Genomic Risk Profiling of Ischemic Stroke: Results of an International Genome-Wide Association Meta-Analysis

IntroductionFamilial aggregation of ischemic stroke derives from shared genetic and environmental factors. We present a meta-analysis of genome-wide association scans (GWAS) from 3 cohorts to identify the contribution of common variants to ischemic stroke risk.
MethodsThis study involved 1464 ischemic stroke cases and 1932 controls. Cases were genotyped using the Illumina 610 or 660 genotyping arrays; controls, with Illumina HumanHap 550Kv1 or 550Kv3 genotyping arrays. Imputation was performed with the 1000 Genomes European ancestry haplotypes (August 2010 release) as a reference. A total of 5,156,597 single-nucleotide polymorphisms (SNPs) were incorporated into the fixed effects meta-analysis. All SNPs associated with ischemic stroke (P<1×10−5) were incorporated into a multivariate risk profile model.
ResultsNo SNP reached genome-wide significance for ischemic stroke (P<5×10−8). Secondary analysis identified a significant cumulative effect for age at onset of stroke (first versus fifth quintile of cumulative profiles based on SNPs associated with late onset, ß = 14.77 [10.85,18.68], P = 5.5×10−12), as well as a strong effect showing increased risk across samples with a high propensity for stroke among samples with enriched counts of suggestive risk alleles (P<5×10−6). Risk profile scores based only on genomic information offered little incremental prediction.
DiscussionThere is little evidence of a common genetic variant contributing to moderate risk of ischemic stroke. Quintiles based on genetic loading of alleles associated with a younger age at onset of ischemic stroke revealed a significant difference in age at onset between those in the upper and lower quintiles. Using common variants from GWAS and imputation, genomic profiling remains inferior to family history of stroke for defining risk. Inclusion of genomic (rare variant) information may be required to improve clinical risk profiling.

引言 缺血性脑卒中的家族聚集性源于共同的遗传与环境因素。本研究针对3个队列的全基因组关联分析（GWAS）开展荟萃分析，以明确常见变异对缺血性脑卒中发病风险的贡献。
方法 本研究共纳入1464例缺血性脑卒中病例与1932例对照。病例组采用Illumina 610或660基因分型阵列进行基因分型，对照组则采用Illumina HumanHap 550Kv1或550Kv3基因分型阵列。以2010年8月发布的1000基因组计划欧洲血统单倍型作为参考面板进行基因型填充。最终共纳入5156597个单核苷酸多态性（SNPs）进行固定效应荟萃分析。将所有与缺血性脑卒中相关（P<1×10^-5）的SNPs纳入多变量风险谱模型。
结果 未发现任何SNP达到缺血性脑卒中的全基因组显著性水平（P<5×10^-8）。后续分析显示，基于与晚发性脑卒中相关SNPs构建的累积风险谱，其最高与最低五分位数分层与脑卒中发病年龄存在显著累积效应（β=14.77，95%CI：10.85~18.68，P=5.5×10^-12）；同时，在携带富集提示性风险等位基因的样本中，脑卒中发病倾向随风险等位基因数量增加而显著升高（P<5×10^-6）。仅基于基因组信息的风险谱评分仅能提供极微弱的增量预测价值。
讨论 目前几乎没有证据表明常见遗传变异可介导缺血性脑卒中的中度发病风险。基于与早发性缺血性脑卒中相关等位基因的遗传负荷构建的五分位数分层，显示最高与最低五分位人群的发病年龄存在显著差异。利用全基因组关联分析及基因型填充得到的常见变异进行基因组风险分型，其风险预测效能仍劣于脑卒中家族史。未来或需纳入基因组（罕见变异）信息以优化临床风险分层模型。