Cardiovascular Disease in Special Populations III: Stroke


Address for correspondence: Philip R. Liebson, MD, Rush University Medical Center, 1635 West Congress Pkwy, Chicago, IL 60612

Stroke is a sudden devastating event that, if it does not kill immediately, leaves the patient with various degrees of disability that affect physical and psychological well-being. It is the third most common cause of death in the United States. Unlike a myocardial infarction, in which the general functioning of ambulation, speech, and thought processes is left intact, stroke produces a physical and psychological dependence on caregivers, leads to a sense of isolation unlike most other cardiovascular (CV) events, and, especially in the case of special populations such as many blacks, puts a powerful strain on economic resources. In addition, there is evidence that the risk for stroke and interventions to the stroke patient in special populations may vary. Blacks, Hispanics, American Indians/Alaska Natives, and Asians have a higher mortality at younger ages than whites.1 In 2005, the prevalence of stroke among Native Americans (6.0%), multiracial persons (4.6%), and blacks (4.0%) was higher than whites (2.3%), with Asians and Hispanics having similar prevalence as whites. Higher educational status is also associated with a history of stroke, from 1.8% in college graduates to 4.4% in persons with less than 12 years of education.

The particularly excessive burden of stroke in blacks is seen especially in the southeastern United States and in those who are relatively young (aged 35–64 years).2 Of stroke deaths in 2002, 12% occurred in persons younger than 65 years and the proportion of stroke deaths in this younger age group was higher among blacks, Native Americans, and Asians compared with whites.2 This amounted to 3400 excess strokes in blacks in this younger population compared with whites. Although mortality rates have decreased in both blacks and whites, there remains an unchanging magnitude of the excess rates in blacks compared with whites.3

Although the mortality from stroke in the United States per population has decreased ever since records were first kept early in the 20th century,4 the vast increase in population and increase in average lifespan has led to many more cases. Trends in the age-, race-, and sex-adjusted prevalence of stroke between 1971 and 1994 from the National Health and Nutrition Examination Surveys (NHANES I to NHANES III) demonstrated an increase from 1.4% to 1.9%, an average increase of 7.5%.4 The noninstitutionalized stroke survivor rate increased by 60% during this period, from 1.5 million to 2.4 million. As for racial differences, 5-year mean change in stroke prevalence in noninstitutionalized survivors of stroke increased by 28% in black women and 12% in white men, but decreased by 3% in black men and white women (Table I). However, in terms of number-estimated noninstitutional stroke survivors, 5-year changes in numbers of survivors increased in all groups, from an increase of 18,000 in black men to 159,000 in white men. These differences of course were accounted for by the 10-fold larger population of whites than blacks.

Table I.   Age-Adjusted Prevalence of Stroke Among Noninstitutionalized White and Black Men and Women in the United States for 3 Time Periods
Population Age, 25–74 yNHANES I (1971–1975)
NHANES II (1976–1980)
NHANES III (1988–1994)
5-Year Mean (95% CI) Change in Stroke Prevalence
  1. Abbreviations: CI, confidence interval; SE, standard error. aNumber of persons self-reporting a history of stroke in the National Health and Nutrition Examination Survey (NHANES). From Muntner et al4 with permission from the American Heart Association, Inc. ©1995.

White men, No.a110167116 
 Adjusted prevalence (SE)1.4 (0.3)1.7 (0.2)2.0 (0.3)+12% (−2%, +29%)
White women, No.a11115697 
 Adjusted prevalence (SE)1.5 (0.3)1.7 (0.2)1.4 (0.2)−3% (−15%, +11%)
Black men, No.a272751 
 Adjusted prevalence (SE)1.9 (0.6)2.6 (0.8)2.1 (0.3)−3% (−15%, 11%)
Black women, No.a423360 
 Adjusted prevalence (SE)0.8 (0.4)2.6 (0.6)2.8 (0.4)+28% (+3%, +56%)

An intriguing finding in terms of possible environmental implications is that the stroke death rate is particularly high in the southeastern United States, as previously mentioned.5 Geographic difference in stroke prevalence may relate to cultural differences in diet and exercise, lack of economic opportunity, and regional differences in health care and preventive services. Both environmental risk factors and new evidence for genetic determinants leading to risk for stroke in special populations have to be taken into consideration in evaluation of preventive interventions.

Incidence, Prevalence, and Risk for Stroke


Let us focus initially on the “stroke belt,” previously described, to discuss the impact of stroke on blacks and whites. The stroke belt was identified in 1965 as a region where a 50% higher stroke mortality rate was found in the southeastern United States.6 The higher stroke mortality in the southeastern United States was evaluated by a 2003 Behavioral Risk Factor Surveillance System (BRFSS) survey.2 A comparison was made among southeastern and non-southeastern states. A total of 95,598 persons responded, including a substantial proportion of both blacks and whites. The highest age-adjusted prevalence of stroke was found in southeastern blacks (3.4%), followed by non-southeastern blacks (2.8%), southeastern whites (2.5%) and non-southeastern whites (1.8%). Increased characteristics of southeastern blacks vs whites that might explain the increased stroke prevalence include lower education levels and higher prevalence of diabetes and high blood pressure. Health care coverage among blacks is also lower in the region than among whites and those in non-southeastern states. Other considerations that have been advanced for regional variations include lower intake of animal protein, potassium and calcium, and higher intake of sodium and complex carbohydrates in the southeastern states.

More recent evaluations of black populations in southern states indicate a black-to-white stroke mortality ratio from 6% to 21% higher in these states than non-southern states, even in southern states that are not part of the stroke belt (Virginia, Florida, and the lower Mississippi region).6

The higher stroke prevalence rate in blacks vs whites is due to higher incidence rates. In a population-based study in carried out in 1993–1994 in the Greater Cincinnati/Northern Kentucky region, fairly representative of demographic and socioeconomic characteristics, stroke incidence in blacks was higher at any age with the greatest risk, 2- to 5-fold, in young and middle-aged blacks compared with whites.7 In terms of the type of stroke experienced (ischemic, intracerebral hemorrhage, and subarachnoid hemorrhage), the race-specific incidence rates were higher for blacks in all categories, with a hazard ratio of 1.8 to 2.0 compared with whites. This racial disparity was accounted for by the higher prevalence of risk factors for stroke in blacks such as hypertension, diabetes, smoking, and alcohol use, as well as decreased access to health care, similar to risk factors in the southern stroke belt.

Further assessment of the hospitalized and autopsied cases of stroke and transient ischemic attacks in this population of 187,000 blacks for first-ever ischemic subtypes demonstrated that strokes of uncertain cause were most common (103), followed by cardioembolic strokes (56 per 100,000), small-vessel infarcts (52), large-vessel infarcts (17), and other causes (17).8 However, the study investigators indicated that the incidence rates of cardioembolic and large-vessel strokes were probably underestimated because of the paucity of studies of carotid ultrasound and cardiac ultrasound, the latter to rule out cardiac sources of emboli (only 23% of patients had both studies). These results compared with a lower rate of 40 per 100,000 cardioembolic strokes in the predominantly white population of Rochester, Minnesota.8,9 The incidence rates of other causes of ischemic stroke in blacks were 4 times that of the Rochester population. There appears to be a particular propensity for middle cerebral artery and supraclinoid carotid artery atherosclerosis compared with whites.10,11


Less study has been accomplished on stroke characterization in Hispanics, although this group is now considered the largest minority population in the United States. Mexican Americans constitute by far the largest subgroup of Hispanic Americans and as such have been exclusively evaluated as representative of Hispanic populations in the NHANES surveys, for example.

In 1990, cerebrovascular disease was the fourth leading cause of deaths in Hispanics in the United States. At that time, stroke deaths were similar between middle-aged (45–64 years) Hispanics and non-Hispanic whites but substantially lower in Hispanics 65 years and older12 (Figure 1). At the time, Hispanics had lower levels of hypertension and hypercholesterolemia than non-Hispanic whites but higher levels of diabetes, smoking, and overweight.

Figure 1.

 National Center for Health Statistics. Graphs show stroke death rates by Hispanic origin, race, and age in men (top) and women (bottom) in 1989–1991. From Gillum et al12 with permission from the American Heart Association, Inc. ©1995.

More recently, the stroke incidence among Mexican Americans and non-Hispanic whites was evaluated in a Texas county from January 2000 to December 2002 as part of the Brain Attack Surveillance in Corpus Christi Project (BASIC).13 Mexican Americans had a 24% higher cumulative incidence of strokes than non-Hispanic whites (168 per 10,000 vs 136 per 10,000). This included a higher incidence of both ischemic and intracerebral stroke. Subarachnoid hemorrhage was also more common but was of borderline statistical significance.

In the BASIC study, subarachnoid hemorrhage disproportionately affected Mexican Americans and women, although no ethnic difference was found in in-hospital mortality or discharge disability rates.14 Aside from increased vascular malformations in Hispanics, other possible reasons include increased prevalence of hypertension, heavy alcohol consumption, and cigarette smoking in Hispanics. For sex differences, possible explanations have included differences in hypertension, smoking, and estrogen effects in women.

In comparative studies of intracerebral hemorrhages in Mexican Americans and non-Hispanic whites in the BASIC project, Mexican Americans were more likely to have smaller, nonlobar hemorrhages.15 The Hispanics had a higher prevalence of diabetes and a lower prevalence of coronary artery disease. The higher characteristic nonlobar stroke incidence in Mexican Americans may relate to a possible increased risk for this pathology in diabetics.

The BASIC project also evaluated stroke burden in Mexican American women.16 Elevated stroke risk was present in both younger and older Mexican American women compared with non-Hispanic white women. At younger ages (45–59 years), stroke risk was increased by 100% and stroke risk between ages 60 and 74 years by 57%. (Table II). Even in persons 75 years and older, stroke risk was increased by 13%, although this risk was of borderline significance. There was no difference in stroke severity or stroke type. The increased stroke risk in Mexican American women was associated with increased hypertension and diabetes.

Table II.   Cumulative Risk of Stroke Among Women by Ethnicity and Age Group (BASIC Project 2000–2006)
Age, yMexican AmericanNon-Hispanic White 
CasesNo.Risk95% CICasesNo.Risk95% CIRR95% CIP Value
  1. Abbreviations: CI, confidence interval; BASIC, Brain Attack Surveillance in Corpus Christi; RR, relative risk. From Lisabeth et al16 with permission from The American Journal of Hypertension©2008.

Ischemic stroke
Intracerebral hemorrhage
Total stroke

Interracial and Intraracial Variations

The relative incidence of strokes among blacks, Hispanics, and whites has been evaluated in the Northern Manhattan Study (NOMAS).17 An evaluation of ischemic stroke between 1993 and 1997 in a section of New York City, demonstrated increased incidences of stroke subtypes in both blacks and Hispanics compared with non-Hispanic whites. This difference was as high as an annual age-adjusted intracranial atherosclerotic stroke risk ratio of 5:1 or above in both blacks and Hispanics compared with non-Hispanic whites but also included increased risk ratios in both minority groups for extracranial atherosclerotic stroke, lacunar stroke, and cardioembolic stroke. A salient feature as well was the younger age of the minority groups at the time of first stroke compared with whites (66 years for Hispanics and 71 years for blacks, compared with 77 years for whites). As reflected in other analyses, the prevalence of stroke risk factors in both minority ethnic groups, including hypertension, hypercholesterolemia, and smoking was much higher than in whites. The only stroke subcategory in which whites had a greater risk was cardioembolic stroke, reflecting the greater prevalence of atrial fibrillation among whites than among blacks and Hispanics.

Native Americans

Relatively few studies have evaluated stroke in Native Americans. This ethnic population has a relatively high rate of tobacco abuse, diabetes, hypertension, and elevated cholesterol levels.18 The rates of stroke are higher than in US whites. Cerebrovascular disease is the sixth leading cause of death in this population, with adjusted rates of 29.6 per 100,000 vs 24.0 for whites.19 Although the risk factors for CV disease are higher in Native Americans than whites, national vital event data suggest a lower mortality for this ethnic/racial group than the general US population. The concern about this finding is the possibility of racial misclassification. Focused epidemiologic studies on Native American populations such as the Strong Heart Study suggest that CV disease incidence and mortality rates in Native Americans are as bad as or worse than that of general US populations.19 A study by the National Center for Health Statistics found an underestimation of death rates of 21% for Native Americans, 11% for Asians, and 2% for Hispanics but an overestimation by 5% for blacks and 1% for whites.20

In general, the prevalence of CV risk factors among US ethnic groups is second highest among Native Americans (47%), compared with blacks (49%).21 In the entire United States, stroke mortality in Native Americans increased from 41 to 45 per 100,000 from 1990 to 2000.22

The recent report from the Strong Heart Study provides an update of incidence and risk factors for stroke in the Native American population.23 This is the largest longitudinal population-based study of CV disease and its risk factors in a diverse group of Native Americans. The study covered the period from 1989 to 1992, when baseline examinations were accomplished, through December 2004. Nonhemorrhagic ischemic infarction occurred in 86% of patients with stroke, with hemorrhagic stroke occurring in the remaining. The incidence rate and case-fatality rate were found to be higher than in US white and black populations of the same age range. Hypertension, diabetes, and smoking were strongly associated with stroke, as was to be expected.

Correlations With Underlying Clinical Conditions

Atrial fibrillation is associated with increased risk for stroke, and part of the pathophysiology of atrial fibrillation is increase in left atrial size. In studies of the Atherosclerosis Risk in Communities (ARIC) black cohort, left atrial size itself was significantly related to ischemic stroke.24 Other factors associated with stroke in addition to left atrial size were female sex, obesity, hypertension, and diabetes. These factors may lead to left ventricular hypertrophy, itself a risk factor, but adjustment for left ventricular hypertrophy and low left ventricular ejection fraction attenuated but did not eliminate the risk from left atrial size.

In patients with atrial fibrillation, the use of anticoagulants could lead to stroke due to intracranial hemorrhage. Non-whites are at greater risk for intracranial hemorrhage than whites.25 In a multiethnic study of nonrheumatic atrial fibrillation in 18,867 hospitalized patients between 1995 and 2000, the magnitude of increased risk for intracranial hemorrhage associated with warfarin use was greater for non-whites than whites.25 The hazard ratio compared with whites was 4.1 for Asians and 2.1 for Hispanics and blacks (Figure 2). Possible explanations for this disparity included polymorphisms of the P450 cytochrome CYP2C9, the enzyme responsible for metabolizing warfarin or for variants in the gene for vitamin K epoxide reductase complex 1, the target enzyme for warfarin. Differences have been found in such haplotypes in African Americans and Asian Americans than in European Americans.26 For example, haplotypes predictive of low maintenance dose of warfarin appear to be present in 89% of Asians, 35% of whites, and only 14% of blacks.27 The Asian population, therefore, appears to require the lowest daily dose of warfarin and the black population the highest daily dose, with the white population in between.26 In terms of the use of warfarin to prevent left atrial thrombi in atrial fibrillation, approximately one third of strokes in atrial fibrillation are nonthromboembolic. Thus, variations of warfarin dosage for appropriate international normalized ratios among different ethnic groups may favor an increase in intracranial hemorrhage or cardioembolic stroke.

Figure 2.

 Adjusted intracerebral hemorrhage (ICH) event-free survival curves by race for patients treated with warfarin for atrial fibrillation. From Shen et al25 with permission from the American College of Cardiology ©2007.


The prognosis of stroke in racial groups varies considerably, as does its epidemiology and prevalence. Much of the variation is due to health disparities related not only to ethnicity but regional variations. Recent evidence bolsters evidence for a greater burden of disease in stroke, greater mortality, and greater severity of strokes in blacks.28Less conclusive is evidence for differences in acute and postacute care as well as disparities among other ethnic groups.

In regard to stroke mortality, for example, US age-adjusted death rates from 1999 were 35% higher in blacks than non-Hispanic whites. In data from the late 1990s, mortality rates for intracerebral hemorrhage were 1.7 times higher in blacks and 1.5 times higher in Asian/Pacific Islanders than whites.28 Although stroke deaths have declined during the past few decades, rates of decline in black men are lower.

Stroke mortality was the third leading cause of death in black women and the sixth in black men in the United States in 1996.29 Age-adjusted death rates for stroke per 100,000 persons were 50.9 for black men, 39.2 for black women, 22.9 for white women, and 26.3 for white men. Stroke death rates for blacks in the 1990s were similar to rates in Japan but lower than those in Eastern Europe. Stroke mortality rates for blacks in Africa and the Caribbean region are also relatively high, but there are concerns about possible death certification inaccuracies and demographic analysis in these regions. Recently, there has been a slowdown in the decline of US stroke mortality rates in both blacks and whites (Figure 3).

Figure 3.

 National Center for Health Statistics. Age-adjusted stroke mortality rates in the United States, 1979–1996. Since the early 1980s, a marked slowdown has occurred in the decline in US stroke mortality in black and white Americans. From Gillum et al29 with permission from the American Heart Association, Inc. ©1999.

A large study of 47,045 elderly blacks and whites selected randomly from a Medicare sample to determine poststroke survival found that blacks were 6% more likely to die after cerebral infarction than whites.30 For those between 65 and 74 years of age, 3-year mortality for blacks was 20% higher than for whites. This survival difference decreased, however, in older age groups.

For patients younger than 65 years, based on a 2002 death certificate surveillance, both the mean ages and stroke mortality rates were higher in blacks, Native Americans, and persons of Asian descent than whites31 (Figure 4). Put another way, age-adjusted estimates of potential life lost before age 75 from stroke were more than twice as high for blacks than for all other racial groups.

Figure 4.

 Age-specific death rates* for stroke among persons younger than 75 years by race and age group: United States, 2002. With permission from the Department of Health and Human Services Centers for Disease Control and Prevention.31

The higher mortality rate in blacks may be attributed to greater incidence. However, in terms of case fatality, although blacks have a higher incidence rate, the case-fatality rate is similar to that of whites.32 In at least one regional study of temporal trend in case fatality rates in the Greater Cincinnati/Northern Kentucky region between 1993 and 1999, case fatality rates did not change in either racial group.32 The lack of substantive decreases in risk factors for stroke could have accounted for the lack of change of stroke incidence in this analysis.

On the other hand, for Mexican Americans, using the BASIC project database, evaluating stroke cases between 2000 and 2002, 28-day all-cause mortality was much lower than for non-Hispanic whites (7.8% vs 13.5%).33 For 3-year outcomes, the lower mortality rates for Mexican Americans vs whites persisted (31.3% vs 47.2%). However, stroke subtypes and severity were similar. These results have been termed the “Mexican paradox.” It should be emphasized that the Mexican Americans studied were fully acculturated, being 2nd- and 3rd-generation US-born persons. However, other investigations have reported a higher Hispanic mortality than non-Hispanic whites.34

Rehabilitation and Secondary Prevention

There are racial disparities in CV risk factors after stroke that reflect risks leading to stroke. A National Health Interview Survey (NHIS) of 96,501 persons between 1999 and 2001 evaluated 2265 stroke survivors.35 Hypertension was 65% more prevalent in black than white survivors and 27% less prevalent in Hispanics than non-Hispanic whites. Hispanics and blacks were less likely to have coronary heart disease than non-Hispanic whites but more likely to have diabetes. Lower physical activity and overweight were considerably more prevalent in blacks and Hispanics than whites (Table III). These considerations are important in targeting secondary prevention measures among different racial groups.

Table III.   Distribution of Selected Self-Reported Cardiovascular Disease Risk Factors Among Persons Living With Stroke by Ethnicity: United States 1999–2001
 Black, Non-Hispanic, % (SE)Hispanic, % (SE)White, Non-Hispanic, % (SE)
  1. aP<.05 compared with whites. bAn affirmative answer on the National Health Interview Survey for either myocardial infarction or coronary heart disease question. cCurrent smoker=smoked ≥100 cigarettes in lifetime and currently smokes every day or some days. dOverweight=body mass index ≥25 kg/m2. eObesity=body mass index ≥30 kg/m2. fInadequate physical activity=performed moderate exercise for <30 minutes at least 3 days per week or vigorous exercise for <20 minutes at least 3 days per week. gBinge drinker=at least 5 drinks of any alcoholic beverage on any 1 occasion. From McGruder et al35 with permission from the American Heart Association, Inc. ©2004.

Medical history
 Hypertension79.4 (2.6)a69.7 (3.2)62.4 (1.4)
 Diabetes32.1 (2.7)a38.0 (3.8)a20.8 (1.2)
 Myocardial infarction21.2 (2.5)23.5 (3.8)24.1 (1.1)
 Total coronary heart diseaseb28.9 (2.8)29.8 (4.1)34.3 (1.3)
Risk behaviors
 Current smokerc25.6 (1.9)13.1 (3.1)20.4 (1.1)
 Overweightd66.9 (3.1)69.6 (3.8)60.1 (1.3)
 Obesitye31.2 (2.7)a28.1 (4.0)23.7 (1.2)
 Inadequate physical activityf85.3 (2.2)a80.4 (3.2)75.5 (1.4)
 Binge drinkerg6.3 (1.3)3.3 (1.3)a6.6 (0.7)

A crucial aspect of secondary stroke prevention is medication access. An NHIS covering 1997 through 2004 determined that, as of 2004, approximately 76,000 stroke survivors were unable to afford medications.36 Those who were less able to afford medication included those younger than 65 years, blacks, women, and those with high comorbidity and poorer health status. Medication affordability varied significantly, with persons without health insurance having the highest percentage of lack of affordability among stroke survivors. Conditions that were associated with lack of medication access included lack of transportation, no health insurance, no usual place of care, low income, and higher out-of-pocket medical expenses. Of interest in the current debate about comprehensive health insurance, inability to afford medications increased significantly from 8.1% to 12.7% from 1997 to 2004. The only region in which there was no increase was the Northeast, which had the lowest regional percentage of those unable to afford medications.


The variations of stroke risk, incidence, prevalence, and prognosis in minority groups compared with whites and sex and age differences in these values have been attributed to differences in both heritable and environmental factors, not the least important being lifestyle. These considerations are important in dealing with interventions to modify stroke risk.

Stroke burden in the United States can be reduced by a combination of interventions including the primary prevention and control of risk factors, catered to the cultural characteristics of individual racial/ethnic groups, public education about the signs and symptoms of stroke and of the need for emergency response and rapid transport to a stroke center, and effective rehabilitation and secondary prevention in stroke patients.35

The major efforts involve control of smoking, hypertension, diabetes, and lipid abnormalities. In this way, it is possible that the proposed US goals for health promotion and disease prevention for the year 2010 in terms of stroke deaths will lead to reduction to no more than 51 per 100,000 persons for each population group.29