Homocysteine and Its Relationship to Stroke Subtypes in a UK Black Population
The South London Ethnicity and Stroke Study
Background and Purpose— Homocysteine is an endothelial toxin and elevated levels have been associated with stroke risk. Stroke, particularly the small vessel disease (SVD) subtype, is increased in U.S. and UK black populations. In white populations elevated homocysteine has been associated with SVD, especially confluent leukoaraiosis, and may be acting through endothelial dysfunction. We determined the association between homocysteine and stroke subtypes, especially SVD, in a well-phenotyped UK cohort of black stroke patients compared to community controls.
Methods— Homocysteine, vitamin B12, folate levels, and renal function were measured in 457 black stroke patients recruited consecutively through the prospective South London Ethnicity and Stroke Study and 179 black community controls. All patients were subtyped using modified TOAST criteria. Leukoaraiosis in SVD patients was graded according to severity, and patients were additionally categorized on the basis of presence or absence of confluent leukoaraiosis. Odds ratios (OR) and 95% confidence intervals (CI) were calculated.
Results— The highest homocysteine levels were seen in SVD patients compared to controls (16.2 [11.6] versus 11.8 [5.7] μmol/L, P<0.001) after adjusting for age, gender, vascular risk factors, vitamin levels, and renal function. Within SVD cases, highest homocysteine levels were found in lacunar infarction with confluent leukoaraiosis (19.6 [14.9] μmol/L) compared to lacunar infarction without leukoaraiosis (13.6 [7.1] μmol/L, P=0.001) and controls (P<0.001). Homocysteine correlated with leukoaraiosis severity (r=0.225, P<0.001).
Conclusions— In this well characterized UK black stroke population homocysteine levels were elevated and highest levels were found in lacunar stroke with leukoaraiosis.
Severe hyperhomocysteinaemia, a feature of inborn errors of methionine metabolism, is associated with atherosclerosis.1 In addition, many studies have reported more modest increases in serum homocysteine in cardiovascular disease, with meta-analyses finding independent graded associations with coronary artery disease and stroke.2,3 The mechanisms underlying these associations are incompletely understood. Homocysteine is a prothrombotic factor affecting coagulation and fibrinolytic cascades.4 A key early event in homocysteine-mediated vascular damage is endothelial injury through direct toxicity and apoptosis,5 and endothelial dysfunction resulting in impaired endothelium-dependent dilatation of blood vessels combined with a proinflammatory and proatherosclerotic endothelial phenotype.6
Stroke is a heterogenous condition, and different stroke subtypes have different pathophysiological mechanisms. Homocysteine has been studied in stroke subtypes in various Asian and European populations.7–14 Although most studies suggest high homocysteine in all stroke subtypes compared with controls, some studies report stronger associations with particular subtypes including cerebral small vessel disease (SVD),7,8 large vessel disease (LVD),9–11 and primary intracerebral hemorrhage (PICH).13 The pathogenesis of SVD is incompletely understood but appears to be distinct from atheroembolism seen with LVD with differing risk factor profiles.15 Pathological and clinical studies have led to the hypothesis that there are two SVD subtypes: single larger lacunar infarcts (isolated lacunar infarction) attributable to microatheroma at the origins of larger perforator arteries (200 to 800 μm diameter), and multiple smaller lacunar infarcts with leukoaraiosis attributable to a diffuse arteriopathy affecting smaller perforator arteries (40 to 200 μm diameter).16,17 Endothelial dysfunction has been implicated in the leukoaraiosis variant of SVD, a subtype referred to as ischemic leukoaraiosis.18 Elevated homocysteine has been reported in this SVD subtype in white populations, the association disappearing after controlling for circulating endothelial markers, implicating a role for endothelial dysfunction.19
The incidence of stroke is increased in black compared to white populations in both the United States and United Kingdom.20,21 The distribution of subtypes also varies with an increased proportion of SVD in black stroke patients.22,23 Attenuated endothelial function has been reported in healthy stroke-free black volunteers, relative to whites.24 Homocysteine has been identified as a risk factor for ischemic stroke in whites but not black Americans in a prospective study where stroke subtyping was not performed.25 Whether elevated homocysteine is particularly associated with the SVD subtype in black populations is uncertain.
In this study, we determined the association between homocysteine and stroke subtypes in a well phenotyped UK cohort of black stroke patients compared to community controls. To explore possible associations with SVD further we compared homocysteine levels between the two SVD subtypes, and correlated levels with leukoaraiosis severity.
The South London Ethnicity and Stroke Study23 is a prospective study that began in 1999 recruiting consecutive black stroke patients from a contiguous catchment area covered by three hospitals in South London. All hospitals have a specialist stroke unit and outpatient TIA clinics. Cases were recruited from both of these sources. In addition the prospective community based South London Stroke Register20 is nested within the catchment area and cases were also identified from this register. Over the period until 2006, 600 black stroke patients were recruited. Of these, 457 patients were included in this study. Of the remaining 143 patients, 58 (9.7%) had no blood taken (12 did not consent, 25 died before phlebotomy and 21 had difficulty with phlebotomy), in 59 (9.8%) the homocysteine assay did not produce a result (because of insufficient sample volume), and 26 patients (4.6%) were on folic acid or vitamin B12 supplementation were excluded because of effects on homocysteine levels.
All patients included in this study underwent brain imaging (CT alone, 282 [61.7%]; MRI with or without CT, 175 [38.3%]) and ECG. Imaging of the extracranial cerebral vessels with Duplex ultrasound, CT contrast angiography, or MR angiography was performed in 97.3% of ischemic strokes, and echocardiography in 55.2%. One hundred seventy-nine age- and sex-matched black community controls free of clinical cerebrovascular disease were also recruited by random sampling from general practices from within the study area. Black ethnicity was defined as black or black British (black Caribbean or black African) according to the National Statistics classification used in the 2001 census in England and Wales (http://www.statistics.gov.uk/about/Classifications/ns_ethnic_ classification.asp). The study protocol was approved by local research ethics committees, and informed consent was obtained from all participants.
Risk factor information and other clinical and investigation details were collected on a standardized proforma. Hypertension was defined as prestroke treatment with antihypertensive drugs or a systolic blood pressure >140 mm Hg or diastolic blood pressure >90 mm Hg persisting at least a week after stroke onset to exclude acute elevation of blood pressure attributable to stroke. Diabetes mellitus was defined as a previous diagnosis of type I or type II diabetes, or at least two random glucose readings of >11.1 mmol/L or fasting blood glucose readings of >7.0 mmol/L after the acute phase of stroke. Hypercholesterolaemia was defined as a serum cholesterol >5.2 mmol/L or prestroke treatment with a cholesterol-lowering agent. Smokers were defined as currently smoking or ever smoked.
One investigator subtyped all strokes with review of original brain imaging. A modification of the TOAST (Trial of Org 10172) subtyping classification was used.26 To avoid bias attributable to different rates of risk factors such as hypertension between the two groups, the presence of hypertension or diabetes was not used as a criterion in the diagnosis of SVD. SVD was defined as a clinical lacunar syndrome with a compatible lesion on MRI or CT or no lesion on CT without another cause of stroke identified. Exclusion criteria included subcortical infarction >15 mm diameter, cortical infarction of any size, carotid, vertebral, or intracranial artery stenosis >50%, and a potential cardiac source of embolism. Of the 152 SVD cases, 144 had a lacunar infarct visualized on CT or MRI, whereas 8 had a lacunar syndrome but no visible infarct on CT and no MRI performed. Large vessel atherosclerotic disease (LVD) was defined as carotid, vertebral, or major intracranial artery stenosis >50%. Cardioembolic stroke was defined based on the presence of a potential source of cardiac embolism categorized as high or moderate risk according to the TOAST criteria. Where no cause of stroke was found, patients were assigned into an “Unknown” category. Patients with more that one potential stroke mechanism was designated a “Tandem” category and rare causes of stroke were defined as “Other.” Primary intracerebral hemorrhage (PICH) was defined as a separate category. Patients with primary subarachnoid hemorrhage were excluded.
Leukoaraiosis Grading and SVD Subtyping
Leukoaraiosis on CT and MRI in all stroke patients was graded using the semiquantitative Fazekas scale which has been correlated with pathological severity in a postmortem validation study,27 modified to separate degrees of confluent leukoaraiosis as previously described18: grade 0, no leukoaraiosis; grade 1, mild leukoaraiosis (>5 white matter hyperintensities); grade 2, moderate confluent leukoaraiosis; and grade 3, severe confluent leukoaraiosis. In addition, on the basis of this leukoaraiosis grade, SVD patients were subtyped into two groups: isolated lacunar infarction (lacunar infarction with absent or mild leukoaraiosis), or ischemic leukoaraiosis (lacunar infarction in the presence of confluent leukoaraiosis) according to a previously validated method.18
Nonfasting blood was collected from both patients and controls. Blood was centrifuged at 4400 r.p.m. and serum stored at −80°C. All biomarker assays were performed by the Department of Chemical Pathology, St George’s Healthcare NHS Trust blinded to subject identity. Homocysteine, B12, and folate were measured on serum using competitive immunoassay. Creatinine was measured on serum using direct colorimetry. Estimated glomerular filtration rate (eGFR) was calculated based on serum creatinine (μmol/L), age (years) and gender using the formula:
All statistical tests were performed using SPSS Version 15. The distributions of all biomarkers except eGFR were skewed and underwent natural logarithmic transformation to normalize distributions. Student t test was used for continuous variables and χ2 testing for categorical variables. Odds ratios (OR) and 95% confidence intervals per 1 μmol/L increase in log homocysteine were calculated using binary logistic regression analysis. Age, gender, hypertension, diabetes, hypercholesterolaemia, smoking, B12, folate, and eGFR were controlled for in multivariate analysis. Unavailable data were defined as missing in SPSS, and these patients and controls excluded from multivariate analysis. Pearson correlation coefficient (r) was calculated for the correlation between homocysteine and other biomarkers, and between homocysteine and leukoaraiosis severity. ANCOVA analysis was used to control for age, gender, vascular risk factors, vitamin, and renal status in the association between homocysteine and leukoaraiosis severity.
Patient and control demographics are given in Table 1. There was no difference in age or gender between strokes and controls. Hypertension and diabetes were more frequent in stroke cases. There was no difference in hypercholesterolaemia and smoking between the two groups.
Homocysteine Levels in Stroke Patients and Controls
Table 1 shows univariate comparisons in biomarkers between strokes and controls. Mean homocysteine concentrations were significantly higher in strokes versus controls. There were no significant differences in B12, folate, and eGFR measurements between the two groups.
Association Between Homocysteine, Age, Gender, Vascular Risk Factors, Vitamin Status, and Renal Function
In both strokes and controls, homocysteine was positively correlated with age (r=0.211, P<0.001) and negatively with B12 (r=−0.311, P<0.001), folate (r=−0.158, P<0.001), and eGFR (r=−0.453, P<0.001). Homocysteine was increased in males compared to females (14.4 [8.9] versus 12.6 [6.9] μmol/L, P=0.002), hypertensives compared to normotensives (14.3 [8.5] versus 11.2 [6.2] μmol/L, P<0.001), and smokers compared to nonsmokers (15.1 [9.6] versus 12.6 [6.8] μmol/L, P<0.001). There was no significant difference in homocysteine levels between diabetics and nondiabetics, or between subjects with elevated or normal cholesterol.
Homocysteine Comparison Between Stroke Patients and Controls
On univariate analysis homocysteine levels were elevated in stroke patients compared to controls (14.3 [8.8] μmol/L versus 11.8 [5.7] μmol/L, P=0.001; Table 1). This difference persisted after adjusting for age and gender (P<0.001), and after additionally adjusting for vascular risk factors, B12, folate, and eGFR (P<0.001). Homocysteine concentrations were separately elevated in both ischemic (14.3 [8.9] μmol/L) and hemorrhagic (14.5 [8.7] μmol/L) stroke compared to controls on univariate analysis (ischemic stroke, P=0.001; PICH, P=0.050), after adjusting for age and gender (ischemic stroke, P<0.001; PICH, P=0.005) and after additionally adjusting for vascular risk factors, B12, folate and eGFR (ischemic stroke, P<0.001; PICH, P=0.006).
Homocysteine levels in different stroke subtypes are shown in Table 2. Compared to controls homocysteine was elevated in all stroke subtypes, except LVD and “Other” categories, both before and after adjusting for age, gender, vascular risk factors, B12, folate, and eGFR. Highest homocysteine levels were seen in SVD. Homocysteine levels were higher in both SVD and cardioembolic stroke, compared to LVD on univariate analysis. After adjusting for age, gender, and vascular risk factors, B12, folate, and eGFR levels were significantly higher only in SVD.
Homocysteine and Stroke Risk
There was a graded positive relationship between homocysteine levels and stroke risk; OR per 1 μmol/L increase in log homocysteine was 1.93 (95% CI: 1.31 to 2.84). After adjusting for age and gender, the OR became 2.09 (95% CI: 1.40 to 3.12); after additionally adjusting for vascular risk factors, B12, folate, and eGFR, the OR was 4.02 (95% CI: 2.16 to 7.51; Table 3).
In a logistic regression model comparing homocysteine in all stroke patients and controls and after adjustment for age, gender, vascular risk factors, B12, folate, and eGFR, a significant interaction was observed between homocysteine and eGFR (P=0.004). To explore this further, the interaction was assessed in the same model using homocysteine tertiles. The significant interaction between homocysteine and eGFR occurred in the upper homocysteine tertile (P=0.044). Of all the homocysteine tertiles, the lowest mean eGFR (66.48 [31.17] ml/mim/1.73m2), equating to poorest renal function, was recorded in stroke patients in the upper tertile. However, the association between all strokes and homocysteine remained significant after adjusting for this interaction (P<0.001).
On analysis of association with stroke subtypes a graded positive relationship between homocysteine and risk was found for all stroke subtypes, except for LVD (Table 3). The adjusted OR was similar for hemorrhagic stroke (OR 4.65) compared to ischemic stroke (OR 4.02). This analysis was not performed for “Tandem” and “Other” categories because of the small numbers of patients in each subtype. The highest adjusted-risk was seen with SVD (OR 7.72) per 1 μmol/L log homocysteine. An interaction analysis demonstrated that the association between homocysteine and SVD was independent of age, gender, vascular risk factors, B12, folate, and eGFR. Furthermore, the age- and gender-adjusted OR for the association between homocysteine and SVD did not change markedly on adding vascular risk factors to the regression model (data not shown).
Homocysteine in SVD Subtypes
There were significant differences between homocysteine concentrations in the two proposed SVD subtypes. Levels in ischemic leukoaraiosis (19.6 [14.9] μmol/L) were higher than those in both isolated lacunar infarction (13.6 [7.1] μmol/L; univariate P=0.001, and fully-adjusted P=0.013), and controls (univariate, P<0.001; fully-adjusted, P<0.001) Mean homocysteine in isolated lacunar infarction was significantly increased compared to controls (univariate, P=0.001; fully-adjusted, P<0.001).
The OR per 1 μmol/L increase in log homocysteine for ischemic leukoaraiosis–control comparison was 17.10 (95% CI: 5.13 to 56.98), ischemic leukoaraiosis–isolated lacunar infarction comparison was 3.89 (95% CI: 1.52 to 9.96), and isolated lacunar infarction–control comparison was 5.65 (95% CI: 1.92 to 16.63) after full adjustment.
Homocysteine Correlation With Leukoaraiosis Severity
A significant positive correlation was seen between homocysteine and leukoaraiosis severity in all strokes (r=0.225, P<0.001) and SVD alone (r=0.256, P=0.001; Figure). After adjusting for age, gender, vascular risk factors, and vitamin status in an ANCOVA analysis, significant differences were seen between moderate (P=0.050) and severe (P=0.023) leukoaraiosis groups compared to the absent leukoaraiosis group in all strokes and between severe and absent leukoaraiosis groups (P=0.050) in SVD patients (Figure).
Pearson correlation coefficient (r) for the association between homocysteine and leukoaraiosis severity for all strokes, stratified to patients imaged with CT only was 0.182 (P=0.003) and stratified to MRI was 0.299 (P<0.001). For SVD, in patients with CT only r=0.241 (P=0.035), and in SVD patients with MRI imaging r=0.267 (P=0.021).
This study found significantly elevated homocysteine levels in a UK black stroke population compared to community controls, and an independent graded association between homocysteine levels and stroke risk. Similar findings were demonstrated for both hemorrhagic and ischemic stroke. Within ischemic stroke cases homocysteine levels were elevated in all stroke subtypes with the exception of patients with LVD and strokes classified as “Other.” The strongest association was seen with SVD. The association was stronger in patients with the ischemic leukoaraiosis subtype of SVD, and there was a strong positive correlation between homocysteine levels and leukoaraiosis grade.
An independent, graded association has been demonstrated between homocysteine and stroke in prospective and retrospective studies in white populations.2,3 Causality is additionally supported by biologically plausible mechanisms for homocysteine-mediated vessel damage,4 and using the Mendelian randomisation approach.19,28 Folate supplementation trials in patients with preexisting vascular disease have been inconsistent in demonstrating stroke prevention,29 although a recent meta-analysis supports a role for folate supplementation in primary prevention of stroke.30 These studies have failed to address the heterogeneity in stroke, and the possibility that homocysteine may have a greater impact on a particular stroke subtype. Studies in white and Asian populations have looked at homocysteine in stroke subtypes. Some have demonstrated stronger associations with SVD,7,8 consistent with our findings, whereas others report stronger associations with LVD.9–12 Other studies report equally strong associations with all stroke subtypes.13,14
Differences in stroke subtypes, and in particular an increase in SVD and decreased cardioembolic and extracranial LVD, have been reported in black, compared with white populations.22,23 The reasons for these differences are unclear. Hypertension, which is increased in black populations, is an important risk factor for SVD, but fails to account for all the risk.23 Endothelial dysfunction has been implicated in the pathogenesis of SVD. In white, markers of endothelial dysfunction are elevated in SVD especially in the presence of confluent leukoaraiosis.18 A similar association between homocysteine and confluent leukoaraiosis has been demonstrated; this association was attenuated after adjusting for markers of endothelial dysfunction suggesting that the pathogenic affect of homocysteine on small vessels be mediated by endothelial dysfunction.19 Endothelial function in healthy blacks is attenuated relative to healthy whites.24 In the current study, highest homocysteine levels were found in SVD patients who also had confluent leukoaraiosis.
A strong association was also found between homocysteine and PICH. Hemorrhage is increased in individuals with leukoaraiosis, and hypertensive PICH and SVD share similar pathology in the form of lipohyalinosis,31 in at least a proportion of cases. It is possible that homocysteine-mediated small vessel damage may partly account for the association between homocysteine and PICH.
Strengths of this study included the prospective and consecutive recruitment of a well characterized cohort of African and African-Caribbean stroke patients and community controls. All patients had brain imaging and ECG, and almost all ischemic strokes had extracranial vessel imaging. Subtyping was performed by one rater with review of all original imaging using a pathophysiological method. Biochemical analyses were performed blinded to the stroke-control status to avoid bias. Homocysteine levels can be affected by vascular risk factors, vitamin status and renal function. These were adjusted for in analyses.
Differences in vitamin status and renal function failed to account for elevated homocysteine in black strokes, although a detailed dietary history was not recorded. A significant interaction between homocysteine and eGFR was seen especially at high homocysteine concentrations, but the association between homocysteine and stroke persisted after adjusting for this interaction. In young whites, genetic factors account for approximately 9% of variation in homocysteine levels, largely attributed to the methylene tetrahydrofolate reductase (MTHFR) C677T polymorphism.32 Black populations lack this mutation,33 although as yet undiscovered genetic factors may be important.
Although our community control population was free of clinical cerebrovascular disease, brain imaging was not performed in this group. Therefore any relationship between silent brain infarction or asymptomatic leukoaraiosis and homocysteine could not be assessed. Scanning of controls would have markedly reduced their recruitment and could potentially introduce selection bias. If anything, by ensuring that controls were free of radiological cerebral ischemia, this would increase the magnitude of the demonstrated difference in homocysteine between patients and controls.
Patients with CT alone were included in the study to avoid selection bias. Although CT is less accurate than MRI in demonstrating the degree of leukoaraiosis, we have previously shown good agreement between grading of the degree of leukoaraiosis on this scale on CT and MRI,18 and analyses performed separately in CT-only and MRI groups demonstrated similar findings for the association between homocysteine and leukoaraiosis severity.
Intracranial atherosclerosis can give rise to lacunar infarcts indistinguishable from lacunes attributable to hypertensive arteriolosclerosis and may result in SVD misclassification. 17.2% of the South London black ischemic stroke population has intracranial atherosclerosis.23 To exclude ischemic stroke patients without intracranial imaging would have introduced selection-bias and limited patient recruitment. However, most SVD patients had lacunar infarct with leukoaraiosis, and there was a strong relationship between homocysteine and degree of leukoaraiosis. Intracranial stenosis would not be expected to cause leukoaraiosis. Nevertheless further studies are required to determine the role of intracranial stenosis in the pathogenesis of lacunar type infarcts in black stroke cohorts.
In summary, this study of homocysteine in a well phenotyped cohort of black stroke patients demonstrated elevated homocysteine levels in all strokes (ischemic and hemorrhagic) compared to community controls. The strongest associations were seen with SVD. Within SVD, the highest homocysteine levels were seen in the presence of confluent leukoaraiosis. Coupled with the positive correlation between homocysteine and leukoaraiosis severity, this lends support to a pathogenic role for homocysteine in small vessel injury, possibly through endothelial cell dysfunction. Our findings may have important implications for the use of vitamin supplementation to lower homocysteine levels in the prevention of stroke. The large randomized-controlled secondary prevention trials of homocysteine lowering using vitamin-supplementation such as NORVIT, HOPE-2, and VISP have failed to demonstrate a reduction of vascular events although a meta-analysis of these trials,29 and a recent reassessment of the results of HOPE-2 and VISP34 suggest a possible effect of vitamin therapy for stroke compared to myocardial infarction. Our results suggest that the efficacy of this therapy may differ for different stroke subtypes and be most beneficial in small vessel disease stroke. Such benefits will only be detected in trials which include adequate stroke subtyping. There may also be ethnic differences in efficacy, and sufficient black stroke patients should be included in trials to determine efficacy in this ethnic group.
Sources of Funding
This work was supported by a Stroke Association Programme Grant (PROG 3). L.K., A.R., and C.D.A.W. are supported by the Biomedical Research Centre. The South London Stroke Register is funded by the Department of Health.
- Received January 3, 2008.
- Revision received March 31, 2008.
- Accepted April 16, 2008.
Wald DS, Law M, Morris JK. Homocysteine and cardiovascular disease: evidence on causality from a meta-analysis. BMJ. 2002; 325: 1202.
Hossain GS, van Thienen JV, Werstuck GH, Zhou J, Sood SK, Dickhout JG, de Koning AB, Tang D, Wu D, Falk E, Poddar R, Jacobsen DW, Zhang K, Kaufman RJ, Austin RC. TDAG51 is induced by homocysteine, promotes detachment-mediated programmed cell death, and contributes to the development of atherosclerosis in hyperhomocysteinemia. J Biol Chem. 2003; 278: 30317–30327.
Perini F, Galloni E, Bolgan I, Bader G, Ruffini R, Arzenton E, Alba S, Azzini C, Bartolomei L, Billo G, Bortolon F, Dudine P, Garofalo PG, L'Erario R, Morra M, Parisen P, Stenta G, Toso V. Elevated plasma homocysteine in acute stroke was not associated with severity and outcome: stronger association with small artery disease. Neurol Sci. 2005; 26: 310–318.
Iso H, Moriyama Y, Sato S, Kitamura A, Tanigawa T, Yamagishi K, Imano H, Ohira T, Okamura T, Naito Y, Shimamoto T. Serum total homocysteine concentrations and risk of stroke and its subtypes in Japanese. Circulation. 2004; 109: 2766–2772.
Eikelboom JW, Hankey GJ, Anand SS, Lofthouse E, Staples N, Baker RI. Association between high homocyst (e) ine and ischemic stroke due to large- and small-artery disease but not other etiologic subtypes of ischemic stroke. Stroke. 2000; 31: 1069–1075.
Tan NC, Venketasubramanian N, Saw SM, Tjia HT. Hyperhomocyst(e)inemia and risk of ischemic stroke among young Asian adults. Stroke. 2002; 33: 1956–1962.
Li Z, Sun L, Zhang H, Liao Y, Wang D, Zhao B, Zhu Z, Zhao J, Ma A, Han Y, Wang Y, Shi Y, Ye J, Hui R. Elevated plasma homocysteine was associated with hemorrhagic and ischemic stroke, but methylenetetrahydrofolate reductase gene C677T polymorphism was a risk factor for thrombotic stroke: a Multicenter Case-Control Study in China. Stroke. 2003; 34: 2085–2090.
Khan U, Porteous L, Hassan A, Markus H. Risk factor profile of cerebral small vessel disease and its subtypes. J Neurol Neurosurg Psychiatry. 2007; 78: 702–706.
Boiten J, Lodder J, Kessels F. Two clinically distinct lacunar infarct entities? A hypothesis. Stroke. 1993; 24: 652–656.
Hassan A, Hunt BJ, O'Sullivan M, Parmar K, Bamford JM, Briley D, Brown MM, Thomas DJ, Markus HS. Markers of endothelial dysfunction in lacunar infarction and ischaemic leukoaraiosis. Brain. 2003; 126: 424–432.
Hassan A, Hunt BJ, O'Sullivan M, Bell R, D'Souza R, Jeffery S, Bamford JM, Markus HS. Homocysteine is a risk factor for cerebral small vessel disease, acting via endothelial dysfunction. Brain. 2004; 127: 212–219.
Stewart JA, Dundas R, Howard RS, Rudd AG, Wolfe CD. Ethnic differences in incidence of stroke: prospective study with stroke register. BMJ. 1999; 318: 967–971.
Sacco RL, Boden-Albala B, Gan R, Chen X, Kargman DE, Shea S, Paik MC, Hauser WA. Stroke incidence among white, black, and Hispanic residents of an urban community: the Northern Manhattan Stroke Study. Am J Epidemiol. 1998; 147: 259–268.
Sacco RL, Kargman DE, Gu Q, Zamanillo MC. Race-ethnicity and determinants of intracranial atherosclerotic cerebral infarction. The Northern Manhattan Stroke Study. Stroke. 1995; 26: 14–20.
Markus HS, Khan U, Birns J, Evans A, Kalra L, Rudd AG, Wolfe CDA, Jerrard-Dunne P. Differences in stroke subtypes between black and white patients with stroke: the South London Ethnicity and Stroke Study. Circulation. 2007; 116: 2157–2164.
Kalra L, Rambaran C, Chowienczyk P, Goss D, Hambleton I, Ritter J, Shah A, Wilks R, Forrester T. Ethnic differences in arterial responses and inflammatory markers in Afro-Caribbean and Caucasian subjects. Arterioscler Thromb Vasc Biol. 2005; 25: 2362–2367.
Sacco RL, Anand K, Lee HS, Boden-Albala B, Stabler S, Allen R, Paik MC. Homocysteine and the risk of ischemic stroke in a triethnic cohort: the Northern Manhattan Study. Stroke. 2004; 35: 2263–2269.
Adams HP Jr, Bendixen BH, Kappelle LJ, Biller J, Love BB, Gordon DL, Marsh EE III. 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. 1993; 24: 35–41.
Fazekas F, Kleinert R, Offenbacher H, Schmidt R, Kleinert G, Payer F, Radner H, Lechner H. Pathologic correlates of incidental MRI white matter signal hyperintensities. Neurology. 1993; 43: 1683–1689.
Masuda J, Tanaka K, Omae T, Ueda K, Sadoshima S. Cerebrovascular diseases and their underlying vascular lesions in Hisayama, Japan–a pathological study of autopsy cases. Stroke. 1983; 14: 934–940.
Kluijtmans LA, Young IS, Boreham CA, Murray L, McMaster D, McNulty H, Strain JJ, McPartlin J, Scott JM, Whitehead AS. Genetic and nutritional factors contributing to hyperhomocysteinemia in young adults. Blood. 2003; 101: 2483–2488.