Associations Between Incident Ischemic Stroke Events and Stroke and Cardiovascular Disease-Related GWAS SNPs in the Population Architecture Using Genomics and Epidemiology (PAGE) Study.

Publication Type:

Journal Article


Circulation. Cardiovascular genetics, Volume 5, Issue 2, p.210-216 (2012)


2012, Center-Authored Paper, Mar 2012, March 2012, Public Health Sciences Division, Shared Resources, Specimen Processing Core Facility


BACKGROUND: -Genome-wide association studies (GWAS) have identified loci associated with ischemic stroke (IS) and cardiovascular disease (CVD) in European-descent individuals, but their replication in different populations has been largely unexplored. METHODS AND RESULTS: -Nine single-nucleotide polymorphisms (SNPs) selected from GWAS and meta-analyses of stroke and 86 SNPs previously associated with myocardial infarction and CVD risk factors including blood lipids (HDL, LDL, triglycerides), type 2 diabetes and body mass index were investigated for associations with incident IS in European Americans (EA) N=26,276; African Americans (AA) N=8970; and American Indians (AI) N= 3570 from the Population Architecture using Genomics and Epidemiology Study. Ancestry-specific fixed effects meta-analysis with inverse variance weighting was used to combine study-specific log hazard ratios from Cox proportional hazards models. Two of 9 stroke SNPs (rs783396 and rs1804689) were associated with increased IS hazard in AA; none were significant in this large EA cohort. Of 73 CVD risk factor SNPs tested in EA, two (HDL and triglycerides SNPs) were associated with IS. In AA, SNPs associated with LDL, HDL and BMI were significantly associated with IS (3 of 86 SNPs tested). Out of 58 SNPs tested in AI, one LDL SNP was significantly associated with IS. CONCLUSIONS: -Our analyses showing lack of replication in spite of reasonable power for many stroke SNPs and differing results by ancestry highlight the need to follow-up on GWAS findings and conduct genetic association studies in diverse populations. We found modest pleiotropy associations between IS and BMI and lipids SNPs, though these findings require confirmation.