Abstract
Background and purpose:
It is unclear whether disparities in mortality among stroke survivors exist long-term. Therefore, the purpose of the current study is to describe rates of longer term mortality among stroke survivors (i.e. beyond 30 days) and to determine whether socioeconomic disparities exist.
Methods:
This analysis included 1,329 black and white participants, ages 45 and older, enrolled between 2003–2007 in the REasons for Geographic And Racial Differences in Stroke (REGARDS) study who suffered a first stroke and survived at least 30 days following the event. Long-term mortality among stroke survivors was defined in person-years as time from 30 days after a first stroke to date of death or censoring. Mortality rate ratios (MRRs) were used to compare rates of post-stroke mortality by demographic and socioeconomic characteristics.
Results:
Among adults who survived ≥ 30 days post stroke, the age-adjusted rate of mortality was 82.3 per 1,000 person-years (95% CI: 75.4, 89.2). Long-term mortality among stroke survivors was higher in older individuals (MRR for 75+ vs. < 65: 3.2, 95% CI: 2.6, 4.1) and among men than women (MRR: 1.3, 95% CI: 1.1, 1.6). It was also higher among those with less educational attainment (MRR for < high school vs. college graduate: 1.5, 95% CI: 1.1, 1.9), lower income (MRR for < $20k vs. > 50K: 1.4, 95% CI: 1.1, 1.9), and lower neighborhood SES (MRR for low vs. high neighborhood SES: 1.4, 95% CI: 1.1, 1.7). There were no differences in age-adjusted rates of long-term post-stroke mortality by race, rurality, or US region.
Conclusions:
Rates of long-term mortality among stroke survivors were higher among individuals with lower socioeconomic status (SES) and among those residing in neighborhoods of lower SES. These results emphasize the need for improvements in long term care post-stroke, especially among individuals of lower SES.
Keywords: Stroke, mortality, disparities
Stroke is the fifth leading cause of death, accountable for more than 140,000 deaths annually in the US.1 Though rates of stroke mortality were greatly reduced in the 20th century,2 disparities in stroke incidence and mortality by demographic factors and socioeconomic status (SES) still persist.3,4 For example, black individuals having higher rates of stroke incidence and mortality due to stroke than white individuals,1 especially at younger ages.5 Markers of low SES, such as lower educational attainment or income have also been linked to higher rates of stroke4,6 and stroke mortality.7 In addition to such individual level SES factors, there are also marked differences in the geographic distribution of stroke incidence in the US. For example, higher rates of stroke have been well documented in the entire southeastern portion of the US (excluding Florida), also referred to as the “stroke belt.”8 Further, rates of stroke have also been noted as higher in rural vs. urban neighborhoods,9 and in neighborhoods of low vs. high SES.10
Disparities in stroke mortality, such as the higher mortality rates among black vs. white individuals or geographic disparities, have been primarily attributed to higher rates of stroke incidence, as suggested by a number of studies reporting similar case-fatality rates across these groups.9,11–15 For example, short-term (mostly 30-day) mortality rates among stroke cases (case-fatality within 30 days) have been shown to be similar among black and white individuals (particularly at older ages),12,13 among those residing in rural vs. urban neighborhoods,9 as well as among those residing in neighborhoods of low vs. high SES.11 However, studies examining long-term trends in mortality among stroke survivors (i.e. case-fatality 30 days after first stroke) reveal different patterns.11,16–19 In fact, rates of long-term mortality among stroke survivors may actually be lower in black vs. white individuals,17–19 though it is unclear whether this trend differs by age group. Similarly, some studies have shown that rates of long-term mortality among stroke survivors are higher among individuals residing in low vs. high SES neighborhoods, in contrast with the null findings of short-term stroke mortality among stroke cases.11,16
Given these divergent findings and the scarce literature, we sought to describe rates of long-term (i.e. post 30-day) mortality among stroke survivors and clarify whether disparities exist across important demographic and socioeconomic factors such as sex, race (across age groups), education, income, and geographic/neighborhood-level factors. To do so, we used data from the REasons for Geographic and Racial Differences in Stroke (REGARDS), a longitudinal study of stroke and its risk factors.
Materials and Methods
In order to abide by its obligations with NIH/NINDS and the Institutional Review Board of the University of Alabama at Birmingham, REGARDS facilitates data sharing through formal data use agreements. Any investigator is welcome to access the REGARDS data through this process. Requests for data access may be sent to regardsadmin@uab.edu.
Study Population
REGARDS is an on-going cohort study designed to better understand the contributors to the higher stroke mortality in black individuals and residents of the stroke belt. The REGARDS study sample was selected from a commercially available nationwide list of US households which was purchased. Inclusion criteria were: having a name, phone number, and address, self-identifying as black or white, and being at least 45 years or age. Exclusion criteria included: not speaking English, cognitive impairment, living in or being on a wait list for a nursing home, or having a self-reported medical conditions that could prevent future participation. The sample includes 30,239, recruited from 48 contiguous US States between 2003 and 2007, and oversampled for blacks and individuals residing in the stroke belt or stroke buckle (the Southeast US). Full study details have been reported elsewhere.20 In brief, participants were recruited through mail and telephone contact. Those who agreed to participate first completed a computer assisted telephone interview (CATI), which included an assessment of demographic, socioeconomic information, and a cardiovascular risk profile assessment. This interview was followed by an in-home physical assessment conducted by a trained health professional (about 2 weeks later), which included clinical measurements of blood pressure, ECG, anthropometrics, medication inventory. Finally, additional stroke risk factors including diet and residential history were assessed in a self-administered questionnaire. During follow-up CATIs, conducted at six-month intervals, participants and/or their proxies were asked about stroke symptoms, hospitalizations, and their general health status. Written informed consent was obtained from all study participants and the study was approved by institutional review boards from each participating site.
Stroke
During follow-up CATIs, participants and/or their proxies were asked about reasons for any medical encounters. Medical records were obtained for participants or proxies who, during follow-up interviews, reported: suspected stroke, transient ischemic attack (TIA), death, unknown reason for hospital visit, brain aneurysm or hemorrhage, or stroke symptoms (e.g. sudden weakness or vision loss, numbness, difficulty speaking, or headache). Medical records were first reviewed by a stroke nurse to rule out obvious non-stroke medical events. Next, records representing possible stroke events were reviewed by an adjudication panel of at least two physicians with expertise in stroke. Any disagreements were resolved by a full adjudication review.21 Stroke events were defined based on meeting any of the following criteria: the World Health Organization definition for stroke,22 clinical stroke (symptoms lasting less than 24 hours with neuroimaging consistent with stroke), and probable stroke (adjudicated cases of stroke with incomplete data for classification based on the definition of World Health Organization stroke or clinical stroke). The current analysis includes strokes that occurred through September 30th 2016.
Mortality
Our outcome of interest was long-term (i.e. post 30-day) mortality after first stroke. Preliminary date of death was first determined by proxies during follow-up interviews and verified using medical records, death certificates, and administrative databases such as the National Death Index. REGARDS investigators obtained copies of official death certificates for most events from next of kin or the state health department.23 Analyses included deaths that occurred up until March 31st 2017.
Other variables
During the baseline CATI at (between 2003 and 2007), participants self-reported their age, sex, race, educational attainment (less than high school (HS), HS graduate, some college, college graduate or more), current marital status (married, widowed, divorced, or single/other), household income (< $20K, $20K - $35K, $35K - $50K, or > $50K), and health insurance status (insured vs. uninsured). Based on residence by home address at baseline, participants were geocoded and classified as living in the following mutually exclusive localities: stroke buckle (coastal plains of North Carolina, South Carolina, and Georgia), stroke belt (remainder of North Carolina, South Carolina, and Georgia, in addition to Tennessee, Alabama, Mississippi, Louisiana, and Arkansas), or neither (all other contiguous US states).24 Using these geocoded data, participants were also linked to their US residential 2000 census block group and were further classified as living in an urban, rural, or mixed environment, using previously published methods.9 Participant linked US residential 2000 census block group data were also used to construct a neighborhood SES summary score from six census block group indicators relating to: 1) median household income, 2) median value of owner occupied housing units, 3) proportion of households receiving interest, dividend, or rental income, 4) proportion of adults with high school diplomas, 5) proportion of adults with a college degree, and 6) proportion of adults employed in professional occupations.25 The neighborhood SES summary score was then further categorized into tertiles representing: high (most advantageous), middle, and low (least advantageous) neighborhood SES condition.
Analytical sample
Given our interest in characterizing long-term mortality among stroke survivors (i.e. case-fatality rates post 30 days of a first stroke event), we first considered individuals who reported to be stroke-free at baseline who had at least one stroke event during follow-up (n=1,621). We excluded individuals who died within 30 days of first stroke (n=292). The final analytic sample included a total of 1,329 individuals who survived at least 30 days after their first stroke.
Statistical analysis
Participants in this study contributed observed time at risk (in years) beginning 30 days following their first stroke and ending at date of death or censoring. Participants who were still alive by the end of the study period (September 30th 2016—date of last follow-up for stroke event) were censored on October 30th 2016—exactly 30 days later.
Mean follow-up time, demographic, socioeconomic, and geographic factors were described for the study sample, overall and according to long-term mortality status (died vs. not). We further calculated age-standardized mortality density rates among stroke survivors, per 1,000 person-years (PY), overall and according to demographic, socioeconomic and geographic factors of interest, as well as mortality rate ratios (MRRs) and mortality rate differences (MRDs) across these factors. We plotted survival curves by age group, sex, race, education, household income, and neighborhood SES; except across age groups, all survival curves were age-adjusted.
Finally, for comparability with prior literature on short-term mortality among stroke cases across race and age groups,12 Cox proportional hazard models were used to estimate the risk of long-term mortality among stroke survivors in blacks compared with whites, across age groups (despite the lack of a significant race by age interaction, p=0.17). All analyses were conducted using SAS version 9.4.
Results
In the sample comprising 1,329 US adults who survived for at least 30 days after first stroke, mean follow-up time was 5.0 years (Table 1). Mean baseline age in this sample was 69 years. Approximately 50% of the sample was female, 56% was white, 29% had a college education or greater, 55% was married, 20% had a family income > $50,000, and 94% had health insurance. Geographically, 35% lived in the stroke belt, 20% lived in the stroke buckle, and 73% lived in an urban environment compared to only 8% who lived in a rural environment. Compared to stroke survivors who died during the study period (i.e. died at least 30 days after first stroke), those who survived had a longer mean follow-up time, were younger, more likely to be women, have higher levels of education, and higher household income (p values all <0.01).
Table 1.
Baseline characteristics (2003–2007) of study participants, overall and according to long-term mortality among stroke survivors, the REGARDS study.
| Overall (n = 1,329) | Survival status | |||
|---|---|---|---|---|
| Survived (n=780) | Died (n =549) | P-value | ||
| Follow up time, years | 5.0 (3.3) | 6.2 (3.2) | 3.4 (2.8) | <0.01 |
| Socio-demographic factors | ||||
| Age, years | 69.2 (8.5) | 66.9 (7.8) | 72.4 (8.3) | <0.01 |
| Age, n (%) | ||||
| < 65 | 404 (30.4) | 303 (38.9) | 101 (18.4) | <0.01 |
| 65 – 74 | 549 (41.3) | 350 (44.9) | 199 (36.3) | |
| ≥ 75 | 376 (28.3) | 127 (16.3) | 249 (45.4) | |
| Female, n (%) | 655 (49.3) | 412 (52.8) | 243 (44.3) | <0.01 |
| White, n (%) | 737 (55.5) | 435 (55.8) | 302 (55.0) | 0.78 |
| Education, n (%) | ||||
| College graduate and above | 386 (29.0) | 252 (32.3) | 134 (24.4) | <0.01 |
| Some college | 342 (25.7) | 206 (26.4) | 136 (24.8) | |
| High school graduate | 386 (29.0) | 219 (28.1) | 167 (30.4) | |
| Less than high school | 214 (16.1) | 103 (13.2) | 111 (20.2) | |
| Marital status, n (%) | ||||
| Single or other | 82 (6.2) | 51 (6.5) | 31 (5.7) | <0.01 |
| Divorced | 182 (13.7) | 118 (15.1) | 64 (11.7) | |
| Widowed | 336 (25.3) | 169 (21.7) | 167 (30.4) | |
| Married | 729 (54.8) | 442 (56.7) | 287 (52.3) | |
| Household income, n (%) | ||||
| ≥ $50k | 262 (19.7) | 184 (23.6) | 78 (14.2) | <0.01 |
| $35k - $50k | 220 (16.6) | 133 (17.1) | 87 (15.9) | |
| $20k - $35k | 376 (28.3) | 218 (28.0) | 158 (28.8) | |
| < $20k | 296 (22.3) | 149 (19.1) | 147 (26.8) | |
| Insurance, n (%) | 1,254 (94.4) | 737 (94.5) | 517 (94.2) | 0.61 |
| Neighborhood SES, n (%) | ||||
| High SES | 403 (30.3) | 253 (32.4) | 150 (27.3) | 0.17 |
| Medium SES | 403 (30.3) | 237 (30.4) | 166 (30.2) | |
| Low SES | 402 (30.3) | 223 (28.6) | 179 (32.6) | |
| Geographic factors | ||||
| Region, n (%) | ||||
| Non-belt | 601 (45.2) | 349 (44.7) | 252 (45.9) | 0.91 |
| Belt | 463 (34.8) | 275 (35.3) | 188 (34.2) | |
| Buckle | 265 (19.9) | 156 (20.0) | 109 (19.9) | |
| Urban-rural status, n (%) | ||||
| Urban (>=75% urban) | 965 (72.6) | 561 (71.9) | 404 (73.6) | 0.73 |
| Mixed (25–75% urban) | 143 (10.8) | 87 (11.2) | 56 (10.2) | |
| Rural (<=25% urban) | 110 (8.3) | 69 (8.9) | 41 (7.5) | |
Abbreviations: REGARDS: Reasons for Geographic and Racial Differences in Stroke; SES: socioeconomic status.
30-days post stroke
The overall age-adjusted rate of long-term mortality among stroke survivors was 82.3 per 1,000 PYs (95% CI: 75.4, 89.2, Table 2). Long-term mortality rates among stroke survivors increased with age: from 47.0/1,000 PY for ages <65, to 69.2/1,000 PY for ages 65–74 and 151.7/1,000 PY for ages 75+. In other words, compared to stroke survivors ages <65 years, those ages 65–74 years and 75+ years had respectively a 50% (MRR=1.5, 95% CI: 1.2, 1.9) and a 220% (MRR=3.2, 95% CI: 2.6, 4.1) higher risk of long term mortality, and had respectively 22.2 (MRD=22.2, 95% CI: 8.9, 35.5) and 104.7 (MRD=104.7, 95%CI: 83.8, 127.7) additional long term mortality cases per 1,000 PY. Age-adjusted long-term mortality rate ratios and long-term mortality rate differences among stroke survivors were also higher among men vs. women (MRR=1.3, 95% CI: 1.1, 1.6 and MRD=23.6, 95% CI: 9.8, 37.5), among those with less than a HS degree vs. a college education or more (MRR=1.5, 95% CI: 1.1, 1.9 and MRD=32.0, 95% CI: 9.9, 54.2), among those with a household income < $20,000 per year vs. ≥ $50,000 per year (MRR=1.4, 95% CI: 1.1, 1.9 and MRD=31.2, 95% CI: 8.5, 54.0), and among those living in a low vs. high SES neighborhood (MRR=1.4, 95% CI: 1.1, 1.7 and MRD= 24.3, 95% CI: 6.9, 41.7). Finally, age-adjusted rates of long-term mortality among stroke survivors did not differ by race (black vs. white MRR=1.1, 95% CI: 1.0, 1.3 and MRD=10.0, 95% CI: −4.1, 24.1), marital status (married vs. single MRR= 1.2, 95% CI: 0.8, 1.8 and MRD=14.5, 95% CI: −13.4, 42.3), geographic region (stroke buckle vs. non-belt or buckle MRR=1.1, 95% CI: 0.8, 1.3 and MRD=4.7, 95% CI: −13.8, 23.2) or urban-rural status (rural vs. mixed MRR=0.9, 95% CI: 0.7, 1.3 and MRD=−6.5, 95% CI: −31.2, 18.2).
Table 2.
Age-standardized long-term* mortality incidence rates, incidence rate ratios, and incidence rate differences, among stroke survivors, the REGARDS study.
| N | # Events | Person-years | MR (95% CI) | MRR (95% CI) | MRD (95% CI) | |
|---|---|---|---|---|---|---|
| Overall | 1,329 | 549 | 6,669 | 82.3 (75.4, 89.2) | ||
| Socio-demographic factors | ||||||
| Age† | ||||||
| < 65 | 404 | 101 | 2,151 | 47.0 (37.8, 56.1) | ref | ref |
| 65 – 74 | 549 | 199 | 2,877 | 69.2 (59.6, 78.8) | 1.5 (1.2, 1.9) | 22.2 (8.9, 35.5) |
| ≥ 75 | 376 | 249 | 1,642 | 151.7 (132.8, 170.5) | 3.2 (2.6, 4.1) | 104.7 (83.8, 125.7) |
| Gender | ||||||
| Female | 655 | 243 | 3,397 | 70.8 (61.9, 79.7) | ref | ref |
| Male | 674 | 306 | 3,273 | 94.5 (83.9, 105.1) | 1.3 (1.1, 1.6) | 23.6 (9.8, 37.5) |
| Race | ||||||
| White | 737 | 302 | 3,622 | 77.0 (68.2, 85.7) | ref | ref |
| Black | 592 | 247 | 3,048 | 87.0 (75.9, 98.0) | 1.1 (1.0, 1.3) | 10.0 (−4.1, 24.1) |
| Education | ||||||
| College graduate + | 386 | 134 | 1,928 | 68.9 (57.2, 80.5) | ref | ref |
| Some college | 342 | 136 | 1,728 | 79.3 (66.0, 92.6) | 1.2 (0.9, 1.5) | 10.4 (−7.3, 28.1) |
| High school graduate | 386 | 167 | 1,931 | 88.1 (74.7, 101.5) | 1.3 (1.0, 1.6) | 19.2 (1.5, 37.0) |
| Less than high school | 214 | 111 | 1,080 | 100.9 (82.1, 119.7) | 1.5 (1.1, 1.9) | 32.0 (9.9, 54.2) |
| Marital status | ||||||
| Single or other | 82 | 31 | 457.2 | 69.2 (43.0, 95.3) | ref | ref |
| Divorced | 182 | 64 | 933.2 | 75.7 (53.6, 97.8) | 1.1 (0.7, 1.8) | 6.5 (−27.7, 40.7) |
| Widowed | 336 | 167 | 1,657 | 75.4 (62.2, 88.5) | 1.1 (0.7, 1.7) | 6.2 (−23.1, 35.4) |
| Married | 729 | 287 | 3,622 | 83.6 (73.9, 93.4) | 1.2 (0.8, 1.8) | 14.5 (−13.4, 42.3) |
| Household income | ||||||
| ≥ $50k | 262 | 78 | 1,339 | 69.5 (53.6, 85.3) | ref | ref |
| $35k - $50k | 220 | 87 | 1,101 | 79.5 (62.7, 96.2) | 1.1 (0.8, 1.6) | 10.0 (−13.1, 33.1) |
| $20k - $35k | 376 | 158 | 1,901 | 79.7 (67.2, 92.2) | 1.1 (0.9, 1.5) | 10.3 (−9.9, 30.5) |
| < $20k | 296 | 147 | 1,467 | 100.7 (84.4, 117.0) | 1.4 (1.1, 1.9) | 31.2 (8.5, 54.0) |
| Neighborhood SES | ||||||
| High SES | 403 | 150 | 2,071 | 68.1 (57.1, 79.1) | ref | ref |
| Medium SES | 403 | 166 | 2,010 | 83.4 (70.7, 96.1) | 1.2 (1.0, 1.5) | 15.3 (−1.5, 32.1) |
| Low SES | 402 | 179 | 1,989 | 92.4 (78.8, 106.0) | 1.4 (1.1, 1.7) | 24.3 (6.9, 41.7) |
| Geographic factors | ||||||
| Region | ||||||
| Non-belt | 601 | 252 | 3,114 | 78.8 (69.0, 88.5) | ref | ref |
| Belt | 463 | 188 | 2,242 | 87.9 (75.3, 100.5) | 1.1 (0.9, 1.3) | 9.1 (−6.9, 25.1) |
| Buckle | 265 | 109 | 1,313 | 83.5 (67.8, 99.2) | 1.1 (0.8, 1.3) | 4.7 (−13.8, 23.2) |
| Urban-rural status | ||||||
| Urban (>=75% urban) | 965 | 404 | 4,831 | 82.5 (74.4, 90.5) | ref | ref |
| Mixed (25–75% urban) | 143 | 56 | 708.2 | 85.5 (62.7, 108.3) | 1.0 (0.8, 1.4) | 3.0 (−21.2, 27.2) |
| Rural (<=25% urban) | 110 | 41 | 559.9 | 76.0 (52.7, 99.3) | 0.9 (0.7, 1.3) | −6.5 (−31.2, 18.2) |
Abbreviations: MR: mortality rate; MRD: mortality rate difference; MRR: mortality rate ratio; REGARDS: Reasons for Geographic and Racial Differences in Stroke; SES: socioeconomic status
30-days post stroke.
Not age-standardized
Figures 1 and 2 depict long-term survival curves among stroke survivors by categories of age, sex, race, education, household income, and neighborhood SES (all being age-adjusted except when depicted across age categories). As shown in the graphs, clear graded associations exist whereby younger age, more education, higher income, and higher neighborhood SES are associated with prolonged survival among survivors of stroke.
Figure 1.
Age adjusted survival curves according to key demographic characteristics, the REGARDS study.
Figure 2.
Age adjusted survival curves according to key socio-economic characteristics, the REGARDS study.
As shown in Figure 3, although there was no overall difference in long-term mortality among black vs. white individuals, there was a suggestion that differences declined with age (interaction p-value = 0.17): among those ages <65 years (HR: 1.47, 95% CI: 0.98, 2.20), ages 65 – 74 years (HR: 1.23, 95% CI: 0.93, 1.63), or among those ages 75+ years (HR: 0.95, 95% CI: 0.73, 1.24).
Figure 3.
Risk of long-term* mortality among stroke survivors in black vs. white US adults according to age category, the REGARDS study. * 30-days post stroke Age by race interaction p-value = 0.17
Discussion
In a national study, we found marked disparities in long-term mortality among stroke survivors, with greater mortality among older adults, men, individuals with low income, low educational attainment, and those residing in neighborhoods of low SES compared to their counterparts. Given that our analysis was restricted to incident stroke cases, these findings of greater long-term mortality among individuals of lower SES could not be explained by a higher stroke incidence. Additionally, we did not find differences in long-term mortality among stroke survivors by race (overall or across age groups), geographic region, or by rural-urban residence status. These results provide further support for what has previously been described,9,12 that stroke incidence is the main driver of black vs. white and rural vs. urban disparities in short-term mortality post stroke, as we found no evidence of racial or geographical disparities in long-term mortality among stroke survivors.
Not surprisingly, age was a major determinant of long-term mortality among stroke survivors, with a three-fold risk of mortality among subjects aged 75+ years compared with those <65 years. We found that long-term mortality among stroke survivors was higher among men vs. women, a finding consistent with some,26 but not all studies—some of which show no differences by sex.27 As women have longer life expectancy than men, our results add to prior literature showing a higher prevalence of CVD among elderly women.28
It has been noted that stroke incidence and stroke mortality rates are up to three fold higher among Blacks vs. Whites below the age of 75, with a smaller disparity at older ages.12 Prior findings from the REGARDS study12 and others,3,13 attribute these higher mortality rates to higher incidence rates among Blacks vs. Whites. However, most of these findings focused on short-term mortality among stroke cases (i.e. within 30 days of stroke).12,13 In the current study, we extended these prior findings to longer term mortality among survivors of stroke (i.e. beyond 30 days). Further, we did not confirm a protective effect of lower long-term mortality in black vs. white stroke survivors, as reported in prior literature.17–19 In fact, among subjects aged <65 years, there was greater long-term mortality among black vs. white stroke survivors, although the association was not statistically significant and was plausibly a chance finding. Altogether, these results support conclusions from prior studies12 that reducing incident stroke, through reduction of stroke risk factors, especially among Blacks, may be the key to absolute reductions in stroke mortality.
Socioeconomic indicators such as income and education have also been shown to be associated with stroke incidence4,6 and stroke mortality.7 For example, among participants of the Established Populations for the Epidemiologic Studies of the Elderly,6 and participants from the Greater Cincinnati/Northern Kentucky Stroke Study,4 low education and low income were associated with a higher incidence of stroke (fatal or nonfatal). In line with these findings, we found strong graded disparities with higher long-term mortality among stroke survivors with lower education and lower income. Additionally, our findings were consistent11,16 with literature showing that residing in a neighborhood of low vs. high SES is associated with greater long-term mortality among stroke survivors. These data suggest that even among individuals who have had a stroke—i.e. beyond risk of incident stroke—disadvantageous SES is associated with increased risk of mortality. Therefore, among stroke survivors, socioeconomic position and neighborhood SES may be important indicators of mortality risk.
Despite reports of higher stroke incidence9 and stroke mortality29 among residents of rural vs. urban settings, we did not find such differences for long-term mortality among survivors of stroke. This extends previously reported findings from REGARDS showing no difference in short-term mortality among stroke cases for those living in rural vs. urban settings.9 It has also been noted that rates of incident stroke and stroke mortality follow geographical patterns with higher stroke incidence and mortality in the stroke belt or buckle.8,21 For example, results from the REGARDS study showed residents of counties with the highest rates of overall stroke mortality were more likely to be in the stroke belt or stroke buckle than counties with lower stroke mortality rates.14 Further, short-term (within 30 days) mortality among stroke cases was also higher in these same counties even after adjustment for age, sex, and race, suggesting that both higher stroke incidence and higher case-fatality were contributors to geographic disparities in stroke mortality.14 However, in our study, we showed no regional differences in long-term mortality among stroke survivors, suggesting that at least for long-term survival after stroke, the greater rates of stroke mortality among stroke survivors in the stroke belt and buckle might be driven by high rates of stroke incidence (rather than case-mortality). Taken together, these findings emphasize the importance of stroke prevention, which should be reinforced in geographic regions with higher rates of stroke incidence.
The current study has a few limitations that are worth noting. First, our measure of neighborhood SES did not account for regional or statewide differences across the contiguous US. Labor markets and cost of living could vary dramatically by region, potentially limiting the precision of this measure. In addition, we did not account for potential changes in SES or in geographic location over the study period. Updating this information might have strengthened associations in some domains, although both SES and geographic region of residence tend to be strongly correlated across the lifecourse. Further, we did not have a measure of stroke severity—which could impact stroke prognosis and ultimately stroke mortality.30 However, the effect of stroke severity on mortality is more pertinent soon after occurrence of stroke and thus was somewhat mitigated by our definition of long-term (i.e. post 30 days) mortality among stroke survivors. Finally, because we restrict to individuals who had a stroke and survived at least 30 days, these associations may be vulnerable to survival bias. Many of the risk factors we examine likely influence both stroke incidence and short-term survival, which could induce spurious (non-causal) associations among the sample we analyzed. For example, given that blacks have higher rates of mortality than whites, it is possible that the current sample of black individuals who have survived at least 30-days post-stroke are more resilient than their counterparts not included in the sample. Thus future research examining the impact of survival bias on post-stroke mortality may be warranted. Despite such limitations, there are notable strengths. The REGARDS study is a large national study of geographically and racially diverse subjects followed for over ten years with comprehensive collection of relevant demographic and clinical information at baseline prior to first stroke. Our study definition of stroke was clinically verified by an adjudication panel of expert stroke physicians. With these rich data, we were able to describe rates of long-term mortality among stroke survivors across major demographic and socioeconomic factors, including by geocoding residential factors, and compare our results with prior findings, which were focused on short-term mortality among stroke cases.
Summary:
Despite declining rates of stroke mortality in recent years, disparities persist. From a national study of stroke and its risk factors, we showed that rates of long-term mortality among stroke survivors were higher among individuals with disadvantageous SES. Improvements in care post-stroke, especially among individuals of low SES standing or residing in low SES neighborhoods, may be necessary to help address such disparities. Further, our results support prior work highlighting that the main driver of higher rates of stroke mortality among black vs. white adults is likely stroke incidence, as we extended prior findings which showed no racial disparities in short-term mortality (within 30 days) among stroke cases to a longer time frame (i.e. mortality after at least 30 days of first stroke). These data suggest that stroke prevention efforts could help address black vs. white disparities in overall stroke mortality.
Acknowledgements:
The authors thank the other investigators, the staff, and the participants of the REGARDS study for their valuable contributions. A full list of participating REGARDS investigators and institutions and further information about the study can be found at http://www.regardsstudy.org.
Sources of Funding:
This research project is supported by a cooperative agreement U01 NS041588 from the National Institute of Neurological Disorders and Stroke, National Institutes of Health, Department of Health and Human Service. Representatives of the funding agency have been involved in the review of the manuscript but not directly involved in the collection, management, analysis or interpretation of the data. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Neurological Disorders and Stroke or the National Institutes of Health.
Tali Elfassy was supported by the American Heart Association post-doctoral fellowship (17POST32490000) and is currently supported by the University of Miami Clinical and Translational Science Institute, from the National Center for Advancing Translational Sciences and the National Institute on Minority Health and Health Disparities (KL2TR002737). Dr. Zeki Al Hazzouri was supported in part by a grant from the NIH, National Institute on Aging (K01AG047273)
Footnotes
Disclosures: The authors have no conflicts of interest or disclosures to report.
References:
- 1.Benjamin EJ, Virani SS, Callaway CW, Chamerlain AM, Chang AR, Cheng S, et al. Heart Disease and Stroke Statistics-2018 Update: A Report From the American Heart Association. Circulation 20 2018;137:e67–e492. [DOI] [PubMed] [Google Scholar]
- 2.Lackland DT, Roccella EJ, Deutsch AF, Fornage M, George MG, Howard G, et al. Factors influencing the decline in stroke mortality: a statement from the American Heart Association/American Stroke Association. Stroke 2014;45:315–353. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Kleindorfer DO, Khoury J, Moomaw CJ, Alwell K, Woo D, Flaherty ML, et al. Stroke incidence is decreasing in whites but not in blacks: a population-based estimate of temporal trends in stroke incidence from the Greater Cincinnati/Northern Kentucky Stroke Study. Stroke 2010;41:1326–1331. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Kleindorfer DO, Lindsell C, Broderick J, Flaherty ML, Woo D, Alwell K, et al. Impact of socioeconomic status on stroke incidence: a population-based study. Annals of neurology 2006;60:480–484. [DOI] [PubMed] [Google Scholar]
- 5.Howard VJ. Reasons Underlying Racial Differences in Stroke Incidence and Mortality. Stroke; a journal of cerebral circulation 2013;44:S126–S128. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Avendano M, Kawachi I, Van Lenthe F, Boshuizen HC, Mackenbach JP, Van den Bos GA, et al. Socioeconomic status and stroke incidence in the US elderly: the role of risk factors in the EPESE study. Stroke 2006;37:1368–1373. [DOI] [PubMed] [Google Scholar]
- 7.Cox AM, McKevitt C, Rudd AG, Wolfe CDA. Socioeconomic status and stroke. The Lancet Neurology 2006;5:181–188. [DOI] [PubMed] [Google Scholar]
- 8.Howard G Why do we have a stroke belt in the southeastern United States? A review of unlikely and uninvestigated potential causes. The American journal of the medical sciences 1999;317:160–167. [DOI] [PubMed] [Google Scholar]
- 9.Howard G, Kleindorfer DO, Cushman M, Long DL, Jasne A, Judd SE, et al. Contributors to the Excess Stroke Mortality in Rural Areas in the United States. Stroke 2017;48:1773–1778. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Howard VJ, McClure LA, Kleindorfer DO, Cunningham SA, Thrift AG, Diez Roux AV, et al. Neighborhood socioeconomic index and stroke incidence in a national cohort of blacks and whites. Neurology 2016;87:2340. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Brown AF, Liang LJ, Vassar SD, Merikin SS, Longstreth WT Jr, Ovibiagele B, et al. Neighborhood socioeconomic disadvantage and mortality after stroke. Neurology 2013;80:520–527. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Howard G, Moy CS, Howard VJ, McClure LA, Kleindorfer DO, Kissela BM, et al. Where to Focus Efforts to Reduce the Black-White Disparity in Stroke Mortality: Incidence Versus Case Fatality? Stroke 2016;47:1893–1898. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Rosamond WD, Folsom AR, Chambless LE, Wang CH, McGovern PG, Howard G, et al. Stroke incidence and survival among middle-aged adults: 9-year follow-up of the Atherosclerosis Risk in Communities (ARIC) cohort. Stroke 1999;30:736–743. [DOI] [PubMed] [Google Scholar]
- 14.Labarthe DR, Howard G, Safford MM, Howard VJ, Judd SE, Cushman M, et al. Incidence and Case Fatality at the County Level as Contributors to Geographic Disparities in Stroke Mortality. Neuroepidemiology 2016;47:96–102. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Osypuk TL, Ehntholt A, Moon JR, Gilsanz P, Glymour MM. Neighborhood Differences in Post-Stroke Mortality. Circulation. Cardiovascular quality and outcomes 2017;10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Kapral MK, Wang H, Mamdani M, Tu JV. Effect of socioeconomic status on treatment and mortality after stroke. Stroke 2002;33:268–273. [DOI] [PubMed] [Google Scholar]
- 17.Wang Y, Rudd AG, Wolfe CD. Trends and survival between ethnic groups after stroke: the South London Stroke Register. Stroke 2013;44:380–387. [DOI] [PubMed] [Google Scholar]
- 18.Wolfe CD, Smeeton NC, Coshall C, Tilling K, Rudd AG. Survival differences after stroke in a multiethnic population: follow-up study with the South London stroke register. BMJ 2005;331:431. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Xian Y, Holloway RG, Noyes K, Shah MN, Friedman B. Racial differences in mortality among patients with acute ischemic stroke: an observational study. Annals of internal medicine 2011;154:152–159. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Howard VJ, Cushman M, Pulley L, Gomez CR, Go RC, Prineas RJ, et al. The reasons for geographic and racial differences in stroke study: objectives and design. Neuroepidemiology 2005;25:135–143. [DOI] [PubMed] [Google Scholar]
- 21.Howard VJ, Kleindorfer DO, Judd SE, McCLure LA, Safford MM, Rhodes JD, et al. Disparities in Stroke Incidence Contributing to Disparities in Stroke Mortality. Annals of neurology 2011;69:619–627. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Stroke−-1989. Recommendations on stroke prevention, diagnosis, and therapy. Report of the WHO Task Force on Stroke and other Cerebrovascular Disorders. Stroke 1989;20:1407–1431. [DOI] [PubMed] [Google Scholar]
- 23.Perkins M, Howard VJ, Wadley VG, Crowe M, Safford MM, Haley WE, et al. Caregiving Strain and All-Cause Mortality: Evidence From the REGARDS Study. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences 2013;68:504–512. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Howard G, Anderson R, Johnson NJ, Sorlie P, Russell G, Howard VJ. Evaluation of social status as a contributing factor to the stroke belt region of the United States. Stroke 1997;28:936–940. [DOI] [PubMed] [Google Scholar]
- 25.Diez-Roux AV, Kiefe CI, Jacobs DR Jr., Haan M, Jackson SA, Nieto FJ, et al. Area characteristics and individual-level socioeconomic position indicators in three population-based epidemiologic studies. Ann Epidemiol 2001;11:395–405. [DOI] [PubMed] [Google Scholar]
- 26.Li OL, Silver FL, Lichtman J, Fang J, Stamplecoski M, Wengle RS, et al. Sex Differences in the Presentation, Care, and Outcomes of Transient Ischemic Attack: Results From the Ontario Stroke Registry. Stroke 2016;47:255–257. [DOI] [PubMed] [Google Scholar]
- 27.Gargano JW, Wehner S, Reeves M. Sex differences in acute stroke care in a statewide stroke registry. Stroke 2008;39:24–29. [DOI] [PubMed] [Google Scholar]
- 28.Mosca L, Barrett-Connor E, Wenger NK. Sex/gender differences in cardiovascular disease prevention: what a difference a decade makes. Circulation 2011;124:2145–2154. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Koifman J, Hall R, Li S, Stamplecoski M, Fang J, Saltman AP, et al. The association between rural residence and stroke care and outcomes. J Neurol Sci 2016;363:16–20. [DOI] [PubMed] [Google Scholar]
- 30.Andersen KK, Olsen TS, Dehlendorff C, Kammersgaard LP. Hemorrhagic and ischemic strokes compared: stroke severity, mortality, and risk factors. Stroke 2009;40:2068–2072. [DOI] [PubMed] [Google Scholar]