APOL1 Nephropathy Risk Variants and Incident Cardiovascular Disease Events in Community-Dwelling Black Adults
Background: APOL1 renal risk variants are strongly associated with chronic kidney disease in Black adults, but reported associations with cardiovascular disease (CVD) have been conflicting.
Methods: We examined associations of APOL1 with incident coronary heart disease (n=323), ischemic stroke (n=331), and the composite CVD outcome (n=500) in 10 605 Black participants of the REGARDS study (Reasons for Geographic and Racial Differences in Stroke). Primary analyses compared individuals with APOL1 high-risk genotypes to APOL1 low-risk genotypes in Cox proportional hazards models adjusted for CVD risk factors and African ancestry.
Results: APOL1 high-risk participants were younger and more likely to have albuminuria at baseline than APOL1 low-risk participants. The risk of incident stroke, coronary heart disease, or composite CVD end point did not significantly differ by APOL1 genotype status in multivariable models. The association of APOL1 genotype with incident composite CVD differed by diabetes mellitus status (Pinteraction=0.004). In those without diabetes mellitus, APOL1 high-risk genotypes associated with greater risk of incident composite CVD (hazard ratio, 1.67; 95% confidence interval, 1.12–2.47) compared with those with APOL1 low-risk genotypes in multivariable adjusted models. This latter association was driven by ischemic strokes (hazard ratio, 2.32; 95% confidence interval, 1.33–4.07), in particular, those related to small vessel disease (hazard ratio, 5.10; 95% confidence interval, 1.55–16.56). There was no statistically significant association of APOL1 genotypes with incident CVD in subjects with diabetes mellitus. The APOL1 high-risk genotype was associated with higher stroke risk in individuals without but not those with chronic kidney disease in fully adjusted models.
Conclusions: APOL1 high-risk status is associated with CVD events in community-dwelling Black adults without diabetes mellitus.
See Editorial by Waitzman et al
Black Americans experience significant health disparities for kidney and cardiovascular diseases. Much of the excess risk for chronic and end-stage kidney disease in Black Americans is attributable to carriage of 2 coding variants in the apolipoprotein L1 gene (APOL1), which are present only in persons with African ancestry; 13% of African Americans carry APOL1 high-risk genotypes. In patients with chronic kidney disease, cardiovascular disease is the leading cause of death; however, the relationship between APOL1 renal risk variants and cardiovascular disease is unresolved and conflicting results have been reported. In this study, the authors studied 10 605 community-dwelling black participants enrolled in the REGARDS (Reasons for Geographic and Racial Differences in Stroke) to assess the association of APOL1 renal risk variants with incident ischemic stroke and coronary heart disease. The study found that carriage of 2 APOL1 risk alleles was not associated with cardiovascular events in individuals with chronic kidney disease or diabetes mellitus, likely because both are themselves such strong risk factors for cardiovascular disease events. However, in participants without diabetes mellitus, carriage of 2 APOL1 risk alleles was independently associated with ischemic stroke because of small vessel occlusion and to a lesser degree, coronary heart disease. This finding may provide clues into reasons for the higher incidence rate of ischemic stroke among blacks and may also have clinical implications for screening and counseling to reduce modifiable risk factors associated with stroke and coronary heart disease in blacks without the traditional risk factors of chronic kidney disease or diabetes mellitus.
Black Americans experience significant health disparities for kidney and cardiovascular diseases (CVDs), which are only partially explained by access to medical care or other socioeconomic factors.1–3 In patients with chronic kidney disease (CKD), CVD is the leading cause of death.4–7 Much of the excess risk for nondiabetic CKD in Blacks is explained by 2 common coding alleles (termed G1 and G2) in APOL1, encoding APOL1 (apolipoprotein L1). These variants only occur on African ancestry chromosomes and extend protection against Trypanosoma brucei.8–11 APOL1 high-risk genotypes (carriage of 2 APOL1 risk alleles) are associated with higher prevalence and progression of a spectrum of CKDs, characterized by albuminuria and reduced estimated glomerular filtration rate, each of which is risk factors for CVD.2,5,7,12–14 The APOL1 protein is expressed in the systemic and kidney vasculature, and APOL1 variants are associated with more severe kidney arteriosclerosis with aging and hypertension.15,16
The relationship between APOL1 kidney risk variants and clinical CVD remains unresolved, with some studies reporting positive associations and others no association for prevalent or incident heart disease or stroke.12–14,17–19 Among 1959 Black Americans in the Jackson Heart Study and 749 Black Americans in the Women’s Health Initiative, APOL1 high-risk status was associated with having higher risk for incident CVD (both cohorts), whereas individuals with APOL1 high-risk genotypes had lower prevalence of coronary artery calcification in Jackson Heart Study participants.14 Among 707 Black individuals, 65 years of age or older enrolled in the Cardiovascular Heart Study, APOL1 high-risk genotypes were positively associated with subclinical peripheral atherosclerosis and incident myocardial infarction (MI).19 However, a study of 1315 middle-aged Black adults with mostly preserved kidney function enrolled in the CARDIA (Coronary Artery Risk Development in Young Adults) failed to find an association between APOL1 renal risk alleles and subclinical markers of atherosclerotic disease.20Among studies enrolling individuals with known CVD risk factors including referral for cardiac catheterization, hypertensive CKD, or hypertension, APOL1 high-risk status did not associate with overall risk for prevalent or incident CVD.12,13,21,22
Given present uncertainties about the association of APOL1 genetic risk variants with incident CVD events, we obtained APOL1 genotypes for 10 605 Black participants enrolled in the REGARDS (Reasons for Geographic and Racial Differences in Stroke) and assessed the association of APOL1 variants with incident ischemic stroke, coronary heart disease (CHD), and the composite of these outcomes. In addition, in prespecified analyses, we examined whether CKD, advanced age, diabetes mellitus, or sickle cell trait modified APOL1 associations with CVD.
The data, analytic methods, and study materials will be made available to other researchers for purposes of reproducing the results or replicating the procedure. The Institutional Review Board has reviewed and been a part for the REGARDS policies and procedures related to data access. Investigators who would like to request that data will need to first obtain information on the policies and procedures to do so. The person to contact for this information and any questions is the Publications and Presentations program manager for REGARDS, Margaret Stewart (email@example.com).
The REGARDS study is a population-based investigation of stroke incidence in Black and White adults ≥45 years of age. The study design has been reported elsewhere.21 Briefly, participants were recruited from the 48 contiguous US states and the District of Columbia. The study was designed to provide an approximately equal representation of men and women and oversampled Black individuals and persons residing in the stroke belt/buckle regions of the United States. Overall, 30 239 individuals were enrolled between January 2003 and October 2007 (42% Black, 55% women). Participants were contacted every 6 months by telephone to assess new-onset stroke, CHD, and death. Medical records were retrieved for stroke and CHD-related hospitalizations and deaths for further adjudication. The REGARDS study protocol was approved by the Institutional Review Boards governing research in human subjects at the participating centers and all participants provided informed consent.
A total of 10 872 REGARDS Black participants were genotyped for APOL1 renal risk variants. Of those, 196 cases were missing genotype data for rs71785313, 98 cases were missing data for rs73885319, and 27 were missing data for both APOL1 alleles, leaving 10 605 participants for the present analysis. We further excluded participants with prevalent stroke or missing data on prevalent stroke status from the analysis of incident stroke (976) and prevalent CHD or missing data on prevalent CHD status from the analysis of incident CHD (1824). Participants with prevalent stroke and prevalent CHD were also removed from the analysis of the composite CVD outcome. History of CHD at baseline was defined by a self-reported history or electrocardiographic evidence of MI or a self-reported history of a revascularization procedure. History of stroke was defined on the basis of self-report. After removing participants missing follow-up time (216) or who had a hemorrhagic stroke (38), there were 9375 participants in the stroke analysis. After removing participants missing follow-up time (191), there were 8590 participants in the CHD analysis. There were 7908 participants in the combined CVD analysis after removal of participants with a history of CHD or stroke.
APOL1 risk variants (G1 and G2) were genotyped using TaqMan SNP Genotyping Assays (Applied Biosystems/ThermoFisher Scientific).9,22 APOL1 high-risk status was defined as the presence of 2 risk alleles (G1/G1, G2/G2, or G1/G2) versus the low-risk status, defined as having 1 or 0 risk variants (G1/G0, G2/G0, G0/G0), representing a recessive model. Sickle cell trait (HbS) was genotyped as described in Naik et al.23 In secondary analyses, an additive (0, 1, 2 risk alleles) and dominant model (2 or 1 APOL1 risk alleles versus 0 risk alleles) were assessed. A subset of participants had available genomic array data (Illumina exome chip) to estimate population substructure (n=6714) via principal components generated using EIGENSOFT software.24,25 In sensitivity analyses, principal components were used to adjust for African ancestry in the subset with available data.
Incident ischemic stroke was investigated by retrieval of hospital records on self-report of a possible stroke/transient ischemic attack or a positive response to the Questionnaire for Verifying Stroke-Free Status during follow-up telephone contact or report by a proxy of death related to stroke. Stroke was then confirmed by a panel of neurologists according to the World Health Organization definition.26 Events not meeting the World Health Organization definition but characterized by symptoms lasting <24 hours with neuroimaging consistent with acute infarct or hemorrhage were classified as clinical strokes. Additionally, medical records, death certificates, and autopsy reports were retrieved and reviewed to determine if a participant death was stroke-related following guidelines described. Strokes were further subclassified into etiologic subtypes of small vessel occlusion, large vessel atherosclerosis, cardioembolic, or unclassified.27 Stroke subtype classifications were based on the stroke cause as determined during adjudication.28–30 Data on incident stroke was available through April 1, 2016.
In addition to stroke events, CHD events including nonfatal MI or CHD death were investigated during the follow-up telephone interviews. After report of a CHD-related hospitalization or death, medical records were retrieved, and the event was adjudicated by trained clinicians following published guidelines.31,32CHD was confirmed by presence of signs or symptoms suggestive of ischemia; a rising or falling pattern in cardiac troponin or creatine phosphokinase-MB over 6 or more hours with a peak value greater than or equal to twice the upper limit of normal (diagnostic cardiac enzymes); and electrocardiographic changes consistent with ischemia or MI, guided by the Minnesota code. Additionally, medical records in the last year of life, death certificates and autopsy reports were collected and reviewed to determine if the death was a CHD death following published guidelines. Incident CHD events through December 31, 2013, were available for the current analyses.
A composite event was defined by the first incident CHD or ischemic stroke event during follow-up. Secondary outcomes included serum total cholesterol, HDL-C (high-density lipoprotein cholesterol), LDL-C (low-density lipoprotein cholesterol), and triglyceride levels. The serum total cholesterol, HDL-C, and triglyceride levels were measured by colorimetric reflectance spectrophotometry using the Ortho Vitros Clinical Chemistry System 950IRC instrument (Johnson & Johnson Clinical Diagnostics, New Brunswick, NJ). LDL-C was further calculated from these 3 measures using the Friedewald equation.33,34
Data Collection at Baseline
Data on covariates of interest were collected during a baseline in-home visit (including the collection of a blood and spot urine sample) and a separate computer-assisted telephone interview. Information on age, smoking status, lipid-lowering drugs, and antihypertensive medication use was collected by self-report. Systolic and diastolic blood pressure were defined as the average of 2 seated blood pressure measures taken after a 5-minute rest. Diabetes mellitus was defined as fasting serum glucose ≥126 mg/dL, nonfasting serum glucose ≥200 mg/dL, or use of antidiabetes mellitus medications. Serum creatinine was calibrated to an international isotope dilution mass spectroscopic-traceable standard measured by colorimetric reflectance spectrophotometry. We calculated estimated glomerular filtration rate using the CKD-Epidemiology Collaboration equation.35 Albumin and creatinine were measured using the random spot urine specimen by nephelometry (BN ProSpec Nephelometer, Dade Behring, Marburg, Germany) and Modular-P chemistry analyzer (Roche/Hitachi, Indianapolis, IN), respectively. Spot urine albumin-to-creatinine ratio was calculated in mg/g. Abnormal albumin-to-creatinine ratio was defined as urine albumin-to-creatinine ratio >30 mg/g. Serum hsCRP (high-sensitivity C-reactive protein) was measured using a high-sensitivity particle-enhanced immunonephelometric assay.
We compared participant characteristics by APOL1 risk allele count (0, 1, or 2) using the χ2 test for categorical variables and ANOVA for continuous variables. We calculated incidence rates and 95% confidence intervals (CIs) of each outcome for each risk category (0, 1, or 2). After confirming the proportionality assumption, Cox proportional hazards models were used to assess the association of APOL1 high-risk versus low-risk status (recessive model) with incident events. The follow-up time was measured from the baseline interview date to occurrence of the event or last follow-up. We generated hazard ratios (HRs) from 3 models including a crude model, a second model adjusted for age, sex, smoking (former, current, or never), hypertension (defined as baseline systolic blood pressure ≥140 mm Hg, diastolic blood pressure ≥90 mm Hg, or self-reported use of hypertension medications), lipid-lowering medications, and diabetes mellitus, and a third model adjusted for variables in model 2 as well as principal components for ancestry in the subset of individuals with available data. Secondary analyses considered additive and dominant models for testing the association of APOL1 risk variants with CVD outcomes. Stratified models were used to assess effect modification of the relationship between APOL1 genotypes and CVD outcomes by diabetes mellitus status, the presence or absence of CKD, baseline age ≥65 compared with <65 years, and sickle cell trait status. Unstratified models that included an interaction term were used to evaluate statistical significance. Analyses of fasting lipids were assessed in crude and multivariable adjusted linear regression models. All analyses were performed using SAS v. 9.4.
Of 10 605 participants, 4420 had 0 APOL1 risk alleles, 4839 had 1 risk allele (2989 had the G1/0 genotype and 1850 the G2/0 genotype), and 1346 had 2 risk alleles (530 were G1 homozygotes, 205 were G2 homozygotes, and 611 were compound G1/G2 heterozygotes).
Table 1 compares demographic, clinical, and laboratory characteristics of participants by a number of APOL1 G1/G2 risk variants. The participants with higher numbers of risk variants were younger, had higher total cholesterol and HDL-C concentrations, and were more likely to have albuminuria at baseline.
Association of APOL1 Risk Status With Incident CVD
There were 331 ischemic stroke events, 323 CHD events, and 500 composite events. Crude incidence rates of the combined CVD end point and individual components of the end point by APOL1 genotype are depicted in Table 2. In general, incident CVD rates numerically increased with carriage of 1 or 2 APOL1 risk variants.
HRs of CVD events by APOL1 genotype (comparing 2 versus 0 or 1 risk variants) are shown in Table 3. In crude and multivariable adjusted models, there were no statistically significant differences in the risk of incident stroke, CHD, or the composite CVD end point in individuals with 2 as compared with 0 or 1 APOL1 risk variants. The results did not differ when analyses were repeated using additive or dominant models (Tables I and II in the Data Supplement) or when comparing individuals homozygous for G1 or G2 to those with 0 risk variants (Tables III and IV in the Data Supplement).
The association of APOL1 genotype with incident composite CVD differed by diabetes mellitus status with a P value for interaction of 0.004 (Table 4). Among those without diabetes mellitus, individuals with 2 APOL1 risk variants had a higher risk of incident combined CVD than individuals with 0 or 1 APOL1 risk variants in the crude model (HR, 1.42; 95% CI, 1.05–1.92) and the multivariable adjusted model (HR, 1.48; 95% CI, 1.09–2.01). The results strengthened when adjusted for ancestry (HR, 1.67; 95% CI, 1.12–2.47), indicating that APOL1 variants and not African ancestry per se is driving the association. When examining individual components of the composite outcome, the magnitude and strength of the association were greatest for incident stroke events (HR, 2.32; 95% CI, 1.33–4.07). In contrast, among individuals with diabetes mellitus at the baseline visit, there was no significant association of APOL1 genotype with risk of incident CVD in any of the models, and the direction of the association was opposite to what was observed for individuals without diabetes mellitus. The results were similar, but the associations were weaker when analyses were repeated using additive and dominant models (Tables V and VI in the Data Supplement, respectively).
Among those with normal kidney function at baseline, individuals with 2 APOL1 risk variants had a higher hazard of incident stroke (HR, 1.79; 95% CI, 1.04–3.07) than individuals with 0 or 1 APOL1 risk variants in the fully adjusted model including ancestry (Model 2; Table 5). In contrast, risks for incident CHD or composite events did not significantly differ by APOL1 genotype in this group. The associations of APOL1 high-risk genotype with stroke in those with normal kidney function were qualitatively similar in additive and dominant models (Tables VII and VIII in the Data Supplement). Among those with CKD at baseline, there were no statistically significant associations of APOL1 variants with CVD outcomes in any model (Table 5; Tables VII and VIII in the Data Supplement). Results did not materially change when stratified by estimated glomerular filtration rate (< versus ≥60 mL/min per 1.73 m2) or albumin-to-creatinine ratio (≥ versus <30 mg/g), separately (Tables IX and X in the Data Supplement).
Tables XI through XIV in the Data Supplement show models stratified by age and sickle cell trait status. In general, there were no statistically significant differences in the association of APOL1 risk variants with incident CVD events by strata of age or sickle cell trait in fully adjusted models including ancestry principal components.
APOL1 Variants and Ischemic Stroke Subtypes
Associations of APOL1 risk variants with ischemic stroke subtypes are presented in Table XV in the Data Supplement. Carriage of the high-risk genotype was associated with greater risk of strokes because of small vessel disease in fully adjusted models (HR, 3.74; 95% CI, 1.61–8.65). When stratified by diabetes mellitus status, the magnitude and strength of this association were greater in individuals without diabetes mellitus (no diabetes mellitus: HR, 5.10; 95% CI, 1.55–16.56; diabetes mellitus: HR, 2.67; 95% CI, 0.79–9.04). No statistically significant associations of high-risk genotypes with other stroke subtypes were noted in multivariable adjusted models.
Association of APOL1 Variants With Lipid Profiles
Since APOL1 is a constituent of HDL, and HDL-C concentrations correlate with cardiovascular risk, we analyzed the associations of APOL1 risk variants with lipid measures under a recessive model (Table XVI in the Data Supplement). HDL-C and total cholesterol concentrations were significantly higher in individuals with 2 as compared with 0 or 1 APOL1 risk variants in models adjusted for age, sex, smoking, use of lipid-lowering medications, hypertension, and diabetes mellitus. However, these associations were attenuated and no longer statistically significant after adjustment for ancestry in the subset of individuals with available data. Similar results were obtained when examining additive or dominant models (Table XVII and XVIII in the Data Supplement).
In this study of community-dwelling Black adults, we assessed the relationship between APOL1 risk variants and incident ischemic stroke, CHD and a composite outcome comprising these clinical events. We observed a trend for higher incidence rates for stroke, CHD, and composite events with increased with greater carriage of APOL1 risk alleles. Further, the association of APOL1 nephropathy risk variants with CVD risk differed by diabetes mellitus status, such that APOL1 high-risk status was independently associated with higher risk for composite CVD events, primarily driven by ischemic stroke, in patients without diabetes mellitus; whereas in patients with diabetes mellitus and APOL1 high-risk status, the risk of stroke, CHD, and composite events was not statistically significant and in the opposite direction. We also found a less robust, but still significant, an association of APOL1 high-risk status with incident stroke in a subgroup analysis of subjects without CKD. Our findings suggest that conflicting results among prior studies for APOL1 associations with CVD may be because of effect modification by diabetes mellitus and CKD, each of which is a known independent risk factor for CVD.
The association of APOL1 with subclinical and clinical CVD has been inconsistent, showing positive association (risk), negative association (protection), or no association for subclinical atherosclerosis and clinical CVD outcomes. Differences in the direction of association have been attributed to the complexity of shared and reciprocal risk factors for CVD and CKD and the inherent difficulties in adequately adjusting for confounding by severity of kidney disease, all of which are strongly associated with APOL1 high-risk status and are risk factors for CVD.6,18,36,37 Positive APOL1 high-risk genotype associations with incident CVD have been evident in some (Jackson Heart Study, Women’s Health Initiative, and Cardiovascular Health Study) but not all (Atherosclerosis Risk in Communities) population-based cohort studies.14,19,38 In contrast, in studies enrolling participants with known risk factors for CVD, including referral for cardiac catheterization, hypertension, and CKD, the association between APOL1 high-risk status and incident CVD was weak or absent, perhaps because it was masked by the stronger risk conferred by underlying morbidities.12,13,39 Relative proportions of European and African ancestry are also potential confounders. Europeans are at greater risk for atherosclerosis, and similarly, increasing European ancestry in Black Americans is associated with more coronary calcification and atherosclerosis, which are predictors of CVD.40,41
In an effort to determine whether CKD, advanced age, diabetes mellitus, or sickle cell trait might be modifiers of APOL1 associations with CVD, we performed subgroup analyses. APOL1 high-risk status was not associated with CVD in subgroups with CKD or diabetes mellitus at baseline. However, among patients without diabetes mellitus, APOL1 high-risk status was associated with greater risk for the composite outcome in the crude model. Further, the magnitude and strength of the association became stronger in the fully adjusted model accounting for African ancestry, primarily driven by stroke. Additionally, APOL1 high-risk genotypes were significantly associated with 79% greater hazard for incident stroke in patients without CKD in the fully adjusted model. These results indicate that the APOL1 high-risk genotype is an independent risk factor for ischemic stroke and, to a lesser degree, CHD, among individuals without diabetes mellitus or CKD. It is also notable that APOL1 high-risk genotypes were most strongly associated with ischemic strokes because of small vessel occlusion, particularly in individuals without diabetes mellitus. If replicated in future studies, these results may provide novel insights into reasons for racial differences in the incidence of strokes caused by small vessel disease.
The notable interaction between diabetes mellitus and APOL1 for incident CVD events was unexpected, although consistent with CKD studies which report positive APOL1 associations for nondiabetic kidney disease but not for diabetic kidney disease.9,42,43 Additionally, APOL1 high-risk status was associated with greater hemodialysis survival rates, mainly attributed to CVD, for dialysis patients with nondiabetic end-stage kidney disease, but not in patients with diabetic end-stage kidney disease.44 Our results, together with others, suggest that diabetes mellitus may attenuate APOL1 penetrance. High concentrations of glucose result in oxidative imbalance and downregulation of autophagic pathways required for cell homeostasis.18,45 APOL1 variant proteins disrupt normal trafficking of endosomal vesicles in kidney cells causing reduced autophagic flux, oxidative stress, and cytoxicity.46–48 The pathophysiological mechanisms underlying the putative diabetes mellitus interaction with APOL1 risk variants warrants further investigation.
To determine whether the APOL1 associations with CVD followed a recessive model, as it does for nondiabetic CKD, we tested the additive and dominant genetic models in the subgroup analyses. We found that APOL1 high-risk status was significantly associated with stroke and composite events under the additive, but not the dominant genetic model, with each variant allele increasing risk for ischemic stroke by 54%. Ito et al14 previously reported that the G2 deletion allele may have greater penetrance (effect size) than the missense G1 variants; however, we found no evidence of this in our study. A more recent study of older patients enrolled in the Cardiovascular Health Study also found no differences in penetrance sizes for the G1 and G2 renal risk alleles with MI.19
To investigate whether lipids may play a role in the association of APOL1 genotype with CVD outcomes, we examined the association of APOL1 with available lipid parameters in REGARDS. APOL1 risk alleles were associated with total cholesterol, HDL-C, and LDL-C in crude models for additive, dominant, and recessive models, but adjusting for African ancestry abrogated the associations. These results argue against dyslipidemia playing a primary role in the association between APOL1 risk variants and CVD risk. Nonetheless, a previous study of Black participants enrolled in the REGARDS reported that increasing numbers of APOL1 were associated with higher circulating levels of small HDL particles, which may contribute to greater risks of CVD in study participants carrying APOL1 risk variants.49 Mechanistically, APOL1 protein isoform association with small HDL complexes rather than overall HDL-C levels might be more relevant to CVD risk.50
The strengths of our study include the large, community-based sample of Black Americans with extensive data on cardiovascular risk factors, with genotyping for APOL1 renal risk variants, sickle cell trait, and Genome Wide Association Study data for 6714 participants to correct for ancestry substructure. By limiting our study to only incident ischemic CVD events, we reduced phenotype heterogeneity. Our study does have limitations. We did not have ancestry informative markers on the entire cohort, which might have strengthened the APOL1 associations. In addition, we did not have follow-up measures of kidney function, precluding us from determining whether associations of APOL1 risk variants with CVD outcomes was at least partially explained by a faster decline in estimated glomerular filtration rate or a greater increase in albuminuria over time.
In conclusion, in a large, community-based study of Black Americans across the United States, APOL1 was independently associated with CVD in nondiabetics, and to a lesser degree, in persons without CKD, and these associations were strengthened after correction for African ancestry. This study highlights the complexity of APOL1 association with clinical CVD and the difficulties in disentangling multiple interactive factors that contribute to CVD.
We dedicate this manuscript to the memory our dear friend and colleague, Michele Sale. We thank the other investigators, the staff, and the participants of the REGARDS study (Reasons for Geographic and Racial Differences in Stroke) for their valuable contributions. A full list of participating REGARDS investigators and institutions can be found at http://www.regardsstudy.org.
Sources of Funding
This study was supported by a cooperative agreement U01 NS041588 from the National Institute of Neurological Disorders and Stroke, National Institutes of Health. Dr Gutiérrez was supported by R01NS080850 and U01DK102730 from the National Institutes of Health (NIH) and a Strategically Focused Research Network in Disparities in Cardiovascular Disease grant from the American Heart Association (15SFDRN25620022). Dr Kopp was supported by the Intramural Research Program, National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK), NIH, Bethesda, MD. Dr Winkler was funded in part with federal funds from the National Cancer Institute, NIH, under contract HHSN26120080001E and supported in part by the Intramural Research Program of the NIH, National Cancer Institute, Center for Cancer Research. The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the US Government. 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, the NIDDK, the National Cancer Institute or the NIH. Representatives of the funding agency have been involved in the review of the article but not directly involved in the collection, management, analysis, or interpretation of the data. Additional funding was provided by an investigator-initiated grant-in-aid from Amgen Corporation. Amgen did not have any role in the design and conduct of the study, the collection, management, data analysis, or interpretation of the data, or the preparation of the article.
↵* Drs Gutiérrez and Irvin contributed equally to this work.
The Data Supplement is available at http://circgenetics.ahajournals.org/lookup/suppl/doi:10.1161/CIRCGEN.117.002098/-/DC1.
- Received December 7, 2017.
- Accepted April 4, 2018.
- © 2018 American Heart Association, Inc.
- Howard G,
- et al
- van der Velde M
- Matsushita K
- Genovese G,
- et al
- Chen TK,
- et al
- Freedman BI,
- et al
- Ito K,
- et al
- Madhavan SM,
- et al
- Hughson MD,
- et al
- Mukamal KJ,
- et al
- Gutiérrez OM,
- et al
- Kopp JB,
- et al
- Naik RP,
- et al
- 26.↵WHO Task Force on Stroke and other Cerebrovascular Disorders. Recommendations on stroke prevention, diagnosis, and therapy. Report of the WHO Task Force on Stroke and Other Cerebrovascular Disorders. Stroke. 1989;20:1407–1431.
- Adams HP Jr.
- Schneider AT,
- et al
- Luepker RV,
- et al
- Thygesen K,
- et al
- Friedewald WT,
- et al
- Warnick GR,
- et al
- Peralta CA,
- et al
- Matsushita K,
- et al
- Grams ME,
- et al
- Wang H,
- et al
- Wassel CL,
- et al
- Divers J,
- et al
- Granado D,
- et al
- Olabisi OA,
- et al
- Cubedo J,
- et al