Genetics and genomics for the prevention and treatment of cardiovascular disease: update: a scientific statement from the American Heart Association.

Abstract

HomeCirculationVol. 128, No. 25Genetics and Genomics for the Prevention and Treatment of Cardiovascular Disease: Update Free AccessResearch ArticlePDF/EPUBAboutView PDFView EPUBSections ToolsAdd to favoritesDownload citationsTrack citationsPermissions ShareShare onFacebookTwitterLinked InMendeleyReddit Jump toFree AccessResearch ArticlePDF/EPUBGenetics and Genomics for the Prevention and Treatment of Cardiovascular Disease: UpdateA Scientific Statement From the American Heart Association Santhi K. Ganesh, MD, FAHA, Donna K. Arnett, PhD, MSPH, FAHA, Themistocles L. Assimes, MD, PhD, Craig T. Basson, MD, PhD, FAHA, Aravinda Chakravarti, PhD, Patrick T. Ellinor, MD, PhD, FAHA, Mary B. Engler, PhD, FAHA, Elizabeth Goldmuntz, MD, David M. Herrington, MD, MHS, FAHA, Ray E. Hershberger, MD, FAHA, Yuling Hong, MD, PhD, FAHA, Julie A. Johnson, PharmD, FAHA, Steven J. Kittner, MD, FAHA, Deborah A. McDermott, MS, CGC, James F. Meschia, MD, Luisa Mestroni, MD, Christopher J. O’Donnell, MD, MPH, Bruce M. Psaty, MD, PhD, FAHA, Ramachandran S. Vasan, MD, FAHA, Marc Ruel, MD, FAHA, Win-Kuang Shen, MD, FAHA, Andre Terzic, MD, FAHA and Scott A. Waldman, MD, PhD, FAHA Santhi K. GaneshSanthi K. Ganesh *This update was prepared, in part, by Drs Mary Engler, Santhi Ganesh, and Christopher O’Donnell in their personal capacity. The opinions expressed in this article are the authors’ own and do not reflect the view of the National Institutes of Health, the US Department of Health and Human Services, or the US government. The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention. Search for more papers by this author , Donna K. ArnettDonna K. Arnett Search for more papers by this author , Themistocles L. AssimesThemistocles L. Assimes Search for more papers by this author , Craig T. BassonCraig T. Basson Search for more papers by this author , Aravinda ChakravartiAravinda Chakravarti Search for more papers by this author , Patrick T. EllinorPatrick T. Ellinor Search for more papers by this author , Mary B. EnglerMary B. Engler *This update was prepared, in part, by Drs Mary Engler, Santhi Ganesh, and Christopher O’Donnell in their personal capacity. The opinions expressed in this article are the authors’ own and do not reflect the view of the National Institutes of Health, the US Department of Health and Human Services, or the US government. The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention. Search for more papers by this author , Elizabeth GoldmuntzElizabeth Goldmuntz Search for more papers by this author , David M. HerringtonDavid M. Herrington Search for more papers by this author , Ray E. HershbergerRay E. Hershberger Search for more papers by this author , Yuling HongYuling Hong Search for more papers by this author , Julie A. JohnsonJulie A. Johnson Search for more papers by this author , Steven J. KittnerSteven J. Kittner Search for more papers by this author , Deborah A. McDermottDeborah A. McDermott Search for more papers by this author , James F. MeschiaJames F. Meschia Search for more papers by this author , Luisa MestroniLuisa Mestroni Search for more papers by this author , Christopher J. O’DonnellChristopher J. O’Donnell *This update was prepared, in part, by Drs Mary Engler, Santhi Ganesh, and Christopher O’Donnell in their personal capacity. The opinions expressed in this article are the authors’ own and do not reflect the view of the National Institutes of Health, the US Department of Health and Human Services, or the US government. The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention. Search for more papers by this author , Bruce M. PsatyBruce M. Psaty Search for more papers by this author , Ramachandran S. VasanRamachandran S. Vasan Search for more papers by this author , Marc RuelMarc Ruel Search for more papers by this author , Win-Kuang ShenWin-Kuang Shen Search for more papers by this author , Andre TerzicAndre Terzic Search for more papers by this author and Scott A. WaldmanScott A. Waldman Search for more papers by this author Search for more papers by this author and on behalf of the American Heart Association Council on Functional Genomics and Translational Biology, Council on Epidemiology and Prevention, Council on Basic Cardiovascular Sciences, Council on Cardiovascular Disease in the Young, Council on Cardiovascular Surgery and Anesthesia, Council on Clinical Cardiology, Council on Cardiovascular and Stroke Nursing, and Stroke Council Originally published2 Dec 2013https://doi.org/10.1161/01.cir.0000437913.98912.1dCirculation. 2013;128:2813–2851is corrected byCorrectionOther version(s) of this articleYou are viewing the most recent version of this article. Previous versions: January 1, 2013: Previous Version 1 Cardiovascular diseases (CVDs) are a major source of morbidity and mortality worldwide. Despite a decline of ≈30% over the past decade, heart disease remains the leading killer of Americans.1 For rare and familial forms of CVD, we are increasingly recognizing single-gene mutations that impart relatively large effects on individual phenotype. Examples include inherited forms of cardiomyopathy, arrhythmias, and aortic diseases. However, the prevalence of monogenic disorders typically accounts for a small proportion of the total CVD observed in the population. CVDs in the general population are complex diseases, with several contributing genetic and environmental factors. Although recent progress in monogenic disorders has occurred, we have seen a period of intense investigation to identify the genetic architecture of more common forms of CVD and related traits.Genomics serves several roles in cardiovascular health and disease, including disease prediction, discovery of genetic loci influencing CVD, functional evaluation of these genetic loci to understand mechanisms, and identification of therapeutic targets. For single-gene CVDs, progress has led to several clinically useful diagnostic tests, extending our ability to inform the management of afflicted patients and their family members. However, there has been little progress in developing genetic testing for complex CVD because individual common variants have only a modest impact on risk. The study of the genomics of complex CVDs is further challenged by the influence of environmental variables, phenotypic heterogeneity, and pathogenic complexity. Characterization of the clinical phenotype requires consideration of the clinical details of the diseases and traits under study.This update expands the prior scientific statement on the relevance of genetics and genomics for the prevention and treatment of CVDs.2 In the earlier report, we focused on the current status of the field, which consisted of predominantly family-based linkage studies and single-gene or mendelian mutations of relatively large phenotypic effect sizes. The past several years have seen many advances in genetics technologies that have allowed a new generation of genetic association discoveries and an explosion of reports in the literature of these new associations. Approaches such as genome-wide association studies (GWASs) and whole-genome sequencing have led to the discovery of novel genomic determinants of CVDs, and we focus on these findings. Unlike candidate gene association studies, hypothesis-generating GWASs and whole-genome sequencing studies have the potential to identify genes with previously unknown (and even unsuspected) roles in CVD. These findings are in various stages of translation to both the clinical setting and basic functional research aimed at uncovering new mechanisms operative in CVDs.The basic methodology for GWASs is to analyze single-nucleotide polymorphisms (SNPs) across the genome in association with dichotomous (case-control) or continuous traits. SNPs are analyzed by genotyping arrays capable of interrogating millions of SNPs directly and predicting millions more genotypes via imputation. Multiple test corrections and replication of findings in independent samples are done to demonstrate that statistical findings are robust. The remarkable discoveries from GWASs have brought enormous progress to identifying and understanding a portion of the cause of complex CVD. The basis for the remaining portion of genomic cause is uncertain but has been hypothesized to reside in less common, rare, or very rare sequence variants.3–5 Further information on GWAS methodology can be found in several excellent reviews.6–8 In addition to genomic technologies reviewed in the initial statement2 and GWASs, methods such as massively parallel next-generation sequencing9,10 are greatly extending our knowledge of human genetic variation and genetic markers of CVD and related traits. The exome, defined as the ≈1% to 2% of the coding portion of the human genome, encompasses ≈19 000 genes. Exome or genome sequencing provides a novel and powerful research strategy for those conditions driven by uncommon or rare variants.11 In this context, uncommon usually means allele frequencies of 0.5% to 5%; rare, <0.5%; and very rare, <0.1%. “Private” is applied to variants not previously reported, which depends on the number of other DNA sequences available for comparison at that locus. Sponsored by the National Heart, Lung, and Blood Institute, the Exome Sequencing Project (ESP),12 undertaken in 2009, has now made available the exome sequences from >6500 individuals with cardiovascular, lung, or blood phenotypes (http://evs.gs.washington.edu/EVS/), greatly enhancing the power of determining how rare a variant may be. A recent ESP report of the initial 2440 exome sequences from European (n=1351) or African (n=1088) ancestry provided a remarkable snapshot of genetic variation.13 Most variants identified were rare and previously undetected. Of the >500 000 single-nucleotide variants identified, 86% had an allele frequency <0.5%, and 82% were previously unknown; of the average of 13 595 single-nucleotide variants observed in each individual, 2.3% were predicted to affect the protein function in ≈313 genes.13 Large-scale experiments in population samples are currently underway to test the hypothesis that rare variants, which are not polled by GWAS assays or imputation methods, may play a larger role in complex disease phenotypes than previously thought.In this update, we review the advances and current state of knowledge in the key CVD areas previously discussed, including coronary artery disease (CAD), stroke, hypertension, and hypercholesterolemia.2 Additionally, we review the areas of diabetes mellitus, inflammation, and pharmacogenetics. In the discussion of common variant associations, we highlight landmark studies and their findings. A comprehensive listing of associations may be found through resources such as the GWAS catalog.14 This update also presents new research findings for mendelian diseases and the clinical approach to these disorders, focusing on cardiomyopathies, inherited arrhythmias, aortic aneurysms, and congenital heart defects. Policy recommendations are addressed in a separate American Heart Association statement.15 We discuss the functional translation of genetic findings for CVDs and stroke, as well as counseling and clinical translation considerations. We conclude with a set of recommendations identifying future directions and the education that will be required of clinicians, researchers, the general public, and our patients for maximal effect for research and translation of these findings to the clinical setting.Updated State of Knowledge in Cardiovascular GeneticsCoronary Artery DiseaseThe familial clustering of CAD and its heritability are well established, with heritability of CAD estimated at ≈40%.16–18 This estimate includes heritability of CAD-associated risk factors. Recently, GWASs have helped define the genetic architecture of CAD.19–23 Indeed, combined with complementary research using transcriptomics and other components of the “omics” methodologies, many CAD loci have been identified in the past 5 years.GWASs initially heralded the investigation of the genetic underpinnings of myocardial infarction (MI) and other clinical CADs with the successful identification of the locus on chromosome 9p21.24–26 Importantly, the 9p21 locus has been associated with a variety of other vascular traits and diseases, including abdominal aortic and intracranial aneurysms,27,28 platelet reactivity, and stroke, as discussed in the next section. However, it rapidly became clear that large discovery study sample sizes and replication studies were needed to identify the rather modest effect sizes associated with additional genetic loci. In this context, the Coronary Artery Disease Genome Wide Replication and Meta-analysis (CARDIoGRAM) consortium compiled data from 14 GWASs of CAD and MI in individuals of European ancestry,22 with >22 000 cases and 64 000 controls, and identified 13 novel susceptibility loci for CAD.23 In parallel, the Coronary Artery Disease (C4D) Genetics Consortium identified an additional 4 novel loci by pooling together 2 cohorts of European ancestry with others of South Asian ancestry.20More recently, these 2 large consortia joined forces to form the CARDIoGRAMplusC4D consortium and executed a large-scale replication of the top findings from CARDIoGRAM using the Illumina Metabochip, which assays ≈200 000 SNP markers in genes of interest for metabolic and CVD traits.29,30 Metabochip data in 34 independent sample collections of European or South Asian descent comprising 41 513 cases and 65 919 controls were combined with existing CARDIoGRAM GWASs to uncover an additional 15 loci, bringing the total number of GWAS susceptibility loci for CAD to 46.30 A further 104 independent variants strongly associated with CAD at a 5% false discovery rate and, together with the 46 loci reaching genome-wide variants, explain ≈11% of the additive heritability of CAD. Together, CARDIoGRAM and C4D20,23,30 evaluated >200 000 individuals and have identified and replicated >30 novel loci for CAD susceptibility (Table 1).Table 1. Susceptibility Loci for CAD Uncovered Through GWASsChromosomal LocationGenes* (Name)Reference1p13.3SORT1 (sortilin 1)20, 23, 251p32.2PPAP2B (phosphatidic acid phosphatase type 2B)231p32.3PCSK9 (proprotein convertase subtilisin/kexin type 9)19, 20, 231q21.3IL6R (interleukin 6 receptor)301q41MIA3 (melanoma inhibitory activity family, member 3)20, 23, 252p11.2VAMP5(vesicle-associated membrane protein 5), VAMP8 (vesicle-associated membrane protein 8)302p21ABCG8 [ATP-binding cassette, subfamily G (WHITE), member 8]302p24.1WDR35 (WD protein repeats domain 35)312p24.1APOB (apolipoprotein B)302q22.3ZEB2 (zinc finger E-box binding homeobox 2)302q33.1WDR12 (WD protein repeats domain 12)19, 20, 233q22.3MRAS (muscle RAS oncogene homolog)20, 21, 234q31.22EDNRA (endothelin receptor type A)304q32.1GUCY1A3 (guanylate cyclase 1, soluble, alpha 3)30, 315q31.1SLC22A4 (solute carrier family 22, member 4)306p21.2KCNK5 (potassium channel, subfamily K, member 5)306p21.31ANKSIA (ankyrin repeat and sterile α motif domain containing 1A)236p21.32C6orf10 (chromosome 6 open reading frame 10)316p24.1PHACTR1 (phosphatase and actin regulator 1)19, 20, 23, 316p24.1C6orf105 (chromosome 6 open reading frame 105)†326q23.2TCF21 (transcription factor 21)23, 316q25.3LPA [lipoprotein(a)]23, 32, 336q26PLG (plasminogen)307p21.1HDAC9 (histone deacetylase 9)307q22BCAP29 (B-cell receptor–associated protein 29)20, 327q32.2ZC3HC1 (zinc finger, C3HC-type containing 1)238p21.3LPL (lipoprotein lipase)308q24.13TRIB1 (tribbles homolog 1)309p21CDKN2A/2B (cyclin-dependent kinase inhibitor 2A/2B)20, 23–26, 319q34.2ABO (ABO blood group)2310p11.23KIAA14622010q11.1CXCL12 (chemokine (C-X-C motif, ligand 12)20, 23, 2510q23LIPA (lipase A, lysosomal acid, cholesterol esterase)2010q24.3CYP17A1 (cytochrome P450, family 17, subfamily A, polypeptide 1)2311q22.3PDGFD (platelet-derived growth factor D)2011q23.3ZNF259 (zinc finger protein 259)2312q21.33ATP2B1 (ATPase, Ca2+ transporting, plasma membrane 1)3112q24.12SH2B3 (SH2B adaptor protein 3)20, 23, 31, 3413q12.3FLT1 (FMS-related tyrosine kinase 1)3013q34COL4A1 (collagen, type IV, α1)2314q32.2HHIPL1 (hedgehog interacting protein-like 1)2315q25.1ADAMTS7 (a disintegrin-like and metallopeptidase with thrombospondin type 1, motif 7)20, 2315q26.1FURIN (paired basic amino acid cleaving enzyme)3017p11.2SMCR3 (Smith-Magenis syndrome chromosome region, candidate 3)2317p13.3SMG6 (SMG6 nonsense mediated mRNA decay factor)2317q21.32UBE2Z (ubiquitin-conjugating enzyme E2Z)2319p13.2LDLR (low-density-lipoprotein receptor)19, 20, 2321q22.11MRPS6 (mitochondrial ribosomal protein S6)19, 23CAD indicates coronary artery disease; and GWAS, genome-wide association study.*Genes are provided for those genes near or within the locus of the primary association signal.†This locus was not replicated in a subsequent larger GWAS.These GWAS investigations were comprehensive in that they also assessed gene expression (for genes in the implicated loci) in tissues using human cell lines, evaluated genome-wide maps of allelic expression imbalance, compared “hits” for CAD with those for common CVD risk factors, and performed complex network analyses.20,23,30 The functional impact of the identified loci was interpreted on the basis of the association of CAD-associated alleles with traditional CAD risk factors, through which the genetic risk for CAD is likely imparted,35 Interestingly, in these analyses adjusted only for age and sex, the lead SNP in 12 of the 46 validated CAD susceptibility loci was found to be significantly associated with a lipid trait in the Global Lipids Genetics Consortium data set (predominantly low-density lipoprotein [LDL] and triglycerides) and 5 with blood pressure (BP) in the International Consortium of Blood Pressure data set, underscoring the strong causal influence of these risk factors on CAD risk.30 Network analysis also indicated a causal association between pathways of inflammation and CAD.30 The functional impact of these loci has also been pursued through the use of molecular methods in transgenic mouse models for the 9p21 association with CAD.35Finally, a recent modest-sized GWAS of ≈1500 Han Chinese subjects with CAD and 5019 control subjects followed by a large-scale replication of top findings in 15 460 cases and 11 472 controls replicated 4 loci uncovered in Europeans and implicated 4 novel loci.31 Two of the novel loci have also been implicated in hypertension in Europeans.31 This GWAS was unable to replicate an association between rs6903956 at 6p24.1 (c6orf105) and CAD uncovered by a previous smaller GWAS in a Chinese Han population despite adequate power to detect it.31,32These studies highlight several key points. First, many CAD loci have consistent effects across studies of various ethnicity and for both men and women,20,31 although data on African Americans and other non-European ancestry ethnic groups remain relatively limited. Second, associations were generally stronger for early-onset CAD compared with later-onset CAD, suggesting that selecting individuals with premature CAD phenotypes enriches for genetic factors and may yield additional valuable insights in future studies.23 Third, associations were stronger for hard end points such as MI and angiographically defined CAD, again emphasizing the importance of careful phenotyping for GWASs.23 Fourth, as noted in other GWASs, the effect sizes in these studies typically were modest (often with odds ratios <1.2) with a 5% to 20% excess risk associated with select loci (Table 1). Consequently, these genetic loci explain only a small proportion of the heritability of CAD. Presumably, additional common variants with very modest effects, gene-environment and gene-gene interactions, rare genetic variants (GWASs focused on SNPs with a minor allele frequency typically >5%), structural variation, epigenetic modifications, microRNA, altered splicing, and posttranslational modification of proteins also contribute to the residual unexplained heritability.3,23,36 These observations underscore the importance of a multifaceted investigation strategy harnessing the tools of systems biology to better elucidate the genetic basis of CAD. Fifth, a minority of the loci implicated in CAD were associated with traditional CVD risk factors, including lipids and BP, whereas others were associated with myriad other common diseases and traits, including hematologic and biochemical traits and intracranial aneurysm, thereby pointing to the pleiotropic effects of these loci.15,20,23 Sixth, whereas the list of loci implicated in CAD is long, the exact causal variants and mechanisms by which these candidate variants and candidate genes mediate their effects remain to be elucidated for most loci. Such additional work involves careful experiments like those accomplished for the SORT1 locus, for which there is some evidence for causal associations with LDL cholesterol (LDL-C) and CAD.37 Finally, the contribution of these newer loci to risk prediction and their clinical utility remain to be clearly demonstrated in future studies, although improvement in this respect has been noted in some cohorts.38–41Ischemic StrokeStroke is a heterogeneous and complex disease composed of 3 main types: ischemic stroke, intracerebral hemorrhage, and subarachnoid hemorrhage. Within the phenotype of ischemic stroke, various subtypes appear to differ in degree of heritability.42 Here, we focus on ischemic stroke and its subtypes.GWASs have provided the strongest evidence for loci associated with the common form of ischemic stroke (Table 2). To date, all such loci have had 2 characteristics. First, most of the loci have been discovered in other mechanistically related conditions such as atrial fibrillation, CAD, and coagulation, with subsequent confirmation of associations with ischemic stroke. This is not unexpected, given the pathogenic associations between stroke and these other traits and because the GWASs of these other conditions have far greater sample sizes and statistical power compared with the GWASs of ischemic stroke. Second, the associations are specific to subtypes of ischemic stroke. For example, a locus on chromosome 4q25 adjacent to the transcription factor PITX2 (codes for pituitary homeobox 2), found to be associated with atrial fibrillation, was subsequently found to be associated with cardioembolic stroke.52 Additionally, a locus on chromosome 16q22 involving ZFHX3 (the zinc finger homeobox protein 3) has been associated with both atrial fibrillation and cardioembolic stroke.53 Similarly, a pooled analysis demonstrated that 6 SNPs in the chromosome 9p21 locus, first identified through a heart disease GWAS,26 are associated with atherosclerotic stroke independently of demographic variables, CAD, MI, and other vascular risk factors.54 The 9p21 locus seems to affect stroke risk independently of the risk of MI. Genetic variants identified through GWASs of coagulation/fibrin phenotypes were subsequently examined for association with ischemic stroke. The rs505922 in the ABO gene was found to have a replicated association with large-vessel and cardioembolic stroke.28,55Table 2. Common Genes Associated With Hemorrhagic and Ischemic StrokeLocusGene Symbol*Gene Name*ReferencesFamilial cerebral amyloid angiopathies 21q21.3APPAmyloid β (A4) precursor protein43, 44 20p11.21CST3Cystatin C45 18q12.1TTRTransthyretin46Sporadic intracerebral hemorrhage 19q13.2APOEApolipoprotein E47Familial ischemic stroke or stroke-like syndromes MitochondrialtRNA leucine48 19p13.2-p13.1NOTCH3Notch349 10q26.3HTRA1HtrA serine peptidase 150Sporadic cardioembolic ischemic stroke 4q25PITX2Paired-like homeodomain 251 16q22.3ZFHX3Zinc finger homeobox 351Sporadic large-vessel atherosclerotic stroke 9p21TBDTBD51 7p21.1HDAC9Histone deacetylase 951TBD indicates to be determined.*Genes are provided for those genes near or within the locus of the primary association signal.GWASs have yet to identify a locus that imparts risk by modifying tissue sensitivity to ischemic injury. A genetic variant that imparts sensitivity of the brain parenchyma to focal ischemia might be expected to have an association across ischemic stroke subtypes while lacking an association with conventional atherosclerotic risk factors for stroke such as atrial fibrillation and hypertension (parenchymal hypothesis).More recently, GWASs of ischemic stroke have identified novel loci that were not previously identified through GWASs of other CVDs. Variation in HDAC9 (codes for histone deacetylase 9) has been associated with large-vessel atherosclerotic stroke.51 In addition to confirming ischemic stroke loci initially identified in nonstroke GWASs (PITX2, ZFHX3, and the 9p21 locus), a meta-analysis that involved 12 389 cases of ischemic stroke confirmed the HDAC9 locus.56HDAC9 variants seem to relate to the predisposition to atherosclerosis.57 As with HDAC9, a novel locus associated with large-vessel ischemic stroke was first discovered in a stroke GWAS.58Mitochondrial DNA has been shown to harbor variants associated with stroke risk. Mitochondrial DNA is another important part of the human genome, separate from nuclear chromosomal DNA. Mitochondrial DNA is inherited directly from mother to child because of the presence of mitochondria in the oocyte, which gives rise to all mitochondria in the developing tissues. Because of the association of mitochondrial DNA variants with the MELAS (mitochondrial encephalopathy, lactic acidosis, and stroke-like episodes) syndrome, a monogenic cause of ischemic stroke,59 the role of mitochondrial variants in the common form of stroke has been investigated. A mitochondrial genetic risk profiling study that included all common variants has also shown that common variants are associated with ischemic stroke and with white matter hyperintensity volume.60 Consistent with the GWASs for CAD and MI, the effect sizes of the associated mitochondrial variants for stroke are modest, impeding our ability to discern which particular loci are most useful for profiling in the clinical setting.60Family studies continue to provide new insights into monogenetic causes of cerebrovascular disease and novel disease mechanisms and pathways. Variants in ACTA2, the vascular smooth muscle cell–specific isoform of α-actin, smooth muscle myosin gene (MYH11),61–63 and multiple transforming growth factor-β signaling genes, including TGFBR1,64–67TGFBR2,68-70TGFB2,71,72 and SMAD3,73,74 have been shown to cause familial aortic aneurysms and dissection and to have cerebrovascular manifestations.64–70,75 More recently and similar to the pleiotropic effects of the 9p21 locus, some of the same variants associated with premature CAD are also associated with premature ischemic stroke, including moyamoya disease, which is also associated with other variants such as RNF213.76,77 Laboratory studies demonstrate that these mutations are associated with increased proliferation of smooth muscle cells.78 Although these mutations were found not to be associated with sporadic premature CAD and stroke, these findings suggest a new pathway to target in future genetic research.Global gene expression profiling has recently been studied as a potential tool for distinguishing ischemic stroke from other conditions79 and for identifying the underlying mechanism of ischemic stroke.80 Gene expression profiles from RNA isolated from sequential early blood samples were able to distinguish large-artery atherosclerotic stroke from cardioembolic stroke and to distinguish cardioembolic stroke related to atrial fibrillation from non–atrial fibrillation cardioembolic causes. If validated in large independent studies, this approach would be useful for guiding the evaluation of stroke of undetermined origin, currently estimated at 35% to 42% of all ischemic strokes.81Coronary Calcium, Carotid Intima-Media Thickness, and Carotid PlaqueThere has also been strong interest in targeting subclinical vascular disease phenotypes as additional traits for GWAS investigations. The Cohorts for Heart and Aging Research in Genomic Epidemiology Consortium82 conducted a meta-analysis of GWASs for coronary artery calcium measured by thoracic computed tomography in 9961 men and women from 5 community-based cohorts and replicated the most promising candidates in 3 independent cohorts (n=6032).83 Several of the loci significantly associated with coronary artery calcium had previously been shown to be associated with MI. These included loci in the 9p21 region (CDKN2B) and SNPs associated with PHACTR1 and COL4A2. Interestingly, other established MI loci showed considerable statistical evidence (although they failed to meet GWAS significance) for association with subclinical coronary disease, including the SORT1/CELSR2/PSRC1 locus, MRAS, CXCL12, the COL4A1/COL4A2 locus, and ADAMST7. The convergence of GWAS evidence for both clinical events and anatomic coronary disease provides even more compelling validation of many of the genetic variants and strongly suggests that their apparent associations with MI are likely attributable to influences on the pathogenesis of atherosclerosis rather than other mechanisms that contribute to the development of acute coronary syndromes.In a much larger GWAS meta-analysis involving 31 211 subjects in discovery and 11 273 subjects in replication cohorts, the Cohorts for Heart and Aging Research in Genomic Epidemiology consortium also sought genetic variants ass