%,
67%, and 60%, respectively. Fewer than 2% of the population was of
nonwhite origin. In the present study, data from participants at the
second examination were used with follow-up until 1997. There
were a total of 12 698 eligible subjects 20 to 98 years of age. In total,
70 persons (0.5%) were lost to follow-up; of these, 69 were due to
immigration, and there was no information for 1 person. A detailed
description of the CCHS has previously been published.24
Self-Reported Stress
At the second study examination, participants were asked about their
levels of stress in terms of intensity and frequency. In the questionnaire, stress was exemplified as the sensation of tension, nervousness, impatience, anxiety, or sleeplessness. Participants were asked
to report their stress intensity given the following 4 possibilities:
never/hardly ever stressed or, if stressed, light, moderate, or high
intensity. The participants were also asked to report their stress
frequency given the following 4 possibilities: never/hardly ever
stressed and, if they were stressed, monthly, weekly, or daily.
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
Socioeconomic Status
The number of years of education (⬍8, 8 to 11, and ⱖ12 years) was
used as a measure of SES. In preliminary analyses, we found that
years of education provided a better fit of the model than income and
that simultaneous adjustment for both education and income did not
improve the statistical model significantly. Therefore, we included
only years of education in the final model.
Covariates
The following variables were assumed to confound the analyses: sex;
tobacco smoking (never smoker, ex-smoker, or smoker of 1 to 14, 15
to 24, or ⬎24 g/d); body mass index (⬍20, 20 to 24, 25 to 29, or ⱖ30
kg/m2); physical activity in leisure time (physically passive or light
physical activity ⬍2 h/wk, light physical activity 2 to 4 h/wk, light
physical activity ⬎4 h/wk, more exhausting physical activity 2 to 4
h/wk, and exhausting physical activity ⬎4 h/wk); systolic blood
pressure (continuous); antihypertensive treatment (yes/no); weekly
alcohol intake exceeding the Danish National guidelines (a maximum of 14 U/wk for women and 21 U/wk for men); forced
expiratory volume in 1 second (FEV1) in percentage of expected
according to age, sex, and height (continuous)25; history of a
myocardial infarction before the second health examination (yes/no);
and diabetes mellitus (yes/no).
Outcome
Identification of possible stroke events was obtained through linkage
of a personal identification number to 2 national registers (the
National Patient Register and Cause of Death Register),26 which
contain complete data on all hospital admissions in Denmark and
causes of death for all deceased subjects who resided in Denmark.
Both registers have consistently used the International Classification
of Disease (ICD) codes. Patient files and/or discharge letters covering admissions for participants who were registered with cerebrovascular disease (ICD, 9th revision, codes 430 through 438; ICD,
10th revision, codes I60 through I69 and G45) were retrieved, and
stroke was diagnosed according to the World Health Organization
definition: sudden onset of focal (or at times global) neurological
symptoms lasting ⬎24 hours or leading to death of presumably
vascular origin.27 Type of stroke was based on available information
such as neuroimaging results, lumbar puncture, and necropsies. In
addition, at each of the 3 CCHS examinations in 1976 to 1978, 1981
to 1983, and 1991 to 1994, participants were asked if they had
suffered a stroke. If they answered positively, they were examined
further, and information was collected from the general practitioner.
Statistical Analysis
Subjects with stroke before the start of the study were excluded
(n⫽124). Cox regression analyses, with age as the underlying time
Self-Reported Stress and Risk of Stroke
857
axis, were used to estimate the association between self-reported
stress intensity and frequency and risk of all, fatal, and nonfatal
stroke. Subjects who never/hardly ever were stressed were used as
the referent category. The initial analyses included adjustment for
age and sex; subsequently, tobacco smoking, body mass index,
physical activity in leisure time, systolic blood pressure, antihypertensive treatment, and weekly alcohol intake were entered into the
model. In final analyses, further adjustment was done, including
FEV1, history of myocardial infarction, and diabetes mellitus. Fatal
stroke refers to events in which the patients died within the first 28
days after onset of stroke symptoms. Analyses were made with the
SAS/STAT statistical software, version 8.2.28 Comparisons of categorical variables were done by use of the ␹2 test, and continuous
variables were compared by use of Wilcoxon’s rank-sum test.
Fisher’s exact test was used to compare stroke type and stress level.
Results
During follow-up, there was a total of 929 first-ever strokes:
456 (49%) in women and 473 (51%) in men. Of these, 207
(22%) were fatal. There were 396 patients (43%) with
ischemic stroke, 67 (7%) with intracerebral hemorrhage, and
34 (4%) with subarachnoid hemorrhage; in 432 (47%), the
type of stroke was unknown.
Self-Reported Stress and Baseline Characteristics
There were significant differences in risk factor profiles
between stress intensity and frequency groups (Table 1). The
proportion of women increased in the highly stressed groups.
Subjects in the more stressed groups were more likely to
smoke, to have a low income, to be less physically active, to
have a higher intake of alcohol, and to be on antihypertensive
treatment compared with the less stressed groups. Although
the differences between groups with regard to body mass
index and mean systolic blood pressure were statistically
significant, there was no clear trend. The proportion of
subjects reporting to have diabetes mellitus did not differ
significantly. There was no significant difference between
type of stroke and level of stress intensity (P⫽0.33) or
frequency (P⫽0.49).
Self-Reported Stress and Risk of Stroke
In analyses of self-reported stress frequency and intensity
adjusted for age and sex, there was a tendency toward higher
estimates of relative risk (RR) of stroke in the more stressed
groups (Table 2). There was a significantly increased risk of
stroke among subjects reporting daily stress. For the remaining stress categories, there were no significant associations.
Further adjustment for confounding variables attenuated the
estimates, especially in the high stress groups, and none of the
results were statistically significant. Additional adjustment
for history of myocardial infarction, FEV1, and diabetes
mellitus had almost no effect on the estimates and did not
change the overall results.
Self-Reported Stress and Risk of Nonfatal Stroke
In the present study, 78% of the stroke patients survived the
initial 28 days after onset of stroke symptoms. Thus, they
constituted the majority of all stroke events, and the results of
analyses between self-reported stress intensity and frequency
and risk of nonfatal stroke resembled those of all strokes
combined (Table 3). In analyses adjusted for age and sex, the
RR estimates increased with increasing stress level and were
858
TABLE 1.
Stroke
April 2003
Baseline Characteristics Between Self-Reported Stress and Covariables
Stress Intensity
Stress Frequency
P
Never/Hardly
ever
Monthly
Weekly
Daily
48
28
16
9
58 (12)
53 (13)
54 (12)
58 (11)
None
Light
Moderate
High
Study sample, %
39
36
19
6
Mean (SD) age, y
57 (12)
54 (13)
55 (12)
58 (11)
Men, %
53
40
39
29
⬍0.0001
51
40
39
34
⬍0.0001
Current smoking, %
56
57
62
68
⬍0.0001
56
58
61
64
⬍0.0001
⬍0.0001
P
⬍0.0001
Education ⬍8 y, %
48
42
45
57
⬍0.0001
48
40
43
54
⬍0.0001
Low income, %
35
27
35
52
⬍0.0001
34
24
31
51
⬍0.0001
16
15
19
35
⬍0.0001
15
14
19
33
⬍0.0001
Mean (SD) alcohol intake, U/wk
Physical activity in leisure time ⬍4 h/wk, %
8 (12)
8 (11)
9 (14)
9 (20)
⬍0.0001
8 (12)
8 (12)
9 (13)
10 (20)
⬍0.0001
Antihypertensive treatment, %
15
17
20
26
⬍0.0001
15
16
21
26
⬍0.0001
3
2
2
3
0.53
3
2
3
2
0.31
Diabetes mellitus, %
2
Mean (SD) body mass index, kg/m
26 (4)
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
Mean (SD) systolic blood pressure, mm Hg
25 (4)
25 (5)
⬍0.0001
26 (4)
25 (4)
25 (4)
25 (5)
⬍0.0001
139 (21)
140 (23)
⬍0.0001
143 (22)
138 (21)
139 (22)
140 (21)
⬍0.0001
25 (4)
143 (22) 139 (21)
highest in subjects reporting a daily level of stress, but none
of the results were statistically significant. Further adjustment
for confounding variables, including chronic diseases, attenuated the estimates and had no effect on the overall results.
Self-Reported Stress and Risk of Fatal Stroke
There were 207 patients (22%) with a first-ever stroke who
died within the first 28 days after onset of stroke symptoms.
We reexamined the data with regard to the association with
self-reported intensity and frequency of stress and risk of
having a fatal stroke (Table 4). In analyses adjusted for age
and sex only, the risk of fatal stroke was higher for those with
a high level of stress (RR, 1.79; 95% CI, 1.07 to 2.99)
compared with subjects who reported never or hardly ever
having stress. Only weekly stress was significantly associated
with a higher risk of fatal stroke compared with the finding in
subjects who reported never or hardly ever being stressed. It
should be noted that the estimate for daily stress indicated
that there might be a higher risk of stroke, but there were only
TABLE 2.
a few events, and the nonsignificant result may be due to low
statistical power. In analyses adjusted for confounding variables, there were few changes in the estimates, and the risk of
fatal stroke remained statistically significant for a high stress
level and for a weekly feeling of stress, also when controlling
for chronic diseases.
Interactions and Trend Analyses
There were no significant interactions between sex and stress
intensity for all strokes (P⫽0.60), fatal stroke (P⫽0.61), or
nonfatal stroke (P⫽0.80). Also, there were no statistically
significant interactions between sex and stress frequency for
all strokes (P⫽0.71), fatal stroke (P⫽0.69), or nonfatal stroke
(P⫽0.52).
We also tested for possible interaction between the 2
measures of stress and level of education in analyses controlled for sex and risk factors. The tests for interaction
between frequency of stress and educational level were
significant for all stroke (P⫽0.02) and nonfatal stroke
Stress Intensity and Frequency and Risk of Stroke
RR (95% CI)
Strokes, n
Age and Sex
Age, Sex, and Risk
Factors*
Age, Sex, Risk Factors,
and Chronic Disease†
Stress intensity
None (n⫽4931)
408
1.00 (referent)
1.00 (referent)
1.00 (referent)
Light (n⫽4521)
288
0.97 (0.83–1.13)
0.96 (0.82–1.12)
0.96 (0.82–1.13)
Moderate (n⫽2381)
173
1.16 (0.97–1.39)
1.12 (0.93–1.35)
1.10 (0.91–1.33)
59
1.27 (0.96–1.67)
1.13 (0.85–1.50)
1.13 (0.85–1.50)
High (n⫽716)
Stress frequency
Never/hardly ever (n⫽6027)
493
1.00 (referent)
1.00 (referent)
1.00 (referent)
Monthly (n⫽3442)
206
1.00 (0.84–1.18)
0.99 (0.84–1.17)
0.99 (0.83–1.17)
Weekly (n⫽2001)
133
1.10 (0.91–1.34)
1.08 (0.89–1.32)
1.08 (0.88–1.31)
94
1.27 (1.02–1.59)
1.18 (0.94–1.48)
1.14 (0.90–1.44)
Daily (n⫽1061)
*Body mass index, smoking, education, physical activity, alcohol, systolic blood pressure, antihypertensive
treatment.
†Acute myocardial infarction, FEV1, diabetes mellitus.
Truelsen et al
TABLE 3.
Self-Reported Stress and Risk of Stroke
859
Stress Intensity and Frequency and Risk of Nonfatal Stroke
RR (95% CI) of Nonfatal Stroke
Nonfatal
Strokes, n
Age and Sex
Age, Sex, and Risk
Factors*
Age, Sex, Risk Factors,
and Chronic Disease†
Stress intensity
None (n⫽4931)
317
1.00 (referent)
1.00 (referent)
1.00 (referent)
Light (n⫽4521)
221
0.95 (0.79–1.13)
0.93 (0.78–1.10)
0.92 (0.77–1.10)
Moderate (n⫽2381)
142
1.19 (0.98–1.46)
1.14 (0.93–1.40)
1.11 (0.90–1.37)
41
1.12 (0.81–1.56)
0.97 (0.70–1.36)
0.95 (0.67–1.33)
High (n⫽716)
Stress frequency
Never/hardly ever (n⫽6027)
381
1.00 (referent)
1.00 (referent)
1.00 (referent)
Monthly (n⫽3442)
169
1.02 (0.85–1.23)
1.02 (0.85–1.23)
1.00 (0.83–1.21)
Weekly (n⫽2001)
97
1.01 (0.81–1.26)
0.98 (0.78–1.23)
0.97 (0.77–1.23)
Daily (n⫽1061)
73
1.27 (0.99–1.64)
1.17 (0.90–1.51)
1.09 (0.83–1.42)
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
*Body mass index, smoking, education, physical activity, alcohol, systolic blood pressure, antihypertensive
treatment.
†Acute myocardial infarction, FEV1, diabetes mellitus.
(P⫽0.02). The remaining tests were not statistically significant. The RR of stroke according to level of stress frequency
is shown in the Figure, stratified for the 3 levels of education.
For subjects with ⬎11 years of education, self-reported stress
frequency and risk of stroke are U-shaped, and monthly or
weekly self-reported stress is associated with the lowest risk
of stroke. For subjects with fewer years of education, there is
no clear pattern. The results for nonfatal strokes are similar to
those for all strokes (results not shown).
Data were reanalyzed to examine whether there were linear
trends between levels of self-reported stress and risk of all
stroke, nonfatal stroke, and fatal stroke. None of these
analyses were significant; values ranged from P⫽0.08 to
P⫽0.73. This indicates that according to the present data,
there is no constant relationship between level of stress and
risk of stroke.
Discussion
In the present study, self-reported weekly stress and high
stress intensity were significantly associated with an inTABLE 4.
creased risk of fatal stroke, whereas for the remaining stress
categories, there were no significant associations. We did not
find a statistical association between self-reported stress and
risk of all strokes and nonfatal strokes. There were no
significant trends between level of stress and risk of stroke.
Subjects with high stress intensity who were frequently
stressed had the most unfavorable risk factor profile.
The CCHS is a large, prospective, population-based, cohort
study. Linkage of personal identification number to central
registers enabled identification of all hospital admissions for
stroke and all deaths caused by stroke. During the entire
follow-up period, admission to health facilities was free of
charge. Results from the Danish Monitoring Trends and
Determinants in Cardiovascular Disease (MONICA) study
have shown that 94% of all patients with nonfatal stroke are
hospitalized and that 96% of all stroke events were identified
by use of data from the central registers.29 In the present
study, interview of the participants at the study examinations
made it possible to include nonfatal nonhospitalized stroke
Stress Intensity and Frequency and Risk of Fatal Stroke
RR (95% CI) of Fatal Stroke
Fatal
Strokes, n
Age and Sex
Age, Sex, and Risk
Factors*
Age, Sex, Risk Factors,
and Chronic Disease†
Stress intensity
None (n⫽4931)
91
1.00 (referent)
1.00 (referent)
1.00 (referent)
Light (n⫽4521)
67
1.07 (0.78–1.48)
1.08 (0.77–1.49)
1.14 (0.82–1.59)
Moderate (n⫽2381)
31
1.01 (0.67–1.53)
1.01 (0.66–1.54)
1.04 (0.67–1.61)
High (n⫽716)
18
1.79 (1.07–2.99)
1.74 (1.03–2.95)
1.89 (1.11–3.21)
1.00 (referent)
1.00 (referent)
1.00 (referent)
Stress frequency
Never/hardly ever (n⫽6027)
112
Monthly (n⫽3442)
37
0.89 (0.61–1.30)
0.88 (0.60–1.29)
0.92 (0.62–1.35)
Weekly (n⫽2001)
36
1.46 (1.00–2.15)
1.49 (1.01–2.19)
1.49 (1.00–2.23)
Daily (n⫽1061)
21
1.28 (0.80–2.04)
1.25 (0.77–2.01)
1.33 (0.82–2.16)
*Body mass index, smoking, education, physical activity, alcohol, systolic blood pressure, hypertensive treatment.
†Acute myocardial infarction, FEV1, diabetes mellitus.
860
Stroke
April 2003
Risk of nonfatal stroke according to education level and stress
frequency. Adjusted for age, sex, body mass index, smoking,
physical activity, alcohol, systolic blood pressure, and antihypertensive treatment.
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
events that were not registered. A few nonfatal nonhospitalized stroke events may have remained unidentified, but the
overall effect on estimates is likely to be negligible.
There is no consensus on how stress should be measured.18
Different studies using different definitions and methods have
examined the relation between stress and stroke and reached
different conclusions. In a Finnish study of 2303 middle-aged
men followed up for 11 years, during which 113 first-time
strokes occurred, elevations in blood pressure during the
anticipation phase before a maximal exercise bicycle ergometer test relative to the average resting baseline blood pressure
were considered a measure of cardiovascular activation in
response to psychological and behavioral stress.23 Subjects
with excessive reactivity had an almost doubled risk of
stroke, and men who were high reactors and poorly educated
had almost a 3-times-higher risk of stroke compared with
better-educated, less reactive men. The same study found that
there was an association between level of stress and intimamedia thickening,30 and stress has also been associated with
increased progression of carotid artery disease.31 A Swedish
study of 238 hypertensive men with a mean follow-up period
of 9.9 years, during which 43 experienced stroke, showed that
poor adaptation to stress measured with a psychological
stressor (the color-word test) was associated with a 3-timeshigher in the risk of stroke.6 In a large intervention study of
7495 middle-aged men (47 to 55 years of age at baseline),
stroke incidence was estimated during 11.8 years of
follow-up in Göteborg, Sweden.8 Psychological stress was
defined as feelings of tension or anxiety on a scale from 1
(never experienced stress) to 6 (feeling permanent stress
during the last 5 days). Subjects with the highest level of
stress had doubled risk of stroke compared with nonstressed
subjects. The stress definition in the latter study is very
similar to ours in that self-reported stress was used in the
analyses. However, it is not surprising that there is divergence
of the results between the studies because we included both
men and women, used a larger age range, and had no
intervention group. Furthermore, the Göteborg study did not
provide separate results for fatal and nonfatal stroke.
A returning question in studies of stress and stroke is the
validity of the stress markers. For example, the generalizability of reactions to laboratory challenges to reactions to
everyday life stress exposures has been questioned.32 Addi-
tionally, the validity of self-reported stress is questionable. A
high validity would necessitate that the interviewed subjects
interpret the meaning of the questions identically and report
identically. Stress could even refer to both physical and
psychological stress. Although we have no proof of how the
participants in the CCHS understood the stress questions, it is
our belief that the majority would interpret it as referring to
psychological and not physical stress because stress in the
questionnaire was exemplified in psychological terms. However, it is likely that the perceived level of stress for the same
exposure may differ from one person to another because
individual expectations about what is stressful differs. Thus,
what is actually measured is a complex conglomerate of both
exposure to “stressors” and the subject’s resources for dealing
with and reporting it. We can therefore expect that as long as
there is no established definition of stress, results from
different studies are likely to differ considerably. The proportion of women increased markedly in the higher stress
groups. It remains unknown whether this results because
women are more exposed to stressors than men, they feel
more stress given the same exposure, or they report the
sensation of stress differently than men. Results from our
study suggested that data from men and women could be
combined in the analyses because the interaction tests between stress and sex were not significant. We acknowledge
that this is a purely statistical approach to examine whether
the effect of stress and risk of stroke are similar for men and
women, and future studies may provide an opportunity to
examine this relation in more detail.
Despite the problem with the validity of self-reported
stress, it was clearly associated with a less favorable risk
factor profile. Many of the considered confounding variables
were both well-established risk factors for stroke and modifiable. Thus, by asking a simple question about stress intensity and frequency, it was possible to identify a group of
subjects that potentially could benefit from intensified preventive efforts. The effect of controlling for the confounding
variables resulted in estimates for different levels of stress
and risk of all stroke that were attenuated, indicating that an
association between self-reported stress and risk of stroke
may be due, at least in part, to differences in risk factor
profile. However, when we adjusted for the same variables in
analyses with only fatal strokes as the end point, the estimates
decreased only slightly and even increased for high stress and
weekly stress. This may indicate that there could be an
independent relation between self-reported stress intensity
and frequency and risk of fatal stroke. However, there were
no clear patterns in any of the analyses and no significant
trends for either stress intensity or frequency. Also, there was
no indication of a threshold, and residual confounding may be
another potential explanation for the findings. Even if there is
an association, only ⬇10% of the subjects with fatal stroke
were in the highly stressed groups; thus, according to our
data, self-reported stress cannot be regarded as an important
risk factor for fatal stroke at the population level.
One of the main purposes of the study was to examine
whether different levels of self-reported stress were associated with different levels of systolic blood pressure. If
subjects feeling stressed were more likely to have higher
Truelsen et al
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
blood pressure levels than nonstressed subjects, this could
explain some of the relationship. The mean systolic blood
pressure values differed significantly for both stress intensity
and frequency, but there was no clear trend between the stress
groups. However, there was an increasing proportion of
subjects on antihypertensive treatment, and that pattern was
found for both stress intensity and frequency. From these
data, it is not possible to address whether stressed subjects
become hypertensive or hypertensive subjects become
stressed, and the increase could also be explained by stressed
people being more frequently in contact with health professionals and therefore more likely to undergo treatment for
increased blood pressure. We reanalyzed data with and
without systolic blood pressure and antihypertensive treatment, but the estimates remained almost the same and the
overall results did not change (data not shown). Therefore,
according to the present data, systolic blood pressure and
antihypertensive treatment cannot explain the relation between self-reported stress and risk of stroke.
There are several possible explanations why stress may be
associated with only a higher risk of fatal stroke. Subjects
who experience stress may be more likely to have more
severe strokes, which would increase the risk of death.
Regrettably, in the present study, we have no information on
severity of stroke. Another possibility is that stressed subjects
who suffer stroke more often than nonstressed subjects are
more susceptible to complications. However, the biological
background for this is unclear. An animal model examining
the relation between stress and stroke with male mice
examined the expression of ischemia-induced Bcl-2 expression.33 The Bcl-2 protooncogene promotes cell survival and
protects against apoptosis and cellular necrosis. The authors
found that male mice that were stressed before occlusion of
cerebral artery occlusion expressed ⬇70% less Bcl-2 than
unstressed mice after ischemia. Whether these findings would
be similar in humans is unknown; however, they suggest that
there may be (unidentified) biological mechanisms between
stress and risk of stroke.
A previous study has shown that men with high reactivity
to stress and little education were at greater risk of stroke than
more educated men.23 Previous studies have shown that there
is an inverse socioeconomic gradient for stroke incidence and
mortality,14 and SES may be related to cardiovascular responses to stress.16,34 The effect of stress on risk of stroke
could be moderated by the coping possibility and strategy
used by the subject; stress in individuals who do not feel in
control could be more harmful than stress that is self-inflicted
or of which the person feels in control. Hence, some degree
of psychological stress may even be perceived as a positive
challenge if the person feels in control of the situation.35 More
educated people will often be more in control of their jobs and
consequently more likely to be able to cope with stressful
situations in the job compared with people with low-control
jobs; thus, the effect of stress may vary according to education level. That some level of stress may be beneficial is in
agreement with the finding of an interaction between educational levels of self-reported stress frequency in the present
study, showing a U-shaped relation between risk of stroke
and stress in subjects with ⬎11 years of education.
Self-Reported Stress and Risk of Stroke
861
Several baseline characteristics and RR estimates showed a
J-shaped relationship between levels of stress and risk of
stroke. This could hypothetically indicate a systematic misclassification of stressed individuals into the “none” and
“never” stressed groups. If that happened, it likely attenuated
the difference between the referent category and the other
stress levels. There are different possible explanations for
stressed individuals being misclassified as not stressed. They
might have felt that the definition of stress was not applicable
to their situation, or they may have been unaware of their high
levels of stress. In addition, underreporting of stress might be
related to lifestyle and may differ between categories of risk
factors, which provides an alternative explanation for the
results between years of education and risk of stroke. Similar
interaction tests for age (⬍65 or ⱖ 65) and smoking (yes/no)
were not significant (results not shown).
The categorization of stress intensity and frequency raises
the issue of multiple comparisons because each stress group
was compared with the referent category. A formal test for
multiple comparisons was not made because the present
results relate mainly to the pattern between self-reported
stress and risk of stroke rather than between specific stress
levels and risk of stroke.
In conclusion, in the present study, self-reported high stress
and weekly stress were associated with a higher risk of fatal
stroke. However, there were no significant trends, and the
present data do not provide strong evidence that self-reported
stress is an independent risk factor for fatal stroke. Even if
such an association exists, the biological background is
unclear. There was no significant association between stress
and risk of all strokes and nonfatal strokes. Stressed subjects
had a more unfavorable risk factor profile than subjects not
reporting to be stressed; thus, asking a relatively simple
question about perceived level of stress may identify subjects
who could benefit from intensified prevention. Although
level of systolic blood pressure differed significantly between
the stress groups, it could not explain the association between
stress and stroke.
Acknowledgments
This study was supported by grants from the Hede Nielsen Foundation. The Danish Heart Foundation supports the CCHS. We are
grateful to G. Jensen, J. Nyboe, A.T. Hansen, and P. Schnohr, who
initiated and conducted the CCHS. We acknowledge M. Appleyard
for her support and assistance.
References
1. Kattapong VJ, Eaton OM, Becker TM. Stroke risk factor knowledge in
university students. J Am Coll Health. 1996;44:236 –238.
2. Pancioli AM, Broderick J, Kothari R, Brott T, Tuchfarber A, Miller R,
Khoury J, Jauch E. Public perception of stroke warning signs and
knowledge of potential risk factors. JAMA. 1998;279:1288 –1292.
3. Kothari R, Sauerbeck L, Jauch E, Broderick J, Brott T, Khoury J, Jauch
E. Patients’ awareness of stroke signs, symptoms, and risk factors. Stroke.
1997;28:1871–1975.
4. Sug YS, Heller RF, Levi C, Wiggers J, Fitzgerald PE. Knowledge of
stroke risk factors, warning symptoms, and treatment among an Australian urban population. Stroke. 2001;32:1926 –1930.
5. Yoon SS, Byles J. Perceptions of stroke in the general public and patients
with stroke: a qualitative study. BMJ. 2002;324:1065.
6. Andre-Petersson L, Engstrom G, Hagberg B, Janzon L, Steen G. Adaptive
behavior in stressful situations and stroke incidence in hypertensive men:
862
7.
8.
9.
10.
11.
12.
13.
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
14.
15.
16.
17.
18.
19.
20.
Stroke
April 2003
results from prospective cohort study “men born in 1914” in Malmö,
Sweden. Stroke. 2001;32:1712–1720.
House A, Dennis M, Mogridge L, Hawton K, Warlow C. Life events and
difficulties preceding stroke. J Neurol Neurosurg Psychiatry. 1990;53:
1024 –1028.
Harmsen P, Rosengren A, Tsipogianni A, Wilhelmsen L. Risk factors for
stroke in middle-aged men in Göteborg, Sweden. Stroke. 1990;21:
223–229.
Macko RF, Ameriso SF, Barndt R, Clough W, Weiner JM, Fisher M.
Precipitants of brain infarction: roles of preceding infection/inflammation
and recent psychological stress. Stroke. 1996;27:1999 –2004.
Carroll D, Smith GD, Sheffield D, Shipley MJ, Marmot MG. Pressor
reactions to psychological stress and prediction of future blood pressure:
data from the Whitehall II Study. BMJ. 1995;310:771–776.
Widgren BR, Wikstrand J, Berglund G, Andersson OK. Increased
response to physical and mental stress in men with hypertensive parents.
Hypertension. 1992;20:606 – 611.
Krantz DS, Manuck SB. Acute psychophysiologic reactivity and risk of
cardiovascular disease: a review and methodologic critique. Psychol Bull.
1984;96:435– 464.
Light KC, Cardiovascular responses to effortful active coping: implications for the roles of stress in hypertension development. Psychophysiology. 1981;18:216 –225.
Modan B, Wagener DK. Some epidemiological aspects of stroke: mortality/morbidity trends, age, sex, race, socioeconomic status. Stroke.
1992;23:1230 –1236.
Barnett PA, Spence JD, Manuck SB, Jennings JR. Psychological stress
and the progression of carotid artery disease. J Hypertens. 1997;15:
49 –55.
Lynch JW, Everson SA, Kaplan GA, Salonen R, Salonen JT. Does low
socioeconomic status potentiate the effects of heightened cardiovascular
responses to stress on the progression of carotid atherosclerosis? Am J
Public Health. 1998;88:389 –394.
Castillo-Richmond A, Schneider RH, Alexander CN, Cook R, Myers H,
Nidich S, Haney C, Rainforth M, Salerno J. Effects of stress reduction on
carotid atherosclerosis in hypertensive African Americans. Stroke. 2000;
31:568 –573.
Schneck MJ. Is psychological stress a risk factor for cerebrovascular
disease? Neuroepidemiology. 1997;16:174 –179.
Engstrom G, Jerntorp I, Pessah-Rasmussen H, Hedblad B, Berglund G,
Janzon L. Geographic distribution of stroke incidence within an urban
population: relations to socioeconomic circumstances and prevalence of
cardiovascular risk factors. Stroke. 2001;32:1098 –1103.
Jakovljevic D, Sarti C, Sivenius J, Torppa J, Mahonen M,
Immonen-Raiha P, Kaarsalo E, Alhainen K, Kuulasmaa K, Tuomilehto J,
Puska P, Salomaa V. Socioeconomic status and ischemic stroke. Stroke.
2001;32:1492–1498.
21. Lindenstrom E, Boysen G, Nyboe J. Risk factors for stroke in
Copenhagen, Denmark, I: basic demographic and social factors. Neuroepidemiology. 1993;12:37– 42.
22. Boysen G, Nyboe J, Appleyard M, Sorensen PS, Boas J, Somnier F,
Jensen G, Schnohr P. Stroke incidence and risk factors for stroke in
Copenhagen, Denmark. Stroke. 1988;19:1345–1353.
23. Everson SA, Lynch JW, Kaplan GA, Lakka TA, Sivenius J, Salonen JT.
Stress-induced blood pressure reactivity and incident stroke in
middle-aged men. Stroke. 2001;32:1263–1270.
24. Appleyard M, Hansen T, Schnohr P, Jensen G, Nyboe J. The Copenhagen
City Heart Study: Østerbroundersøgelsen. A book of tables with data
from the first examination (1976 –78) and a five year follow-up
(1981– 83). Scand J Soc Med. 1989;170(suppl 41):1–160. 1989.
25. Truelsen T, Prescott E, Lange P, Schnohr P, Boysen G. Lung function and
risk of fatal and non-fatal stroke: the Copenhagen City Heart Study. Int J
Epidemiol. 2001;30:145–151.
26. Sundhedsstyrelsen. Nye tal fra sundhedsstyrelsen. Dødsårsagsregisteret
1994 og 1995. Munkgaards Forlag; 1997.
27. WHO MONICA Project Investigators. The World Health Organization
MONICA Project (Monitoring Trends and Determinants in Cardiovascular Disease). J Clin Epidemiol. 1988;41:105–114.
28. SAS Institute Inc. SAS/STAT software: the PHREG Procedure. Release
8.02. Cary, NC: SAS Institute Inc; 1999.
29. Thorvaldsen P, Davidsen M, Bronnum-Hansen H, Schroll M. Stable
stroke occurrence despite incidence reduction in an aging population:
stroke trends in the Danish Monitoring Trends and Determinants in
Cardiovascular Disease (MONICA) population. Stroke. 1999;30:
2529 –2534.
30. Everson SA, Lynch JW, Chesney MA, Kaplan GA, Goldberg DE, Shade
SB, Cohen RD, Salonen R, Salonen JT. Interaction of workplace demands
and cardiovascular reactivity in progression of carotid atherosclerosis:
population based study. BMJ. 1997;314:553–558.
31. Lynch J, Kaplan GA, Salonen R, Salonen JT. Socioeconomic status and
progression of carotid atherosclerosis: prospective evidence from the
Kuopio Ischemic Heart Disease Risk Factor Study. Arterioscler Thromb
Vasc Biol. 1997;17:513– 419.
32. Pickering T, Gerin W. Cardiovascular reactivity in the laboratory and the
role of behavioral factors in hypertension: a critical review. Ann Behav
Med. 1990;12:3–16.
33. DeVries AC, Joh HD, Bernard O, Hattori K, Hurn PD, Traystman RJ,
Alkayed NJ. Social stress exacerbates stroke outcome by suppressing
Bcl-2 expression. Proc Natl Acad Sci U S A. 2001;98:11824 –11828.
34. Gump BB, Matthews KA, Raikkonen K. Modeling relationships among
socioeconomic status, hostility, cardiovascular reactivity, and left ventricular mass in African American and white children. Health Psychol.
1999;18:140 –150.
35. Kaplan MK, Sallis JF, Patterson TL. Health and Human Behavior. New
York, NY: McGraw-Hill, Inc; 1993.
Self-Reported Stress and Risk of Stroke: The Copenhagen City Heart Study
Thomas Truelsen, Naja Nielsen, Gudrun Boysen and Morten Grønbæk
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
Stroke. 2003;34:856-862; originally published online March 13, 2003;
doi: 10.1161/01.STR.0000062345.80774.40
Stroke is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 2003 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/34/4/856
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/