ORIGINAL ARTICLE

Racial and ethnic disparities in stroke subtypes: a multiethnicsample of patients with stroke

Jose Gutierrez • Sebastian Koch • Chuanhui Dong • Teresita Casanova •

Royya Modir • Michael Katsnelson • Gustavo A. Ortiz •

Ralph L. Sacco • Jose G. Romano • Tatjana Rundek

Received: 20 August 2013 / Accepted: 3 October 2013 / Published online: 15 October 2013

� Springer-Verlag Italia 2013

Abstract Stroke subtypes have been reported to differ by

race and ethnic subgroups and have not been adequately

explained. We aim to evaluate if the prevalence of vascular

risk factors accounts for differences observed in stroke

subtypes by race/ethnicity. Patients with acute stroke were

prospectively enrolled in the Miami Stroke Registry.

Patients’ demographic, clinical and radiological charac-

teristics were systematically collected. Stroke subtypes

were ascertained using TOAST criteria. The sample was

divided into Non-Hispanic Whites (NHW), Hispanics,

African Americans (AA), and Non-Hispanic Black Carib-

bean (NHBC). Univariable and multivariable logistic

regression analyses were performed to assess differences

among groups. Among 473 stroke patients (mean age

64 ± 14 years; 63.7 % were men) of which 52.9 % were

Hispanic, 22.6 % were AA, 13.5 % NHBC and 11.0 %

were NHW. Large artery atherosclerosis was more pre-

valent in NHBC (OR 1.74, 95 % CI 1.02–2.97) than in the

other groups. Adjusting for covariates rendered the asso-

ciation not significant (OR 1.71, 95 % CI 0.93–3.16).

Cardioembolism was more frequent in Hispanics (OR 1.94,

95 % CI 1.28–2.96) and NHW (OR 2.66, 95 % CI

1.42–4.96) as compared to NHBC and AA combined.

Adjusting for covariates, the association was no longer

significant for Hispanics but was further strengthened for

NHW (OR 3.02, 95 % CI 1.42–6.42). Our results suggest

that the vascular risk factors prevalence among different

racial and ethnic groups partially explains disparities found

in the prevalence of some stroke subtypes. Addressing

health disparities remains an important public health aspect

of stroke prevention.

Keywords African American � Atherothrombotic �Hispanic � Ischemic stroke � Stroke subtypes

Introduction

Effective preventive measures are needed to curb the

growing epidemic of vascular diseases in the world. Since

stroke is a heterogeneous condition, a better characteriza-

tion of stroke subtypes can lead to improved preventive

strategies [1]. Assuming that stroke subtypes occurred at a

different rate in certain population subgroups, then the

strategies to prevent stroke need to be tailored to each sub-

group to maximize efficacy.

A subject of interest and controversy has been the role

that race and ethnicity play in conferring different risks to

develop a particular stroke subtypes [2–4]. For example,

the Northern Manhattan Study (NOMAS) included three

J. Gutierrez

Department of Neurology, Columbia University Medical Center,

New York, NY, USA

S. Koch � C. Dong � T. Casanova � M. Katsnelson �G. A. Ortiz � R. L. Sacco � J. G. Romano � T. Rundek

Department of Neurology, University of Miami, Miami,

FL, USA

R. Modir

Department of Neurology, University of California San Diego,

San Diego, CA, USA

R. L. Sacco � T. Rundek (&)

Department of Epidemiology and Public Health, University

of Miami, Miami, FL, USA

e-mail: trundek@med.miami.edu

R. L. Sacco

Department of Human Genetics, University of Miami,

Miami, FL, USA

123

Neurol Sci (2014) 35:577–582

DOI 10.1007/s10072-013-1561-z

ethnic groups with, Hispanics having a higher age-adjusted

relative risk of cardioembolism (CE) and intracranial large

artery atherosclerosis (LAA) as compared to NHB or NHW

[1]. The strength of association between intracranial LAA

for both NHB and Hispanics attenuated after controlling

for the different proportion of risk factors [2]. The differ-

ential effect of race and ethnicity on stroke subtypes may

not be entirely explained by the presence of traditional risk

factors.

The objective of this study is to evaluate if in our sample,

adjusting for prevalence of vascular risk factors modifies

the association of race/ethnicity with stroke subtypes.

Methods

The Miami Stroke Registry at the University of Miami/

Jackson Memorial Hospital collects prospective data on

every consecutive patient with stroke (ischemic or hem-

orrhagic) and transient ischemic attack (TIA) admitted to

our hospital that provides informed consent. The data

included in this analysis were collected from November

2008 to February 2011. For this analysis, only participants

diagnosed with ischemic stroke were included. The local

Institutional Review Board approved the stroke registry.

The demographic information collected included age, sex,

race, ethnicity, and country of origin. Race and ethnicity

were based on self-identification. For the purpose of this

analysis, blacks who had the US as country of origin were

considered AA, non-Hispanic blacks who had a Caribbean

country of origin were assigned to the Non-Hispanic Black

Caribbean (NHBC) category, Hispanics remained as one

category independent of the listed race and NHW were

white participants born in the US. The reason behind

including NHBC as a different ethnic category was the

different vascular profile that we have found in this pop-

ulation in previous analysis [5].

Current smoking was defined as use of cigarettes within

the last 6 months. Current alcohol use was defined as two

or more drinks per day. Information about illicit drug use

was obtained from patient’s report, family member report

or results from urine toxicology screening. Hypertension

(HTN) was defined as a systolic blood pressure above

140 mmHg or diastolic blood pressure above 90 mmHg,

the self-report of having high blood pressure or the use of

antihypertensive drugs. Dyslipidemia was defined as hav-

ing an LDL [100 mg/dl, self-report of having high cho-

lesterol or the use of statins. Diabetes mellitus (DM) was

defined as having glycosylated hemoglobin above 6.5 %,

self-report of DM or the use of hypoglycemic drugs or any

form of insulin. Information about previous medical history

of atrial fibrillation (afib), angina, cardiac stent, myocardial

infarction, congestive heart failure, and peripheral vascular

disease was obtained from the patients or family members

and/or from the medical record corresponding to the hos-

pital admission. The NIHSS on admission was obtained

from the initial evaluation by the stroke team. All variables

were dichotomized.

The ischemic stroke subtypes were defined by the

TOAST criteria as LAA, CE, SVO, stroke of other deter-

mined etiology, stroke of undetermined etiology (crypto-

genic), and conflicting mechanisms [6]. The workup

thoroughness was deferred to the fellowship-trained vas-

cular neurologist who staffed the case, while the participant

was hospitalized. The mechanism of stroke was determined

a posteriori by the same or another fellowship-trained

vascular neurologist after reviewing all available infor-

mation from the medical records.

Statistical analysis

Analysis of variance (ANOVA) was used to compare eth-

nic groups and stroke subtypes with respect to age, and post

hoc comparisons were made with Bonferroni correction if

the overall P value was significant. Chi squared test was

used to evaluate overall categorical differences among

groups. Age was categorized to 65 years or older to be used

in logistic regression models. This cutoff corresponds to

the median of our population. Choosing a fixed referent

group might neglect important differences observed among

the compared groups. To maximize the identification of

important differences among all ethnic groups, we decided

to change the referent group based on the most dissimilar

prevalence rate of each of the studied characteristic, except

for age that was analyzed with ANOVA. Simple logistic

regression was used to estimate the OR, the confidence

intervals and the P value. To evaluate for confounding

effect, we created two models: Model 1 included age and

sex as covariates. Model 2 included model 1 plus vascular

risk factors as covariates. A P value of 0.05 or less was

considered statistically significant. The Statistical software

used for the analysis was PASW Statistics 18 (Release

18.0.0), Copyright IBM Corporation 2010.

Results

The sample studied included 473 patients (mean age

64 ± 14 years; 63.7 % were men) of which 52.9 % were

Hispanic (62.5 % from Cuba), 22.6 % were AA, 13.5 %

NHBC (89.7 % from Haiti) and 11.0 % were NHW. The

mean admission NIHSS in our sample was 9; range 1–34;

24 % of patients had NIHSS [14. The mean NIHSS was

16 in Hispanics vs. 13 in non-Hispanic participants. The

ethnic groups were different in several characteristics

(Table 1). When compared with NHW and Hispanics, AA

578 Neurol Sci (2014) 35:577–582

123

and NHBC were more likely to be younger. Non-Hispanic

whites were more frequently men as compared to other

groups (OR 1.81, 95 % CI 1.00–3.34). Also, AA had

higher prevalence of cocaine abuse (OR 6.37, 95 % CI

3.35–12.11) than other groups. Dyslipidemia (OR 0.40,

95 % CI 0.22–0.72) and DM (OR 0.50, 95 % CI

0.23–1.00) were less prevalent in NHW than in other

groups. When compared to other groups, NHBC had lower

odds of smoking (OR 0.39, 95 % CI 0.22–0.70), but higher

odds of HTN (OR 4.11, 95 % CI 1.46–11.6). Atrial

fibrillation was more prevalent in Hispanics than in all

other groups (OR 1.71, 95 % CI 1.00–2.95) and non-Afib

cardiac history less frequently reported in NHBC as com-

pared to the rest of the sample (OR 0.47, 95 % CI

0.21–1.00). Ischemic stroke was diagnosed in 82.3 % of

the patients.

The most common mechanism of ischemic stroke was

CE (33.4 %), followed by LAA (22.4 %), SVO (20.7 %)

cryptogenic (12.9 %), and other etiologies or conflicting

mechanism (10.6 %).

There were differences noted in the prevalence of stroke

subtypes by racial/ethnic groups (Tables 1, 2). In univari-

ate analysis, LAA was more prevalent in NHBC (OR 1.74,

95 % CI 1.02–2.97) than in all the other groups. The odds

remained unchanged after adjusting for demographics

(model 1) but became not significant with further adjusting

for vascular risk factors (model 2, OR 1.71, 95 % CI

0.93–3.16). In unadjusted analysis, CE was more frequent

in Hispanics (OR 1.94, 95 % CI 1.28–2.96) and NHW

(2.66, 95 % CI 1.42–4.96) as compared to NHBC and AA

combined. In model 1, the strength of association was

decreased for Hispanics (OR 1.59, 95 % CI 1.02–2.47) and

NHW (OR 2.30, 95 % CI 1.20–4.39) but in model 2, the

association became even stronger for NHW (OR 3.02,

95 % CI 1.42–6.42) but not for Hispanics (OR 1.62, 95 %

CI 0.97–2.72). A trend was noted for an increase in SVO

among NHBC in the unadjusted analyses (OR 1.56, 95 %

CI 0.88–2.79), but there was no difference observed after

adjustment. In Hispanics, SVO was less prevalent than in

all other groups, but adjusting for demographics and vas-

cular risk factors decreased the statistical significance.

Further data analysis showed important statistical

interactions. Concomitant presence of dyslipidemia and

HTN was more frequent in LAA (OR 2.08, 95 % CI

1.23–3.52) than in other mechanisms regardless of the

racial/ethnic group. Cardioembolism was more often

diagnosed in individuals with both, Afib and HTN (OR

24.57, 95 % CI 5.81–104.00) with no difference by ethnic

or racial group. Hypertension was associated with SVO in

NHBC (OR 1.76, 95 % CI 1.00–3.16), while in Hispanics,

Table 1 Characteristic of the stroke patients stratified by Racial and Ethnic Group

Non-Hispanic

White

Hispanic Non-Hispanic

Black Caribbean

Black African

American

Overall P value

for differences

among groups?

N = 52 N = 250 N = 64 N = 107

Age (in years [mean ± SD]) 65.0 ± 15.9 66.5 ± 14.1 61.1 ± 10.6* 59.5 ± 14.0* \0.001

Men (%) 75.0 63.0 59.3* 62.6 0.324

Risk factors (%)

Current smoking 55.0 44.0 26.5� 53.0 0.003

Current alcohol use 23.1 13.6* 7.8� 21.5 0.033

Cocaine abuse 5.8 3.2 3.1 20.5� \0.001

Hypertension 73.0 78.8 93.8� 80.4 0.024

Diabetes 17.3� 28.4 28.1 32.7 0.247

Dyslipidemia 44.2� 64.8 73.4 65.4 0.010

Atrial fibrillation 13.4 16.0� 9.4 9.3 0.230

Non-atrial fibrillation vascular disease 17.3 21.2 10.9 * 21.5 0.274

Stroke subtypes (%)

Large artery atherosclerosis 15.4 21.6 31.3� 22.4 0.217

Cardioembolism 44.2 38.0 23.4� 23.4� 0.005

Small vessel occlusion 19.2 17.2 28.1* 25.2 0.144

Cryptogenic 11.5 13.2 12.5 13.1 0.999

Other causes/conflicting mechanisms 9.6 10.0 4.7 15.9� 0.128

The subgroups analysis was carried out comparing each group to all other groups combined as a reference and the P values represent these

univariate analyses

* P value 0.1–0.05, � P value 0.05–0.01, � P value 0.01–0.001? For age, two-tailed t test was used to obtain P values. Chi squared test was used for all other comparisons

Neurol Sci (2014) 35:577–582 579

123

HTN had the opposite effect (OR 0.57, 95 % CI

0.36–0.91). In our group, the patients were studied with

multiple diagnostic methods to ensure a thorough workup

(Table 3).

Discussion

Our results suggest that the stroke mechanisms vary by

ethnic and racial groups. These variations can be partially

Table 2 Ischemic stroke subtype model comparisons by race/ethniciy

Race/ethnicity Stroke subtype* Unadjusted analysis

OR (95 % CI)

Model 1 OR

(95 % CI)

Model 2 OR

(95 % CI)

Hispanic LAA 0.91 (0.60–1.39) 0.90 (0.58–1.38) 0.84 (0.53–1.36)

CE 1.94 (1.28–2.96) 1.59 (1.02–2.47) 1.61 (0.97–2.71)

SVO 0.65 (0.42–1.00) 0.72 (0.46–1.13) 0.70 (0.44–1.09)

Cryptogenic 1.06 (0.62–1.80) 1.09 (0.58–1.38) 1.06 (0.61–1.83)

African American LAA 0.97 (0.59–1.60) 1.01 (0.61–1.70) 1.09 (0.63–1.89)

CE 0.54 (0.33–0.87) 0.63 (0.38–1.04) 0.60 (0.33–1.07)

SVO 1.33 (0.81–2.18) 1.17 (0.71–1.94) 1.34 (0.79–2.27)

Cryptogenic 1.00 (0.53–1.86) 1.00 (0.51–1.84) 1.04 (0.54–2.02)

Non-Hispanic black Caribbean LAA 1.74 (1.02–2.97) 1.71 (1.00–3.02) 1.71 (0.93–1.71)

CE 0.59 (0.33–1.07) 0.70 (0.38–1.29) 0.69 (0.34–1.41)

SVO 1.56 (0.88–2.79) 1.40 (0.78–2.52) 1.24 (0.67–2.29)

Cryptogenic 0.96 (0.44–2.10) 0.94 (0.42–2.06) 0.89 (0.39–2.02)

Non-Hispanic white LAA 0.66 (0.30–1.42) 0.59 (0.27–1.30) 0.62 (0.27–1.41)

CE 2.66 (1.42–4.96) 2.30 (1.20–4.39) 3.02 (1.42–6.42)

SVO 0.97 (0.47–2.00) 1.06 (0.51–2.18) 1.12 (0.53–2.37)

Cryptogenic 0.93 (0.38–2.26) 0.93 (0.38–2.27) 0.93 (0.37–2.35)

Model 1: Adjusted by age and gender

Model 2: Adjusted by age, gender and vascular risk factors

LAA large artery atherosclerosis, CE cardioembolism, SVO small vessel occlusion

* Participants in the subtype category ‘‘other causes/conflicting mechanism’’ were included in the reference category but the category was not

used as dependent variable due to heterogeneity of condition encompassed

Table 3 Diagnostic procedures performed per stroke subtype

LAA CE SVO Other Cryptogenic P value*

N 106 158 98 50 61

Brain CT (%) 94.3 95.6 96.9 94.0 93.4 0.230

Brain MRI (%) 86.8 66.5 89.8 86.0 90.2 \0.001

Brain MRA (%) 84.9 64.6 90.8 82.0 80.3 \0.001

Brain CTA (%) 42.5 27.2 11.2 28.0 23.0 \0.001

Neck MRA (%) 21.7 12.7 8.2 18.0 29.5 \0.001

Neck CTA (%) 38.7 17.1 7.1 26.0 21.3 \0.001

Neck vessel Doppler (%) 85.8 81.6 93.9 84.0 85.2 \0.001

Trans-thoracic echocardiogram (%) 96.2 95.6 94.6 90.0 90.2 0.007

Trans-esophageal echocardiogram (%) 24.5 27.2 16.3 30.0 70.5 \0.001

Transcranial Doppler (%) 68.9 50.6 32.7 46.0 60.7 \0.001

TCD-EDS (%) 38.7 17.7 5.1 28.0 34.4 \0.001

DSA (%) 34.0 21.5 3.1 32.0 18.0 \0.001

LAA large artery atherosclerosis, CE Cardioembolism, SVO small vessel occlusion, CT computed tomography, MRI magnetic resonance imaging,

MRA magnetic resonance angiography, CTA computed tomographic angiography, TCD-EDS transcranial Doppler with emboli detection study,

DSA digital subtracted angiography

* P value for differences in workup for each stroke subtype. Chi squared test was used to obtain the P values

580 Neurol Sci (2014) 35:577–582

123

explained or accounted by the different prevalence of

vascular risk factors among the groups. For example,

NHBC have a higher prevalence of LAA as compared to

other groups, but controlling for covariates attenuated the

association of LAA with NHBC. This suggests that the

observed racial disparity can be explained in part by the

different prevalence of vascular risk factors. We found an

elevated prevalence of CE in NHW and Hispanics. Con-

trolling for vascular risk factors attenuated the association

in Hispanics but not in NHW which suggest a differential

effect of the prevalence of vascular risk factors by eth-

nicity. Previous reports confirm that NHW have higher

incidence and/or prevalence of CE than blacks [2], and

Hispanics [2], but less than Maori/Pacific islanders [4].

Small vessel occlusions were more often found in NHBC,

but controlling for covariates decreased the strength of

association. In a previous report from our group, an asso-

ciation between HTN and SVO in NHBC was observed,

suggesting that the observed trend was mediated by the

increased prevalence of HTN in NHBC [5]. We confirmed

a higher prevalence of SVO in AA, but it did not reach

statistical significance as compared to the other groups.

The rate of cryptogenic stroke in our study is one of the

lowest among reported studies (12.9 vs. 41.4 % (OH/KY,

US) [7], 24.2 % (South London, UK) [3], 36 % (MN, US)

[8], 27.4 % (Singapore) [9], 21 % (NY, US) [2], and 51 %

(Auckland, Australia) [4], which we believe is a strength of

our report. We did not specifically focus on the ‘‘other

stroke subtypes’’ or ‘‘conflicting mechanisms’’ categories

since they represent a heterogeneous group.

The racial/ethnic disparities in stroke subtypes could be

related to the excess of vascular risk factors and stroke

incidence in minorities as compared to NHW in the US and

other countries [10]. Methodological issues could explain

these disparities. Some authors have opted for subdividing

risk factors in categories that can express different degrees

of risk severity; however, this approach has not erased the

asymmetry observed in stroke subtype susceptibilities [11].

Others have used continuous variables and although the

disparity in stroke subtypes attenuates in some cases, it has

not fully explained the differences among different ethnic

and racial groups [12]. Differences in socioeconomic status

(SES) can also partially account for the different suscep-

tibilities. Unfortunately, we did not collect SES-related

information. However, adjustment for SES has not been

sufficient to explain these differences in other studies [1, 4,

10].

An alternative explanation is that the disparities

observed in stroke subtypes might be attributed to genetic

differences among studied groups. There is evidence of

genetic clustering that can differentiate self-defined racial

or ethnic groups [13]. Some alleles associated with

increased vascular disease are unequally distributed among

ethnic and racial groups [14]. There are also differences in

mediators of vascular disease among racial and ethnic

groups, such as homocysteine [15], C-reactive protein [16],

intima-media thickness [17], waist-to-hip ratio [12], carotid

artery geometry [18], and vasomotor reactivity [19], that,

although, partially related to the environment, genetic

polymorphisms that increase or modulate these markers

cannot be completely ruled out. However, the role of

genetics may not be as significant since the differences in

polymorphisms between races are small and, therefore, not

sufficient to explain the disparity [20], the usual race

classifier (i.e., skin color) does not relate to a differential

risk of disease [21], and there is equal distribution of

genetic polymorphism considered important for stroke risk

[22]. The geographical ancestral distance is more accurate

than skin color to predict genetic variations among races

[23] although the admixture between races and ethnic

groups may limit the applicability of this clinal concept.

[24].

The selection bias in our study is a limitation to

extrapolate the results to the general population. In addi-

tion, the lack of precision in the quantification of risk

factors and socioeconomic characteristics hinders a firm

conclusion on the stroke subtypes disparities observed.

Over fitting of the data is also a risk when modeling

probability. All the limitations of a cross-sectional design

are applicable to our study.

In summary, our results suggest that the asymmetric

proportion of vascular risk factors among racial and ethnic

groups accounts for most of the differences observed in

stroke subtypes in these groups. There is a need for studies

that can capture the majority of the confounding variables

including genetic polymorphisms that may increase or

moderate the risk for stroke.

References

1. White H, Boden-Albala B, Wang C, Elkind MS, Rundek T,

Wright CB, Sacco RL (2005) Ischemic stroke subtype incidence

among whites, blacks, and Hispanics: the Northern Manhattan

Study. Circulation 111(10):1327–1331. doi:10.1161/01.cir.

0000157736.19739.d0

2. Sacco RL, Kargman DE, Gu Q, Zamanillo MC (1995) Race-

ethnicity and determinants of intracranial atherosclerotic cerebral

infarction. The Northern Manhattan Stroke Study. Stroke

26(1):14–20

3. Hajat C, Dundas R, Stewart JA, Lawrence E, Rudd AG, Howard

R, Wolfe CD (2001) Cerebrovascular risk factors and stroke

subtypes: differences between ethnic groups. Stroke 32(1):37–42

4. Feigin V, Carter K, Hackett M, Barber PA, McNaughton H, Dyall

L, Chen MH, Anderson C, Auckland Regional Community

Stroke Study Group (2006) Ethnic disparities in incidence of

stroke subtypes: Auckland Regional Community Stroke Study,

2002–2003. Lancet Neurol 5(2):130–139. doi:10.1016/S1474-

4422(05)70325-2

Neurol Sci (2014) 35:577–582 581

123

5. Koch S, Pabon D, Rabinstein AA, Chirinos J, Romano JG, For-

teza A (2005) Stroke etiology among Haitians living in Miami.

Neuroepidemiology 25(4):192–195. doi:10.1159/000087449

6. Adams HP Jr, Bendixen BH, Kappelle LJ, Biller J, Love BB,

Gordon DL, Marsh EE, 3rd (24) Classification of subtype of acute

ischemic stroke. Definitions for use in a multicenter clinical trial.

TOAST. Trial of Org 10172 in Acute Stroke Treatment. Stroke

24(1):35–41

7. Woo D, Gebel J, Miller R, Kothari R, Brott T, Khoury J, Salis-

bury S, Shukla R, Pancioli A, Jauch E, Broderick J (1999) Inci-

dence rates of first-ever ischemic stroke subtypes among blacks: a

population-based study. Stroke 30(12):2517–2522

8. Petty GW, Brown RD Jr, Whisnant JP, Sicks JD, O’Fallon WM,

Wiebers DO (1999) Ischemic stroke subtypes: a population-based

study of incidence and risk factors. Stroke 30(12):2513–2516

9. Sharma VK, Tsivgoulis G, Teoh HL, Ong BK, Chan BP (2012)

Stroke risk factors and outcomes among various Asian ethnic

groups in Singapore. J Stroke Cerebrovasc Dis 21(4):299–304.

doi:10.1016/jjstrokecerebrovasdis2010.08.010

10. Wolfe CD, Rudd AG, Howard R, Coshall C, Stewart J, Lawrence

E, Hajat C, Hillen T (2002) Incidence and case fatality rates of

stroke subtypes in a multiethnic population: the South London

Stroke Register. J Neurol Neurosurg Psychiatry 72(2):211–216

11. Bang OY, Saver JL, Liebeskind DS, Pineda S, Yun SW, Ovbiagele

B (2009) Impact of metabolic syndrome on distribution of cervic-

ocephalic atherosclerosis: data from a diverse race-ethnic group.

J Neurol Sci 284(1–2):40–45. doi:10.1016/j.jns.2009.03.033

12. Yatsuya H, Yamagishi K, North KE, Brancati FL, Stevens J,

Folsom AR (2010) Associations of obesity measures with sub-

types of ischemic stroke in the ARIC Study. J Epidemiol

20(5):347–354

13. Tang H, Quertermous T, Rodriguez B, Kardia SL, Zhu X, Brown

A, Pankow JS, Province MA, Hunt SC, Boerwinkle E, Schork NJ,

Risch NJ (2005) Genetic structure, self-identified race/ethnicity,

and confounding in case-control association studies. Am J Hum

Genet 76(2):268–275. doi:10.1086/427888

14. Mak W, Cheng TS, Chan KH, Cheung RT, Ho SL (2005) A

possible explanation for the racial difference in distribution of

large-arterial cerebrovascular disease: ancestral European settlers

evolved genetic resistance to atherosclerosis, but confined to the

intracranial arteries. Med Hypotheses 65(4):637–648. doi:10.

1016/j.mehy.2005.05.017

15. Khan U, Crossley C, Kalra L, Rudd A, Wolfe CD, Collinson P,

Markus HS (2008) Homocysteine and its relationship to stroke

subtypes in a UK black population: the south London ethnicity

and stroke study. Stroke 39(11):2943–2949. doi:10.1161/

strokeaha.107.513416

16. Cushman M, McClure LA, Howard VJ, Jenny NS, Lakoski SG,

Howard G (2009) Implications of increased C-reactive protein for

cardiovascular risk stratification in black and white men and

women in the US. Clin Chem 55(9):1627–1636. doi:10.1373/

clinchem.2008.122093

17. Ohira T, Shahar E, Iso H, Chambless LE, Rosamond WD,

Sharrett AR, Folsom AR (2011) Carotid artery wall thickness and

risk of stroke subtypes: the atherosclerosis risk in communities

study. Stroke 42(2):397–403. doi:10.1161/strokeaha.110.592261

18. Ruan L, Chen W, Srinivasan SR, Sun M, Wang H, Toprak A,

Berenson GS (2009) Correlates of common carotid artery lumen

diameter in black and white younger adults: the Bogalusa Heart

Study. Stroke 40(3):702–707. doi:10.1161/strokeaha.108.531608

19. Kalra L, Rambaran C, Chowienczyk P, Goss D, Hambleton I,

Ritter J, Shah A, Wilks R, Forrester T (2005) Ethnic differences

in arterial responses and inflammatory markers in Afro-Caribbean

and Caucasian subjects. Arterioscler Thromb Vasc Biol

25(11):2362–2367. doi:10.1161/01.ATV.0000183887.76087.6a

20. Gravlee CC (2009) How race becomes biology: embodiment of

social inequality. Am J Phys Anthropol 139(1):47–57. doi:10.

1002/ajpa.20983

21. Gravlee CC, Dressler WW, Bernard HR (2005) Skin color, social

classification, and blood pressure in southeastern Puerto Rico.

Am J Public Health 95(12):2191–2197. doi:10.2105/AJPH.2005.

065615

22. Luke MM, O’Meara ES, Rowland CM, Shiffman D, Bare LA,

Arellano AR, Longstreth WT Jr, Lumley T, Rice K, Tracy RP,

Devlin JJ, Psaty BM (2009) Gene variants associated with

ischemic stroke: the cardiovascular health study. Stroke

40(2):363–368. doi:10.1161/strokeaha.108.521328

23. Keita SO, Kittles RA, Royal CD, Bonney GE, Furbert-Harris P,

Dunston GM, Rotimi CN (2004) Conceptualizing human varia-

tion. Nat Genet 36(11 Suppl):S17–S20. doi:ng145510.1038/

ng1455

24. Wang S, Ray N, Rojas W, Parra MV, Bedoya G, Gallo C, Poletti

G, Mazzotti G, Hill K, Hurtado AM, Camrena B, Nicolini H,

Klitz W, Barrantes R, Molina JA, Freimer NB, Bortolini MC,

Salzano FM, Petzl-Erler ML, Tsuneto LT, Dipierri JE, Alfaro EL,

Bailliet G, Bianchi NO, Llop E, Rothhammer F, Excoffier L,

Ruiz-Linares A (2008) Geographic patterns of genome admixture

in Latin American Mestizos. PLoS Genet 4(3):e1000037. doi:10.

1371/journal.pgen.1000037

582 Neurol Sci (2014) 35:577–582

123