Contribution of Obesity and Abdominal Fat Mass to Risk of Stroke and Transient

Ischemic Attacks
Yaroslav Winter, Sabine Rohrmann, Jakob Linseisen, Oliver Lanczik, Peter A. Ringleb,
Johannes Hebebrand and Tobias Back
Stroke. 2008;39:3145-3151; originally published online August 14, 2008;
doi: 10.1161/STROKEAHA.108.523001
Stroke is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright 2008 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/39/12/3145

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 September 16, 2012

Contribution of Obesity and Abdominal Fat Mass to Risk of

Stroke and Transient Ischemic Attacks
Yaroslav Winter, MD; Sabine Rohrmann, PhD; Jakob Linseisen, PhD; Oliver Lanczik, MD;
Peter A. Ringleb, MD; Johannes Hebebrand, MD; Tobias Back, MD
Background and PurposeWaist circumference has been shown to be a better predictor of cardiovascular risk than body
mass index (BMI). Our case-control study aimed to evaluate the contribution of obesity and abdominal fat mass to the
risk of stroke and transient ischemic attacks (TIA).
MethodsWe recruited 1137 participants: 379 cases with stroke/TIA and 758 regional controls matched for age and sex.
Associations between different markers of obesity (BMI, waist-to-hip ratio, waist circumference and waist-to-stature
ratio) and risk of stroke/TIA were assessed by using conditional logistic regression adjusted for other risk factors.
ResultsBMI showed a positive association with cerebrovascular risk which became nonsignificant after adjustment for
physical inactivity, smoking, hypertension, and diabetes (odds ratio 1.18; 95% CI, 0.77 to 1.79, top tertile versus bottom
tertile). Markers of abdominal adiposity were strongly associated with the risk of stroke/TIA. For the waist-to-hip ratio,
adjusted odds ratios for every successive tertile were greater than that of the previous one (2nd tertile: 2.78, 1.57 to 4.91;
3rd tertile: 7.69, 4.53 to 13.03). Significant associations with the risk of stroke/TIA were also found for waist
circumference and waist-to-stature ratio (odds ratio 4.25, 2.65 to 6.84 and odds ratio 4.67, 2.82 to 7.73, top versus
bottom tertile after risk adjustment, respectively).
ConclusionsMarkers of abdominal adiposity showed a graded and significant association with risk of stroke/TIA,
independent of other vascular risk factors. Waist circumference and related ratios can better predict cerebrovascular
events than BMI. (Stroke. 2008;39:3145-3151.)
Key Words: stroke transient ischemic attack obesity body mass index waist circumference risk factors

besity has become one of the most prevalent conditions

making a significant impact on public health worldwide.
In the United States, 65.7% of adults are either overweight or
obese, and 30.4% are obese.1 In Germany, currently 49.6% of
inhabitants are overweight, among those 13.6% are obese.2
The unfavorable association of obesity with coronary heart
disease3 and myocardial infarction4 is well recognized. Largescale prospective studies have documented that abdominal
obesity measured by waist-to-hip ratio (WHR) is more
strongly associated with cardiovascular risk than body mass
index (BMI).4,5 However, the relationship between increased
relative body weight and stroke risk is controversial. There
are studies showing that increasing BMI is associated with a
graded elevated risk of stroke.6,7 In other studies, however, no
relation was found between BMI and stroke risk.8 10 Possibly, BMI is not an appropriate indicator to assess the risk of
stroke.11 Markers of abdominal obesity have rarely been
studied in cerebrovascular disease. In 2 of those studies,
WHR was more strongly associated with the risk of ischemic

stroke than BMI, but the strength of this association was

attenuated after adjustment for cardiovascular risk factors.11,12 Other studies included small numbers of cases13,14 or
concentrated on cardiovascular risk.4 Thus, data on the role of
abdominal obesity for stroke are limited and completely
lacking for transient ischemic attacks (TIA), which represent
important cerebrovascular events often preceeding major
strokes.15 In order to provide evidence for the impact of
body-fat distribution on the risk of stroke and TIA, we
conducted a case-control study in a well-defined population
of central Western Europe.

Patients and Methods

The study included 1137 participants (379 cases and 758 controls).
Consecutive cases of ischemic stroke (n301, 79%), intracerebral
hemorrhage (n37, 10%) or TIAs (n41, 11%) were recruited in the
Departments of Neurology of the Klinikum Mannheim and Klinikum
Heidelberg between February 1, 2005 and January 31, 2006. Cases
with prior cerebrovascular events were not excluded. Of 401 initially
recruited cases, detailed clinical records were unavailable for 22

Received April 15, 2008; final revision received May 6, 2008; accepted May 20, 2008.
From the Department of Neurology (Y.W., O.L., T.B.), Klinikum Mannheim, University of Heidelberg, Germany; the Division of Cancer
Epidemiology (S.R., J.L.), German Cancer Research Center, Heidelberg, Germany; the Department of Child and Adolescent Psychiatry (J.H.), Rheinische
Kliniken, University of Duisburg-Essen, Germany; the Department of Neurology (P.A.R.), Klinikum Heidelberg, University of Heidelberg, Germany; the
Department of Neurology (T.B.), Saxon Hospital Arnsdorf, Arnsdorf/Dresden, Germany; and the Center for Mental Health (Y.W.), Klinikum Stuttgart,
Germany.
Correspondence to Prof Tobias Back, MD, Department of Neurology, Saxon Hospital Arnsdorf, Hufelandstr. 15, D-01477 Arnsdorf/Dresden, Germany.
E-mail tobias.back@skhar.sms.sachsen.de
2008 American Heart Association, Inc.
Stroke is available at http://stroke.ahajournals.org

DOI: 10.1161/STROKEAHA.108.523001

3145
Downloaded from http://stroke.ahajournals.org/
by guest on September 16, 2012

3146

Stroke

December 2008

patients, who were therefore excluded from analysis. Each index

patient was matched with 2 controls without a history of cerebrovascular disease. The study was approved by the local ethics
committee and all patients gave informed consent.
Each patient received a physical and neurological examination,
CT and/or MRI of the head. Stroke was defined according to the
World Health Organization (WHO).16 The obesity phenotype was
characterized by anthropometric measures, such as BMI, WHR and
waist circumference. BMI was calculated as weight in kilograms
divided by height in meters squared.17 WHR was defined as waist
divided by hip circumference.11 Waist circumference was measured
in centimeters at the level of the umbilicus,18 hip circumference at
the level of the bilateral greater trochanters.19 The role of body height
was also investigated by using the waist-to-stature ratio (WSR),
defined as waist circumference divided by body height.20 The
anthropometric measurements were performed in less than 48 hours
after admission.
We used threshold categories for obesity measures defined by
expert groups from the WHO.19 In BMI categories, we distinguished
between normal weight (BMI 25.0 kg/m2) and overweight (BMI

Table 1.

25.0 kg/m2), including preobesity (BMI 25.0 to 29.9 kg/m2) and

obesity (BMI 30 kg/m2). Obese women had WHR 0.85 and
obese men WHR 1.0. Threshold categories for waist circumference
in men were 94.0 cm (normal weight), 94.0 to 101.9 cm (overweight) and 102.0 cm (obesity). In women they were 80.0 cm,
80.0 to 87.9 cm and 88.0 cm, respectively.21 Hypertension was
defined as systolic blood pressure 140 mm Hg and/or diastolic
blood pressure 90 mm Hg or treatment with antihypertensive
agents. Hyperlipidemia was defined as total serum cholesterol level
240 mg/dL or use of antihyperlipidemic agents. Diabetes was
defined as fastening blood glucose 126 mg/dL or use of insulin or
oral hypoglycemic agents.
Regional controls were matched for age and sex from a database
with 25 540 participants of the population-based cohort study EPICHeidelberg. There were no individuals younger than 45 or older than
75 years among controls. The age of cases in our study cohort ranged
from 25 to 90 years. Age groups 50, 50 to 55, 56 to 60, 61 to 65,
66 to 70 and 70 were used for matching. Cases younger than 45
years (5.3%, n20) were matched within the age group 50 years;
cases older than 75 (24.3%, n92) were matched within the age

Descriptive Variables Comparing Cases and Controls Stratified by Gender

Men
Controls

Cases
SD

Mean

Women

Mean

Controls
SD

P Value*

Mean

Cases
SD

Mean

SD

P Value*

All cases and controls

N

476

238

282

141

Age (years)

65.0

7.6

66.9

10.9

0.017

65.0

9.4

68.0

14.1

0.020

Weight (kg)

83.4

12.3

84.7

12.7

0.172

69.0

11.8

74.8

18.4

0.001

Height (cm)

175.5

7.0

173.9

7.1

0.005

162.8

5.8

161.7

6.4

0.097

96.9

10.1

103.6

11.1

.001

83.8

11.6

98.2

15.5

.001

102.2

7.0

103.5

10.6

0.091

102.1

8.5

105.3

12.1

Waist circ. (cm)

Hip circ. (cm)
WHR
BMI
WSR

0.95

0.06
3.7

27.1
0.55

0.06

1.00
28.0
0.60

0.08

.001

3.7

0.002

0.07

.001

0.82
26.1
0.52

0.09
4.5
0.07

0.93
28.6
0.61

0.10

0.004
.001

6.5

0.001

0.10

.001

PIN (% yes)

44.3

82.5

.001

52.1

87.2

.001

Smoke (% yes)

15.8

30.2

.001

12.1

24.8

0.001

AHTN (% yes)

53.4

84.3

.001

50.0

80.4

.001

Diabetes (% yes)

11.8

30.9

.001

7.1

30.7

.001

HLP (% yes)

56.7

50.4

0.111

51.8

53.9

0.680

Only cases aged 4575 years and 2 controls per case

N

352

176

162

81

Age (years)

64.0

7.0

64.3

7.4

0.688

63.6

8.3

63.9

8.7

0.853

Weight (kg)

83.1

12.4

85.5

12.8

0.041

70.0

12.2

76.7

19.0

0.005

Height (cm)

175.9

6.9

174.1

6.9

0.006

163.8

5.8

162.7

5.8

0.192

96.2

9.8

104.2

10.9

.001

83.8

11.8

100.4

16.3

.001

102.0

6.8

103.3

9.1

0.083

102.6

8.6

105.9

13.6

0.08

.001

3.8

0.001

0.06

.001

Waist circ. (cm)

Hip circ. (cm)
WHR
BMI
WSR

0.94

0.06
3.8

26.9
0.55

0.06

1.01
28.2
0.60

0.81
26.1
0.51

0.10
4.7
0.07

0.95
28.9
0.62

0.052

0.10

.001

6.7

0.001

0.10

.001

PIN (% yes)

40.3

80.9

.001

51.8

86.4

Smoke (% yes)

15.9

34.7

.001

14.2

34.6

.001
0.001

AHTN (% yes)

51.1

85.7

.001

48.8

81.3

.001

Diabetes (% yes)

11.4

31.2

.001

5.6

30.9

.001

HLP (% yes)

56.5

54.0

0.577

48.8

56.8

0.238

*t test for continuous variables; 2 test for categorial variables.

Circ indicates circumference; AHTN, arterial hypertension; HLP, hyperlipidemia; PIN, physical inactivity.

Downloaded from http://stroke.ahajournals.org/ by guest on September 16, 2012

Winter et al
Table 2.

Abdominal Obesity and Stroke Risk

Associations Between Anthropometric Variables and Stroke Using All Cases and Controls
All Cases and Controls
Cases/
Controls

Model 1
OR (95% CI)

Model 2
OR (95% CI)

86/252

1.00 (reference)

1.00 (reference)

128/254

1.52 (1.05, 2.20)

1.42 (0.95, 2.12)

165/252

2.34 (1.63, 3.34)

Waist

40/248

Circ*

2
3
1
2

WHR*

WSR*

Only Cases Aged 45 to 75 and 2 Controls Per Case

Cases/Controls

Model 1
OR (95% CI)

Model 2
OR (95% CI)

1.00 (reference)

53/188

1.00 (reference)

1.00 (reference)

1.00 (reference)

1.17 (0.77, 1.79)

83/164

1.81 (1.21, 2.71)

1.79 (1.14, 2.79)

1.49 (0.93, 2.40)

1.86 (1.26, 2.74)

1.18 (0.77, 1.79)

121/162

2.67 (1.81, 3.93)

2.11 (1.38, 3.24)

1.37 (0.86, 2.17)

1.00 (reference)

1.00 (reference)

1.00 (reference)

24/176

1.00 (reference)

1.00 (reference)

1.00 (reference)

89/266

2.32 (1.47, 3.65)

2.23 (1.38, 3.60)

1.98 (1.21, 3.26)

63/178

2.71 (1.62, 4.55)

2.73 (1.57, 4.73)

2.26 (1.28, 4.01)

250/244

7.13 (4.65, 10.94)

5.83 (3.71, 9.16)

4.25 (2.65, 6.84)

170/160

8.34 (5.12, 13.56)

7.04 (4.19, 11.83)

4.84 (2.81, 8.34)

29/263

1.00 (reference)

1.00 (reference)

1.00 (reference)

16/188

1.00 (reference)

1.00 (reference)

1.00 (reference)

78/229

3.66 (2.17, 6.19)

3.34 (1.93, 5.80)

2.78 (1.57, 4.91)

48/158

3.67 (2.01, 6.72)

3.25 (1.73, 6.10)

2.59 (1.35, 4.95)

272/266

12.78 (7.83, 20.86)

10.11 (6.07, 16.83)

7.69 (4.53, 13.03)

193/168

14.56 (8.34, 25.45)

10.86 (6.09, 19.34)

7.91 (4.35, 14.36)

35/246

1.00 (reference)

1.00 (reference)

1.00 (reference)

20/179

1.00 (reference)

1.00 (reference)

85/241

2.99 (1.85, 4.84)

2.68 (1.61, 4.45)

2.39 (1.42, 4.04)

59/168

3.42 (1.96, 5.96)

3.05 (1.70, 5.46)

2.56 (1.40, 4.69)

259/271

8.66 (5.50, 13.64)

6.67 (4.13, 10.75)

4.67 (2.82, 7.73)

178/167

10.95 (6.48, 18.49)

8.46 (4.88, 14.66)

5.60 (3.14, 10.00)

Tertile*
BMI*

3147

Model 3
OR (95% CI)

Model 3
OR (95% CI)

1.00 (reference)

*Cutpoints for tertiles. Men: waist circumference (cm) 91.8; 100.0; BMI (kg/m2) 25.19; 28.09; WHR 0.92; 0.97; WSR 0.52; 0.57.
Women: waist circumference (cm) 76.5; 89.5; BMI (kg/m2) 23.83; 27.45; WHR 0.78; 0.84; WSR 0.475; 0.54.
Model 1: matched for age and sex; Model 2: matched for age and sex and adjusted for physical inactivity, smoking; Model 3: matched for age and sex and adjusted
for physical inactivity, smoking, history of hypertension, history of diabetes.
Circ indicates circumference.

group 70 years. The plausibility proof was performed by analyzing

only the exactly matched set of cases and controls aged 45 to 75
years.

Statistical Analysis
Statistical analysis was performed with SAS version 9.1 (SAS
Institute Inc). In the univariate analysis, categorical variables were
compared by 2 test. Continuous variables expressed as meanSD
were compared by the t test. Conditional logistic regression models
were used to calculate the odds ratio (OR) and 95% CI for BMI,
WHR, WSR, and waist circumference with stratification by sex and
age groups as described above. Adjustment was performed for the
following stroke risk factors: arterial hypertension (yes/no), diabetes
mellitus (yes/no), smoking (smoker during previous 5 years; yes/no)
and physical inactivity (at least 2 hours of physical activity per week;
yes/no). Hyperlipidemia was not statistically significant in the
univariate analysis and was, thus, not included in the conditional
logistic regression model.
Three approaches were chosen to assess the role of obesity
markers in predicting the risk of stroke/TIA: first, comparisons of
ORs across the tertiles of BMI, WHR, WSR and waist circumference, respectively, using the bottom tertile as a reference category;
second, estimation of the OR for 1 SD change in BMI, WHR, WSR
or waist circumference; and third, comparisons of the receiveroperator curves (ROC) in relation to stroke or TIA for obesity
measures. The ROC is a plot of test sensitivity versus its falsepositive rate (or 1specificity). The area under the ROC is a measure
of the accuracy of a diagnostic test. A test with an area under the
ROC of 1.0 is perfectly accurate, and a test with the accuracy of 50%
(random guessing) has an area of 0.5, and a test with an area of 0.0
is completely inaccurate. ROCs were compared by using the method
of DeLong and coworkers.22

Results
A total of 379 patients with stroke (n338) or TIA (n41)
and 758 age- and sex-matched regional controls were evaluated. Of these cases, 37.2% (n141) were female. The mean
age of controls was slightly lower than the age of cases
(65.08.3 versus 67.312.2, P0.02) because there were no
individuals older than 75 years among the controls. Table 1
shows the demographics and distribution of stroke risk
factors in the study population stratified by sex. Patients aged

45 to 75 years were separated within the study cohort for the

reasons mentioned above. The prevalence of obesity, defined
by BMI, was higher in cases (29.8%, n113) than in controls
(20.1%, n152; P0.01). Among cases, 24.5% (n93) had
a history of a previous stroke.
The results of conditional logistic regression analysis are
shown in Table 2. In a model adjusted for sex and age, BMI
showed a positive association with risk of stroke or TIA (OR
2.34; 95% CI, 1.63 to 3.34; P0.001, top versus bottom
tertile). After adjustment for other risk factors (models 2 and
3), this association was attenuated (model 3: OR 1.18; 95%
CI 0. 0.77 to 1.79) and lost its significance (P0.45; Figure).
The risk of stroke or TIA increased in a graded manner with
increasing WHR. In a model adjusted for sex and age,
patients in the highest tertile had a 12.78-fold (95% CI 7.83
to 20.86) elevated risk of cerebrovascular disease (P0.001)
compared with the lowest tertile (Table 2). This association
was attenuated after adjustment for other risk factors (models
2 and 3), but still remained significant (model 3: OR 7.69;
95% CI, 4.53 to 13.03; P0.001). It was consistent both in
men and women, with higher risks in the latter (Table 3). The
plausibility proof with cases/controls aged 45 to 75 years,
confirmed the strong association between WHR and stroke/
TIA risk (OR 7.91; 95% CI 4.35 to 14.36 top versus bottom
tertile, fully adjusted model).
Increased waist circumference was also related to higher
risk of stroke or TIA. In the highest tertile group this risk was
7.13-fold (95% CI 4.65 to 10.94; P0.001) compared with
the bottom tertile (Table 2). This strong positive association
remained significant after risk adjustment (model 3: OR 4.25;
95% CI 2.65 to 6.84; P0.001). The plausibility proof with
patients aged 45 to 75 years confirmed the strong association
between waist circumference and risk of stroke or TIA
(model 3: OR 4.84; 95% CI, 2.81 to 8.34; P0.001).
Cases were on average 1.4 cm shorter than controls
(169.369.03 cm versus 170.748.94 cm; P0.05; Table

Downloaded from http://stroke.ahajournals.org/ by guest on September 16, 2012

3148

Stroke

December 2008

100

Odds Ratio

10

Waist circumference
(tertile)

0.1
Cases
Controls

40
248

89
266

250
244

WSR
(tertile)
35
246

85
241

WHR
(tertile)
259
259

29
263

78
229

BMI
(tertile)
272
266

87
252

128
254

165
252

black: matched for age and sex; white: adjusted for stroke risk factors

Figure. 1 Association of specific measures of obesity with risk of stroke and TIA. Vertical bars indicate 95% CI; WSR, waist-to-stature
ratio; WHR, waist-to-hip ratio; BMI, body-mass index. Filled symbols indicate matched for age and sex; open symbols, adjusted for
physical inactivity, smoking, history of hypertension, and history of diabetes.

1), but after adjusting for other covariables, the results of the
logistic regression models were not statistically significant
(data not shown). Body height is a component of WSR, which
is another marker of abdominal obesity. In a model adjusted
for sex and age, increased WSR was associated with an
elevated cerebrovascular risk (OR 8.66; 95% CI 5.50 to
Table 3.

OR of Stroke/TIA Risk Between Threshold Categories of Specific Obesity Measures Using the Entire Study Cohort
Sex
2

BMI (kg/m )

Men

Women

Waist circumference (cm)

Men

Women

WHO Categories
25.0

Men

Model 1* OR (95% CI)

Model 2 OR (95% CI)

Model 3 OR (95% CI)

45/145

1.00

1.00

1.00

133/238

1.95 (1.26 to 3.03)

1.72 (1.07 to 2.76)

1.36 (0.82 to 2.25)

30.034.9

60/93

2.77 (1.52 to 4.24)

1.77 (1.02 to 3.08)

0.99 (0.55 to 1.80)

25.0

45/123

1.00

1.00

1.00

25.029.9

44/103

1.39 (0.77 to 2.51)

1.43 (0.75 to 2.70)

1.17 (0.60 to 2.28)

30.034.9

52/56

2.97 (1.60 to 5.52)

2.35 (1.19 to 4.64)

1.63 (0.78 to 3.37)

94.0

38/186

1.00

1.00

1.00

94.0101.9

60/160

2.35 (1.41 to 3.90)

2.09 (1.22 to 3.58)

1.80 (1.03 to 3.15)

102.0

3.71 (2.18 to 6.32)

140/130

6.22 (3.85 to 10.03)

5.07 (3.05 to 8.41)

80.0

16/120

1.00

1.00

1.00

80.087.9

15/56

1.48 (0.59 to 3.70)

1.11 (0.41 to 2.98)

1.29 (0.47 to 3.53)

110/106

7.53 (3.85 to 14.71)

5.69 (2.79 to 11.61)

4.49 (2.13 to 9.46)

1.0
1.0

Women

Cases/Controls

25.029.9

88.0
WHR

13.64; P0.001, top versus bottom tertile; Table 2). After

adjustment for risk factors (models 2 and 3) this relation
remained significant (model 3: OR 4.67; 95% CI 2.82 to 7.73;
P0.001).
We also compared the effect of 1 SD increase in different
obesity measures (Table 4). The increase in the OR calculated

0.85
0.85

103/377

1.00

1.00

1.00

135/99

5.81 (3.95 to 8.54)

4.68 (3.11 to 7.02)

4.13 (2.70 to 6.30)

39/199
102/83

1.00

1.00

1.00

9.93 (5.40 to 18.28)

9.56 (4.92 to 18.57)

7.77 (3.87 to 15.61)

*Model 1: matched for age and sex.

Model 2: matched for age and sex and adjusted for physical inactivity, smoking.
Model 3: matched for age and sex and adjusted for physical inactivity, smoking, history of hypertension, history of diabetes.

Downloaded from http://stroke.ahajournals.org/ by guest on September 16, 2012

Winter et al
Table 4.

Abdominal Obesity and Stroke Risk

3149

Comparative Effect of 1 SD Increase in a Specific Measure of Obesity

Obesity Marker

1 SD

Model 1* OR (95% CI)

Model 2 OR (95% CI)

Model 3 OR (95% CI)

3.75 (m), 5.33 (f)

1.45 (1.26 to 1.67)

1.30 (1.12 to 1.51)

1.07 (0.91 to 1.26)

10.90 (m), 14.68 (f)

2.46 (2.09 to 2.90)

2.21 (1.86 to 2.63)

1.92 (1.60 to 2.30)

Waist-to-hip ratio

0.08 (m), 0.10 (f)

3.50 (2.87 to 4.27)

3.03 (2.47 to 3.71)

2.92 (2.35 to 3.62)

Waist-to-stature ratio

0.07 (m), 0.09 (f)

2.68 (2.11 to 3.02)

2.42 (1.98 to 2.87)

2.01 (1.78 to 2.32)

BMI (kg/m2)
Waist circumference (cm)

*Model 1: matched for age and sex; Model 2: matched for age and sex and adjusted for physical inactivity, smoking; Model 3: matched for age and sex and
adjusted for physical inactivity, smoking, history of hypertension, history of diabetes.
m indicates male; f, female.

for 1 SD increase in WHR was the largest among all

anthropometric indices studied. A moderate increase in the
OR was found for waist circumference. The association
between WSR and risk of stroke or TIA was slightly weaker
than the association with WHR, but stronger than the association with waist circumference. The OR associated with 1
SD increase in BMI was the weakest one.
The area under the ROC of WHR (0.774) was slightly
larger than that of WSR (0.730) or waist circumference
(0.721), showing that the WHR was most accurate in predicting the risk of stroke or TIA. The smallest area under the
ROC was observed for BMI (0.595). All values presented
were significantly increased (P0.01). The differences between single values were statistically significant (P0.01),
except for the comparison between WSR and waist circumference (P0.09).
Regarding threshold categories for BMI, ORs were 1.95
for preobese men and 2.77 for obese men compared with
males of normal BMI (P0.01; Table 3). Corresponding
ORs in women were 1.39 (preobese females, P0.28) and
2.97 (obese females, P0.01). After adjustment for cerebrovascular risk factors, OR in BMI categories lost their
significance. With respect to threshold categories for
WHR, ORs adjusted for risk factors were 4.13 in men and
7.77 in women (P0.01 each). This gender-related difference was not significant in the multivariate model. For
waist circumference, ORs in the highest category were
strongly attenuated after adjustment for risk factors, but
still remained significant (3.71 in men, 4.49 in women,
P0.01 each).

Discussion
The present case-control study evaluated the predictive value
of different markers of adiposity and abdominal body fat for
stroke and TIA in a well-defined region of southwestern
Germany. Three different statistical approaches uniformly
showed that various markers of abdominal adiposity were
superior to the BMI in predicting the risk of stroke or TIA.
The waist-to-hip ratio served as the best predictor among the
obesity markers studied.
The association of abdominal obesity with increased atherosclerotic and cardiovascular risk has been shown in
previous studies.23,24 For example, a large international casecontrol study proved the superiority of abdominal body fat
markers compared to the BMI and demonstrated a strong
association with the risk of myocardial infarction.4 In our
study, comparable results were obtained for the stroke/TIA
risk by using a similar statistical approach. Complementary to

the findings of other investigators,4,25 the waist-to-hip ratio

had the strongest predictive value in our cohort of patients
with cerebrovascular disease. By contrast, BMI did not show
a consistent graded relation to stroke risk and became
nonsignificant after adjustment for other risk factors in both
models (with and without inclusion of arterial hypertension
and diabetes), which possibly mediate the association between BMI and stroke risk. In the case-control Northern
Manhattan Stroke Study, abdominal obesity increased the risk
for the ischemic type of stroke by factor 3.0.11 In contrast to
our findings, the association of WHR with stroke risk tended
to be stronger in men than in women (OR 3.8; 95% CI, 1.8 to
5.0 versus OR 2.5; 95% CI 1.6 to 4.0). However, hemorrhagic
stroke and TIA were excluded, as well as patients having a
history of stroke.11 In the longitudinal Atherosclerosis Risk in
Communities (ARIC) study,25 abdominal adiposity was associated with an increased risk for nonlacunar, but not for
lacunar stroke. This finding may result from the adjustment
for risk factors that are relevant for microangiopathic brain
lesions.
The results of studies investigating obesity as a risk factor
of hemorrhagic stroke are inconsistent. There are studies
demonstrating no association,26 decreasing risk6 or increasing
risk7 of hemorrhagic stroke with increasing BMI. In a
longitudinal study including 28 000 male US health professionals, age-adjusted relative risk of total stroke was 2.33
(95% CI 1.25 to 4.37) in a comparison between top and
bottom quintiles of WHR.14 A large-scale Scandinavian
cohort study found that abdominal obesity (measured by
WHR) was weakly associated with total stroke risk only in
men (OR 1.55; 95% CI 1.06 to 2.26, top versus bottom
tertiles), but not in women.12 TIA was not recorded as an
outcome event and participants were free of coronary heart
disease at baseline. This may help to explain, besides the
different study designs applied and populations studied, the
difference to our results. Case-control studieslike the present onein which obesity markers are measured closely to
the time point of the vascular event4,11 frequently show a
stronger association between (abdominal) obesity and risk for
stroke/TIA compared to longitudinal studies.10,12,14,25 Based
on the study design, the predictive value of case-control
studies may focus on a short-term perspective, the one of
longitudinal studies more on a long-term prediction.
In comparison, our study detected a strong and graded
association of abdominal fat markers with the risk of ischemic and hemorrhagic stroke or TIA for both genders. The
risks presented here tend to be higher than in previous reports,
possibly because of the inclusion of TIA as an important

Downloaded from http://stroke.ahajournals.org/ by guest on September 16, 2012

3150

Stroke

December 2008

cerebrovascular event.15 The inclusion of patients with prior

history of cerebrovascular disease may also account for this
difference. However, the consistent trend of higher risks in
females compared to males should be regarded with some
caution. Female cases were slightly under-represented in the
cohort and more than 1 year older on average compared to
male patients. In comparison to the corresponding controls,
female patients tended to present with a higher degree of
obesity than the male counterparts which may have altered
their overall vascular risk.
There is growing evidence that the waist-to-stature ratio, a
marker of abdominal obesity including body height, may
serve as a reliable predictor of cardiovascular risk. Recent
data4,27 show that the WSR is a weaker indicator of an
increased risk of coronary heart disease than the WHR, but at
least as strong as waist circumference and stronger than BMI.
Ho et al28 considered WSR to be even the best predictor of
cardiovascular risk among other simple anthropometric indices. Our results underline that WSR is an appropriate measure
to assess the risk of stroke and TIA comparable to waist
circumference, but further studies are needed to clarify which
marker is the most robust to predict total or subtype specific
cerebrovascular risks.
There are controversies concerning the impact of body
height in cardiovascular and cerebrovascular disease. Some
authors have described an association between body height
and cardiovascular risk.29 Others have reported their inverse
relation and observed no association between height and risk
of stroke.30 Walker et al14 reported that taller male health
professionals tended to carry a lower risk of stroke. There are
data showing an increased incidence of fatal stroke in shorter
people.31 Interestingly, cases in our patient cohort were
significantly shorter than controls, but the results of the
logistic regression models were not statistically significant
after adjustment for other covariables.
There are several limitations in this study. First, very old
patients or patients with very severe strokes or global aphasia
were possibly not recruited because of unavailable informed
consent or limited capacity of stroke-unit care. Their inclusion may have modified the risk analysis. In the opposite
direction, the fact that nearly one quarter of cases reported on
former strokes may have led to an overestimation of the
stroke risk, because per definition the history of controls was
free of previous stroke. Second, controls included only
individuals aged 45 to 75 years limiting the range that was
available for an exact age match. In order to avoid a
systematic error of data, the plausibility proof was conducted
by using only the exact age match that confirmed all main
results as robust. Both effect modification by age and a biased
control selection cannot be fully excluded to explain the
trends to higher risk measures in the exact match cohort.
Although the source population for cases and controls was
identical, participants of the EPIC study potentially were
above average concerning their health-oriented lifestyle.
Third, the number of cases with intracerebral hemorrhage or
TIA was too small for a detailed subgroup analysis. Fourth,
we did not record dietary variables in our sample, which
could represent potential confounders. Not only normal body
weight, but also a healthy diet may prevent stroke.32

Summary
Markers of abdominal adiposity showed a graded and significant association with risk of stroke and TIA, independent of
other vascular risk factors. The redefinition of obesity based
on the waist-to-hip ratio or waist circumference instead of
BMI increases considerably the estimate of cerebrovascular
events attributable to obesity. There is a trend for women to
be more strongly affected, which needs further investigation.
Waist circumference and related ratios, such as waist-to-hip
ratio and waist-to-stature ratio, can better predict cerebrovascular events than BMI in a population of central Western
Europe.

Sources of Funding
This study was funded by the German Ministry of Education and
Research (BMBF)/National Genome Research Network (NGFN)
research grant 01GS0491 (to project leader: T.B.).

Disclosures
None.

References
1. Hedley AA, Ogden CL, Johnson CL, Carroll MD, Curtin LR, Flegal KM.
Prevalence of overweight and obesity among us children, adolescents,
and adults, 1999 2002. JAMA. 2004;291:28472850.
2. Beneke A, Vogel H. Overweight and obesity. National Health Survey;16.
Berlin: Robert Koch Institute, 2005.
3. Hubert HB, Feinleib M, McNamara PM, Castelli WP. Obesity as an
independent risk factor for cardiovascular disease: a 26-year follow-up of
participants in the Framingham Heart Study. Circulation. 1983;67:
968 977.
unpuu S, Bautista L, Franzosi MG, Commerford P,
4. Yusuf S, Hawken S, O
Lang CC, Rumboldt Z, Onen CL, Lisheng L, Tanomsup S, Wangai JP,
Razak F, Sharma AM, Anand SS; on behalf of the INTERHEART Study
Investigators. Obesity and the risk of myocardial infarction in 27 000
participants from 52 countries: a case-control study. Lancet. 2005;366:
1640 1649.
5. Rimm EB, Stampfer MJ, Giovannucci E, Ascherio A, Spiegelman D,
Colditz GA, Willett WC. Body size and fat distribution as predictors of
coronary heart disease among middle-aged and older US men. Am J
Epidemiol. 1995;141:11171127.
6. Rexrode KM, Hennekens CH, Willett WC, Colditz GA, Stampfer MJ,
Rich-Edwards JW, Speizer FE, Manson JE. A prospective study of body
mass index, weight change, and risk of stroke in women. JAMA. 1997;
277:1539 1545.
7. Kurth T, Gaziano JM, Berger K, Kase CS, Rexrode KM, Cook NR,
Buring JE, Manson JE. Body mass index and the risk of stroke in men.
Arch Intern Med. 2002;162:25572562.
8. Khaw KT, Barrett-Connor E, Suarez L, Criqui MH. Predictors of strokeassociated mortality in the elderly. Stroke. 1984;15:244 248.
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. Lu M, Ye W, Adami HO, Weiderpass E. Prospective study of body size
and risk for stroke amongst women below age 60. J Intern Med. 2006;
260:442 450.
11. Suk SH, Sacco RL, Boden-Albala B, Cheun JF, Pittman JG, Elkind MS,
Paik MC. Abdominal obesity and risk of ischemic stroke: the Northern
Manhattan Stroke Study. Stroke. 2003;34:1586 1592.
12. Hu G, Tuomilehto J, Silventoinen K, Sarti C, Mannisto S, Jousilahti P.
Body mass index, waist circumference, and waist-hip ratio on the risk of
total and type-specific stroke. Arch Intern Med. 2007;167:1420 1427.
13. Lapidus L, Bengtsson C, Larsson B, Pennert K, Rybo E, Sjostrom L.
Distribution of adipose tissue and risk of cardiovascular disease and
death: a 12 year follow up of participants in the population study of
women in Gothenburg, Sweden BMJ. 1984;289:12571261.
14. Walker SP, Rimm EB, Ascherio A, Kawachi I, Stampfer MJ, Willett WC.
Body size and fat distribution as predictors of stroke among US men.
Am J Epidemiol. 1996;144:11431150.

Downloaded from http://stroke.ahajournals.org/ by guest on September 16, 2012

Winter et al
15. Daffertshofer M, Mielke O, Pullwitt A, Felsenstein M, Hennerici M.
Transient ischemic attacks are more than Ministrokes. Stroke. 2004;35:
24532458.
16. WHO MONICA Project Principal Investigators. The World Health Organization MONICA project (monitoring trends in cardiovascular disease):
a major international collaboration. J Clin Epidemiol. 1988;41:105114.
17. World Health Organization. Physical status: The use and interpretation of
anthropometry. Report of a WHO Expert Committee. Who Technical
Report Series; 854. Geneva: World health organization, 1995.
18. Lohman T, Roche AF, Martorell R. Anthropometric standardization
manual. Champaign, IL: Human Kinetics; 1998.
19. World Health Organization. Obesity: preventing and managing the global
epidemic. Report of a WHO consultation. WHO Technical Report Series;
894. Geneva: World Health Organization, 2000.
20. Bosy-Westphal A, Geisler C, Onur S, Korth O, Selberg O, Schrezenmeir
J, Muller MJ Value of body fat mass vs anthropometric obesity indices in
the assessment of metabolic risk factors. Int J Obes (Lond). 2006;30:
475 483.
21. Lean ME, Han TS, Morrison CE. Waist circumference as a measure for
indicating need for weight management. BMJ. 1995;311:158 161.
22. DeLong R, DeLong DM, Clarke-Pearson DL. Comparing the areas under
two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics. 1988;44:837 845.
23. Daniels SR, Morrison JA, Sprecher DL, Khoury P, Kimball TR. Association of body fat distribution and cardiovascular risk factors in children
and adolescents. Circulation. 1999;99:541545.

Abdominal Obesity and Stroke Risk

3151

24. See R, Abdullah SM, McGuire DK, Khera A, Patel MJ, Lindsey JB,
Grundy SM, de Lemos JA. The association of differing measures of
overweight and obesity with prevalent atherosclerosis: The Dallas Heart
Study. J Am Coll Cardiol. 2007;50:752759.
25. Ohira T, Shahar E, Chambless LE, Rosamond WD, Mosley TH Jr,
Folsom AR. Risk factors for ischemic stroke subtypes: the Atherosclerosis Risk in Communities Study. Stroke. 2006;37:24932498.
26. Jood K, Jern C, Wilhelmsen L, Rosengren A. Body mass index in mid-life
is associated with a first stroke in men: a prospective population study
over 28 years. Stroke. 2004;35:2764 2769.
27. Welborn TA, Dhaliwal SS. Preferred clinical measures of central obesity
for predicting mortality. Eur J Clin Nutr. 2007;61:13731379.
28. Ho SY, Lam TH, Janus ED. Waist to stature ratio is more strongly
associated with cardiovascular risk factors than other simple anthropometric indices. Ann Epidemiol. 2003;13:683 691.
29. Samaras TTE, H., Storms LH. Is short height really a risk factor for
coronary heart disease and stroke mortality? A review. Med Sci Monit.
2004;10.
30. Hebert PR, Rich-Edwards JW, Manson E. Height and incidence of cardiovascular disease of male physicians. Circulation. 1993;88:14371443.
31. Goldbourt U, Tanne D. Body height is associated with decreased
long-term stroke but not coronary heart disease mortality? Stroke. 2002;
33:743748.
32. Fung TT, Stampfer MJ, Manson JAE, Rexrode KM, Willett WC, Hu FB.
Prospective study of major dietary patterns and stroke risk in women.
Stroke. 2004;35:2014 2019.

Downloaded from http://stroke.ahajournals.org/ by guest on September 16, 2012