RACIAL DISPARITIES IN STROKE AFFECTING BLACK AMERICANS

ABSTRACT

Stroke is the fifth leading cause of death in the United States and disproportionately affects Black Americans. The Reasons for Geographic and Racial Disparities in Stroke (REGARDS) cohort study is investigating why this disparity exists by comparing Black and White adults age 45+. Seminal findings included that higher stroke incidence in Black than in White persons, but not higher case fatality, drives the disparity. Higher prevalence of stroke risk factors in Black participants, especially hypertension and diabetes, explained 50% of the racial disparity. Elevated lipoprotein(a) was three times more frequent in Black participants and was a race-specific stroke risk factor only among Black participants. Higher interleukin-6 was a strong stroke risk factor in both Black and White participants with a hazard ratio of 2.0 for concentrations in the top versus bottom quartile. In mediation analysis, higher interleukin-6 with presence of stroke risk factors party explained the race disparity in stroke. Findings highlight the potential that prevention of stroke risk factors, treatment of higher Lp(a), and inflammation reduction have in reducing the race disparity in stroke.

BACKGROUND

After years of declining mortality, over the last several years, stroke mortality has been rising for reasons that are unclear (1). At the same time, Black Americans suffer disproportionately over all other race or ethnic groups in the United States and have a twofold higher risk of dying of stroke compared with White Americans. Further, after decades of a steady disparity, the Black–White disparity in stroke is now worsening. This review will discuss the reasons for these disparities and the research done over the past 20 years by our group to elucidate them. Race is considered in this review as a social construct, not a biological one. When genetic factors related to stroke are considered in this context, factors related to African ancestry will then be considered as potential causes of this race disparity but do not represent race as a social construct.

Traditional stroke risk factors include older age, hypertension, diabetes, heart disease, cigarette smoking, and atrial fibrillation. Other than age and atrial fibrillation, all of these risk factors are more common in Black than White adults.

THE REGARDS STUDY

The REGARDS cohort study was initiated in 2003 to improve understanding of the above issues. REGARDS broke the paradigm of epidemiological cohort studies, casting away use of field centers to use marketing techniques to enroll participants by telephone, visit them in their homes, and follow them exclusively by telephone to collect information. A company capable of doing in-home examinations was engaged to conduct the home visits. From 2003–2007, the study enrolled 30,239 Black and White adults ages 45–100 from across the United States (2). The Stroke Belt region of the southeast comprised half the sample; 42% were Black and 58% White and 55% female and 45% male. The Stroke Belt is an area where stroke mortality is higher than in other parts of the country (North Carolina, South Carolina, Georgia, Mississippi, Alabama, Louisiana, Arkansas, and Tennessee). The baseline examination, including a computer-assisted telephone interview, in-home physical measurements, electrocardiogram, and blood collection and analysis, was followed by phone calls every six months to ascertain possible stroke events. Blood samples were centrifuged near the participants’ homes and shipped overnight for reprocessing, analysis, or storage at a central laboratory at the University of Vermont (3). When possible, stroke was identified, and medical records were collected, reviewed, and adjudicated to validate and classify stroke type as ischemic (and its subtypes) or as primary hemorrhage.

DISPARITIES IN STROKE RISK FACTORS IN REGARDS

In cross-sectional analysis assessing the Framingham Stroke Risk Scale, REGARDS established that Black participants had a greater age- and sex-adjusted mean 10-year predicted stroke probability than White participants: 12.0% versus 9.2%, respectively (p < 0.001) (4). Race differences were largest for the risk score components of hypertension, systolic blood pressure, diabetes, smoking, and left ventricular hypertrophy (a measure of long-term blood pressure control).

CONTRIBUTORS TO STROKE DISPARITY AFFECTING BLACK AMERICANS

REGARDS investigators observed that stroke incidence was the cause of greater stroke mortality affecting Black Americans (5). Importantly, relative to White participants, as shown in Figure 1, Black participants aged 45–54 had a threefold greater incidence; this difference was 2.5-fold at age 55–64, it was 1.5-fold greater at age 65–74, and there was no excess risk at older ages. This finding is particularly problematic since stroke is a leading cause of disability with the disparity concentrated in the peak productive years of life.

Fig. 1.

Association of race with stroke incidence in REGARDS differs by age.

We then observed that the traditional risk factors listed above accounted for 50% of the excess stroke risk in Black participants; higher rates of hypertension and diabetes were the primary risk factors responsible (6). Further, hypertension was a more potent risk factor. Specifically, for every 10 mm Hg increment of systolic blood pressure, the risk of stroke was threefold greater in Black than White participants (HR 1.24 vs 1.08), even after accounting for other risk factors (7).

These findings can help reduce the stroke disparity burdening Black Americans. Better control of hypertension and diabetes may be key. However, the findings for hypertension suggest that more aggressive treatment targets may be needed. Data from the SPRINT trial support a target blood pressure of 120/80, and efforts to achieve this should be redoubled for Black Americans (8).

DISPARITIES IN HYPERTENSION AND DIABETES IN REGARDS

The study next aimed to understand why these risk factors disparately impacted Black individuals. To address this, a second large telephone survey and in-home visit were conducted about 10 years after baseline to update health history and especially to classify new onset of hypertension and diabetes since baseline.

We observed that Black individuals had a twofold greater risk of developing hypertension over 10 years. Risk factors mediating this were adverse levels by race of poor diet, high dietary sodium, lower education and income, and in women only, measures of adiposity (greater waist circumference and body mass index) (9). Black individuals also had a twofold greater risk of developing diabetes over 10 years. The main mediators of this were poor diet, lower socioeconomic status, and for women only, the adiposity measures (10).

The above findings suggest actionable steps to narrow the stroke disparity affecting Black Americans. Primordial prevention of hypertension and diabetes probably represents the most actionable step that could have a long-term impact on this condition.

MOLECULAR EPIDEMIOLOGY OF THE RACIAL DISPARITY IN STROKE IN REGARDS

Molecular epidemiology is the study of laboratory-based biomarkers in relation to disease risk. The purpose is to identify underlying factors related to pathophysiology, which in the case of this review, might explain racial disparities in stroke. The biomarkers discussed include circulating proteins and genetic markers. Biomarkers were studied two ways. They can be considered as mediators of the association of race and stroke risk, which means they reflect factors related to the social construct of race that represent factors in the causal pathway to stroke. They can also be considered as moderators, which means they are related to stroke in Black persons and not in White persons or vice versa. We tested these hypotheses by performing mediation analysis and interaction testing for these two questions.

Case Cohort Study to Assess Biomarkers of Stroke Risk

The study design embedded in REGARDS to address these research questions was a case cohort study of stroke risk. In this design, case groups and a cohort random sample were identified, and blood samples (serum, plasma DNA components) from those participants were retrieved from our study biorepository and analyzed for risk markers of interest (subsequently referred to as biomarkers). Biomarkers were selected in general because they have more adverse levels in Black than White persons and have biological rationale as stroke risk markers. Cox proportional hazards models, including sequential models adjusting for demographics, socioeconomic factors, and stroke risk factors, were then fit to estimate hazard ratios of stroke based on the biomarkers of interest. This cost-efficient design approach is highly efficient with results approximating what would be found if the biomarker measurements had been done on the full 30,239 participants. It also allowed consideration of other case groups not discussed here including incident cognitive impairment, coronary heart disease events, and venous thromboembolism.

At 5.4 years of follow-up, 572 cases of incident (first-time) ischemic stroke were identified. A cohort random sample of 1,104 participants was selected using stratification into equal groups by race and sex, and specific distributions of age groups to assure representation across the age spectrum (given the differential stroke risk by age across race).

Genetic Markers of Stroke Risk

The presentation discussed elevated lipoprotein(a) [Lp(a)] as a genetic-related risk factor for stroke. Lp(a) is a proinflammatory lipoprotein found in atherosclerotic plaques and is thought to contribute to destabilization of plaques. An American Heart Association scientific statement thoroughly reviewed its epidemiology and associations with risk of coronary heart disease and concluded higher levels are a causal risk factor for atherosclerotic disease (11). Higher levels and gene variation were associated with risk of coronary heart disease events with fewer data on stroke, or on race differences in associations. Circulating concentration of Lp(a) is stable over the life course, little influenced by environmental or other cardiovascular risk factors, and strongly influenced by genetic differences. One in four individuals have high levels based on currently accepted thresholds. As such, cardiovascular prevention guidelines recommend one-time testing for adults to assess cardiovascular risk (12). However, there is not yet an approved intervention for reducing risk among those with higher levels. As such, optimizing cardiovascular health using proven means like healthy lifestyle promotion, statins, and blood pressure control in patients with high levels is recommended. Finally, investigational drugs that target Lp(a) lowering are in phase 3 trials now for secondary prevention of atherosclerotic disease.

Among race and ethnic groups, concentrations of Lp(a) vary, and this is thought to be driven by genetic variation in a number of genes. People of African ancestry have by far the highest concentration but are less represented in epidemiology studies and clinical trials. Given this, Lp(a) becomes a possible contributor to racial disparities in stroke affecting Black Americans.

We measured Lp(a) in plasma from baseline samples in the REGARDS case cohort study (13). Median (interquartile range) values for Lp(a) in the cohort random sample were about threefold higher in Black compared to White participants. Specifically, values were 32.8 (16.8–55.0) mg/dL in Black women, 26.7 (12.6–56.4) mg/dL in Black men, 11.0 (3.5–39.5) mg/dL in White women, and 8.8 (3.3–30.6) mg/dLin White men. After adjusting for demographics and stroke risk factors, the hazard ratio (HR) of stroke was 2.0 (95% CI 1.1–3.4) among Black participants and 1.1 (95% CI 0.6–2.0) among White participants. The difference between these two HRs was not statistically significant with p=0.12, but this was close to the a priori p value threshold for detecting differences of 0.10. Importantly, race- and sex-specific thresholds were used to define the quartiles, such that for Black participants the top quartile values were >57 mg/dL in Black men, >56 mg/dL in Black women, and >32 mg/dL and >40 mg/dL in White men and women, respectively. Based on these results, we concluded that it appears Lp(a) was a race-specific risk factor for stroke in Black participants of REGARDS. Thus, it may contribute to the racial disparity in stroke affecting Black Americans.

REGARDS findings for Lp(a) suggest that new drugs might have greater efficacy to reduce stroke risk in persons of African ancestry. Whether these findings would hold for overall efficacy is unknown, but it is unlikely that the landmark studies of these drugs will have sufficient power to answer this question.

Circulating Protein Markers of Stroke

The presentation then discussed markers of inflammation and risk of stroke. The prototype inflammation marker used in clinical practice to identify cardiovascular risk is C-reactive protein (CRP), a nonspecific inflammation marker. In REGARDS, CRP was measured in all participants at baseline. Black participants, especially Black women, had much higher levels than White participants (14). Over 6.9 years of follow-up, the association of CRP with stroke risk was modest—the adjusted HR per standard deviation increment of baseline CRP was 1.18 (95% CI 1.09–1.28) overall, 1.14 (95% CI 1.00–1.29) in Black participants, and 1.22 (95% CI 1.10–1.35) in White participants, so there was no difference in the association by race (15). However, to predict the same level of stroke risk associated with CRP >3 mg/L in a White participant, CRP would need to be >9 mg/L in a Black participant.

Since CRP is not a causal risk factor but rather a marker of inflammation, the REGARDS case cohort study measured three cytokines to further study the relationship of inflammation with stroke, the pro-inflammatory cytokines interleukin-6 (IL-6) and IL-8, and the anti-inflammatory cytokine IL-10. Like CRP, IL-6 is an established risk marker for coronary heart disease, but unlike CRP it appears to be a causal risk factor (16). We observed no associations of IL-8 and IL-10 with stroke risk in REGARDS, but IL-6 was strongly associated with risk (17). In the fully adjusted model, the HR was 2.0 (95% CI 1.2–3.1) comparing the top to the bottom quartile of IL-6, and this did not differ by race (p >0.2). In mediation analysis to assess whether IL-6 explained the association of race with stroke risk, we observed that IL-6 mediated the racial disparity via the effects of stroke risk factors such that the pro-inflammatory effects of risk factors might explain part of the race disparity in stroke. This leads us to a hypothesis that inflammation lowering in those with risk factors might reduce stroke more in Black than White persons and also reduce the disparity.

CONCLUSIONS

REGARDS has identified actionable steps that might reduce race disparities in stroke affecting Black Americans. These might include targeted interventions to reduce the incidence of hypertension and diabetes but might also include novel interventions against inflammation in people with risk factors for stroke. Agents are available for this such as canakinumab, a fully human monoclonal antibody against IL-1 beta, which lowers IL-6 and CRP and is indicated for treatment of certain autoimmune diseases. For common diseases, while there is efficacy in reducing cardiovascular disease in those with established cardiovascular disease, it is primarily being developed as an anticancer treatment (16). Further research on racial differences in the association of Lp(a) with stroke and other cardiovascular diseases is needed, and it is hoped that clinical trials of Lp(a) inhibition include sufficient representation of Black individuals, as they have the highest concentrations of Lp(a).