African Americans and Women Have the Highest Stroke Mortality in Texas
Background and Purpose Hispanic American (HA), African American (AA), and non-Hispanic white (NHW) populations are well represented in Texas. The Texas HA population is 95% Mexican American, affording comparison with other Hispanic populations. From risk factor profiles we expected stroke mortality to be highest in AAs and HAs. We also expected stroke mortality to be considerably higher for men than for women based on previous data.
Methods We used International Classification of Diseases, 9th Revision, codes 430 to 438 to search Texas vital statistics data for the 5-year period from 1988 through 1992. Race/ethnic differences are presented in age- and sex-specific format to avoid masking the important interaction of age and sex with stroke mortality.
Results Women constituted 61% of the 40 346 stroke deaths in Texas during this period. The ratio of stroke deaths for women versus men approximates the ratio of women to men in the population. AAs had a threefold to fourfold increased stroke mortality relative to NHWs at young ages. At older ages, when stroke mortality is the highest, the stroke mortality rate in NHWs approached the stroke mortality rate of AAs. HAs had a significantly higher rate of stroke mortality at younger ages relative to NHWs but a significantly lower rate at older ages.
Conclusions Measures to prevent stroke mortality should emphasize its predilection for young AAs and women. A rigorous surveillance project is needed to determine whether stroke mortality is underestimated in the HA population.
Texas is an ideal place to study race/ethnicity issues since the state has large populations of non-Hispanic whites (NHWs), African Americans (AAs), and Hispanic Americans (HAs). The Hispanic population has received little attention in regard to stroke risk. Considerable controversy exists regarding the definition of “Hispanic” for research purposes. The Hispanic population in Texas is 95% Mexican American1 and affords an excellent comparison with data from other Hispanic populations. As we approach the year 2000, we are reminded of a key mission statement of Healthy People 20002 : to provide better public health planning for the United States' largest minority groups, AAs and HAs.
Previous epidemiological studies have suggested that AAs have a large burden of stroke and that HAs are relatively protected; however, few data are available on age-specific rates. Howard et al,3 using the National Longitudinal Mortality Study, demonstrated that the excess stroke mortality in AAs is greatest at younger ages. Of the stroke deaths, 6% occurred before age 60 years in NHWs, whereas more than 15% of stroke deaths in AAs and HAs occurred before age 60. However, few events in their study occurred in HAs. Texas vital statistics data allow population-based validation of stroke mortality rates with significantly large event rates in minority groups.
We sought to describe the pattern of race and ethnic differences in stroke mortality. Based on previous studies, we expected higher stroke mortality rates in AAs than NHWs and sought to confirm these results while examining the ethnic differences between HAs and NHWs. Since HAs have a high prevalence of diabetes and smoking,4 we expected that they would also have a higher stroke mortality than NHWs. Lower rates of hypertension and total cholesterol4 5 may complicate this relationship.
In addition to race/ethnicity issues, male sex has traditionally been considered an unmodifiable risk factor for stroke.6 We wondered whether this was still true and whether women should receive attention as a high-risk group for stroke death. A focus on race/ethnic and sex issues is important to guide public health efforts to target groups at highest risk.
Subjects and Methods
Texas law requires that all deaths be recorded with the State Registrar for Vital Statistics. The Texas Department of Health creates computer files containing information abstracted from death certificates available for research. The 1990 census7 reports that of the Texas population aged 45 years or older, 72.6% were NHW, 9.4% were AA, and 16.7% were HA. Other racial groups constituted less than 2% of the Texas population in 1990 and were excluded from this study. The Texas death certificate records the race of the deceased and whether the deceased was of Hispanic origin. Hispanic origin as designated on the death certificate was used to separate Hispanic whites and NHWs. Black Hispanics were considered AAs. Age is calculated from the date of birth reported on the death certificate.
In the period from January 1, 1988, to December 31, 1992, 627 509 resident deaths were recorded in Texas. Of these deaths, 541 838 were aged 45 years or older. Deaths coded ICD-9 430 to 438 (cerebrovascular disease)8 were considered stroke deaths. Of deaths to persons aged 45 years or older, 40 346 were attributed to stroke. These stroke deaths were 76% NHW, 12% AA, and 12% HA.
We calculated 5-year stroke mortality rates using the total number of deaths during a 5-year period divided by the total population at risk for the period (ie, the sum of the estimated population aged 45 years and older across the 5 years) multiplied by 100 000. CIs were calculated for the rates for each age-race/ethnicity-sex category. Epigram,9 a population and mortality analysis program, was used to produce the rates, CIs, and counts used in the tables and figures. NHWs are used as the referent group for rate ratio. The sex-race/ethnicity-age rate ratio was calculated by dividing the rate for each race/ethnicity group by the NHW mortality rate. The ratio of male stroke to female stroke mortality was also calculated. CDC Wonder10 was used to delineate stroke mortality rates for AAs and combined NHWs and HAs aged 75 years and older. All rates shown are per 100 000 per year and are expressed in age-specific fashion. Age adjustment was not used in this report to avoid masking the important interaction of age and stroke mortality.
The overall stroke mortality rate in Texas from 1988 to 1992 in people aged 45 years and older was 175 per 100 000 (95% CI, 173 to 177). Table 1⇓ shows that the stroke mortality rate increased in geometric proportion across advancing age groups in both men and women of all race/ethnic groups. In the group aged 45 to 59 years, stroke mortality was four times greater in AA men than in NHW men (Table 2⇓) and three times greater in AA women than in NHW women. HAs also had higher rates than NHWs in this age category. Both AA and HA men had significantly higher stroke mortality rates than women. There were no sex differences among NHWs in the group aged 45 to 59 years.
In the group aged 60 to 74 years, AA men and women had an approximately 2.5-fold greater stroke mortality than either NHWs or HAs who had similar rates (Table⇑s 1 and 2). The relative gap among the groups was considerably less than that seen in the younger group (Fig 1⇓). In this age group, stroke mortality rates were significantly higher in men than in women in all race/ethnic groups.
In the group aged 75 years and older, AAs still had a significantly higher rate than NHWs; however, the rate ratio was lower than in the younger groups: 1.26 for men and 1.10 for women (Table⇑s 1 and 2). The rate for NHWs approached that for AAs (Fig 1⇑). Stroke mortality was higher in NHW women than AA women older than 85 years. AA men older than 85 years still had a higher stroke mortality rate than NHW men, but the gap was smaller. US vital statistics data confirm the mortality rate crossover above age 85 years for both men and women. In Texas, AA men older than 75 years had the highest stroke mortality rate of all the sex-race/ethnic groups. HAs had a significantly lower risk of stroke mortality at this age than NHWs; rate ratios were 0.77 among men and 0.63 among women.
Women accounted for 61% of the 40 346 stroke deaths observed in the population aged 45 years and older during this period. This is a greater percentage than expected based on their representation in the population aged 45 years and older, which is 55% (Fig 2⇓). The apparent excess is due to the greater risk of stroke at older ages and the greater representation of women in the older age groups of the population (Fig 2⇓). Age-specific data confirm that the interaction of age and sex accounts for the higher overall stroke mortality rate in women than men older than 75 years by showing that men still have a higher stroke mortality rate than women in the group aged 75 to 84 years.
In this report we examined the burden of stroke mortality in HAs, the fastest growing minority population in the United States. Generally, we found results consistent with those of Howard et al3 in an analysis of data from the National Longitudinal Mortality Study. Across successively older ages, stroke mortality rates increased more rapidly in NHWs than in HAs. In contrast to the results of Howard et al, we found significantly higher stroke mortality rates in young HA men (RR, 1.73) and women (RR, 1.30) than in NHWs. This difference may be the result of more stable rates attributable to the larger numbers of events that we observed in young HAs.
Why were stroke mortality rates among HAs significantly higher than those among NHWs for ages 45 to 59 years but significantly lower for those aged 75 years and older? HAs have a higher prevalence of diabetes mellitus but a similar or slightly lower prevalence of hypertension4 5 than NHWs. These risk factors for stroke may operate differently for the various subtypes of stroke, which may show different ethnic patterns across age groups.
The Corpus Christi Heart Project is a population-based epidemiological surveillance project that has challenged the notion that HAs are relatively protected from death due to myocardial infarction as suggested by vital statistics data.11 These authors have found higher hospitalization and case-fatality rates for myocardial infarction among Mexican Americans than among NHWs.12 13 We believe that a rigorous surveillance project is needed to determine whether stroke mortality is underestimated in the Hispanic population.
There are several different “Hispanic” populations in the United States. The Hispanic population in Texas is 95% Mexican American1 and therefore provides a good comparison with other well-defined Hispanic populations. The Northern Manhattan Study14 15 documents results from a predominantly Dominican Hispanic population. Their community-based mortality data are presented as age-adjusted rates.15 We believe that age-adjusted rates hide the intricate relationship of age and stroke mortality as outlined in Fig 1⇑. These authors do report age-specific stroke incidence rates, which confirm our findings that HAs had a higher rate of stroke at younger ages than NHWs, and this relationship changed across successively older age groups.15
There are limitations of using vital statistics data to study stroke mortality in HAs. The denominator data (population) for Hispanic origin are from self-reports on the census questionnaire. The numerator data (deaths) are from reports of the person who completes the death certificate. Poe et al16 showed that failure to classify a decedent as Hispanic on death certificates may contribute to a 19.6% underestimation of death in Hispanics. Another possible bias from using death certificate data for HAs is that some may choose to return to Mexico to die after a severe stroke.
It has been recognized for some time that AAs have a high stroke mortality rate; age-specific data allow us to target the proper age range for public health intervention. Howard et al3 demonstrated an RR of 3.67 of stroke mortality for AA men aged 45 to 59 years compared with NHW men. Similar results were found for women (RR, 3.00). Texas data are remarkably alike, with an RR of 4.06 for men and 3.22 for women (Table 2⇑). Although AAs have significantly higher stroke mortality at all ages relative to NHWs (Fig 1⇑), the effect diminishes over successively older age groups. For both men and women the rate ratio is 3.60 at 45 to 59 years, 2.42 at 60 to 74 years, and 1.15 for those 75 years and older.
A possible explanation for why NHWs have a relatively higher stroke mortality rate than AAs in the group aged 75 years and older is the issue of competing mortality. If AAs with atherosclerotic disease died young from coronary artery disease, this may reflect why the rate ratio of stroke mortality in AA males relative to NHWs decreases from 4.0 to 1.3 in the groups aged 45 to 59 years and 75 years and older, respectively. However, the ischemic heart disease mortality rate per 100 000 people in the group aged 45 to 59 years for the 1988 to 1992 time period in Texas9 was 173 for NHWs and 245 for AAs; in the group aged 60 to 74 years, this rate was 688 for NHWs and 705 for AAs. These relative differences are small compared with the excess stroke mortality seen in young AAs. It seems unlikely that competing mortality from myocardial disease would account for the relative decline of stroke mortality of AAs compared with NHWs in the group aged 75 years and older.
Using Texas vital statistics data, we have also shown that women have higher overall stroke mortality rates than men. This finding is due to the high stroke mortality rate in NHW women aged 75 years and older. This group experienced more than twice the number of stroke deaths as men in this age group. However, as shown in Fig 2⇑, the percentage of deaths closely follows the prevalence of women in the population. Age-specific data, including data for the group aged 75 to 84 years, also illustrate that the higher stroke mortality rate in women aged 75 years and older is explained by their increased prevalence in older ages compared with men. Nevertheless, women still account for more than 60% of all stroke deaths in Texas. Zhang et al,6 using World Health Organization data, described age-specific changes over time in the sex ratio of men to women. They reported that men had a higher stroke mortality relative to women in the United States as of 1988. In none of the 27 countries studied did women have a higher stroke mortality rate than men. Zhang et al did show that the sex ratio (ratio of men to women) for stroke mortality rate was lowest at older ages, which is consistent with our findings.
Stroke mortality data from Texas suggest that women and men have similar stroke mortality rates and that women have the largest public health impact from stroke mortality, contributing greater than 60% of all stroke deaths. While incidence rates may still be higher in men,17 perhaps women have more severe stroke or receive less intense medical care. Since stroke is thought of as a “male disease,” physicians may have a low index of suspicion to recognize and treat cerebrovascular disease in women. This could lead to higher mortality from stroke. Indeed, an increased case-fatality rate for women relative to men has been found in myocardial infarction as well.12 We know that specific stroke therapies proven effective for men may not be beneficial in women. The Asymptomatic Carotid Endarterectomy Trial demonstrated a benefit of surgery in men with carotid artery stenosis but not in women.18 We suggest that public health efforts, including physician education, focus on women as a high-risk group for stroke mortality.
Selected Abbreviations and Acronyms
|RR||=||relative risk or rate ratio|
This study was supported in part by an American Heart Association Clinician Scientist Award (96004040) (Dr Morgenstern).
- Received May 15, 1996.
- Revision received September 3, 1996.
- Accepted September 5, 1996.
- Copyright © 1997 by American Heart Association
Plan and Operation of the Hispanic Health and Nutrition Examination Survey 1982-84. Hyattsville, Md: US Public Health Service; 1985. US Dept of Health and Human Services publication PHS 85-1321.
Healthy People 2000. Hyattsville, Md: US Public Health Service; 1990. US Dept of Health and Human Services publication PHS 91-50213.
Howard G, Anderson R, Sorlie P, Andrews V, Backlund E, Burke GL. Ethnic differences in stroke mortality between non-Hispanic whites, Hispanic whites, and blacks: the National Longitudinal Mortality Study. Stroke. 1994;25:2120-2125.
Gillum RF. Epidemiology of stroke in Hispanic Americans. Stroke. 1995;26:1707-1712.
Burt VL, Whelton P, Rocella EJ, Brown C, Cutler JA, Higgins M, Horan MJ, Labarthe D. Prevalence of hypertension in the US adult population: results from the Third National Health and Nutrition Examination Survey, 1988-1991. Hypertension. 1995;25:305-313.
Zhang XH, Sasaki S, Kesteloot H. Changes in the sex ratio of stroke mortality in the period of 1955 through 1990. Stroke. 1995;26:1774-1780.
US Bureau of the Census. 1990 Census of Population and Housing. Washington, DC: Government Printing Office; 1991. US Dept of Commerce publication 1990 CPH-1-45.
International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM). Los Angeles, Calif: Practice Management Information Corp; 1995.
Goldman D. Epigram: a computer program for analyzing Texas population and mortality data. Austin, Tex: Texas Dept of Health, Bureau of Chronic Disease Prevention and Disease Control; 1992.
CDC Wonder. http://wonder.cdc.gov/. (1 aug. 1996).
Goff DC Jr, Varas C, Ramsey DJ, Wear ML, Labarthe DR, Nichaman MZ. Mortality after hospitalization for myocardial infarction among Mexican Americans and non-Hispanic whites: the Corpus Christi Heart Project. Ethn Dis. 1993;3:53-63.
Goff DC Jr, Ramsey DJ, Labarthe DR, Nichaman MZ. Greater case-fatality after myocardial infarction among Mexican Americans and women than among non-Hispanic whites and men: the Corpus Christi Heart Study. Am J Epidemiol. 1994;139:474-483.
Sacco RL, Shi T, Zamanillo MC, Kargman DE. Predictors of mortality and recurrence after hospitalized cerebral infarction in an urban community: the Northern Manhattan Stroke Study. Neurology. 1994;44:626-634.
Sacco RL, Hauser WA, Mohr JP. Hospitalized stroke in blacks and Hispanics in Northern Manhattan. Stroke. 1991;22:1491-1496.
Poe GS, Powell-Griner E, McLaughlin JK, Placek PJ, Thompson GB, Robinson K. Comparability of the death certificate and the 1986 National Mortality Followback Survey. Vital Health Stat 2 (118); 1993.