The relationship between higher blood pressure and ischaemic,haemorrhagic stroke among Chinese and Caucasians: meta-analysis

Abstract

Background and purpose

Previous studies have suggested that blood pressure is a particularly important risk factor for stroke in Chinese, and that the magnitude of the effect may be greater than in Caucasians. We performed a meta-analysis in order to define the magnitude of the risk of stroke associated with hypertension among Chinese, and to compare the magnitude of this risk with Caucasians.

Methods

We searched Medline from 1966 to 2004, plus Chinese Medical Literature databases from 1977 to 2003 for Caucasian and Chinese studies. Results were pooled using the random effects model, and heterogeneity and publication bias were checked.

Results

For a 10 mmHg increase in systolic blood pressure, the pooled risk ratio was 1.44 [95% confidence interval (CI)] 1.39–1.51 for ischaemic stroke, and 1.5 (95% CI 1.45–1.57) for haemorrhagic stroke in Chinese, versus 1.19 (95% CI 1.15–1.23) for total stroke in Caucasians. The pooled odds ratio for hypertension measured dichotomously (defined as >160/95 mmHg) was 5.8 (95% CI 4.7–7.2) among Chinese versus 1.93 (95% CI 1.7–2.2) among Caucasians for ischaemic stroke; and 7.2 (95% CI 5.3–9.7) in Chinese versus 3.1 (95% CI 2.5–3.9) in Caucasians for haemorrhagic stroke.

Conclusion

The risk of stroke associated with hypertension is consistently and significantly greater in Chinese than Caucasians. This may help genetic epidemiologists to dissect the cause of stroke, and emphasizes the particular importance of hypertension control in the Chinese population. Eur J Cardiovasc Prev Rehabil 13:429–437 © 2006 The European Society of Cardiology

Keywords

stroke ischaemic stroke haemorrhagic stroke hypertension meta-analysis blood pressure Chinese Caucasians

Introduction

It is increasingly apparent that the spectrum of cardiovascular disease (CVD) among Chinese is very different than in Caucasians. Whereas coronary heart disease (CHD) predominates in Caucasians, stroke predominates in Chinese. Intracerebral haemorrhage occurs between two and three times more frequently in Chinese than in Caucasians. Only 6–18% of strokes in European populations are reported as intracerebral haemorrhages compared with 23–52% haemorrhagic strokes in Chinese [1, 2].

The mortality from stroke is commensurately high. Among the Chinese population, the mortality rate for stroke is three times higher than that for coronary heart disease [3], and haemorrhagic strokes also have a higher case-fatality rate than ischaemic strokes [3].

The reasons for this ethnic difference are slowly being teased out. A population-based incident case-control study found that the prevalence, odds ratio (OR), and population attributable risk (PAR) for stroke risk factors varied between ethnic groups [4]. Our previous work also observed that there were differences in the prevalence of hypertension, the magnitude of stroke risk associated with hypertension, and the PAR for hypertension between Chinese and Caucasians [5]. Other studies also confirm that the PAR of stroke associated with hypertension is higher in Chinese [6], and in East Chinese [7] than in Caucasians. Diet might explain at least part of this difference. For example, the International Study on Macronutrients and Blood Pressure (INTERMAP) found that macronutrient intakes were much higher in Western diets for total fat, saturated and trans fatty acids, key dietary lipid scores and sugars, but lower in total carbohydrate and starch. Dietary sodium was higher, but potassium lower, hence the sodium/potassium ratio was higher in the Asian diet. Higher salt intake may be part of the mechanism for hypertension among Chinese [8].

A comprehensive and quantitative summary of this data can be obtained by performing a meta-analysis. Indeed, a meta-analysis of the literature up to 1994 supported the view that hypertension had a more severe effect on stroke in Chinese than in Caucasians, although this was heterogeneous [6]. In the present study, we update this meta-analysis with two aims: first, to define the magnitude of the risk of ischaemic and haemorrhagic stroke separately associated with hypertension among Chinese; and second, to compare the magnitude of this risk with Caucasians.

Methods

Data collection

Selection of studies for review

For Chinese studies, firstly, we searched Medline from 1966 to April 2004 with the following strategy: stroke (MeSH) or ischaemic stroke (MeSH) or haemorrhagic stroke (MeSH) AND risk factors (MeSH) AND hypertension (MeSH) or blood pressure (key words); limited to Chinese. Secondly, we updated the previous meta-analysis of Chinese studies [6] by searching the Chinese Medical Literature database to 2003 (strategy available from authors on request).

For Caucasian studies we searched Medline data from 1966 to 2004 with the following strategy: stroke (MeSH) or ischaemic stroke (MeSH) or haemorrhagic stroke (MeSH) AND risk factors (MeSH) AND hypertension (MeSH) or blood pressure (key word) AND cohort or prospective study (MeSH) AND incidence (MeSH); limited to human and English. For case-control studies, the strategy was similar except that the term case-control was used instead of cohort. The searches were supplemented by personal files and expert knowledge.

Inclusion/exclusion criteria

The inclusion criteria were cohort or case-control design; outcome was reported for ischaemic or haemorrhagic stroke separately; multiple regression analysis was used and coefficients or odds ratios reported. Duplicate studies, or studies lacking any adjustment for confounders were excluded.

Statistical methods

Results were weighted and pooled using the inverse variance method [6]. Heterogeneity was assessed using a Q-test with a significance level of 0.1. Publication bias was assessed using funnel plots and the Egger test. Risk ratios (RR) or ORs and 95% confidence intervals (CIs) are presented after antilogarithmic transformation. We calculated the pooled measure of effect for hypertension measured in two ways: continuously, per 10mmHg increase in systolic blood pressure (SBP) or dichotomously using a threshold of SBP >160 mmHg or DBP >95 mmHg. Ischaemic or haemorrhagic stroke were analysed separately. If the study reported the coefficient and standard error for a 1 mmHg increase in SBP this was converted to estimate the effect of a 10 mmHg increase by multiplying the initial estimate and standard error by 10. Pooled ORs were obtained for Chinese and Caucasians separately. If heterogeneity was present, a meta-regression model including age, cholesterol, diabetes, smoking and follow-up years (where available) was used to try to ascertain the reasons for this. If the source of the heterogeneity could not be identified, data were pooled using a random effects model. STATA statistical software 7.0 (STATA, College Station, Texas, USA) was used.

Results

Only one of the 67 articles on Medline was potentially eligible [9], but it only reported the association of total fatal stroke with hypertension defined dichotomously (160/95 mmHg), and was therefore excluded [9]. From a previous meta-analysis for stroke [6], 12 studies were identified; six of these were excluded for the following reasons: reporting only fatal stroke (n = 2) [10, 11]; not reporting 95% CI (n = 3) [12–14]; different cut-off for hypertension (n = 1) [15]. This search was updated using the Chinese Medical Literature database: 223 references were identified, of which 12 were new, potentially eligible studies [16–27]. Six studies were excluded because of the following reasons: they did not adjust for other co-variables (n = 4) [17, 20, 23, 27]; they used a different cutoff for hypertension (n = 1) [24]; they reported total stroke only (n = 1) [19], leaving six eligible studies. Hence, the six studies identified from the Chinese Medical Literature database [16, 18, 21, 22, 25, 26] and the six studies identified from the previous meta-analysis [6], plus our own previous study [5], led to 13 studies being included in this meta-analysis.

Table 1

Characteristics of Chinese studies treating hypertension as a continuous variable

Study (outcome)	Sample size (stroke events)	Age range (years)	Average follow-up (years)	Men (%)	RR (total stroke)	RR (ischaemic stroke)	RR (haemorrhagic stroke)
Zhou [16] (incidence)	9111 (259 cases)	35–54	13	50	1.48 (1.42–1.54)	1.46 (1.38–1.55)	1.49 (1.38–1.61)
Zhang [18] (incidence)	24734 (256 cases)	35–59	9	50	1.49 (1.45–1.55)	1.47 (1.40–1.54)	1.54 (1.46–1.62)
Zhang [5] (incidence)	4400 (152 cases)	18–74	13.5	100	1.38 (1.29–1.49)	1.34 (1.21–1.48)	1.43 (1.28–1.60)

Adjusted Cox proportional model hazard ratio (95% confidence interval) of stroke associated with a 10 mmHg increase in systolic blood pressure. RR, risk ratio.

Table 2

Characteristics of Chinese studies treating hypertension as a dichotomous variable

Study	Design	Sample size and matching variables	Age range (years)	Men (%)	OR (95% CI)
Ischaemic stroke
Zhang [5]	Cohort	4400 (87 cases during 13.5 years, Cox model)	18–74	100	3.43 (2.05–5.8)
Wang [22]	Case-control	Hospital cases, 1:1 match for age, sex; 102 pairs	45–74	66	8.3 (3.4–20.2)
Niu [21]	Case-control	Hospital cases, 1:1 match for age, sex; 91 pairs	28–50	63	7.5 (1.6–35.5)
Wang [25]	Case-control	Hospital cases, 1:1 match for age, sex, and site; 148 pairs	27–84	80	8.2 (4.6–14.7)
Yuan [6]	Case-control	Hospital cases, 1:1 match for age, sex, and ethnicity, occupation and site; 808 pairs	NA	NA	6.19 (4.46–8.60)
Qin [6]	Case-control	Hospital cases, 1:1 match for age, sex, and ethnicity, occupation and income; 110 pairs	40–75	71	7.57 (2.41–23.82)
Qiu [6]	Case-control	Hospital cases, 1:1 match for age, sex, and ethnicity; 158 pairs	45–74	72	8.08 (3.19–20.51)
Li [6]	Case-control	Prevalent cases, and community controls, match for age, sex, ethnicity, and occupation; 449 pairs	NA	NA	4.9 (3.2–7.5)
Haemorrhagic stroke
Zhang [5]	Cohort	4400 (58 cases during 13.5 years, Cox model	18–74	100	5.42 (2.9–10)
Wang [22]	Case-control	Hospital cases, 1:1 match for age, sex; 102 pairs	45–74	66	11.8 (4.2–19.8)
Zhao [26]	Case-control	Hospital cases, 1:1 match for age, sex, and site; 126 pairs	NA	NA	9.0 (4.0–20.0)
Xu [6]	Case-control	Hospital cases, 1:1 match for age, sex, ethnicity; 162 pairs	45–74	59	8.72 (4.18–18.19)
Yuan YH [6]	Case-control	Hospital cases, 1:1 match for age, sex, ethnicity, occupation, education and site; 192 pairs	56 (mean age)	63	5.27 (2.73–10.16)
Li [6]	Case-control	Prevalent cases, and community controls, matched for age, sex, and ethnicity, and occupation; 136 pairs	NA	NA	6.8 (3.0–16.9)

Adjusted logistic odds ratio (OR) (95% confidence interval; CI) of ischaemic and haemorrhagic stroke associated with hypertension (>160/95 mmHg). NA, not available.

Eleven cohort studies [28–38] were potentially eligible from 771 references identified for hypertension treated as a continuous variable. Among these, four studies reported only fatal stroke [29, 31, 34, 35] leaving seven cohort studies for inclusion in this meta-analysis.

Out of 280 references, 20 studies [39–55] reporting hypertension as a dichotomous variable (>160/95 mmHg) were potentially eligible. Eight of these did not identify the type of stroke [29, 36, 48–50, 52–54] leaving 12 studies for inclusion. One previous meta-analysis [56] identified a further three studies meeting inclusion criteria [57–59]. Thus, in total, seven cohorts and 15 case-controls were included.

Chinese studies

Tables 1 and 2 present the details of the studies on mainland China. The three cohorts and 10 case-control studies report 3160 stroke events in total.

Hypertension as a continuous variable

Ischaemic stroke Figure 1 shows the Peto plot for the three Asian cohort studies assessing the association between a 10mmHg increase in SBP and ischaemic stroke. There was no heterogeneity (P=0.256) and no publication bias (P = 0.12) among these three individual studies. The pooled coefficient indicated a risk of 1.44 (95% CI 1.39–1.51) for each 10mmHg rise in SBP.

Haemorrhagic stroke Figure 1 also shows the Peto plot for the three Asian cohort studies assessing the association between a 10mmHg increase in SBP and haemorrhagic stroke. There was no heterogeneity (P = 0.454) but there was evidence of publication bias (P = 0.019), although with only three studies this significance is questionable. The pooled coefficient indicated a risk of 1.51 (95% CI 1.45–1.57) for each 10mmHg rise in SBP.

Hypertension as a dichotomous variable

Ischaemic stroke Figure 2 shows the Peto plot for eight studies assessing hypertension (as a dichotomous variable) and the OR of ischaemic stroke. There was no heterogeneity (P = 0.381), and no publication bias (P = 0.41). The pooled OR of ischaemic stroke associated with hypertension was 5.8 (95% CI 4.7–7.2).

Haemorrhagic stroke Figure 3 shows the Peto plot for six studies assessing hypertension as a dichotomous variable and odds ratio of haemorrhagic stroke. Again, there was no heterogeneity (P = 0.579) and no publication bias (P=0.157). The pooled OR of haemorrhagic stroke associated with hypertension was 7.2 (95% CI 5.3–9.7).

Fig. 1

(a) Risk ratio (95% confidence interval) of ischaemic and haemorrhagic stroke for hypertension as 10 mmHg increase in systolic blood pressure (SBP) with non-heterogeneity among Asian cohort studies. The Q tests for heterogeneity were 0.256 for ischaemic stroke, P = 0.454 for haemorrhagic stroke. Egger's test for publication bias: P = 0.12 for ischaemic stroke, P =0.019 for haemorrhagic stroke. (b) Risk ratio (95% confidence interval) of stroke for a 10mmHg increase in SBP with non-heterogeneity among Caucasian cohort studies by two age groups. The Q test values for heterogeneity were 0.236 for the younger group and 0.996 for the older group. Egger's test for publication bias results were 0.263 for the younger group, and 0.713 for the elder group.

Caucasian studies

Tables 3 and 4 present the details of Caucasian studies. The seven cohorts and 15 case-control studies report over 7000 stroke events in total.

Hypertension as a continuous variable

Seven cohort studies assessed hypertension (as a continuous variable) and risk of total stroke. This estimate showed heterogeneity (P = 0.024) but no evidence of publication bias (P = 0.751). The pooled OR was 1.19 (95% CI 1.15–1.23) for each 10mmHg increase in SBP. It was not possible to sub-divide this into ischaemic or haemorrhagic strokes because no studies reported these data separately. It is well established, however, that in Caucasians the majority of these (>80%) would be ischaemic stroke [1].

Meta-regression did not demonstrate reasons for heterogeneity but we suspected age, and hence divided studies into two age groups: average age ≤45 and ≥70 years old. A test for heterogeneity was not significant (P = 0.236), and there was no indication of publication bias (P = 0.263) among the three younger cohort studies; the pooled OR was 1.24 (95% CI 1.20–1.29) for each 10 mmHg increase in SBP (Fig. 1). For the four older Caucasian cohort studies, there was no indication of heterogeneity (P = 0.996), or publication bias (P = 0.713) (Fig. 1) and the pooled OR was 1.15 (95% CI 1.12–1.18) for each 10mmHg increase in SBP.

Hypertension as a dichotomous variable

Ischaemic stroke Figure 2 shows the Peto plot for 10 studies assessing hypertension (as a dichotomous variable) and risk of ischaemic stroke. There was no heterogeneity (P = 0.12), and no publication bias (P = 0.31). The pooled OR for ischaemic stroke risk was 1.96 (95% CI 1.66–2.30).

Haemorrhagic stroke Figure 3 shows the Peto plot for seven studies assessing hypertension as a dichotomous variable and risk of haemorrhagic stroke. There was no heterogeneity (P = 0.11), and no publication bias (P=0.56). The pooled OR for haemorrhagic stroke was 3.11 (95% CI 2.39–4.05).

Fig. 2

(a) Odds ratio (95% confidence interval) of ischaemic stroke for hypertension as defined by 160/95mmHg in Asian case-control studies. Q test for heterogeneity: P = 0.381. Egger's test for publication bias: P = 0.41. (b) Odds ratio (95% confidence interval) of ischaemic stroke for hypertension as defined as 160/95 mmHg in Caucasian case-control studies. Q test for heterogeneity: P = 0.12. Egger's test for publication bias: P = 0.31.

Discussion

Our meta-analysis is the first to address both Caucasian and Asian populations, and the most significant finding is that the magnitude of the effect of hypertension, whether measured continuously or dichotomously, on the risk of ischaemic or haemorrhagic stroke is significantly higher in Chinese than in Caucasians. These results are consistent with other studies: for example, a meta-analysis in the Asia Pacific region indicated that a 10mmHg lower SBP was associated with 41% lower stroke risk in Asia and 30% lower stroke risk in Australia [60]. This has implications at the population level.

The population attributable risks for stroke associated with hypertension in Chinese are 31% for ischaemic and 42% for haemorrhagic [5] compared with 25 and 34%, respectively, in Caucasians [4, 46]. This highlights the need to target hypertension as a public health issue and preventive measure in China. Current evidence indicates that the rate of control of hypertension is lower than in Caucasians. For example, among hypertensive patients, only 45% were aware of their high blood pressure, 28% were taking antihypertensive medicine, and 8% achieved blood pressure control (140/90 mmHg) [61]. Decreasing the incidence of hypertension may entrain certain reductions in stroke beyond that expected from the PAR. For example, in China, the prevalence of smoking is much higher than in Western countries [3, 5], and the deleterious impact on stroke, particularly for ischaemic stroke, may be potentiated in combination with hypertension.

Fig. 3

(a) Odds ratio (95% confidence interval) of haemorrhagic stroke for hypertension as defined by 160/95 mmHg in Asian case-control studies. Q test for heterogeneity: P = 0.579. Egger's tests for publication bias: P = 0.157. (b) Odds ratio (95% confidence interval) of haemorrhagic stroke for hypertension as defined by 160/95 mmHg in Caucasian case-control studies. Q test for heterogeneity: P = 0.11. Egger's test for publication bias: P = 0.56.

The reasons for this difference between blood pressure and stroke risk among Chinese and Caucasians are uncertain. It could be partially confounded by other factors, such as a higher proportion of the study population on antihypertensive medications in Caucasians versus Chinese, or the differential effect and distribution of hypercholesterolaemia on ischaemic and haemorrhagic stroke. It could also indicate a difference in genetic make-up or environmental factors between ethnic groups. For example, Chinese in New York City had a greater proportion of deaths due to haemorrhagic stroke than did whites in New York although there were similar rates of total stroke for both New York Chinese and whites. Haemorrhagic stroke deaths among New York City Chinese were intermediate between Chinese in China (highest) and New York City Caucasians (lowest) [62].

Table 3

Characteristics of Caucasian studies treating hypertension as a continuous variable

Study (outcome)	Sample size (stroke events)	Age range (years)	Mean age	Average follow-up (years)	Men %	RR for SBP 10mmHg increase (95% CI)
Finnmark [28] (incidence)	7002 (144)	35–52	41.5	14.5	100	1.31 (1.19–1.42)
Oslo [30] (incidence)	14403 (81)	40–49	45	12	100	1.25 (1.19–1.31)
Copenhagen [35] (incidence)	3015 (286)	55–84	62.9	10	100	1.15 (1.1–1.21)
Framingham [36] (incidence)	2372 (472)	55–84	65.4	10	100	1.16 (1.1–1.24)
CHS [32] (incidence)	4902 (385)	Over 65	72.6	6.7	40	1.15 (1.09–1.20)
North East of England [33] (incidence)	4351 (329)	Over 65	76	5	46	1.15 (1.06–1.24)
Britain [37] (incidence)	7735 (467)	40–59	50	20	100	1.21 (1.16–1.26)

Adjusted Cox proportional model hazard ratio (95% confidence interval; CI) of total stroke associated with a 10 mmHg increase in systolic blood pressure (SBP). RR, risk ratio; CHS, the Cardiovascular Health Study.

Table 4

Characteristics of Caucasian studies treating hypertension as a dichotomous variable

Study	Design	Sample size, and matched variables	Age range (years) or average age ± SD	Ischaemic stroke	Haemorrhagic stroke
Cohort studies
Woo [46]	Population-based case-control study	188 cases of intracerebral haemorrhage, 366 control subjects, matched by age, sex and race	65 ± 16	2.2 (1.5–3.5)
Sacco [4]	Population-based 1:2 matched case-control study	688 cases of ischaemic stroke, matched by age, sex and race-ethnicity	70 ± 13	2.0 (1.5–2.5)
Feigin [47]	Population-based case-control study	237 ischaemic stroke, 237 controls, matched by age, sex	68 ± 9	3.8 (2.12–6.7)
Rodriguez [39]	Framingham cohort study	89 ischaemic stroke, 18 haemorrhagic stroke among 1216 men followed-up 20 years	56 ± 6	1.8(1.2–2.9)	3.4(1.3–8.6)
Simons [40]	Cohort study	306 stroke events among 2808 subjects over 60 years with 8 years follow-up	60–80+	1.37 (1.04–1.80)
Rohr [43]	Population-based case-control study	296 events among 1220 adults	18–44	1.6 (0.7–3.2)
Whisnant [44]	Nested case-control study	1444 events during 25 years follow-up, match 1:1 with controls	NA	2.0 (1.61–2.52)
Al-Roomi [55]	A population-based case-control	167 stroke events (91 ischaemic, 31 haemorrhagic, 45 unspecified), 480 controls	35–69	2.53 (1.48–4.34)	5.50 (2.36–12.8)
Case-control studies
Henrich [41]	Case-control	90cases (56 men), 174 controls (106 men)	NA	3.4 (1.9–6.0)
Haapaniemi [42]	Hospitalized case-control study	506 events (366 men, 140 women), 345 controls (219 men, 126 women)	16–60 2.7 (1.6–4.6) for men
Whisnant [45]	Case-control study	931 cases match 1:1 with controls	NA	1.8 (1.4–2.2)
Thrift [51]	Case-control	Intracerebral haemorrhage 331 case (200 men), 331 controls (200 men)	18–80	2.45 (1.61–3.73)

Multivariate risk factor-adjusted hazard ratio (95% confidence interval) or odds ratio (95% confidence interval) of stroke associated with hypertension (>160/95 mmHg) among Caucasian cohort studies. NA, not available.

Conclusion

This is first meta-analysis to compare the effect of hypertension on ischaemic and haemorrhagic stroke between Chinese and Caucasians, respectively. The ORs and RRs of both ischaemic and haemorrhagic stroke are consistently and significantly greater in Chinese than Caucasians. Such ethnic differences may assist genetic epidemiologists to dissect the causes of stroke, but also emphasizes that public health interventions targeting hypertension control among Chinese are critical in order to avoid the high morbidity and mortality of cardiovascular diseases.

References

journal

Yusuf

Reddy

Ôunpuu

Anand

Global Burden of cardiovascular diseases part II: variations in cardiovascular disease by specific ethnic groups and geographic regions and prevention strategies. Circulation 2001; 104:2855–2864.

journal Zhang

Yang

Hong

Yuan

Zou

Zhao

. Proportion of different subtypes of stroke in China. Stroke 2003; 34:2091–2096.

journal

Chen

Collins

Peto

Early health effects of the emerging tobacco epidemic in China. JAMA 1997; 278:1500–1504.

journal Sacco

Boden-Albala

Abel

Lin

Elkind

Hauser

. Raceethnic disparities in the impact of stroke risk factors: the Northern Manhattan Stroke Study. Stroke 2001; 32:1725–1731.

journal

Zhang

Attia

D'Este

XH.

Prevalence and magnitude of classical risk factors for stroke in a cohort of 5092 Chinese steelworkers over 13.5 years follow-up. Stroke 2004; 35:1052–1056.

journal

Klag

Whelton

PK.

Stroke in the People's Republic of China: II. Meta-analysis of hypertension and risk of stroke. Stroke 1995; 26: 2228–2232.

journal Eastern Stroke and Coronary Heart Disease Collaborative Research Group. Blood pressure, cholesterol, and stroke in eastern Asia. Lancet 1998; 352:1801–1807.

journal Zhou

Stamler

Dennis

Moag-Stahlberg

Okuda

Robertson

. Nutrient intakes of middle-aged men and women in China, Japan, United Kingdom, and United States in the late 1990s: The INTERMAP study. J Hum Hypertens 2003; 17:623–630.

journal

Ross

Yuan

Henderson

Park

Gao

MC.

Prospective evaluation of dietary and other predictors of fatal stroke in Shanghai, China. Circulation 1997; 96:50–55.

10.

journal

Chen

WC.

Blood pressure and mortality rates in some rural communities of Shanghai [in Chinese]. Zhong Hua Liu Xing Bing Xue Za Zhi 1987; 8:80–83.

11.

journal

Zhang

Zhou

GD.

Indoor air pollution of coal fumes as risk factors of stroke, Shanghai. Am J Public Health 1988; 78:975–977.

12.

journal

Yen

Wang

SY.

A cohort study of relation between hypertension and stroke [in Chinese]. Zhong Hua Liu Xing Bing Xue Za Zhi 1985; 6:36–38.

13.

journal

Xien

Gan

Tang

SH.

Risk factors of cerebral infarction [in Chinese]. Zhong Hua Shen Jing He Jing Shen Ji Bing Za Zhi 1990; 16:216–220.

14.

journal Xu

Zheng

Wang

Shen

Chen

. Cox regression analysis of risk factors in Shanghai urban residents [in Chinese]. Zhong Hua Liu Xing Bing Xue Za Zhi 1994; 15:94–98.

15.

journal

Chen

Lai

Wang

Yen

SH.

A prospective study of risk factors for stroke in the steel workers in Beijing: case-control analysis [in Chinese]. Zhong Hua Xin Xue Guan Bing Za Zhi 1987; 15:239.

16.

journal Zhou

Liu

Yang

Yao

. Pulse pressure in prediction of incidence of cardiovascular disease in middle-aged Chinese cohort [in Chinese]. Chinese Journal of Cardiology 2002; 30:687–691.

17.

journal Guo

Wnag

Zhang

Qian

Wang

Zhou

. A population-based prospective cohort study on risk factors of stroke in Jintan. Chinese Journal of Cardiology 2002; 30:563–566.

18.

journal

Zhang

Yang

Zhou

Tao

SQ.

A prospective study of risk factors for stroke in 10 Chinese group populations [in Chinese]. Chinese Journal of Chronic Disease Prevention and Control 1996; 4:150–172.

19.

journal Wu

Yao

Zhao

Wang

Liu

. A prospective cohort study on cardiovascular disease incidence in 11 provinces of China I. Associations between risk factor level and cardiovascular disease incidence [in Chinese]. Chinese Journal of Cardiology 1999; 27:5–8.

20.

journal

Huang

Zhang

Zhou

TS.

The attributable risk analysis of ischaemic stroke [in Chinese]. Gui Zhou Yi Yao 1998; 22:144–146.

21.

journal

Niu

Yang

Wang

SP.

Logistic regression analysis on the risk factors for cerebral infarction patients aged 18 to 80 in Shanxi Province, China [in Chinese]. Chinese Journal of Chronic Disease Prevention and Control 2003; 11:102–103.

22.

journal

Wang

Geng

Qiu

WL.

Comparison of risk factors between haemorrhagic and ischaemic stroke [in Chinese]. Chinese Journal of Chronic Disease Prevention and Control 1996; 4:9–10, 17.

23.

journal

Zhu

Ding

Song

Logistic regression of risk factors in haemorrhagic stroke and ischaemic stroke [in Chinese]. Chinese Journal of Chronic Disease Prevention and Control 2001; 9:243–269.

24.

journal

Yang

Lin

Dai

Jin

Yang

Huang

XS.

Logistic regression of risk factors in haemorrhagic and ischaemic stroke [in Chinese]. Chinese Journal of Chronic Disease Prevention and Control 1995; 3:3–6.

25.

journal

Wang

Xing

Han

XL.

Comparison risk factors between ischaemic and haemorrhagic stroke [in Chinese]. China Public Health 1997; 13:335–337.

26.

journal

Zhao

Zhang

Fang

Wang

Gao

TY.

Study on case-control of risk factors of cerebral haemorrhage. China Public Health 2000; 16:99–100.

27.

journal

Qiu

Yan

Liu

LX.

The relation between smoking, alcohol, hypertension and haemorrhage stroke. Zhong Guo Gong Gong Wei Sheng Xue Bao 1994; 13:286–288.

28.

journal

Njolstad

Arnesen

Lund-Larsen

PG.

Body height, cardiovascular risk factors, and risk of stroke in middle-aged men and women: a 14–year follow-up of the Finnmark study. Circulation 1996; 94:2877–2882.

29.

journal

Rastenyte

Tuomilehto

Domarkiene

Cepaitis

Reklaitiene

Risk factors for death from stroke in middle-aged Lithuanian men: results from a 20–year prospective study. Stroke 1996; 27:672–676.

30.

journal

Haheim

Holme

Hjermann

Leren

risk factors of stroke incidence and mortality: a 12–year follow-up of the Oslo study. Stroke 1993; 24:1484–1489.

31.

journal

Hart

Hole

Smith

GD.

Risk factors and 20–year stroke mortality in men and women in the Renfrew/Paisley study in Scotland. Stroke 1999; 30:1999–2007.

32.

journal Psaty

Furberg

Kuller

Cushman

Savage

Levine

. Association between blood pressure level and the risk of myocardial infarction, stroke, and total mortality: The Cardiovascular Health Study. Arch Intern Med 2001; 161:1183–1192.

33.

journal

Rodgers

Greenaway

Davies

Wood

Steen

Thomson

Risk factors for first-ever stroke in older people in the north east of England: a population-based study. Stroke 2004; 35:7–11.

34.

journal

Knuiman

HTV.

Risk factors for stroke mortality in men and women: the Busselton Study. J Cardiovasc Risk 1996; 3:447–452.

35.

journal

Truelsen

Lindenstrøm

Boysen

Comparison of probability of stroke between the Copenhagen City Heart Study and the Framingham Study. Stroke 1994; 25:802–807.

36.

journal

D'Agostino

Wolf

Belanger

Kannel

WB.

Stroke Risk Profile: Adjustment for Antihypertensive Medicine. The Framingham Study. Stroke 1994; 25:40–43.

37.

journal

Morris

Whincup

Emberson

Lampe

Walker

Shaper

AG.

North-South gradients in Britain for stroke and CHD: are they explained by the same factors? Stroke 2003; 34:2604–2611.

38.

journal Miura

Dyer

Greenland

Daviglus

Hill

Liu

. Pulse pressure compared with other blood pressure indexes in the prediction of 25–year cardiovascular and all-cause mortality rates: the Chicago Heart Association Detection Project in Industry Study. Hypertension 2001; 38:232–237.

39.

journal Rodriguez

D'Agostino

Abbott

Kagan

Burchfiel

Yano

. Risk of hospitalized stroke in men enrolled in the Honolulu Heart Program and the Framingham study: A comparison of incidence and risk factor effects. Stroke 2002; 33:230–237.

40.

journal

Simons

Friedlander

McCallum

Cholesterol and other lipids predict coronary heart disease and ischaemic stroke in the elderly, but only in those below 70 years. Atherosclerosis 2001; 159:201–208.

41.

journal

Henrich

Horwitz

RI.

The contributions of individual factors to thromboebolic stroke. J Gen Intern Med 1989; 4:195–201.

42.

journal

Haapaniemi

Hillbom

Juvela

Lifestyle-associated risk factors for acute brain infarction among persons of working age. Stroke 1997; 28:26–30.

43.

journal Rohr

Kittner

Feeser

Hebel

Whyte

Weinstein

. Traditional risk factors and ischaemic stroke in young adults: the Baltimore-Washington Cooperative Young Stroke Study. Arch Neurol 1996; 53:603–607.

44.

journal

Whisnant

Wiebers

O'Fallon

Sicks

Frye

RL.

A population-based model of risk factors for ischaemic stroke: Rochester, Minnesota. Neurology 1996; 47:1420–1428.

45.

journal

Whisnant

Brown

Petty

O'Fallon

Sicks

Wiebers

DO.

Comparison of population-based models of risk factors for TIA and ischaemic stroke. Neurology 1999; 53:532–536.

46.

journal Woo

Sauebeck

Kissela

Khoury

Szaflarski

Gebel

. Genetic and environmental risk factors for intracerebral haemorrhage: preliminary results of a population-based study. Stroke 2002; 33: 1190–1196.

47.

journal

Feigin

Wiebers

Nikitin

O'Fallon

Whisnant

JP.

Risk factors for ischaemic stroke in a Russian community: a population-based case-control study. Stroke 1998; 29:34–39.

48.

journal

Berger

Schulte

Stögbauer

Assmann

Incidence and risk factors for stroke in an occupational cohort: the PROCAM study. Stroke 1998; 29:1562–1566.

49.

journal

Herman

Schmitz

PIM

Leyten

ACM

van Luijk

Frenken

CGM

Coul

AAW

Schulte

BPM.

Multivariate logistic analysis of risk factors for stroke in Tilburg, the Netherlands. Am J Epidemiol 1983; 118:514–525.

50.

journal

Beghi

Bogliun

Cosso

Fiorelli

Lorini

Mandelli

Bellini

Stroke and alcohol intake in a hospital population: a case-control study. Stroke 1995; 26:1691–1696.

51.

journal

Thrift

McNeil

Forbes

Donnan

GA.

Three important subgroups of hypertensive persons at greater risk of intracerebral haemorrhage. Hypertension 1998; 31:1223–1229.

52.

journal

McNamee

Cruickshank

Stroke risk from multiple risk factors combined with hypertension: a primary care based case-control study in a defined population of northwest England. Ann Epidemiol 2000; 10:380–388.

53.

journal

Weinehall

Stegmayr

Boman

Hallmans

Wall

What else adds to hypertension in predicting stroke? An incident case-referent study. J Intern Med 2000; 248:475–482.

54.

journal

Martin-Gonzalez

Romero-Escobar

Vioque

Peris

Matias-Guiu

Hyperuricemia as a risk factor for cerebrovascular accident: a case-control study. Rev Neurol 2000; 31:8–13.

55.

journal

Al-Roomi

Heller

Holland

Floate

Wlodarczyk

The importance of hypertension in the aetiology of infarctive and haemorrhagic stroke: the lower Hunter Stroke Study. Med J Aust 1992; 157:452–455.

56.

journal

Ariesen

Clasus

Rinkel

GJE

Algra

Risk factors for intracerebral haemorrhage in the general population: a systematic review. Stroke 2003; 34:2060–2066.

57.

journal

Monforte

Estruch

Graus

Nicolas

Urbano-Marquez

High ethanol consumption as risk factor for intracerebral hemorrhage in young and middle-aged people. Stroke 1990; 21:1529–1532.

58.

journal

Juvela

Hillborn

Palomaki

Risk factors for spontaneous intracerebral haemorrhage. Stroke 1995; 26:1558–1564.

59.

journal

Saloheimo

Juvela

Hillborn

Use of aspirin, epistaxis, and untreated hypertension as risk factors for primary intracerebral haemorrhage in middled-aged and elder people. Stroke 2001; 32:339–404.

60.

journal Asia Pacific Cohort Studies Collaboration. Blood pressure and cardiovascular disease in the Asia Pacific region. J Hypertens 2003; 21:707–716.

61.

journal Gu

Reynolds

Chen

Duan

Muntner

. Prevalence, awareness, treatment, and control of hypertension in China. Hypertension 2002; 40:920–927.

62.

journal

Fang

Madhavan

Alderman

Cardiovascular mortality of Chinese in New York City. J Urban Health 1999; 76:51–61.