Association of Genome-Wide Variation With the Risk of Incident Heart Failure in Adults of European and African Ancestry
Although genetic factors contribute to the onset of heart failure (HF), no large-scale genome-wide investigation of HF risk has been published to date. We have investigated the association of 2 478 304 single-nucleotide polymorphisms with incident HF by meta-analyzing data from 4 community-based prospective cohorts: the Atherosclerosis Risk in Communities Study, the Cardiovascular Health Study, the Framingham Heart Study, and the Rotterdam Study.
Methods and Results—
Eligible participants for these analyses were of European or African ancestry and free of clinical HF at baseline. Each study independently conducted genome-wide scans and imputed data to the ≈2.5 million single-nucleotide polymorphisms in HapMap. Within each study, Cox proportional hazards regression models provided age- and sex-adjusted estimates of the association between each variant and time to incident HF. Fixed-effect meta-analyses combined results for each single-nucleotide polymorphism from the 4 cohorts to produce an overall association estimate and P value. A genome-wide significance P value threshold was set a priori at 5.0×10−7. During a mean follow-up of 11.5 years, 2526 incident HF events (12%) occurred in 20 926 European-ancestry participants. The meta-analysis identified a genome-wide significant locus at chromosomal position 15q22 (1.4×10−8), which was 58.8 kb from USP3. Among 2895 African-ancestry participants, 466 incident HF events (16%) occurred during a mean follow-up of 13.7 years. One genome-wide significant locus was identified at 12q14 (6.7×10−8), which was 6.3 kb from LRIG3.
We identified 2 loci that were associated with incident HF and exceeded genome-wide significance. The findings merit replication in other community-based settings of incident HF.
Heart failure (HF) is a serious cardiovascular condition that affects 1 in 5 individuals during their lifetime.1 Failure arises primarily from the burden of systemic hypertension, coronary ischemia, or cardiac valvular disease.2,3 Regardless of the clinical mechanism responsible for reduced cardiac output in HF, long-term strain triggers maladaptive cellular function and growth, possibly resulting in irreversible myocardial injury.4,5
Editorial on p 226
Clinical Perspective on p 266
Genetic factors contribute to HF onset, and an estimated ≈18% of the risk of HF may be attributable to parental HF.6 To date, approaches to identify genetic variants associated with HF risk have relied on candidate genes, primarily those of the renin-angiotensin-aldosterone system, adrenergic receptors, sarcomeres, and cytoskeletal proteins.7,8 No large-scale genome-wide investigation of HF risk has been published.9 This report presents findings from an investigation of ≈2.5 million single-nucleotide polymorphisms (SNPs) and their associations with incident HF among 23 821 adults of European or African ancestry.
The setting for this meta-analysis is the Cohorts for Heart and Aging Research in Genomic Epidemiology (CHARGE) Consortium and includes data from 4 prospective, population-based cohorts of adults in the USA and the Netherlands: the Atherosclerosis Risk in Communities (ARIC) Study, the Cardiovascular Health Study (CHS), the Framingham Heart Study (FHS), and the Rotterdam Study (RS).10 The designs of these studies have been described elsewhere.11–16 In brief, ARIC recruited 15 792 participants 45 to 64 years of age from 1987 to 1989 from 4 US communities. The CHS recruited participants 65 years of age and older from 4 US communities in 2 waves: 5201 participants in 1989–1990 and an additional 687 African Americans in 1992–1993. The FHS recruited 2 generations of participants 28 to 72 years of age in 2 time periods: 5209 Original Cohort participants were recruited in 1948 from Framingham, Massachusetts, and 5124 Offspring Cohort participants were recruited in 1971–1975. The RS recruited 7983 participants 55 years of age and older from 1990 to 1993 from Rotterdam, the Netherlands. The ARIC and CHS studies included African American participants, 27% and 15%, respectively, whereas FHS and RS included almost exclusively participants of European ancestry. Each study independently conducted genome-wide scans of study participants, and within-study analyses were conducted according to a prespecified plan. The results were combined for this prospective meta-analysis.
Eligible participants for these analyses were of European or African ancestry and free of clinical HF. Details about cohort-specific methods for ascertaining prevalent HF at baseline have been published elsewhere.6,17–20 Race was self-reported. Each study received an institutional review board approval, and all participants provided written informed consent for the use of their DNA for research.
For this analysis, baseline measures of clinical and demographic characteristics were obtained at the time of cohort entry for ARIC, CHS, and RS, and at the time of DNA collection for FHS. Measures were taken using standardized methods as specified by each study and included the following: in-person measures of height, weight, systolic and diastolic blood pressure, and total cholesterol; self-reported history and treatment of hypertension and diabetes; and self-reported history of prevalent coronary artery disease (including a history of myocardial infarction [MI], angina, or coronary revascularization) and cerebrovascular disease (stroke or transient ischemic attack). For CHS, FHS, and RS, prevalent cardiovascular disease was validated by medical record review.
Follow-Up and HF Diagnosis
Potential incident HF events were identified during follow-up by annual or semiannual self-report, from administrative data, or from periodic in-study examinations. In brief, ARIC relied on International Classification of Diseases (ICD)-9 codes collected from hospital discharge summaries and death certificates.20,21 For CHS, FHS, and RS, HF events identified by self-report or administrative data were validated by physician's review of medical records. CHS and Framingham applied their published criteria,22,23 and RS applied the European Society of Cardiology criteria.18,24 Details can be found in supplemental Table I and in previously published reports.
Genotyping and Imputation
Genotyping used DNA collected from phlebotomy. Genome-wide assays of SNPs were conducted independently in each cohort using various technologies: Affymetrix 6.0 for ARIC, Illumina 370CNV for CHS, Affymetrix 500K and 50K for FHS, and Illumina 550K v3 for RS. Genotype quality control and data cleaning were conducted independently by each study and are summarized in supplemental Table II. Methods were very similar across studies and reflect those used in other published reports on genome-wide associations.25,26
We investigated genetic variation in the 22 autosomal chromosomes. For European-ancestry participants, each study independently imputed their genotype data to the ≈2.5 million SNPs identified in HapMap Centre d'Etude du Polymorphisme Humain from Utah (CEU) samples.27,28 For African-ancestry participants, CHS imputed to combined CEU and Yoruban (YRI) samples and ARIC did not impute. Supplemental Table II provides details about the imputation methods for each cohort and the quality control criteria applied to the imputed genotypes. Imputation results were summarized as an “allele dosage” defined as the expected number of copies of the minor allele at that SNP (a continuous value between 0 and 2) for each genotype. Each cohort calculated a ratio of observed versus expected variance of the dosage statistic for each SNP. This value, which generally ranges from 0 to 1 (poor to excellent), is a measure of imputation quality. High-signal SNPs that had low observed versus expected variance values in CHS (<0.8) were directly measured to improve the quality of the signal. No other cohort conducted additional genotyping.
Investigators from all cohorts jointly developed the prespecified analytic plan. Within each study, Cox proportional hazards regression models were used to test the association between each SNP and time to incident HF, while adjusting for sex and baseline age. Time-to-event models are standard for longitudinal study designs. CHS adjusted for study site and FHS adjusted for generation, and for ancestry using principal components.29 Time of entry to the analysis was the date of cohort entry (ARIC, CHS, RS) or DNA collection (FHS), and failure time was the time of HF diagnosis. Participants without incident HF were censored at the time of death, last date of contact, or at the end of follow-up, whichever came first. For each SNP, additive genetic models were used to estimate the regression coefficient for the hazard ratio (HR) for allele dosage and its respective standard error. In ARIC and CHS, analyses were conducted separately in 2 ancestry groups, European and African. For each analysis, a genomic control coefficient (λ) was calculated, which estimated the extent of underlying population structure.
Meta-analyses were performed using MetABEL software (http://mga.bionet.nsc.ru/≈yurii/ABEL/).30 For the populations of European ancestry, fixed-effect meta-analyses combined regression β coefficients and λ-adjusted standard errors across the 4 cohorts for each SNP to produce an overall β coefficient, standard error, and P value. We found that asymptotic assumptions of survival analysis models were not accurate when the number of cases carrying the variant allele was small. Hence, we excluded SNPs for which the postmeta-analysis population-size-weighted minor allele frequency (MAF) was <0.015.
We created a quantile-quantile (Q-Q) plot of the distribution of the observed and expected P values associated with the successful SNPs, along with a plot of all the P values according to location within each of the 22 chromosomes. The a priori threshold of genome-wide significance was set at 5.0×10−7, also used by the Wellcome Trust Case Control Consortium.31 For 2.5 million tests, this threshold limits the expected number of genome-wide false-positives to approximately 1. We also reported high-signal SNPs that were associated with P values <1.0×10−5, the point where observations deviate from those expected on the Q-Q plot. When more than 1 SNP clustered at a locus, we picked the SNP with the smallest P value as the locus marker. Forest plots that included cohort-specific findings were created for genome-wide significant SNPs.
In sensitivity analyses, we subjected our top associations to time-to-event analyses that excluded participants with a baseline history of clinical MI and censored those with an incident MI during follow-up. This approach allowed us to estimate the association of selected SNPs and HF onset among participants whose HF was less likely to be ischemic in origin.
For participants of African ancestry, a similar analytic approach was undertaken in ARIC and CHS. The ARIC cohort did not have imputed data on those of African ancestry, so testing was restricted to successfully genotyped SNPs on their Affymetrix 6.0 chip that had an average weighted MAF of at least 0.075 in meta-analyses. Genome-wide significance was set at 5.0×10−7. High-signal SNPs from the meta-analysis in the European- and African-ancestry populations were compared.
A total of 20 926 eligible participants of European ancestry and 2895 participants of African ancestry were free of HF at cohort entry and had genome-wide data that met cohort-specific quality control measures. Characteristics of the 23 821 cohort participants included in these analyses are provided in Table 1. The average age ranged from 53.3 to 72.6 for the 4 cohorts, and the majority of the participants were women (57%). Other demographic and clinical characteristics were similar across the 4 cohorts. A total of 2526 incident HF events (12.1%) were identified over a weighted average of 11.5 years of follow-up among those of European ancestry and 466 events (16.1%) over a weighted average of 13.7 years among those of African ancestry. The average age at the time of HF onset was 73.6 years (range, 44 to 101 years) and 52% of the 2992 events occurred among women. Among European- and African-ancestry participants, 73% and 78%, respectively, of the events were diagnosed in the absence of a prior MI.
|Mean age, y||54.2||53.3||72.3||72.6||64.6||69.1|
|Age range, y||45 to 64||45 to 64||65 to 98||65 to 92||30 to 100||55 to 99|
|Systolic blood pressure, mm Hg||118.3||128.7||135.2||142.5||130.7||139.3|
|Diastolic blood pressure, mm Hg||71.7||80.0||70.4||76.9||74.1||73.8|
|Total cholesterol, mg/dL||214.8||215.9||212.9||209.4||202.4||255.2|
|Treated hypertension, %||25.3||54.4||30.2||51.7||35.6||23.2|
|Treated diabetes, %||8.0*||19.0*||5.4||11.2||6.5||3.8|
|Prevalent CAD, %||8.8||6.9||0||0||12.1||13.8|
|Myocardial infarction, %||3.6||3.1||0||0||4.9||10.3|
|Coronary revascularization, %||3.1||1.8||0||0||4.9||2.5|
|Prevalent CBD, %||2.1||2.9||0||0||5.0||4.1|
|Current smoker, %||24.9||29.9||11.3||16.6||14.2||22.8|
|Mean follow-up, y||15.3||14.6||11.8||10.1||5.5||10.8|
|Incident HF, n (%)||691 (9.0)||331 (14.3)||838 (25.4)||135 (23.5)||249 (5.8)||748 (13.1)|
Participants of European Ancestry
A total of 2 478 304 SNPs were investigated across the 4 cohorts. The λ coefficients for analyses of European participants in each of the 4 cohorts were small (≤1.03) and suggested negligible test statistic inflation. Supplemental Figure IA presents the Q-Q plot of meta-analyzed P values derived from the 2 478 304 tests performed plotted against the expected distribution. Overall, evidence for population admixture was small (λ=1.02). At the lowest P values, the observed exceeded the expected.
Figure 1A presents all 2 478 304 P values organized by chromosome and genomic position. Among these SNPs, 5 exceeded the genome-wide significance threshold (5.0× 10−7) and marked 2 loci, 1 on chromosome 13 (13q22) and another on chromosome 15 (15q22) (see supplemental Table III). Top SNPs at both loci were poorly imputed in CHS, so each was genotyped. After the inclusion of newly measured SNPs in CHS, only the 15q22 locus retained genome-wide significance. An additional 12 loci on 9 chromosomes were marked by 28 SNPs with P values <1.0×10−5 (see supplemental Table III). Top SNPs for 7 loci were poorly imputed in CHS, so these were genotyped. Among the 7 SNPs, 3 retained a P value <1.0×10−5 after inclusion of newly measured SNPs. Table 2 lists the top SNP for each of the 10 loci, after genotyping in CHS.
|POP||rs No.||CHR||Region||Position||Variant||MAF||P||HR (CI)*||Closest Gene|
|Genome-wide significant findings (<5.0×10−7)|
|EA||rs10519210||15||15q22||61524978||T→G||0.032||1.4×10−8||1.53 (1.05 to 2.24)||58.8 kb from USP3|
|AA||rs11172782||12||12q14||57545895||A→G||0.290||6.7×10−8||1.46 (1.03 to 2.09)||6.3 kb from LRIG3|
|Other high-signal findings (<1.0×10−5)|
|EA||rs11118620||1||1q41||219095131||T→C||0.292||7.3×10−6||1.15 (0.98 to 1.35)||0.0 kb from LOC100129376(intron)|
|EA||rs1523288||3||3q26.1||166762903||T→C||0.354||6.1×10−6||0.87 (0.74 to 1.02)||210.5 kb from BCHE|
|EA||rs13225783||7||7p15||27296581||C→T||0.048||6.8×10−6||1.38 (0.96 to 1.99)||43.9 kb from EVX1|
|EA||rs6473383||8||8q21.1||83831675||G→A||0.145||3.1×10−6||1.19 (0.99 to 1.44)||918.3 kb from SNX16|
|EA||rs10812610||9||9p21||27523984||A→C||0.496||5.1×10−6||1.14 (0.99 to 1.31)||4.1 kb from MOBKL2B|
|EA||rs11203032||10||10q23||90954594||T→G||0.103||8.3×10−6||1.22 (0.97 to 1.53)||1.1 kb from CH25H|
|EA||rs1520832||12||12q12||41539682||G→T||0.036||1.2×10−6||1.39 (0.99 to 1.95)||269.9 kb from PRICKLE1|
|EA||rs548097||13||13q22||74674269||T→G||0.018||6.4×10−7||1.62 (1.00 to 2.63)||82.5 kb from TBC1D4|
|EA||rs11880198||19||19p13.3||3110769||A→G||0.134||5.6×10−6||1.23 (0.98 to 1.54)||0.0 kb from GNA15 (intron)|
|AA||rs13418717||2||2q20||127379367||T→A||0.202||3.4×10−6||1.46 (0.97 to 2.20)||4.8 kb from LOC339760|
|AA||rs2210327||9||9p22||18099235||T→A||0.185||6.8×10−7||3.14 (0.99 to 10.0)||282.0 kb from SH3GL2|
|AA||rs563519||11||11q24||125533927||C→T||0.344||3.3×10−6||0.69 (0.46 to 1.03)||43.6 kb from RPUSD4|
|AA||rs2046383||12||12p11.2||29995409||T→G||0.302||3.3×10−6||1.39 (0.97 to 1.97)||166.5 kb from TMTC1|
|AA||rs17019682||12||12q22||90829401||A→T||0.165||2.0×10−6||1.47 (0.98 to 2.22)||231.6 kb from BTG1|
Genome-Wide Significant Locus in European-Ancestry Participants
Chromosomal position 15q22 was marked by 1 SNP, rs10519210, which had a MAF of 0.032 and was associated with a P value of 1.4×10−8. The HR for this SNP was 1.53 per G allele. This SNP is located between 2 genes: 58.8 kb from USP3 (ubiquitin-specific peptidase 3) and 63.9 kb from CA12 (carbonic anhydrase XII). Figure 2A presents detailed information about the linkage disequilibrium (LD) and P values of regional markers for this locus. This SNP was genotyped in 1 population (ARIC) and imputed in the other 3 populations. The cohort-specific estimates are provided in Table 3 and Figure 3A. Risk estimates were generally consistent across cohorts.
|Genome-wide significant findings (<5.0×10−7)|
|Other high-signal findings (<1.0×10−5)|
There were 19 195 European-ancestry participants at the risk of HF who did not have a history of clinical MI at baseline. Among these participants, 1854 had an incident HF event without a prior incident MI during follow-up. For locus 15q22 (rs10519210), the P value and the HR were diminished slightly but similar to the findings for all HF events: P=2.8×10−6 and HR=1.51 (see supplemental Table IV).
Loci With a P Value <1.0×10−5 in European-Ancestry Participants
Nine additional loci on chromosomes 1, 3, 7, 8, 9, 10, 12, 13, and 19 were associated with SNPs that had P values <1.0×10−5 (Table 2 and supplemental Table III). Among the 9 loci, 2 were intronic: rs11118620 (MAF=0.292) in LOC100129376 and rs11880198 (MAF=0.134) in GNA15 (guanine nucleotide binding protein [G protein], alpha 15 [Gq class]). Four loci were close to known genes: rs13225783 (MAF=0.048) within 43.9 kb of EVX1 (even-skipped homeobox 1); rs10812610 (MAF=0.496) within 4.1 kb of MOBKL2B (MOB1, Mps One Binder kinase activator-like 2B) and 4.1 kb of IFNK (interferon, kappa); rs11203032 (MAF=0.103) within 1.1 kb of CH25H (cholesterol 25-hydroxylase) and 8.7 kb of LIPA (lipase A, lysosomal acid, cholesterol esterase); and rs548097 (MAF=0.018) within 82.5 kb of TBC1D4 (TBC1 domain family, member 4). The remaining 3 loci were not within 100 kb of known genes, and the nearest genes were BCHE (butyrylcholinesterase) for rs1523288, SNX16 (sorting nexin 16) for rs6473383, and PRICKLE1 (prickle homolog 1) for rs1520832. Supplemental Figure IIA through III present detailed information about the LD and P values of regional markers for these 9 loci. Cohort-specific details on these loci are presented in Table 3, and sensitivity analyses restricted to those without a prior MI are presented in supplemental Table IV.
Participants of African Ancestry
A total of 692 330 SNPs were investigated across the 2 cohorts that included participants of African ancestry. The λ coefficients were small (<1.02) and suggested negligible population structure. Supplemental Figure IB presents the Q-Q plot from the 692 330 tests performed and Figure 1B presents all P values organized by chromosome and position. No SNP had a P value that exceeded the genome-wide significance threshold (5.0×10−7). Seven loci on 4 chromosomes were marked by 10 SNPs with P values <1.0×10−5 (see supplemental Table III). Five loci were marked by poorly imputed SNPs in CHS and 2 of these were genotyped. Among the 2 loci, 1 reached genome-wide significance (12q14) and the other did not retain a P value <1.0×10−5. Table 2 lists the top SNP for the 6 loci with a P value <1.0×10−5.
Genome-Wide Significant Locus in African-Ancestry Participants
Chromosomal position 12q14 was marked by 1 SNP, rs11172782, which had a MAF of 0.29 and was associated with a P value of 6.7×10−8. The HR for this SNP was 1.46 per G allele. This SNP is located 6.3 kb from LRIG3 (leucine-rich repeats and immunoglobulin-like domains 3). Figure 2B presents detailed information about the LD and P values of regional markers for this locus. The cohort-specific estimates are provided in Table 3 and Figure 3B. Risk estimates were consistent across both cohorts.
There were 2631 participants of African ancestry at the risk of HF who did not have a history of clinical MI at baseline. Among these participants, 372 had an incident HF event without a prior incident MI during follow-up. For locus 12q14 (rs11172782), the P value and the HR were similar to the findings for all HF events: P value=6.5×10−8 and HR=1.53 (see supplemental Table IV).
Loci With a P Value <1.0×10−5 in African-Ancestry Participants
Five loci on chromosomes 2, 9, 11, and 12 were associated with SNPs that had P values <1.0×10−5 (Table 2). Two loci were close to hypothesized or known genes: rs13418717 (MAF=0.202) within 4.8 kb of LOC339760 and rs563519 (MAF=0.344) within 43.6 kb of RPUSD4 (RNA pseudouridylate synthase domain containing 4). The remaining 3 loci were not within 100 kb of known genes and the nearest genes were SH3GL2 (SH3-domain GRB2 [growth factor receptor-bound protein 2]-like 2) for rs2210327, TMTC1 (transmembrane and tetratricopeptide repeat containing 1) for rs2046383, and BTG1 (B-cell translocation gene 1, antiproliferative) for rs17019682. Supplemental Figure IIJ through IIN presents detailed information about the LD and P values of regional markers for the 5 loci. Cohort-specific details on these loci are presented in Table 3, and sensitivity analyses restricted to those without a prior MI are presented in supplemental Table IV.
The genome-wide significant SNP in European-ancestry participants, rs10519210, was not associated with HF in those of African ancestry: 7.6×10−1 (HR=1.06; MAF=0.179). The genome-wide significant SNP in African-ancestry participants, rs11172782, was uncommon and not associated with HF in those of European ancestry: 5.6×10−1 (HR=0.91; MAF=0.018).
From among the 9 high-signal SNPs identified in European-ancestry participants, none was genotyped in ARIC; therefore, none was included in African-ancestry genome-wide analyses. When these 9 SNPs were examined in CHS participants of African ancestry, 1 of the 9 SNPs was associated with a P value <0.05: rs10812610 had a P value of 9.1×10−3 (HR=1.52; MAF=0.766 [C allele was not a minor allele in these participants]). When meta-analyzed with the European-ancestry findings, the resulting combined-ancestry P value for this locus was 7.5×10−7. From among the 5 high-signal SNPs in African-ancestry participants, none was associated with HF risk in European-ancestry participants (all P values >5.0×10−2).
Among 20 926 European-ancestry participants with 2526 incident HF events, we identified 1 locus with a SNP whose P value (1.4×10−8) exceeded the genome-wide statistical significance threshold of 5×10−7 and was associated with a 53% increase in HF risk. Among 2895 African-ancestry participants who had 466 incident events, we identified 1 locus with a SNP whose P value (6.7×10−8) exceeded the genome-wide statistical significance and was associated with a 46% increase in HF risk. For both loci, risk estimates were similar for those subjects who HF was not attributable to MI. We identified an additional 14 loci in European-ancestry and African-ancestry participants that were marked by high-signal SNPs, P value <1.0×10−5, but that did not reach genome-wide significance. For most loci, risk estimates were modest and did not seem to differ in a subset of participants without an MI preceding HF onset.
Genes Associated With Genome-Wide Significant and High-Signal Findings
Our results suggest that there may be several genomic regions associated with new-onset HF in older adults, and support the hypothesis that common genetic variation, regardless of the clinical mechanism responsible for reduced cardiac output in HF, contributes to risk. The 2 loci with marker SNPs that reached genome-wide significance were within 60 kb of known genes: USP3 and CA12 in European-ancestry participants and LRIG3 in African-ancestry participants. The product of the USP3 gene is an ubiquitin-specific protease (USP). Ubiquitin is a highly conserved 76-amino acid protein involved in the regulation of intracellular protein breakdown, cell-cycle regulation, and stress response. It is released from degraded proteins by the disassembly of polyubiquitin chains, which is mediated by USPs.32 The USP3 gene may also have a role in genome stability.33 Several papers have recently been published that describe the role of ubiquitin proteasome dysfunction in human hypertrophic and dilated cardiomyopathies and ubiquitin myocardial expression in pathogenic cardiac hypertrophy.34,35 The gene product from CA12 is a membrane protein highly expressed in many tissue beds and may have a role in cancer prognosis.36 The LRIG3 gene is a member of the LRIG family of genes that is widely expressed and impacts tissue development and tumor progression.37–39
Among the 14 high-signal SNPs, 2 were intronic, 1 of which was located in a known gene, GNA15, and the other in a hypothetical gene LOC100129376. The GNA15 gene, marked by rs11880198, is a guanine nucleotide binding protein of the Gq class, found in some G-protein-coupled receptors. This family of receptors include adrenergic, endothelin, and angiotensin II receptors, which regulate cardiac output, arterial pressure, and blood volume.40,41 Three important amino acid substitutions (1 nonsense and 2 missense) and 1 frameshift variant have been identified in the GNA15 gene. None of these variants were included in the imputed SNP panel, and LD with the HF SNP in our data are not known. Two additional high-signal SNPs were within 10 kb of known genes, which included MOBKL2B and IFNK, and CH25H and LIPA. None seems to have known functional qualities related to cardiac processes.42,43 Three additional high-signal SNPs were within 100 kb of known genes including EVX1, TBC1D4, and RPUSD4. Several papers have been published on the protein product of the TBC1D4 gene, known as AS160, and its role in handling insulin in skeletal muscles.44–46 The remaining 6 SNPs were not within 100 kb of genes.
Although we have described genes nearest to high-signal markers as potential culprits in the observed associations, it is possible that the true casual variant, if one exists, lies in neighboring genes or within the non-protein–coding transcripts in intergenic regions.45 In this article, we have identified the genes nearest to the high-signal marker as good candidates, but identifying the true causal genes will require further scientific investigation.
The consistency of findings across the European-ancestry and African-ancestry populations was modest. For the 2 genome-wide significant findings, the MAFs differed greatly between African-ancestry and European-ancestry participants. These high-signal markers are probably not themselves causal and are likely to be in linkage disequilibrium with the causal variants. In view of the large differences in MAFs and the different patterns of LD between the populations, it is perhaps not surprising that the associations are not consistent across ancestral groups. Indeed, the optimal replication sample for the findings in those of European ancestry would be other similar populations of European ancestry.
Strengths and Limitations
To our knowledge, this is the first large-scale attempt to discover genetic risk variants for incident HF, and included more than 23 000 individuals and ≈2.5 million markers spread throughout the genome. Data came from 4 large, population-based cohort studies that included cardiovascular outcomes as primary study endpoints. Over 2500 incident HF events among European-ancestry participants were included in our analyses, which allowed 80% power (2-sided α=5×10−7) to detect HRs as small as 1.6, 1.4, and 1.3 for variants with a MAF of 0.05, 0.1, and 0.2, respectively. For African-ancestry participants, minimal detectable HRs were larger: 2.6, 2.2, and 1.9, respectively. Statistical power for African-ancestry analyses was limited. We found 1 SNP of genome-wide significance for each ancestry population, on par with what one would expect by chance alone. Replication of associations between these SNPs and incident HF will clarify what role, if any, these SNPs have in the genesis of clinical HF. Each cohort relied on slightly different criteria to classify HF events, and it was not feasible to reclassify all 2992 events using a standard protocol. The between-study variation in HF classification likely introduced some heterogeneity that may have diminished our power to detect associations. Furthermore, we did not have comparable and complete data across cohorts on the type of failure. Previously published data from CHS and FHS indicated that half of the patients with new-onset HF had cardiac systolic dysfunction at the time of clinical onset.48,49 Not all 2 478 304 variants tested in patients of European ancestry were directly genotyped, and the imputation quality varied across SNPs. For poorly imputed SNPs, there was reduced statistical power to detect an association. We excluded SNPs with low MAF from meta-analyses because asymptotic assumptions of survival analysis models may not be accurate if the number of cases with the variant allele is low. For European-ancestry participants, the MAF threshold of 0.015 translates to ≈75 participants with incident HF who are carriers of 1 or more variant alleles. This relatively large number of events is not expected to violate asymptotic assumptions, and P values are likely accurate. Most of the across-ancestry comparisons were limited to data from CHS, where imputation quality for African-ancestry participants was generally poor, because the Illumina 370CNV chip was not designed to optimize coverage in this ancestry.
Using data from 4 prospective cohort studies that included 23 821 participants and prospectively identified 2992 incident HF events in older adults, we identified 2 SNPs that reached genome-wide significance, 1 in European-ancestry and 1 in African-ancestry participants. An additional 14 subthreshold but high-signal loci were identified and included intronic markers in attractive candidate genes. Our findings merit replication in other community-based settings of incident HF.
We acknowledge the essential role of the CHARGE Consortium in developing this article and for the support. CHARGE members include National Heart, Lung, and Blood Institute's ARIC Study, NIA's Iceland Age, Gene/Environment Susceptibility Study, National Heart, Lung, and Blood Institute's Cardiovascular Health Study and Framingham Heart Study, and the Netherland's Rotterdam Study. The authors also acknowledge the thousands of study participants who volunteered their time to help advance science and the scores of research staff and scientists who have made this research possible. The Rotterdam Study authors thank Mila Jhamai, Pascal Arp, Dr Michael J. Moorhouse, Marijn Verkerk, and Sander Bervoets for their help in creating the database and Maxim Struchalin for his contributions to the imputations of the data.
Sources of Funding
The ARIC Study is carried out as a collaborative study supported by
The CHS is supported by contract numbers
The FHS was supported by
The RS is supported by the
The funding sources had no role in the study design, analyses, or drafting of the article. The NHLBI reviews all articles submitted for publication but it was not involved in the decision to publish. Drs NL Smith, JF Felix, AC Morrison, and S Demissie had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Lloyd-Jones DM, Larson MG, Leip EP, Beiser A, D'Agostino RB, Kannel WB, Murabito JM, Vasan RS, Benjamin EJ, Levy D. Lifetime risk for developing congestive heart failure: the Framingham Heart Study. Circulation. 2002; 106:3068–3072.LinkGoogle Scholar
Ho KK, Pinsky JL, Kannel WB, Levy D. The epidemiology of heart failure: the Framingham Study. J Am Coll Cardiol.1993; 22:6A–13A.CrossrefMedlineGoogle Scholar
Gottdiener JS, Arnold AM, Aurigemma GP, Polak JF, Tracy RP, Kitzman DW, Gardin JM, Rutledge JE, Boineau RC. Predictors of congestive heart failure in the elderly: the Cardiovascular Health Study. J Am Coll Cardiol. 2000; 35:1628–1637.CrossrefMedlineGoogle Scholar
Mann DL, Bristow MR. Mechanisms and models in heart failure: the biomechanical model and beyond. Circulation. 2005; 111:2837–2849.LinkGoogle Scholar
Mudd JO, Kass DA. Tackling heart failure in the twenty-first century. Nature. 2008; 451:919–928.CrossrefMedlineGoogle Scholar
Lee DS, Pencina MJ, Benjamin EJ, Wang TJ, Levy D, O'Donnell CJ, Nam BH, Larson MG, D'Agostino RB, Vasan RS. Association of parental heart failure with risk of heart failure in offspring. N Engl J Med. 2006; 355:138–147.CrossrefMedlineGoogle Scholar
Kitsios G, Zintzaras E. Genetic variation associated with ischemic heart failure: a HuGE review and meta-analysis. Am J Epidemiol. 2007; 166:619–633.CrossrefMedlineGoogle Scholar
Karkkainen S, Peuhkurinen K. Genetics of dilated cardiomyopathy. Ann Med. 2007; 39:91–107.CrossrefMedlineGoogle Scholar
Larson MG, Atwood LD, Benjamin EJ, Cupples LA, D'Agostino RB, Fox CS, Govindaraju DR, Guo CY, Heard-Costa NL, Hwang SJ, Murabito JM, Newton-Cheh C, O'Donnell CJ, Seshadri S, Vasan RS, Wang TJ, Wolf PA, Levy D. Framingham Heart Study 100K project: genome-wide associations for cardiovascular disease outcomes. BMC Med Genet. 2007; 8(suppl 1):S5.CrossrefMedlineGoogle Scholar
Psaty BM, O'Donnell CJ, Gudnason V, Lunetta KL, Folsom AR, Rotter JI, Uitterlinden AG, Harris TB, Witteman JC, Boerwinkle E. Cohorts for Heart and Aging Research in Genomic Epidemiology (CHARGE) Consortium: design of prospective meta-analysis of genome-wide association studies from 5 cohorts. Circ Cardiovasc Genet. 2009; 2:73–80.LinkGoogle Scholar
- 11. The Atherosclerosis Risk in Communities (ARIC) Study: design and objectives. The ARIC investigators. Am J Epidemiol. 1989; 129:687–702.CrossrefMedlineGoogle Scholar
Fried LP, Borhani NO, Enright P, Furberg CD, Gardin JM, Kronmal RA, Kuller LH, Manolio TA, Mittelmark MB, Newman A, O'Leary DH, Psaty BM, Rautaharju P, Tracy RP, Weiler PG. The Cardiovascular Health Study: design and rationale. Ann Epidemiol. 1991; 1:263–276.CrossrefMedlineGoogle Scholar
Dawber TR, Meadors GF, Moore FE. Epidemiological approaches to heart disease: the Framingham Study. Am J Public Health Nations Health. 1951; 41:279–281.CrossrefMedlineGoogle Scholar
Feinleib M, Kannel WB, Garrison RJ, McNamara PM, Castelli WP. The Framingham Offspring Study. Design and preliminary data. Prev Med. 1975; 4:518–525.CrossrefMedlineGoogle Scholar
Hofman A, Breteler MM, van Duijn CM, Janssen HL, Krestin GP, Kuipers EJ, Stricker BH, Tiemeier H, Uitterlinden AG, Vingerling JR, Witteman JC. The Rotterdam Study: 2010 objectives and design update. Eur J Epidemiol. 2009; 24:553–572.CrossrefMedlineGoogle Scholar
Hofman A, Grobbee DE, de Jong PT, van den Ouweland FA. Determinants of disease and disability in the elderly: the Rotterdam Elderly Study. Eur J Epidemiol. 1991; 7:403–422.CrossrefMedlineGoogle Scholar
Psaty BM, Kuller LH, Bild D, Burke GL, Kittner SJ, Mittelmark M, Price TR, Rautaharju PM, Robbins J. Methods of assessing prevalent cardiovascular disease in the Cardiovascular Health Study. Ann Epidemiol. 1995; 5:270–277.CrossrefMedlineGoogle Scholar
Bleumink GS, Knetsch AM, Sturkenboom MC, Straus SM, Hofman A, Deckers JW, Witteman JC, Stricker BH. Quantifying the heart failure epidemic: prevalence, incidence rate, lifetime risk and prognosis of heart failure: the Rotterdam Study. Eur Heart J. 2004; 25:1614–1619.CrossrefMedlineGoogle Scholar
Mosterd A, Hoes AW, de Bruyne MC, Deckers JW, Linker DT, Hofman A, Grobbee DE. Prevalence of heart failure and left ventricular dysfunction in the general population: the Rotterdam Study. Eur Heart J. 1999; 20:447–455.CrossrefMedlineGoogle Scholar
Loehr LR, Rosamond WD, Chang PP, Folsom AR, Chambless LE. Heart failure incidence and survival (from the Atherosclerosis Risk in Communities study). Am J Cardiol. 2008; 101:1016–1022.CrossrefMedlineGoogle Scholar
White AD, Folsom AR, Chambless LE, Sharret AR, Yang K, Conwill D, Higgins M, Williams OD, Tyroler HA. Community surveillance of coronary heart disease in the Atherosclerosis Risk in Communities (ARIC) Study: methods and initial two years' experience. J Clin Epidemiol. 1996; 49:223–233.CrossrefMedlineGoogle Scholar
Ives DG, Fitzpatrick AL, Bild DE, Psaty BM, Kuller LH, Crowley PM, Cruise RG, Theroux S. Surveillance and ascertainment of cardiovascular events: the Cardiovascular Health Study. Ann Epidemiol. 1995; 5:278–285.CrossrefMedlineGoogle Scholar
McKee PA, Castelli WP, McNamara PM, Kannel WB. The natural history of congestive heart failure: the Framingham study. N Engl J Med. 1971; 285:1441–1446.CrossrefMedlineGoogle Scholar
Remme WJ, Swedberg K. Guidelines for the diagnosis and treatment of chronic heart failure. Eur Heart J. 2001; 22:1527–1560.CrossrefMedlineGoogle Scholar
Levy D, Ehret GB, Rice K, Verwoert GC, Launer LJ, Dehghan A, Glazer NL, Morrison AC, Johnson AD, Aspelund T, Aulchenko Y, Lumley T, Kottgen A, Vasan RS, Rivadeneira F, Eiriksdottir G, Guo X, Arking DE, Mitchell GF, Mattace-Raso FU, Smith AV, Taylor K, Scharpf RB, Hwang SJ, Sijbrands EJ, Bis J, Harris TB, Ganesh SK, O'Donnell CJ, Hofman A, Rotter JI, Coresh J, Benjamin EJ, Uitterlinden AG, Heiss G, Fox CS, Witteman JC, Boerwinkle E, Wang TJ, Gudnason V, Larson MG, Chakravarti A, Psaty BM, van Duijn CM. Genome-wide association study of blood pressure and hypertension. Nat Gen. 2009; 41:677–687.CrossrefMedlineGoogle Scholar
Ganesh SK, Zakai NA, van Rooij FJ, Soranzo N, Smith AV, Nalls MA, Chen MH, Kottgen A, Glazer NL, Dehghan A, Kuhnel B, Aspelund T, Yang Q, Tanaka T, Jaffe A, Bis JC, Verwoert GC, Teumer A, Fox CS, Guralnik JM, Ehret GB, Rice K, Felix JF, Rendon A, Eiriksdottir G, Levy D, Patel KV, Boerwinkle E, Rotter JI, Hofman A, Sambrook JG, Hernandez DG, Zheng G, Bandinelli S, Singleton AB, Coresh J, Lumley T, Uitterlinden AG, Vangils JM, Launer LJ, Cupples LA, Oostra BA, Zwaginga JJ, Ouwehand WH, Thein SL, Meisinger C, Deloukas P, Nauck M, Spector TD, Gieger C, Gudnason V, van Duijn CM, Psaty BM, Ferrucci L, Chakravarti A, Greinacher A, O'Donnell CJ, Witteman JC, Furth S, Cushman M, Harris TB, Lin JP. Multiple loci influence erythrocyte phenotypes in the CHARGE Consortium. Nat Genet. 2009; 41:1191–1198.CrossrefMedlineGoogle Scholar
Servin B, Stephens M. Imputation-based analysis of association studies: candidate regions and quantitative traits. PLoS Genet. 2007; 3:e114.CrossrefMedlineGoogle Scholar
Li Y, Abecasis GR. Mach 1.0: rapid haplotype reconstruction and missing genotype inference. Am J Hum Genet2006; S79:2290.Google Scholar
Price AL, Patterson NJ, Plenge RM, Weinblatt ME, Shadick NA, Reich D. Principal components analysis corrects for stratification in genome-wide association studies. Nat Genet. 2006; 38:904–909.CrossrefMedlineGoogle Scholar
Aulchenko YS, Ripke S, Isaacs A, van Duijn CM. GenABEL: an R library for genome-wide association analysis. Bioinformatics. 2007; 23:1294–1296.CrossrefMedlineGoogle Scholar
- 31. Genome-wide association study of 14,000 cases of seven common diseases and 3,000 shared controls. Nature. 2007; 447:661–678.CrossrefMedlineGoogle Scholar
Sloper-Mould KE, Eyre HJ, Wang XW, Sutherland GR, Baker RT. Characterization and chromosomal localization of USP3, a novel human ubiquitin-specific protease. J Biol Chem. 1999; 274:26878–26884.CrossrefMedlineGoogle Scholar
Nicassio F, Corrado N, Vissers JH, Areces LB, Bergink S, Marteijn JA, Geverts B, Houtsmuller AB, Vermeulen W, Di Fiore PP, Citterio E. Human USP3 is a chromatin modifier required for S phase progression and genome stability. Curr Biol. 2007; 17:1972–1977.CrossrefMedlineGoogle Scholar
Predmore JM, Wang P, Davis F, Bartolone S, Westfall MV, Dyke DB, Pagani F, Powell SR, Day SM. Ubiquitin proteasome dysfunction in human hypertrophic and dilated cardiomyopathies. Circulation. 2010; 121:997–1004.LinkGoogle Scholar
Wohlschlaeger J, Sixt SU, Stoeppler T, Schmitz KJ, Levkau B, Tsagakis K, Vahlhaus C, Schmid C, Peters J, Schmid KW, Milting H, Baba HA. Ventricular unloading is associated with increased 20s proteasome protein expression in the myocardium. J Heart Lung Transplant. 2010; 29:125–132.CrossrefMedlineGoogle Scholar
Pastorekova S, Zatovicova M, Pastorek J. Cancer-associated carbonic anhydrases and their inhibition. Curr Pharm Des. 2008; 14:685–698.CrossrefMedlineGoogle Scholar
Guo D, Holmlund C, Henriksson R, Hedman H. The LRIG gene family has three vertebrate paralogs widely expressed in human and mouse tissues and a homolog in Ascidiacea. Genomics. 2004; 84:157–165.CrossrefMedlineGoogle Scholar
Abraira VE, Del Rio T, Tucker AF, Slonimsky J, Keirnes HL, Goodrich LV. Cross-repressive interactions between Lrig3 and netrin 1 shape the architecture of the inner ear. Development. 2008; 135:4091–4099.CrossrefMedlineGoogle Scholar
Yi W, Haapasalo H, Holmlund C, Jarvela S, Raheem O, Bergenheim AT, Hedman H, Henriksson R. Expression of leucine-rich repeats and immunoglobulin-like domains (LRIG) proteins in human ependymoma relates to tumor location, WHO grade, and patient age. Clin Neuropathol. 2009; 28:21–27.CrossrefMedlineGoogle Scholar
Rockman HA, Koch WJ, Lefkowitz RJ. Seven-transmembrane-spanning receptors and heart function. Nature. 2002; 415:206–212.CrossrefMedlineGoogle Scholar
Kilts JD, Grocott HP, Kwatra MM. G. alpha (q)-coupled receptors in human atrium function through protein kinase C epsilon and delta. J Mol Cell Cardiol. 2005; 38:267–276.CrossrefMedlineGoogle Scholar
Bea S, Salaverria I, Armengol L, Pinyol M, Fernandez V, Hartmann EM, Jares P, Amador V, Hernandez L, Navarro A, Ott G, Rosenwald A, Estivill X, Campo E. Uniparental disomies, homozygous deletions, amplifications and target genes in mantle cell lymphoma revealed by integrative high-resolution whole genome profiling. Blood. 2008; 113: 3059–3069.CrossrefMedlineGoogle Scholar
Nardelli B, Zaritskaya L, Semenuk M, Cho YH, LaFleur DW, Shah D, Ullrich S, Girolomoni G, Albanesi C, Moore PA. Regulatory effect of IFN-kappa, a novel type I IFN, on cytokine production by cells of the innate immune system. J Immunol. 2002; 169:4822–4830.CrossrefMedlineGoogle Scholar
Miinea CP, Sano H, Kane S, Sano E, Fukuda M, Peranen J, Lane WS, Lienhard GE. AS160, the Akt substrate regulating GLUT4 translocation, has a functional Rab GTPase-activating protein domain. Biochem J. 2005; 391:87–93.CrossrefMedlineGoogle Scholar
Kramer HF, Taylor EB, Witczak CA, Fujii N, Hirshman MF, Goodyear LJ. Calmodulin-binding domain of AS160 regulates contraction- but not insulin-stimulated glucose uptake in skeletal muscle. Diabetes. 2007; 56: 2854–2862.CrossrefMedlineGoogle Scholar
Witteles RM, Fowler MB. Insulin-resistant cardiomyopathy clinical evidence, mechanisms, and treatment options. J Am Coll Cardiol. 2008; 51:93–102.CrossrefMedlineGoogle Scholar
Birney E, Stamatoyannopoulos JA, Dutta A, Guigo R, Gingeras TR, Margulies EH, Weng Z, Snyder M, Dermitzakis ET, Thurman RE, Kuehn MS, Taylor CM, Neph S, Koch CM, Asthana S, Malhotra A, Adzhubei I, Greenbaum JA, Andrews RM, Flicek P, Boyle PJ, Cao H, Carter NP, Clelland GK, Davis S, Day N, Dhami P, Dillon SC, Dorschner MO, Fiegler H, Giresi PG, Goldy J, Hawrylycz M, Haydock A, Humbert R, James KD, Johnson BE, Johnson EM, Frum TT, Rosenzweig ER, Karnani N, Lee K, Lefebvre GC, Navas PA, Neri F, Parker SC, Sabo PJ, Sandstrom R, Shafer A, Vetrie D, Weaver M, Wilcox S, Yu M, Collins FS, Dekker J, Lieb JD, Tullius TD, Crawford GE, Sunyaev S, Noble WS, Dunham I, Denoeud F, Reymond A, Kapranov P, Rozowsky J, Zheng D, Castelo R, Frankish A, Harrow J, Ghosh S, Sandelin A, Hofacker IL, Baertsch R, Keefe D, Dike S, Cheng J, Hirsch HA, Sekinger EA, Lagarde J, Abril JF, Shahab A, Flamm C, Fried C, Hackermuller J, Hertel J, Lindemeyer M, Missal K, Tanzer A, Washietl S, Korbel J, Emanuelsson O, Pedersen JS, Holroyd N, Taylor R, Swarbreck D, Matthews N, Dickson MC, Thomas DJ, Weirauch MT, Gilbert J, Drenkow J, Bell I, Zhao X, Srinivasan KG, Sung WK, Ooi HS, Chiu KP, Foissac S, Alioto T, Brent M, Pachter L, Tress ML, Valencia A, Choo SW, Choo CY, Ucla C, Manzano C, Wyss C, Cheung E, Clark TG, Brown JB, Ganesh M, Patel S, Tammana H, Chrast J, Henrichsen CN, Kai C, Kawai J, Nagalakshmi U, Wu J, Lian Z, Lian J, Newburger P, Zhang X, Bickel P, Mattick JS, Carninci P, Hayashizaki Y, Weissman S, Hubbard T, Myers RM, Rogers J, Stadler PF, Lowe TM, Wei CL, Ruan Y, Struhl K, Gerstein M, Antonarakis SE, Fu Y, Green ED, Karaoz U, Siepel A, Taylor J, Liefer LA, Wetterstrand KA, Good PJ, Feingold EA, Guyer MS, Cooper GM, Asimenos G, Dewey CN, Hou M, Nikolaev S, Montoya-Burgos JI, Loytynoja A, Whelan S, Pardi F, Massingham T, Huang H, Zhang NR, Holmes I, Mullikin JC, Ureta-Vidal A, Paten B, Seringhaus M, Church D, Rosenbloom K, Kent WJ, Stone EA, Batzoglou S, Goldman N, Hardison RC, Haussler D, Miller W, Sidow A, Trinklein ND, Zhang ZD, Barrera L, Stuart R, King DC, Ameur A, Enroth S, Bieda MC, Kim J, Bhinge AA, Jiang N, Liu J, Yao F, Vega VB, Lee CW, Ng P, Shahab A, Yang A, Moqtaderi Z, Zhu Z, Xu X, Squazzo S, Oberley MJ, Inman D, Singer MA, Richmond TA, Munn KJ, Rada-Iglesias A, Wallerman O, Komorowski J, Fowler JC, Couttet P, Bruce AW, Dovey OM, Ellis PD, Langford CF, Nix DA, Euskirchen G, Hartman S, Urban AE, Kraus P, Van Calcar S, Heintzman N, Kim TH, Wang K, Qu C, Hon G, Luna R, Glass CK, Rosenfeld MG, Aldred SF, Cooper SJ, Halees A, Lin JM, Shulha HP, Zhang X, Xu M, Haidar JN, Yu Y, Ruan Y, Iyer VR, Green RD, Wadelius C, Farnham PJ, Ren B, Harte RA, Hinrichs AS, Trumbower H, Clawson H, Hillman-Jackson J, Zweig AS, Smith K, Thakkapallayil A, Barber G, Kuhn RM, Karolchik D, Armengol L, Bird CP, de Bakker PI, Kern AD, Lopez-Bigas N, Martin JD, Stranger BE, Woodroffe A, Davydov E, Dimas A, Eyras E, Hallgrimsdottir IB, Huppert J, Zody MC, Abecasis GR, Estivill X, Bouffard GG, Guan X, Hansen NF, Idol JR, Maduro VV, Maskeri B, McDowell JC, Park M, Thomas PJ, Young AC, Blakesley RW, Muzny DM, Sodergren E, Wheeler DA, Worley KC, Jiang H, Weinstock GM, Gibbs RA, Graves T, Fulton R, Mardis ER, Wilson RK, Clamp M, Cuff J, Gnerre S, Jaffe DB, Chang JL, Lindblad-Toh K, Lander ES, Koriabine M, Nefedov M, Osoegawa K, Yoshinaga Y, Zhu B, de Jong PJ. Identification and analysis of functional elements in 1% of the human genome by the ENCODE pilot project. Nature. 2007; 447:799–816.CrossrefMedlineGoogle Scholar
Smith NL, Chan JD, Rea TD, Wiggins KL, Gottdiener JS, Lumley T, Psaty BM. Time trends in the use of beta-blockers and other pharmacotherapies in older adults with congestive heart failure. Am Heart J. 2004; 148:710–717.CrossrefMedlineGoogle Scholar
Vasan RS, Larson MG, Benjamin EJ, Evans JC, Reiss CK, Levy D. Congestive heart failure in subjects with normal versus reduced left ventricular ejection fraction: prevalence and mortality in a population-based cohort. J Am Coll Cardiol. 1999; 33:1948–1955.CrossrefMedlineGoogle Scholar
Genetic factors contribute to heart failure (HF) onset and, to date, most approaches to identify genetic variants associated with HF risk have relied on candidate genes. No large-scale genome-wide investigation of HF risk has been published. We investigated the association of ≈2.5 million single-nucleotide polymorphisms with incident HF by meta-analyzing data from 4 community-based prospective cohorts: the Atherosclerosis Risk in Communities Study, the Cardiovascular Health Study, the Framingham Heart Study, and the Rotterdam Study. Among 20 926 European-ancestry participants with 2526 incident HF events, we identified 1 locus with a single-nucleotide polymorphism, whose P value (1.4×10−8) exceeded the genome-wide statistical significance threshold of 5×10−7 and was associated with a 53% increase in HF risk. Among 2895 African-ancestry participants who had 466 incident events, we identified 1 locus with a single-nucleotide polymorphism, whose P value (6.7×10−8) exceeded the genome-wide statistical significance and was associated with a 46% increase in HF risk. We identified an additional 14 loci in European-ancestry and African-ancestry participants that were marked by high-signal single-nucleotide polymorphisms with a P value <1.0×10−5. For most loci, risk estimates were modest and did not seem to differ in subjects without a myocardial infarction preceding HF onset. Our results suggest that there may be several genomic regions associated with new-onset HF in older adults, and support the hypothesis that common genetic variation, regardless of the clinical mechanism responsible for reduced cardiac output in HF, contributes to the risk. These findings merit replication in other community-based settings of incident HF.