Combined Effects of Socioeconomic Position, Smoking, and
Hypertension on Risk of Ischemic and Hemorrhagic Stroke
Helene Nordahl, MS, PhD; Merete Osler, MD, PhD;
Birgitte Lidegaard Frederiksen, MD, PhD; Ingelise Andersen, MS, PhD; Eva Prescott, MD, PhD;
Kim Overvad, MD, PhD; Finn Diderichsen, MD, PhD; Naja Hulvej Rod, MS, PhD
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
Background and Purpose—Combined effects of socioeconomic position and well-established risk factors on stroke
incidence have not been formally investigated.
Methods—In a pooled cohort study of 68 643 men and women aged 30 to 70 years in Denmark, we examined the combined
effect and interaction between socioeconomic position (ie, education), smoking, and hypertension on ischemic and
hemorrhagic stroke incidence by the use of the additive hazards model.
Results—During 14 years of follow-up, 3613 ischemic strokes and 776 hemorrhagic strokes were observed. Current smoking
and hypertension were more prevalent among those with low education. Low versus high education was associated with
greater ischemic, but not hemorrhagic, stroke incidence. The combined effect of low education and current smoking was
more than expected by the sum of their separate effects on ischemic stroke incidence, particularly among men: 134 (95%
confidence interval, 49–219) extra cases per 100 000 person-years because of interaction, adjusted for age, cohort study,
and birth cohort. There was no clear evidence of interaction between low education and hypertension. The combined
effect of current smoking and hypertension was more than expected by the sum of their separate effects on ischemic and
hemorrhagic stroke incidence. This effect was most pronounced for ischemic stroke among women: 178 (95% confidence
interval, 103–253) extra cases per 100 000 person-years because of interaction, adjusted for age, cohort study, and birth
cohort.
Conclusions—Reducing smoking in those with low socioeconomic position and in those with hypertension could potentially
reduce social inequality stroke incidence. (Stroke. 2014;45:2582-2587.)
Key Words: hypertension ◼ smoking ◼ socioeconomic position ◼ stroke
A
socioeconomic gradient in stroke and stroke risk factors
have been observed in a variety of settings.1,2 It has been
documented that some stroke risk factors may interact or act
synergistically with each other.3–5 Because people in lower
socioeconomic groups are frequently exposed to several risk
factors and the effect of one of these factors on risk of stroke
is likely to be stronger in the lower socioeconomic groups than
in the higher—these people are more vulnerable.6 Addressing
such differential vulnerability and interactions between risk
factors are relevant to the public health and clinical agenda
because it allow us to quantify in which subgroups intervention could potentially prevent the most cases.7,8
The mechanisms driving social inequality in stroke have
been described as being partly a result of differential exposure to well-established risk factors, such as hypertension and
smoking, during a life course.9–12 In previous studies, authors
have suggested that smoking has a larger effect on ischemic
stroke, hemorrhagic stroke, and heart disease if high blood
pressure is also present.3,5,13–15 However, much less attention
has been given to the extent to which these risk factors interact
with socioeconomic exposures.4
In a pooled cohort study including 68 643 men and women,
we examined the combined effects and interaction among
socioeconomic position, smoking, and hypertension on ischemic and hemorrhagic stroke incidence.
Methods
Study Population
The cohort consortium, the Social Inequality in Cancer cohort study,
combined questionnaire data on 83 006 participants from 7 existing
population-based cohort studies from Denmark: the Copenhagen City
Heart Study; the 1936 Cohort Study; Monica I, II, and III; the Diet
Cancer and Health Study; the Inter99 Study. Detailed descriptions
of this consortium and the enrolled studies have previously been
described.16
Received February 25, 2014; final revision received June 13, 2014; accepted June 20, 2014.
From the Section of Social Medicine, Department of Public Health, University of Copenhagen, Copenhagen, Denmark (H.N., I.A., F.D., N.H.R.);
Research Center for Prevention and Health, Glostrup University Hospital, Glostrup, Denmark (M.O.); Clinical Research Centre, Hvidovre University
Hospital, Copenhagen, Denmark (B.L.F.); Department of Cardiology, Bispebjerg University Hospital and the Copenhagen City Heart Study, Bispebjerg,
Denmark (E.P.); and Section for Epidemiology, Department of Public Health, Aarhus University, Aarhus, Denmark, and Department of Cardiology, Aalborg
University Hospital, Aalborg, Denmark (K.O.).
Correspondence to Helene Nordahl, MS, PhD, Department of Public Health, University of Copenhagen, Øster Farimagsgade 5A, Post Box 2099, 1014
Copenhagen, Denmark. E-mail [email protected]
© 2014 American Heart Association, Inc.
Stroke is available at http://stroke.ahajournals.org
DOI: 10.1161/STROKEAHA.114.005252
2582
Nordahl et al Combined Effects of Risk Factors on Stroke 2583
For the current study, we excluded 492 participants aged ≤29
years (assuming that some of these participants had not yet reached
their final level of education); 4424 participants born before 1921
(information on education was unavailable); 1796 participants with
missing information on education; 7020 participants with preexisting cardiovascular disease (defined as diagnosed before study entry
in the Danish National Registry of Patients using the International
Classification of Diseases [ICD], ICD8: 390–458; ICD10: I00–I99,
G45); and 631 participants with missing information on smoking or
systolic blood pressure; leaving 68 643 participants aged 30 to 70
years for the analyses.
Outcomes
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
Information on stroke cases was obtained through linkage with the
Danish National Registry of Patients and the Causes of Death Registry
using the ICD-Eighth Revision from 1981 to 1994 and the Tenth Revision
thereafter. The cases were classified as ischemic stroke (IDC8:433–434;
ICD10:I63), unspecified stroke (ICD8:436; ICD10:I64), subarachnoid
hemorrhagic stroke (ICD8:430; ICD10:I60) and intracerebral hemorrhagic stroke (ICD8:431; ICD10:I61). Participants with ischemic
strokes and unspecified strokes were analyzed together because Krarup
et al17 in their validation study confirmed that the majority of the unspecified stroke diagnosis in the Danish National Registry of Patients
was diagnosed as ischemic stroke in their clinical evaluations. To increase the power of the limited number of hemorrhagic stroke cases,
patients with subarachnoid hemorrhage and intracerebral hemorrhage
were combined for the analysis. Follow-up was from date of study entry until date of first primary diagnosis of fatal or nonfatal stroke, death,
emigration, or end of follow-up (December 31, 2009). Fewer than 0.1%
were lost to follow-up because of emigration.
Exposures
Socioeconomic position was assessed by linkage to the high-quality
data from central and administrative registries using highest attained
education, which was measured 1 year before study entry. Education
was categorized in 3 groups: low (primary and lower secondary
education), medium (upper secondary, vocational or technical education, as well as short-cycle higher nonuniversity programs), and
high (medium-cycle university or nonuniversity programs, as well as
long-cycle university programs). Information on smoking was based
on questions covering smoking status and levels of current smoking, which were combined into a 4-category smoking variable: never
smoker, former smoker, smoker of 1 to 15 g/d, and smoker of ≥16
g/d. Blood pressure was measured by trained nurses according to the
World Health Organization guidelines. We constructed a 4-category
variable for systolic blood pressure; <120 mm Hg, 120 to 139 mm Hg,
140 to 159 mm Hg, and ≥160 mm Hg. Hypertension was defined as
systolic blood pressure ≥140 mm Hg.
Covariates
Other covariates included age, sex, cohort study, birth cohort (from
1921 to 1970 in 5-year intervals), alcohol intake (continuous in
grams per day), physical activity in leisure time (sedentary, <2 hours/
wk of light activity; light, 2–4 hours/wk of light activity or <2 hours/
wk of high-level activity; moderate, >4 hours/wk of light activity or
2–4 hours/wk of high-level activity; and high, >4 hours/wk of highlevel activity), and body mass index (≤18.5, 18.6–24.9, 25.0–29.9,
or ≥30 kg/m2).
Statistical Analysis
The concept of differential vulnerability was defined as deviation
from additivity of absolute effect analogous to assessing additive
interaction in statistical terms.18,19 We applied the additive hazards
model that was recently proposed as a technique to assess additive
interaction in survival analyses.19
In this model, the hazard of ischemic and hemorrhagic stroke was
modeled as a linear function of the exposure variables (ie, education,
smoking, and hypertension) adjusted for confounders (ie, sex, cohort
study, and birth cohort) plus an unspecified baseline hazard using age
as the underlying time scale.20 The effect estimate was rate difference
interpreted as the extra number of cases per 100 000 person-years at
risk in a specific exposure category when compared with the reference. The analysis for ischemic stroke was stratified into men and
women because of marked sex-based differences. For hemorrhagic
stroke, men and women were analyzed together as we found no indication of sex-based differences.
The magnitude of the deviation from additivity (or additive interaction) was directly assessed by including product terms of the exposure variables.19 We also defined 4 combined exposure variables:
education-by-smoking, education-by-hypertension, smoking-by-hypertension, and education-by-smoking-by-hypertension. These were
entered one at a time into the model.
In sensitivity analyses, we (1) restricted analysis to patients with
an ICD code for ischemic stroke only and for intracerebral hemorrhagic stroke only, (2) assessed study-specific effects by including the
exposure-by-cohort interaction terms, (3) evaluated potential measurement error by applying different cutoff points for smoking and
hypertension and by testing interactions using 4-category variables,
(4) evaluated potential age-dependent effects of education, smoking,
or hypertension by the use of standard techniques,21 (5) evaluated
the influence of other health-related behaviors by adjusting for body
mass index, alcohol intake, and physical activity.
Results
During 14 years of follow-up, we observed 2346 unspecified
strokes and 1951 ischemic strokes, which were combined
as ischemic strokes. Furthermore, we observed 776 hemorrhagic strokes (with 70% of these being diagnosed as intracerebral hemorrhages and the remaining as subarachnoidal
hemorrhages).
Smoking was clearly more frequent in those with low education, and the level of blood pressure was only slightly higher
among those with low versus high education (Table 1). The
main effects of education, smoking, and blood pressure on
ischemic and hemorrhagic stroke are shown in Table 2. Clear
educational differences were observed for ischemic stroke but
only slightly for hemorrhagic stroke. In absolute terms among
men, low versus high education was associated with 181 (95%
confidence interval, 127, 235) extra cases per 100 000 personyears at risk of ischemic stroke. Correspondingly, among
women there were 93 (56, 129) extra cases of ischemic stroke.
Smoking and high blood pressure were markedly associated
with incidence of both ischemic and hemorrhagic stroke.
As presented in Table 3, the combined effect of low education and current smoking was more than expected by the sum
of their separate effects on ischemic stroke but negligible for
hemorrhagic stroke. This was most pronounced among men,
where the separate effect of having low education was associated with 42 (−5, 90) extra cases per 100 000 person-years at
risk of ischemic stroke and the separate effect of being current smoker was associated with 112 (38, 186) extra cases.
However, the combination of current smoking and low education was associated with 289 (238, 340) extra cases per
100 000 person-years at risk of ischemic stroke. Thus, 134 (49,
219) extra cases per 100 000 person-years at risk of ischemic
stroke could be ascribed to the interaction between smoking and education. We could not confirm evidence of similar patterns among women. There was no clear evidence of
interaction with respect to the combination of low education
and hypertension on risk of ischemic stroke (men, P=0.89;
2584 Stroke September 2014
Table 1. Characteristics of 68 643 Participants From the SIC Cohort
Educational Level
Total (n=68 643)
Women, %
Low
Medium
High
(n=20 867)
(n=33 670)
(n=14 106)
53
63
50
47
54 (40, 63)
54 (40, 64)
53 (35, 63)
53 (40, 63)
Never smoker
34
28
35
39
Former smoker
26
23
26
31
Smoker (1–15 g/d)
21
25
20
17
Heavy smoker (≥16 g/d)
19
24
19
13
<120 mm Hg
21
20
19
24
120–139 mm Hg
42
40
42
44
140–159 mm Hg
26
27
27
24
≥160 mm Hg
11
12
12
9
12 (0, 62)
9 (0, 61)
13 (0, 63)
15 (1, 62)
26 (4)
26 (5)
26 (4)
25 (4)
17
20
17
16
Age at study entry, y; median (5%,
95% percentile)
Tobacco smoking, %
Systolic blood pressure, %
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
Alcohol intake, g/d; median (5%, 95%
percentile)
Body mass index, kg/m2; mean (SD)
Sedentary leisure time*, %
Cohort study, y; %
CCHS, 1981–1983
10
50
38
12
1936-COHORT, 1981–1982
1
36
52
12
MONICA I, 1982–1984
5
37
51
13
MONICA II, 1986–1987
2
36
50
14
MONICA III, 1991–1992
3
34
54
13
71
28
50
23
8
24
55
22
100
30
49
21
DCH, 1993–1997
INTER99, 1999–2001
Pooled SIC cohort 1981–2001
CCHS indicates Copenhagen City Heart Study; DCH, Diet Cancer and Health study; INTER99, Randomized
Nonpharmacological Intervention Study for Prevention of Ischemic Heart Disease; MONICA (I, II, III),
Multinational Monitoring of Trends and Determinants in Cardiovascular Disease; and SIC, Social Inequality
in Cancer cohort study.
*Sedentary leisure time: <2 h/wk of light activity.
women, P=0.05) or hemorrhagic stroke (P=0.53). However,
the combined effect of current smoking and hypertension was
more than expected by the sum of their separate effects on
ischemic and hemorrhagic stroke. This was most pronounced
among women, where 178 (103, 253) extra cases per 100 000
person-years at risk of ischemic stroke could be ascribed to
the interaction between smoking and hypertension.
The combined effect of exposure to all 3 risk factors on
ischemic stroke was associated with 566 extra cases among
men and 438 extra cases among women when compared
with no exposure (high education, no current smoking, and
no hypertension), as presented in the Figure. This was more
than expected from the sum of their separate effects, ie, among
women 308 extra cases of ischemic stroke per 100 000 personyears (calculated as 438–[−10+107+33]).
The sensitivity analyses indicated (data not shown) (1)
that restricting the analysis to patients with an ICD code for
ischemic stroke only and for intracerebral hemorrhagic stroke
only did not change the conclusions, (2) no notable interaction
between cohort study and the 3 exposure variables, (3) that
using 4-category variables for smoking and systolic blood
pressure with different cutoff point did not affect the presented
results, (5) no indication of age-dependent effects, (6) that
adjusting for body mass index, alcohol intake, and physical
activity did not change the conclusions.
Discussion
In agreement with other studies,1,2,12,22 our results demonstrate
that low versus high education is associated with greater ischemic but not of hemorrhagic, stroke incidence, and that stroke
risk factors are more prevalent in low educational groups.
The present study provides further insight into the combined
effects and interactions among socioeconomic position, smoking, and hypertension.
In previous studies, authors have suggested that the existence of an interaction between the 2 well-established risk
factors, smoking and hypertension, in relation to stroke in
general5,13 and hemorrhagic stroke.3 Our data confirmed this
Nordahl et al Combined Effects of Risk Factors on Stroke 2585
Table 2. RD per 100 000 Person-Years at Risk of Ischemic Stroke (2005 Cases in Men, 1608 Cases in Women) and Hemorrhagic
Stroke (776 Cases in Men and Women Together) Associated With Education, Smoking, and Hypertension, Among 68 643 Participants
From the Social Inequality in Cancer Cohort
Ischemic Stroke
Hemorrhagic Stroke
Men, n=32 126
No. of Cases
Women, n=36 517
RD* (95% CI)
No. of Cases
0 (Ref.)
Men and Women, n=68 643
RD* (95% CI)
No. of Cases
RD† (95% CI)
Educational level
High
Medium
Low
197
0 (Ref.)
125
0 (Ref.)
1037
339
91 (52, 131)
647
38 (6, 70)
386
13 (0, 25)
629
181 (127, 235)
764
93 (56, 129)
265
11 (−4, 25)
Tobacco smoking
Never smoker
365
0 (Ref.)
460
Former smoker
545
−15 (−60, 29)
303
13 (−20, 46)
0 (Ref.)
199
178
−1 (−13, 12)
0 (Ref.)
Smoker (1–15 g/d)
418
145 (95, 195)
545
177 (143, 211)
210
31 (17, 45)
Heavy smoker (≥16 g/d)
677
270 (221, 319)
300
266 (221, 311)
189
42 (27, 57)
<120
197
−32 (−72, 9)
253
−53 (−83, −24)
104
−8 (−19, 3)
120–139
619
0 (Ref.)
529
140–159
696
136 (93, 179)
483
71 (31, 110)
253
26 (12, 40)
≥160
493
461 (386, 536)
343
276 (214, 338)
175
80 (57, 103)
Systolic blood pressure, mm Hg
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
0 (Ref.)
244
0 (Ref.)
CI indicates confidence interval; Ref, reference category; and RD, rate difference.
*RD per 100 000 person-years at risk adjusted for age, cohort study, and birth cohort.
†RD per 100 000 person-years at risk adjusted for sex, age, cohort study, and birth cohort.
interaction and showed that the combined effect of current
smoking and hypertension exceeded the sum of their separate effects on incidence of both ischemic and hemorrhagic
stroke. Interestingly, an equally important interaction between
education and smoking on incidence of ischemic stroke was
observed, suggesting that people, particularly men, in lower
socioeconomic groups could be more vulnerable to the effect
of smoking than those in higher socioeconomic groups in
terms of ischemic stroke. This finding may be of benefit to the
public health and clinical agenda because health interventions
and preventive strategies should be aimed at patients or population subgroups in which the most cases could potentially be
prevented.
In support of our findings, we have previously shown that
the same level of smoking has different effects across socioeconomic groups in relation to death from cardiovascular
disease.23 Such apparent differential vulnerability could be
a consequence of clustering and mutual interaction between
other active determinants.24 Although others have described
the important role of smoking and hypertension as contributors to the social gradients in stroke,9–12 our data indicated
only sparse differential exposure to hypertension across
socioeconomic groups. Likewise, in previous trend studies
during the 1980s and 1990s, no marked socioeconomic differences in blood pressure have been reported.25,26 Our results
indicate that the combined effect of low socioeconomic position, current smoking, and hypertension exceeded the sum
of their separate effects. However, the apparent vulnerability
to the effect of smoking in the low socioeconomic group did
not seem to be purely a consequence of clustering or mutual
interaction with hypertension. Actually, the stronger effect of
smoking in the low socioeconomic group seemed regardless
of the presence of hypertension. In a US population-based
study, Liu et al4 found a combined effect of socioeconomic
position and chronic conditions, such as hypertension, heart
problems, and diabetes mellitus, in relation to risk of stroke.
Furthermore, they suggested potential cumulative effects of
childhood social conditions, adulthood socioeconomic position, and adult chronic conditions on risk of stroke. Thus, to
understand the role of differential vulnerability better, it might
be particularly relevant to address the interaction between earlier exposures and those that occur later in life because the
presence of risk factors during childhood has been shown
to increase the risk of stroke.27 However, we were unable to
apply a life course perspective because information on childhood socioeconomic circumstances was not available in the
present study.
To our knowledge, this is the first cohort study to evaluate
interactions among socioeconomic position, smoking, and
hypertension in relation to stroke incidence comprehensively.
The large population sample and long follow-up provided
sufficient power to test the hypotheses of interaction even in
subgroups of stroke. Furthermore, the linkage to nationwide
register data allowed for nearly complete follow-up. However,
some limitations of the present analysis should be pointed out.
First, there may be variations in the quality of register data for
some diagnostic subgroups of stroke, which could compromise the estimation of absolute risk measures. Two validation
studies from Denmark have confirmed high predictive values
(>80%) of the overall stroke and ischemic stroke diagnosis in
Danish National Registry of Patients.17,28 However, Krarup et al17
showed that for unspecified stroke more than two thirds were
diagnosed as ischemic stroke by clinical evaluators. Thus, in
the current study, we combined ischemic and unspecified stroke
2586 Stroke September 2014
Table 3. RD per 100 000 Person-Years at Risk of Ischemic Stroke (2005 Cases in Men, 1608 Cases in Women) and Hemorrhagic
Stroke (776 Cases in Men and Women Together) Associated With 2-Way Combined Effects of Education, Smoking, and
Hypertension, Among 68 643 Participants From the Social Inequality in Cancer Cohort
Ischemic Stroke
Hemorrhagic Stroke
Men, n=32 126
No. of Cases
Women, n=36 517
RD* (95% CI)
No. of Cases
RD* (95% CI)
Men and Women, n=68 643
No. of Cases
RD† (95% CI)
Combined effects of education and smoking
High education and no smoking
200
0 (Ref.)
112
Low education and no smoking
710
42 (−5, 90)
651
23 (−11, 57)
High education and current smoking
139
112 (38, 186)
85
156 (91, 221)
46
17 (−6, 40)
Low education and current smoking
956
289 (238, 340)
760
231 (192, 270)
353
40 (24, 56)
P value for interaction
...
<0.01
...
No. of extra cases because of interaction
...
134 (49, 219)
...
0 (Ref.)
0.15
53 (−20, 125)
79
298
0 (Ref.)
0 (−14, 15)
...
0.08
...
23 (−3, 49)
0 (Ref.)
Combined effects education and hypertension
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
High education and no hypertension
124
0 (Ref.)
112
0 (Ref.)
60
Low education and no hypertension
692
107 (66, 148)
670
28 (−2, 59)
288
6 (−7, 19)
High education and hypertension
215
232 (159, 305)
85
83 (11, 154)
65
36 (10, 61)
Low education and hypertension
189 (147, 231)
363
51 (35, 67)
974
333 (283, 383)
741
P value for interaction
...
0.89
...
No. of extra cases because of interaction
...
−6 (−91, 78)
...
0.05
78 (−1, 157)
...
...
0.53
9 (−20, 39)
Combined effects smoking and hypertension
No smoking and no hypertension
323
165
0 (Ref.)
Current smoking no hypertension
493
176 (137, 214)
0 (Ref.)
326
456
156 (125, 187)
0 (Ref.)
183
19 (8, 31)
No smoking and hypertension
587
178 (130, 226)
437
101 (65, 137)
212
25 (11, 38)
Current smoking hypertension
602
506 (444, 568)
389
435 (375, 495)
216
98 (77, 120)
P value for interaction
...
<0.01
...
<0.01
...
<0.01
No. of extra cases because of interaction
...
151 (69, 233)
...
178 (103, 253)
...
54 (28, 80)
CI indicates confidence interval; Ref, reference category; and RD, rate difference.
*RD per 100 000 person-years at risk adjusted for age, cohort study, and birth cohort.
†RD per 100 000 person-years at risk adjusted for sex, age, cohort study, and birth cohort.
because we did not find any indication of different effects across
these subgroups. Second, during the past decade, the diagnostic
strategy and management of patients with stroke have undergone important changes, and improved hypertension control
might have contributed to a reduction in stroke incidence.28,29
Therefore, we controlled for such potential calendar effects by
adjusting for birth cohort in 5-year intervals. Third, we were able
to control for several important potential confounders, but we
are aware that these may not be sufficient to avoid residual confounding. Because our focus in this study was initiated by a specific interest in mechanisms driving social inequality in stroke,
we defined both behavioral risk factors (eg, alcohol intake and
physical activity) and biological risk factors (eg, diabetes mellitus and atrial fibrillation) as intermediate variables rather than
Figure. Rate difference per 100 000 person-years
at risk of ischemic stroke (2005 cases in men and
1608 cases in women) associated with 3-way combined effects of education, smoking, and hypertension, among 68 643 men and women from the
Social Inequality in Cancer cohort adjusted for age,
cohort study, and birth cohort.
Nordahl et al Combined Effects of Risk Factors on Stroke 2587
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
confounders of the education–stroke relationship. Fourth, the
observed interactions between education and smoking in this
study may be partly explained by the crude categorization of
the smoking variable, which might result in an interaction if the
residual exposure to smoking within each category depends on
education. Another explanation of the interaction is educationdependent under-reporting of smoking or smoking- and education-dependent residual confounding by unobserved variables.
However, our sensitivity analysis showed that the observed interaction was robust to the use of a 4-category smoking variable
tested with different cutoff points for heavy smoking.
These caveats considered, the public health message of the
current study is that targeted interventions with a given reduction in smoking among people from lower socioeconomic
groups would yield a greater reduction in ischemic stroke than
interventions targeting people from higher socioeconomic
groups. Similarly, universal interventions (eg, legislation or
taxes) with the same reduction of smokers in both groups
would have the strongest effect on ischemic stroke in people
from lower socioeconomic groups.
Conclusions
We showed that the combined effect of socioeconomic position and smoking exceeded the sum of their separate absolute
effect on risk of ischemic stroke, particularly among men. We
also found evidence of an interaction between smoking and
hypertension on risk of both ischemic and hemorrhagic stroke.
Because both the prevalence and the effect of smoking and
hypertension are socially skewed, reducing smoking in lower
socioeconomic group and in those with hypertension may help
reduce social inequality in risk of stroke.
Acknowledgments
We thank the collaborators behind the Social Inequality in Cancer
Cohort Study: the Copenhagen City Heart Study, the Diet, Cancer and
Health Study, and the cohorts at the Research Center for Prevention
and Health.
Sources of Funding
This work was supported by the Danish Cancer Society, Commission
of Social Inequality in Cancer (grant number SU08004).
Disclosures
None.
References
1. Addo J, Ayerbe L, Mohan KM, Crichton S, Sheldenkar A, Chen R,
et al. Socioeconomic status and stroke: an updated review. Stroke.
2012;43:1186–1191.
2. Cox AM, McKevitt C, Rudd AG, Wolfe CD. Socioeconomic status and
stroke. Lancet Neurol. 2006;5:181–188.
3.Lindekleiv H, Sandvei MS, Romundstad PR, Wilsgaard T, Njølstad I,
Ingebrigtsen T, et al. Joint effect of modifiable risk factors on the risk of aneurysmal subarachnoid hemorrhage: a cohort study. Stroke. 2012;43:1885–1889.
4. Liu L, Xue F, Ma J, Ma M, Long Y, Newschaffer CJ. Social position and
chronic conditions across the life span and risk of stroke: a life course
epidemiological analysis of 22,847 American adults in ages over 50. Int
J Stroke. 2013;8(suppl A100):50–55.
5. Shaper AG, Phillips AN, Pocock SJ, Walker M, Macfarlane PW. Risk factors for stroke in middle aged British men. BMJ. 1991;302:1111–1115.
6. Diderichsen F, Hallqvist J. Social inequality in health: some methodological considerations for the study of social position and social context.
In: Arve-Páres B, ed. Inequality in Health—A Swedish Perspective.
Strockholm: Swedish Council for Social Research; 1998:25–39.
7. Greenland S. Interactions in epidemiology: relevance, identification, and
estimation. Epidemiology. 2009;20:14–17.
8. Rothman KJ, Greenland S, Walker AM. Concepts of interaction. Am J
Epidemiol. 1980;112:467–470.
9. Avendano M, Kawachi I, Van Lenthe F, Boshuizen HC, Mackenbach JP, Van
den Bos GA, et al. Socioeconomic status and stroke incidence in the US elderly:
the role of risk factors in the EPESE study. Stroke. 2006;37:1368–1373.
10. Kerr GD, Slavin H, Clark D, Coupar F, Langhorne P, Stott DJ. Do vascular risk factors explain the association between socioeconomic status and
stroke incidence: a meta-analysis. Cerebrovasc Dis. 2011;31:57–63.
11. McFadden E, Luben R, Wareham N, Bingham S, Khaw KT. Social class,
risk factors, and stroke incidence in men and women: a prospective study
in the European prospective investigation into cancer in Norfolk cohort.
Stroke. 2009;40:1070–1077.
12. Kuper H, Adami HO, Theorell T, Weiderpass E. The socioeconomic
gradient in the incidence of stroke: a prospective study in middle-aged
women in Sweden. Stroke. 2007;38:27–33.
13. Yamagishi K, Iso H, Kitamura A, Sankai T, Tanigawa T, Naito Y, et al.
Smoking raises the risk of total and ischemic strokes in hypertensive
men. Hypertens Res. 2003;26:209–217.
14. Baldassarre D, Castelnuovo S, Frigerio B, Amato M, Werba JP, De Jong
A, et al. Effects of timing and extent of smoking, type of cigarettes, and
concomitant risk factors on the association between smoking and subclinical atherosclerosis. Stroke. 2009;40:1991–1998.
15. Janzon E, Hedblad B, Berglund G, Engström G. Tobacco and myocardial
infarction in middle-aged women: a study of factors modifying the risk. J
Intern Med. 2004;256:111–118.
16. Nordahl H, Hvidtfeldt UA, Diderichsen F, Rod NH, Osler M, Frederiksen
B, et al. Cohort profile: the Social Inequality in Cancer (SIC) cohort
study [published online ahead of print]. Int J Epidemiol. 2014. http://
ije.oxfordjournals.org/content/early/2014/02/17/ije.dyu003.abstract.
Accessed February 18, 2014.
17. Krarup LH, Boysen G, Janjua H, Prescott E, Truelsen T. Validity of
stroke diagnoses in a National Register of Patients. Neuroepidemiology.
2007;28:150–154.
18. Hallqvist J, Ahlbom A, Diderichsen F, Reuterwall C. How to evaluate
interaction between causes: a review of practices in cardiovascular epidemiology. J Intern Med. 1996;239:377–382.
19. Rod NH, Lange T, Andersen I, Marott JL, Diderichsen F. Additive
interaction in survival analysis: use of the additive hazards model.
Epidemiology. 2012;23:733–737.
20. Lange T, Hansen JV. Direct and indirect effects in a survival context.
Epidemiology. 2011;22:575–581.
21. Martinussen T, Scheike TH. Dynamic Regression Models for Survival
Data. New York: Springer; 2006.
22. Li C, Hedblad B, Rosvall M, Buchwald F, Khan FA, Engström G. Stroke
incidence, recurrence, and case-fatality in relation to socioeconomic
position: a population-based study of middle-aged Swedish men and
women. Stroke. 2008;39:2191–2196.
23. Nordahl H, Lange T, Osler M, Diderichsen F, Andersen I, Prescott E, et
al. Education and cause-specific mortality: the mediating role of differential exposure and vulnerability to behavioral risk factors. Epidemiology.
2014;25:389–396.
24. Diderichsen F, Andersen I, Manuel C, Andersen AM, Bach E, Baadsgaard
M, et al; Working Group of Danish Review on Social Determinants of
Health. Health inequality–determinants and policies. Scand J Public
Health. 2012;40(suppl 8):12–105.
25. Galobardes B, Costanza MC, Bernstein MS, Delhumeau C, Morabia A.
Trends in risk factors for lifestyle-related diseases by socioeconomic
position in Geneva, Switzerland, 1993-2000: health inequalities persist.
Am J Public Health. 2003;93:1302–1309.
26. Colhoun HM, Hemingway H, Poulter NR. Socio-economic status and
blood pressure: an overview analysis. J Hum Hypertens. 1998;12:91–110.
27. Galobardes B, Lynch JW, Davey Smith G. Childhood socioeconomic circumstances and cause-specific mortality in adulthood: systematic review
and interpretation. Epidemiol Rev. 2004;26:7–21.
28. Johnsen SP, Overvad K, Sørensen HT, Tjønneland A, Husted SE. Predictive
value of stroke and transient ischemic attack discharge diagnoses in The
Danish National Registry of Patients. J Clin Epidemiol. 2002;55:602–607.
29. Thorvaldsen P, Davidsen M, Brønnum-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.
Combined Effects of Socioeconomic Position, Smoking, and Hypertension on Risk of
Ischemic and Hemorrhagic Stroke
Helene Nordahl, Merete Osler, Birgitte Lidegaard Frederiksen, Ingelise Andersen, Eva Prescott,
Kim Overvad, Finn Diderichsen and Naja Hulvej Rod
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
Stroke. 2014;45:2582-2587; originally published online August 14, 2014;
doi: 10.1161/STROKEAHA.114.005252
Stroke is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 2014 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/45/9/2582
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/