Coagulation Factors and the Risk of Ischemic Heart Disease: A Mendelian Randomization Study
Circulation: Genomic and Precision Medicine
Abstract
Background:
Coagulation plays a role in ischemic heart disease (IHD). However, which coagulation factors are targets of intervention is unclear. We assessed how genetically predicted vWF (von Willebrand factor), ETP (endogenous thrombin potential), FVIII (factor VIII), d-dimer, tPA (tissue-type plasminogen activator), and PAI (plasminogen activator inhibitor)-1 affected IHD. We similarly estimated effects on lipids to determine whether any associations were independent of lipids.
Methods and Results:
Separate sample instrumental variable analysis with genetic instruments, that is, Mendelian randomization, was used to obtain unconfounded estimates of effects on IHD using extensively genotyped studies of coronary artery disease/myocardial infarction, CARDIoGRAMplusC4D Metabochip (64 374 cases, 130 681 controls) and CARDIoGRAMplusC4D 1000 Genomes (60 801 cases, 123 504 controls), and on lipids using the Global Lipids Genetics Consortium Results (n=196 475). Genetically predicted ETP was positively associated with IHD (odds ratio, 1.05 per log-transformed SD; 95% confidence interval, 1.03–1.07) based on 15 single-nucleotide polymorphisms, as were vWF (odds ratio, 1.05 per SD; 95% confidence interval, 1.02–1.08) and FVIII (odds ratio, 1.06 per SD; 95% confidence interval, 1.03–1.09) based on 16 and 6 single-nucleotide polymorphisms, respectively, but the latter associations were null after considering pleiotropy. vWF and FVIII were associated with higher LDL (low-density lipoprotein) cholesterol, but not after considering pleiotropy. Genetically predicted d-dimer, tPA, and PAI-1 were not clearly associated with IHD or lipids based on 3, 3, and 5 single-nucleotide polymorphisms, respectively.
Conclusions:
ETP may affect IHD. Assessing the role of its drivers in more precisely phenotyped studies of IHD could be worthwhile.
Introduction
See Editorial by Aslibekyan and Lawler
Clinical Perspective
The role of coagulation in ischemic heart disease (IHD) is gaining increasing attention, but which coagulation factors are targets of intervention is unclear. Here, using Mendelian randomization which provides unconfounded estimates, we examined comprehensively how genetically predicted vWF (von Willebrand factor), thrombin generation potential (indicated by ETP [endogenous thrombin potential]), FVIII (factor VIII), d-dimer, tPA (tissue-type plasminogen activator) and PAI (plasminogen activator inhibitor)-1 affected IHD. We also assessed whether any of the key genes relevant to these coagulation factors were associated with IHD and myocardial infarction (MI). Our findings do not clearly corroborate the observed positive associations of d-dimer, tPA, and PAI-1 with coronary artery disease/MI. These results, however, do not rule out the possibility of a role in specific IHD- or MI-related phenotypes, but suggest that ETP and thromboxane A synthase might potentially play a role in IHD and MI. Whether vWF and FVIII affect IHD and MI or whether associations could be an unidentified attribute of ABO blood group remains to be clarified. Assessment of the role of the drivers and consequences of ETP and thromboxane A2 and of factors targeting their genes in IHD and MI, as well as clarifying the role of ABO blood group, should be worthwhile, with relevance to prevention and treatment of IHD.
Cardiovascular disease is the leading cause of mortality globally.1 Despite substantial progress in prevention and control, the cause of cardiovascular disease is not completely understood,2 as evidenced by the challenges of developing new cardiovascular therapeutics.3 Thrombosis can be an important step in the pathogenesis of ischemic heart disease (IHD), involving platelets and a coagulation cascade.4 Specifically, after injury, platelets adhere to collagen and vWF (von Willebrand factor) in subendothelial tissue, where vWF acts as a bridge between endothelial collagen and platelet surface receptors. Activated platelets release stored granules, such as ADP and thromboxane A2, stimulating platelet aggregation.4 Coagulation activation proceeds through activation of proteases; factor Xa (activated factor X) initiates the final common pathway in the cascade and results in the formation of thrombin.4 Thrombin is central to the clotting process: it converts soluble fibrinogen to fibrin; activates FV, FVIII, and FXI generating more thrombin; stimulates platelets; and favors the formation of cross-linked bonds among the fibrin molecules by activating FXIII, which stabilizes the clot.5 Cofactors are also involved in the cascade, including vitamin K, and in the fibrinolysis, such as PAI (plasminogen activator inhibitor), fibrinogen, and d-dimer.4
Systematic reviews and meta-analyses of observational studies have shown that coagulation and fibrinolysis factors, including fibrinogen, vWF, d-dimer, tPA (tissue-type plasminogen activator) antigen, and PAI-1, are associated with IHD.6–8 Observational studies are open to residual confounding, making them difficult to use as a basis for action.9 Moreover, coagulation factors are closely correlated and related to inflammation,10 so separating their roles is challenging in observational studies. Candidate gene studies have shown that genetic variants, such as rs6025 in the FV gene (F5) and rs1799963 in the prothrombin gene (F2), increase circulating thrombin generation and are associated with IHD.6 However, candidate gene studies are open to selective reporting. Currently, the clinical effectiveness of anticoagulation agents, including the antithrombin DNA aptamer NU172,11 the anti-FIXa aptamer REG1,11 and anti-vWF agents,12 is being tested in clinical trials, although these trials were not preceded by genetic validation.
In this situation, comparing cardiovascular events according to naturally occurring genetic variants related to coagulation factors, that is, Mendelian randomization (MR), provides a means of examining causal effects, without any potentially harmful intervention.9 To examine the role of coagulation factors, we searched systematically for genome-wide association studies (GWAS) of major coagulation factors, that is, vWF, thrombin generation potential, FV, FVII, FVIII, FXa, tPA, PAI-1, d-dimer, and thromboxane. We examined their role in coronary artery disease (CAD)/myocardial infarction (MI), a composite of stable CAD and acute MI, and MI, using very large case–control studies of CAD/MI and MI with extensive genotyping.13–15 To determine whether any associations were independent of lipids, we similarly examined their role in lipids using a very large study providing genetic associations with lipids.16 Finally, we assessed whether any of the key genes relevant to these coagulation factors, that is, potential genetic targets of intervention, were strongly associated with IHD.
Methods
The data and methods are publicly available. We searched Pubmed and LD Hub systematically for GWAS giving single-nucleotide polymorphisms (SNPs) reaching GWAS significance (5×10−8) for the coagulation factors, that is, vWF, thrombin generation potential, FV, FVII, FVIII, FXa, tPA, PAI-1, d-dimer, and thromboxane. Novel SNPs showing promising associations (P value <5×10−5) were also used, but excluded in sensitivity analysis. To ensure that the selected SNPs were solely associated with CAD/MI or MI via the relevant exposures, we checked for pleiotropy, that is, genetic associations with CAD/MI or MI via other phenotypes, using 2 comprehensive genetic cross-reference systems, Ensembl (http://www.ensembl.org/index.html) and PhenoScanner (www.phenoscanner.medschl.cam.ac.uk), which provide all known well-established associations of SNPs with phenotypes, including subgenome-wide associations.17 We also checked for linkage disequilibrium (LD), using Ensembl (http://www.ensembl.org/index.html). For SNPs in strong LD (r2≥0.8), the SNP with the smaller P value was selected. For SNPs with r2 <0.8, all SNPs were retained, but their correlation, obtained from Ensembl, was taken into account in the analysis.
Data on CAD/MI and MI have been contributed by CARDIoGRAMPLUSC4D investigators and downloaded from www.CARDIOGRAMPLUSC4D.ORG. CARDIoGRAMplusC4D Metabochip and CARDIoGRAMplusC4D 1000 Genomes are consortia of case–control studies with detailed phenotyping of CAD, MI, or both, based on medical records, clinical diagnosis, procedures that indicate CAD (such as revascularization), and angiographic evidence of stenosis, and sometimes based on medications or symptoms that indicate angina or self-report of CAD.13–15 CARDIoGRAMplusC4D Metabochip is a case (n=63,746)–control (n=130,681) study of CAD/MI, largely of people of European descent, with limited genotyping, imputed using HapMap.13 CARDIOGRAM is a subset of CARDIoGRAMplusC4D Metabochip (22 233 cases, 64 762 controls) with more extensive genotyping,14 was used as a substitute for CARDIoGRAMplusC4D Metabochip as necessary. CARDIoGRAMplusC4D 1000 Genomes partly overlaps with CARDIoGRAMplusC4D and has 60 801 cases of CAD/MI (≈70% MI) and 123 504 controls largely of European descent, imputed using 1000 Genomes phase 1.15 To reduce variability from different genetic imputations, we used CARDIoGRAMplusC4D Metabochip13 (or CARDIOGRAM14) for exposures genetically predicted with HapMap imputation and CARDIoGRAMplusC4D 1000 Genomes15 for exposures with 1000 Genomes imputation. In a sensitivity analysis, we repeated the analysis using the other case–control study. Genetic associations with lipids (as inverse normal transformed effect sizes), including HDL (high-density lipoprotein) cholesterol, LDL (low-density lipoprotein) cholesterol and triglycerides, adjusted for age, age2 and sex, were obtained from the Global Lipids Genetics Consortium Results (http://csg.sph.umich.edu//abecasis/public/lipids2013/), which has 188 577 participants of European descent and 7898 participants of non-European descent.16
We obtained the associations of coagulation factors with CAD/MI and MI from separate sample instrumental variable analysis. Where a SNP predicting a coagulation factor was not available for the outcomes, we sought a highly correlated proxy (r2≥0.8). We aligned SNPs based on allele letter and frequency. We obtained SNP-specific Wald estimates (ratio of the genetic association with CAD/MI or MI to the genetic association with exposure) and the ratio SD using Fieller’s theorem.18 We combined SNP-specific estimates using inverse variance weighting with fixed effects, for uncorrelated SNPs,19 and weighted generalized linear regression, for correlated SNPs.19 We checked for replication in independent studies for each SNP and conducted sensitivity analysis only using the replicated SNPs. To account for potential bias from unknown pleiotropy, we conducted sensitivity analyses using a weighted median (in uncorrelated SNPs) and MR Egger (if >3 SNPs predicted the exposure),20 which are more robust to pleiotropy. We checked whether the intercept from MR Egger was nonzero because this indicates that some genetic predictors might be acting other than via the exposure (ie, directional pleiotropy).20
We also checked whether coagulation-related genes (vWF for vWF; F2, F2R, F2RL1, F2RL2, and F2RL3 for thrombin; FGB for fibrinogen; F5 for FV; F7 for FVII; F8 for FVIII; F10 for FXa; F11 for FXI; F13A1 for FXIII; PLAT for tPA; SERPINE1 for PAI-1; and TBXAS1 for thromboxane A synthase) were associated with IHD and MI, as potential targets of intervention. For each gene, we assessed whether the distribution of the P values for the associations of the SNPs with CAD/MI or MI differed from the uniform distribution expected by chance, using all the SNPs available in the CARDIOGRAMplusC4D 1000 Genomes for each gene and taking into account correlations between SNPs to obtain an overall P value for the association of the gene with CAD/MI or MI. We used a gene-based association test using extended Simes, which is a Simes test adjusted for the LD of the P values.21 Gene-based association test using extended Simes has the advantage of not requiring simulations and provides a validated approximation to other methods.22 The LD was obtained from the 1000 Genomes catalog. All statistical analyses were conducted using R version 3.3.0 (R Foundation for Statistical Computing, Vienna, Austria). This analysis of publicly available data does not require ethical approval.
Results
We obtained genetic predictors, that is, SNPs, for vWF, thrombin generation potential (indicated by ETP [endogenous thrombin potential]), d-dimer, FVIII, tPA, and PAI-1 from GWAS.23–29 We could not find GWAS for FV, FVII, FXa, or thromboxane. We identified 20 SNPs for vWF from 3 GWAS,23–25 18 SNPs for ETP from 1 GWAS,26 3 SNPs for d-dimer from 1 GWAS,27 10 SNPs for FVIII from 3 GWAS,23–25 4 SNPs for tPA from 1 GWAS,28 and 8 SNPs for PAI-1 from 1 GWAS.29 All the GWAS were conducted in people of European ancestry. Table I in the Data Supplement gives associations of these SNPs with the coagulation factors. Genetic predictors for vWF, FVIII, d-dimer, tPA, and PAI-1 used HapMap for imputation23–25,27–29 and genetic predictors for ETP used 1000 Genomes for imputation.26 Table II in the Data Supplement gives the genetic associations with CAD/MI in CARDIoGRAMplusC4D Metabochip and 1000 Genomes and with MI in CARDIoGRAMplusC4D 1000 Genomes. Of the 20 SNPs for vWF, 4 SNPs (rs529565, rs379440, rs1438993, and rs2579103) in strong LD were discarded (rs529565 with rs687621, rs379440 with rs13361927, rs1438993 and rs2579103 with rs10745527), giving 16 SNPs for vWF (Table I in the Data Supplement). One SNP (rs687621) for vWF was in the ABO gene and associated with IL-6 (interleukin-6).30 Given this and the close association of vWF with ABO blood group,31 estimates including and excluding rs687621 are given for vWF. Of the 18 SNPs for ETP, 2 SNPs (rs60206633 and rs11038982) were discarded because of strong LD (with rs77977823 and rs3136516, respectively), and 1 SNP (rs149327057) was not in the Cardiogram Consortia, nor was its proxy (rs78008827), giving 15 SNPs for ETP. Three SNPs were only weakly correlated (r2=0.18 for rs17787912 and rs3136516, r2=0.33 for rs3136516 and rs10838599), which was taken into account. Only 1 SNP (rs17787912) for ETP was available in CARDIoGRAMplusC4D Metabochip or CARDIOGRAM, so we only used CARDIoGRAMplusC4D 1000 Genomes. All 3 SNPs for d-dimer were used. Of the 10 SNPs for FVIII, 2 SNPs (rs12557310 and rs2096362) were not in the Cardiogram Consortia, and no proxy SNPs were identified. Two highly correlated SNPs for FVIII (rs216321 [vWF] and rs9390459 [STXBP5]) were also associated with vWF.23 Two highly correlated ABO SNPs (rs687289 and rs529565) were associated with vWF, ABO blood group,32 and IL-6.30 Given the close associations of FVIII with vWF and ABO blood group,31 estimates including and excluding these SNPs are given for FVIII. Of the 4 SNPs for tPA, 1 SNP (rs2020921) was discarded because of high correlation with rs3136739, giving 3 SNPs for tPA. Of the 8 SNPs for PAI-1, 3 SNPs were discarded because of high correlation (rs6976053 and rs12672665 with rs314376, and rs2075756 with rs3847067), and 2 weakly correlated SNPs (r2=0.37 for rs3847067 and rs314376) were retained, giving 5 SNPs for PAI-1. Rs6486122 (ARNTL) for PAI-1, associated with CRP (C-reactive protein),33 was retained because CRP does not play a causal role in IHD in MR studies.34 All SNPs for vWF, d-dimer, FVIII, tPA, and PAI-1 were available in the Global Lipids Genetics Consortium Results; however, only 5 SNPs (rs117784795, rs17787912, rs3136516, rs77977823 and rs4752807, proxy SNP for rs35800856) for ETP were available in Global Lipids Genetics Consortium Results.
Genetically predicted vWF was positively associated with CAD/MI and MI (Table 1), but the associations were null after excluding rs687621 (ABO; Table 1). Similarly, genetically predicted FVIII was positively associated with CAD/MI and MI (Table 1), but after considering pleiotropy, the association with CAD/MI was null using CARDIoGRAMplusC4D Metabochip, and the association with MI was null using a weighted median and MR Egger (Table 1). Genetically predicted ETP was positively associated with CAD/MI and MI (Table 1). Genetically predicted PAI-1 was positively associated with CAD/MI but was not associated with MI (Table 1). The associations were in the other direction using MR Egger, where the non-null intercept indicated directional pleiotropy using CARDIoGRAMPLUSC4D 1000 Genomes. The influential genetic predictor, rs314376 (SLC12A9), was detected by Cook’s distance and Studentized residual,20 however, excluding it did not change the direction of the MR Egger estimate (odds ratio [OR], 0.36, 95% confidence interval, 0.16–0.80 for CAD/MI; OR, 0.24; 95% confidence interval, 0.10–0.59 for MI). A scatter plot was not consistent with a positive association of PAI-1 with CAD/MI (Figure I in the Data Supplement). Genetically predicted d-dimer and tPA were not clearly associated with CAD/MI or MI (Table 1). Most SNPs for vWF, D-dimer, and PAI-1 have been replicated in independent studies, whereas some SNPs for ETP, FVIII, and tPA located in new loci remain to be replicated (Table I in the Data Supplement). The findings were unchanged using only replicated SNPs, but the associations of FVIII with CAD/MI and MI were null, possibly because of fewer SNPs (Table III in the Data Supplement). The findings were unchanged only using genome-wide significance SNPs (Table IV in the Data Supplement).
Exposure* | Outcome | Method | Outcome Study With the Same Imputation | Outcome Study With Different Imputation | ||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
Outcome Study | No. of SNPs | OR | 95% CI | Intercept P Value | Outcome study | No. of SNPs | OR | 95% CI | Intercept P Value | |||
vWF (SD) | CAD/MI | IVW | Metabochip | 16 | 1.05 | 1.02–1.08 | 1000 G | 16 | 1.05 | 1.03–1.08 | ||
WM | 16 | 1.06 | 1.03–1.09 | 16 | 1.07 | 1.04–1.10 | ||||||
MR Egger | 16 | 1.06 | 1.02–1.11 | 0.38 | 16 | 1.07 | 1.03–1.12 | 0.32 | ||||
CAD/MI without rs687621 (ABO) | IVW | Metabochip | 15 | 1.01 | 0.95–1.08 | 15 | 1.02 | 0.98–1.06 | ||||
WM | 15 | 1.05 | 0.96–1.14 | 15 | 1.01 | 0.96–1.06 | ||||||
MR Egger | 15 | 0.99 | 0.83–1.18 | 0.82 | 15 | 1.003 | 0.91–1.11 | 0.60 | ||||
MI | IVW | Metabochip | NA | NA | NA | 1000 G | 16 | 1.09 | 1.06–1.12 | |||
WM | NA | NA | NA | 16 | 1.13 | 1.09–1.17 | ||||||
MR Egger | NA | NA | NA | NA | 16 | 1.13 | 1.08–1.19 | 0.05 | ||||
MI without rs687621 (ABO) | IVW | Metabochip | NA | NA | NA | 1000 G | 15 | 1.03 | 0.98–1.07 | |||
WM | NA | NA | NA | 15 | 1.02 | 0.96–1.08 | ||||||
MR Egger | NA | NA | NA | NA | 15 | 0.99 | 0.89–1.11 | 0.48 | ||||
ETP (SD) | CAD/MI | WGLM | 1000 G | 15 | 1.05 | 1.03–1.07 | Metabochip† | NA | NA | NA | ||
WM | 14 | 1.04 | 1.01–1.07 | NA | NA | NA | ||||||
MR Egger | 15 | 1.02 | 0.98–1.06 | 0.51 | NA | NA | NA | NA | ||||
MI | WGLM | 1000 G | 15 | 1.08 | 1.06–1.11 | Metabochip† | NA | NA | NA | |||
WM | 14 | 1.07 | 1.04–1.10 | NA | NA | NA | ||||||
MR Egger | 15 | 1.04 | 0.99–1.09 | 0.48 | NA | NA | NA | NA | ||||
d-dimer (SD) | CAD/MI | IVW | Metabochip | 3 | 1.02 | 0.87–1.19 | 1000 G | 3 | 0.99 | 0.92–1.08 | ||
WM | 3 | 1.00 | 0.83–1.21 | 3 | 0.99 | 0.90–1.08 | ||||||
MI | IVW | Metabochip | NA | NA | NA | 1000 G | 3 | 0.97 | 0.89–1.07 | |||
WM | NA | NA | NA | 3 | 0.94 | 0.85–1.05 | ||||||
FVIII (SD) | CAD/MI | IVW | Metabochip | 6 | 1.06 | 1.03–1.09 | 1000 G | 6 | 1.08 | 1.05–1.12 | ||
WM | 6 | 1.06 | 1.03–1.09 | 6 | 1.08 | 1.04–1.12 | ||||||
MR Egger | 6 | 1.05 | 1.004–1.11 | 0.72 | 6 | 1.07 | 1.02–1.12 | 0.47 | ||||
CAD/MI without potentially pleiotropic SNPs | IVW | Metabochip | 4 | 1.07 | 0.90–1.27 | 1000 G | 4 | 1.14 | 1.02–1.28 | |||
WM | 4 | 1.06 | 0.84–1.34 | 4 | 1.13 | 0.96–1.32 | ||||||
MR Egger | 4 | 0.78 | 0.38–1.60 | 0.38 | 4 | 0.93 | 0.58–1.47 | 0.37 | ||||
MI | IVW | Metabochip | NA | NA | NA | 1000 G | 6 | 1.14 | 1.10–1.19 | |||
WM | NA | NA | NA | 6 | 1.14 | 1.10–1.19 | ||||||
MR Egger | NA | NA | NA | NA | 6 | 1.14 | 1.08–1.19 | 0.76 | ||||
MI without potentially pleiotropic SNPs | IVW | Metabochip | NA | NA | NA | 1000 G | 4 | 1.14 | 1.01–1.30 | |||
WM | NA | NA | NA | 4 | 1.14 | 0.96–1.36 | ||||||
MR Egger | NA | NA | NA | NA | 4 | 0.96 | 0.57–1.62 | 0.50 | ||||
tPA, ng/mL | CAD/MI | IVW | Metabochip | 3 | 0.95 | 0.54–1.67 | 1000 G | 3 | 1.12 | 0.75–1.67 | ||
WM | 3 | 1.22 | 0.60–2.48 | 3 | 1.14 | 0.69–1.87 | ||||||
MI | IVW | Metabochip | NA | NA | NA | 1000 G | 3 | 1.41 | 0.90–2.22 | |||
WM | NA | NA | NA | 3 | 1.68 | 0.98–2.90 | ||||||
PAI-1, ng/mL | CAD/MI | WGLM | Metabochip | 5 | 1.18 | 0.95–1.48 | 1000 G | 5 | 1.28 | 1.10–1.48 | ||
WM | 4 | 0.94 | 0.67–1.31 | 4 | 1.21 | 0.97–1.50 | ||||||
MR Egger | 5 | 0.36 | 0.13–1.05 | 0.03 | 5 | 0.50 | 0.24–1.02 | 0.009 | ||||
MI | WGLM | Metabochip | NA | NA | NA | 1000 G | 5 | 1.12 | 0.94–1.32 | |||
WM | NA | NA | NA | 4 | 1.13 | 0.89–1.44 | ||||||
MR Egger | NA | NA | NA | NA | 5 | 0.33 | 0.14–0.78 | 0.005* |
1000 G indicates 1000 Genome; CAD/MI, coronary artery disease/myocardial infarction; CI, confidence interval; ETP, endogenous thrombin potential; FVIII, factor VIII; IVW, inverse variance weighting; OR, odds ratio; NA, not available; PAI, plasminogen activator inhibitor; SNP, single-nucleotide polymorphism; tPA, tissue-type plasminogen activator; vWF, von Willebrand factor; WGLM, weighted generalized linear regression model; and WM, weighted median method.
*
ETP, d-dimer, tPA, and PAI-1 were log transformed in the genome-wide association studies providing genetic associations. Weighted median method was conducted in uncorrelated SNPs, so correlated SNPs (rs3136516 for ETP and rs314376 for PAI-1) were deleted.
†
Only 1 SNP (rs17787912) for ETP was available in CARDIoGRAMplusC4D Metabochip or CARDIOGRAM, so we only estimated the association in CARDIoGRAMplusC4D 1000 Genomes.
Genetically predicted vWF and FVIII were associated with higher LDL cholesterol and HDL cholesterol using inverse variance weighting, but not after excluding rs687621 (ABO) for vWF and potentially pleiotropic SNPs for FVIII. Genetically predicted FVIII was also associated with lower triglycerides, but not after excluding potentially pleiotropic SNPs. The non-null intercept from MR Egger indicated the presence of directional pleiotropy for the associations of vWF and FVIII with LDL cholesterol; the associations were consistent in MR Egger. Genetically predicted ETP was associated with higher HDL cholesterol, lower LDL cholesterol, and lower triglycerides (Table 2). However, the intercept from MR Egger for LDL cholesterol was non-null, and the associations with LDL cholesterol and triglycerides were null using a weighted median. Genetically predicted PAI-1 was positively associated with triglycerides (Table 2).
Exposure | Outcome | Method | No. of SNPs | β | 95% CI | Intercept P Value |
---|---|---|---|---|---|---|
vWF (SD) | HDL cholesterol | IVW | 16 | 0.01* | 0.003 to 0.02* | |
WM | 16 | 0.02* | 0.01 to 0.03* | |||
MR Egger | 16 | 0.025 | 0.007 to 0.04 | 0.07 | ||
Drop ABO SNP | 15 | −0.005 | −0.03 to 0.02 | |||
LDL cholesterol | IVW | 16 | 0.05* | 0.03 to 0.06* | ||
WM | 16 | 0.06* | 0.05 to 0.08* | |||
MR Egger | 16 | 0.08* | 0.05 to 0.11* | 0.002* | ||
Drop ABO SNP | 15 | −0.002 | −0.02 to 0.02 | |||
Triglycerides | IVW | 16 | −0.01 | −0.02 to 0.001 | ||
WM | 16 | −0.01* | −0.03 to −0.004* | |||
MR Egger | 16 | −0.02 | −0.04 to 0.01 | 0.50 | ||
Drop ABO SNP | 15 | 0.01 | −0.01 to 0.03 | |||
ETP (SD) | HDL cholesterol | WGLM | 5 | 0.05* | 0.04 to 0.07* | |
WM | 4 | 0.05* | 0.03 to 0.07* | |||
MR Egger | 5 | 0.01 | –0.06 to 0.07 | 0.18 | ||
LDL cholesterol | WGLM | 5 | −0.02* | −0.04 to −0.01* | ||
WM | 4 | −0.02 | −0.04 to 0.01 | |||
MR Egger | 5 | 0.03 | −0.01 to 0.06 | 0.004* | ||
Triglycerides | WGLM | 5 | −0.02* | −0.03 to −0.01* | ||
WM | 4 | −0.01 | −0.03 to 0.005 | |||
MR Egger | 5 | 0.001 | −0.04 to 0.04 | 0.32 | ||
D-dimer (SD) | HDL cholesterol | IVW | 3 | 0.01 | −0.03 to 0.06 | |
WM | 3 | 0.02 | −0.03 to 0.07 | |||
LDL cholesterol | IVW | 3 | 0.01 | −0.03 to 0.06 | ||
WM | 3 | 0.01 | −0.04 to 0.07 | |||
Triglycerides | IVW | 3 | 0.004 | −0.04 to 0.05 | ||
WM | 3 | −0.003 | −0.05 to 0.05 | |||
FVIII (SD) | HDL cholesterol | IVW | 6 | 0.02* | 0.01 to 0.03* | |
WM | 6 | 0.02* | 0.01 to 0.03* | |||
MR Egger | 6 | 0.025* | 0.007 to 0.04* | 0.61 | ||
Drop potentially pleiotropic SNPs | 4 | −0.01 | −0.06 to 0.05 | |||
LDL cholesterol | IVW | 6 | 0.06* | 0.05 to 0.08* | ||
WM | 6 | 0.07* | 0.05 to 0.08* | |||
MR Egger | 6 | 0.08* | 0.07 to 0.10* | 0.01* | ||
Drop potentially pleiotropic SNPs | 4 | 0.01 | −0.05 to 0.07 | |||
Triglycerides | IVW | 6 | −0.01* | −0.02 to −0.002* | ||
WM | 6 | −0.01* | −0.03 to −0.003* | |||
MR Egger | 6 | –0.02 | −0.05 to 0.02 | 0.84 | ||
Drop potentially pleiotropic SNPs | 4 | 0.04 | −0.03 to 0.10 | |||
tPA, ng/mL | HDL cholesterol | IVW | 3 | 0.06 | −0.13 to 0.26 | |
WM | 3 | 0.05 | −0.17 to 0.28 | |||
LDL cholesterol | IVW | 3 | 0.07 | −0.15 to 0.29 | ||
WM | 3 | 0.03 | −0.22 to 0.29 | |||
Triglycerides | IVW | 3 | −0.04 | −0.23 to 0.15 | ||
WM | 3 | −0.03 | −0.24 to 0.19 | |||
PAI-1, ng/mL | HDL cholesterol | WGLM | 5 | 0.03 | −0.04 to 0.11 | |
WM | 4 | 0.10 | –0.02 to 0.23 | |||
MR Egger | 5 | −0.07 | −0.67 to 0.52 | 0.72 | ||
LDL cholesterol | WGLM | 5 | 0.06 | −0.02 to 0.14 | ||
WM | 4 | 0.01 | −0.10 to 0.12 | |||
MR Egger | 5 | –0.07 | −0.49 to 0.36 | 0.55 | ||
Triglycerides | WGLM | 5 | 0.10 | 0.03 to 0.18 | ||
WM | 4 | 0.11 | 0.001 to 0.22 | |||
MR Egger | 5 | 0.14 | −0.45 to 0.72 | 0.92 |
CI indicates confidence interval; ETP, endogenous thrombin potential; FVIII, factor VIII; HDL, high-density lipoprotein; IVW, inverse variance weighting; LDL, low-density lipoprotein; PAI, plasminogen activator inhibitor; tPA, tissue-type plasminogen activator; vWF, von Willebrand factor; WGLM, weighted generalized linear regression model; and WM, weighted median method.
*
p value<0.05.
The vWF-related gene, vWF, and 1 thrombin-related gene, F2RL1, were associated with IHD, whereas a different thrombin-related gene, F2, was associated with MI (Table 3). The thromboxane gene (TBXAS1) was strongly associated with IHD and MI. d-dimer–related, tPA-related, and PAI-1–related genes were not associated with IHD or MI. F8 is on the X chromosome, which is not included in the Cardiogram Consortia.
Coagulation Factor | Related Genes | Chromosome | No. of SNPs | Gene P Value | ||
---|---|---|---|---|---|---|
CAD/MI | MI | CAD/MI | MI | |||
vWF | vWF | 12 | 724 | 717 | 0.049* | 0.22 |
Thrombin | F2 | 11 | 72 | 72 | 0.24 | 0.02* |
F2R | 5 | 66 | 65 | 0.27 | 0.93 | |
F2RL1 | 5 | 56 | 55 | 0.016* | 0.13 | |
F2RL2 | 5 | 101 | 101 | 0.87 | 0.78 | |
F2RL3 | 19 | 34 | 34 | 0.83 | 0.80 | |
Fibrinogen | FGB | 4 | 52 | 52 | 0.96 | 0.65 |
Factor V | F5 | 1 | 369 | 370 | 0.42 | 0.59 |
Factor VII | F7 | 13 | 73 | 72 | 0.07 | 0.48 |
Factor VIII | F8 | X chromosome | n/a | n/a | n/a | n/a |
Factor Xa | F10 | 13 | 76 | 74 | 0.56 | 0.96 |
Factor XI | F11 | 4 | 113 | 112 | 0.24 | 0.81 |
Factor XIII | F13A1 | 6 | 745 | 739 | 0.96 | 0.45 |
tPA | PLAT | 8 | 79 | 78 | 0.75 | 0.46 |
PAI-1 | SERPINE1 | 7 | 73 | 73 | 0.41 | 0.65 |
Thromboxane A synthase | TBXAS1 | 7 | 797 | 782 | 0.00071* | 0.000096* |
CAD/MI indicates coronary artery disease/myocardial infarction; n/a, not available; PAI, plasminogen activator inhibitor; SNP, single-nucleotide polymorphism; tPA, tissue-type plasminogen activator; and vWF, von Willebrand factor.
*
p value<0.05.
Discussion
In this systematic examination of the role of vWF, ETP, d-dimer, FVIII, tPA, and PAI-1 in IHD and MI using MR consistent with meta-analysis of candidate gene studies,35 genetically predicted ETP was positively associated with IHD and MI, but not with unhealthy lipids. Genetically predicted vWF was positively associated with IHD, MI, and LDL cholesterol possibly because of rs687621 in the ABO gene. Genetically predicted FVIII was associated with higher IHD, MI, and LDL cholesterol; however, these associations might be because of pleiotropy and were not robust to sensitivity analyses. Genetically predicted PAI-1 was positively associated with IHD but not MI, and associations were not robust to sensitivity analysis. Our study also suggests that d-dimer and tPA might be associated with IHD in observational studies6,8 as biomarkers rather than causes of IHD, consistent with their prognostic role.6,8 Correspondingly, the PAI-1 and tPA genes were not associated with IHD or MI, but the vWF gene and a thrombin gene were associated with IHD, and another thrombin gene was associated with MI. The thromboxane A2 gene TBXAS1 was associated with IHD and MI.
This is the first study to assess comprehensively the role of coagulation factors in IHD, MI, and lipids using very large samples and methods which provide unconfounded estimates. Genetic variants are randomly allocated at conception and so are less likely to be influenced by the confounders in observational studies, such as lifestyle, health status, socioeconomic position, and use of medication. We used separate samples for coagulation factors and IHD or MI, which reduced the risk of chance associations from the underlying data structure.36 Nevertheless, several limitations exist. First, genetic predictors for FV, FVII, FXa, and thromboxane A2 have not been reliably identified. We partially addressed this gap by considering associations of the relevant genes with IHD and MI, which suggested that FV and FXa were unlikely to affect IHD or MI. We could not include platelet response to ADP and other agonists37 because the effects of the relevant genetic predictors would only be evident in people taking these agonists, which we cannot identify. Second, confounding by population stratification is possible; however, the genetic associations are from studies conducted in Caucasians, largely of European descent, with genomic control.13–15,23–29 The same imputation catalog, that is, HapMap CEU,13,14,23–25,27–29 or 1000 Genomes,15,26 was used for exposures and IHD. As such, the estimates should not be confounded by population stratification. Confounding by LD is also possible, if the genetic predictor is in LD with a SNP associated with confounders, such as ethnicity. However, our check in 2 comprehensive genetic cross-reference systems did not identify such SNPs. Third, pleiotropy, that is, genetic associations with IHD or MI via other phenotypes than these coagulation factors, might exist. Some SNPs are in functionally relevant genes, such as rs1063857 and rs216321 in vWF for vWF,23,25 but some SNPs are in genes with unclear functions, such as rs150611042 in ORM1 for ETP, and rs9399599 in STXBP5 and rs7301826 in STX2 for tPA, which is typical for SNPs selected statistically from GWAS rather than based on biological pathways.38 Experimental evidence supports associations of the lead SNPs in STXBP5 and STX2 with expression levels of the respective transcripts.28 The assumptions of MR do not require the SNPs to cause the exposure but to be associated with SNPs that do.39 We checked for known potential pleiotropy of the SNPs predicting the coagulation factors in 2 comprehensive genetic cross-reference systems including subgenome-wide associations, but the possibility of currently unknown pleiotropic associations remains. However, the findings were consistent only using SNPs replicated in independent studies, adding some validity. We also checked for statistical evidence of unknown pleiotropy using MR Egger, which raised question about the associations of PAI-1 with IHD and MI. Fourth, our study, limited to Caucasians, might not apply to other populations. However, effects of causal factors are not expected to vary by setting,40 although they might not always be relevant. Fifth, we could not assess whether associations with IHD and MI varied by baseline levels of the coagulation factors. However, observational studies in people with different baseline vWF have shown consistent associations with IHD.8 Sixth, estimates might be biased if all genetic associations are from a single sample.36 However, we used different samples for genetic variants on coagulation factors and on IHD, MI, and lipids; the participants only partly overlap for vWF, FVIII, d-dimer, tPA, and PAI-1. As such, any correlation of the genetic variants with unmeasured confounders in the samples with these coagulation factors is unlikely to be replicated in the samples with the outcomes because of the different data structures.41 Moreover, genetic associations with coagulation factors and with IHD, MI, and lipids all controlled for age and sex (Table I in the Data Supplement), which improved precision of the estimates. Seventh, the small estimate for ETP on IHD and MI may not have clinical significance. However, IHD is not a very clearly defined phenotype because IHD may have arisen for a variety of reasons, be diagnosed in many ways, and its diagnosis affected by use of medications. Imprecision in the IHD phenotype would bias toward the null, making this study, as other MR studies, more suitable for refuting causality than establishing the exact size of a causal effect.42 Our findings may be more relevant to firstly identifying which coagulation factors are the most promising targets and secondly to population health, where a small shift in the overall population distribution may have a large impact.43 Eighth, canalization is possible, that is, influences of genetic predictors might be damped or buffered by compensatory developmental processes.9 However, the extent to which canalization occurs is unknown. Ninth, effects might differ by sex, and the effect of vWF might vary with ABO blood group, which cannot be assessed from the data freely available. Examination of whether the associations were mediated by lipids was not possible, because the publically available genetic data for CAD/MI and MI do not include lipids. Tenth, MR requires large samples because genetic predictors only capture a small proportion of the variance in exposures.44 However, explaining all the variance in the coagulation factors would invalidate an instrument because it would be equivalent to the coagulation factor and be equally confounded. Taking advantage of large publicly available studies enables a cost-efficient, well-powered study.19 CARDIoGRAMplusC4D with 64 374 cases and 130 681 controls has 0.8 power to detect an OR of 1.0344 with an R2 of ≈0.19 for vWF,23–25 OR of 1.10 with an R2 of ≈0.02 for FVIII,24 ETP, and d-dimer,26,27,38 and an OR of 1.15 with an R2 of ≈0.01 for tPA and PAI-1.28,29 Replication in other studies, using different genetic predictors is required.
Our study adds to the limited evidence concerning the role of coagulation factors in IHD and MI and suggests that ETP and thromboxane A synthase might play a role, but d-dimer, tPA, and PAI-1 probably do not. However, IHD is a complex phenotype, so examining the role of these factors, and related factors, such as ADP, in more precisely defined phenotypes would be worthwhile. Given the multiple functions of thrombin,45 several pathways might exist. First, thrombin can activate FV, FVIII, FXI, and FXIII and convert soluble fibrinogen into insoluble fibrin, which forms fibrin clots.11 However, FV, FXI, and FXIII were not associated with IHD or MI using gene-based tests. A previous MR study also suggests that fibrinogen plays no role in IHD.46 Second, prothrombin fragments are positively associated with plasma oxidized LDL, suggesting a potential interplay with lipids.47 However, ETP did not affect lipids detrimentally here. Thrombin is converted from prothrombin, whose biosynthesis is vitamin K dependent.48 ETP is lowered by vitamin K antagonists.49 Consistently, vitamin K is positively associated with IHD in an MR study.50 In contrast, our study does not support fibrinolytic activity as having a role in IHD or MI because fibrinolytic activity is largely determined by the balance between tPA and its natural, fast-acting inhibitor, PAI-1,28 and our findings do not corroborate a role of tPA or PAI-1 in IHD or MI. Enalapril, which improves fibrinolytic activity,51 did not reduce cardiovascular disease events in a randomized controlled trial.52 As regards vWF and FVIII, it is uncertain whether they affect IHD or MI or whether associations are an unidentified attribute of ABO blood group, although the vWF gene was associated with IHD (Table 3). Therapeutics targeting vWF are being investigated.12 Understanding the role of ABO blood group in IHD and MI might explain the findings here and generate additional causal insights. The gene TBXAS1 was associated with IHD and MI and is targeted by existing therapeutics, such as Ridogrel, Dazoxiben, and NM-702. Few trials have assessed their effects on IHD, although Ridogrel was more effective than aspirin in preventing new ischemic events in a randomized controlled trial.53
Conclusions
Our findings do not clearly corroborate observed positive associations of d-dimer, tPA, and PAI-1 with CAD/MI, or rule out the possibility of a role in specific IHD- or MI-related phenotypes, but suggest that ETP and thromboxane A synthase might potentially play a role in IHD and MI. Whether vWF and FVIII affect IHD and MI or whether associations could be an unidentified attribute of ABO blood group remains to be clarified. Assessment of the role of the drivers and consequences of ETP and thromboxane A2 and of factors targeting their genes in IHD and MI, as well as clarifying the role of ABO blood group, should be worthwhile, with relevance to prevention and treatment of IHD.
Supplemental Material
References
1.
Mathers CD, Boerma T, Ma Fat D. Global and regional causes of death. Br Med Bull. 2009;92:7–32. doi: 10.1093/bmb/ldp028.
2.
Ezzati M, Obermeyer Z, Tzoulaki I, Mayosi BM, Elliott P, Leon DA. Contributions of risk factors and medical care to cardiovascular mortality trends. Nat Rev Cardiol. 2015;12:508–530. doi: 10.1038/nrcardio.2015.82.
3.
Jackson N, Atar D, Borentain M, Breithardt G, van Eickels M, Endres M, et al. Improving clinical trials for cardiovascular diseases: a position paper from the Cardiovascular Round Table of the European Society of Cardiology. Eur Heart J. 2016;37:747–754. doi: 10.1093/eurheartj/ehv213.
4.
Palta S, Saroa R, Palta A. Overview of the coagulation system. Indian J Anaesth. 2014;58:515–523. doi: 10.4103/0019-5049.144643.
5.
Di Nisio M, Middeldorp S, Büller HR. Direct thrombin inhibitors. N Engl J Med. 2005;353:1028–1040. doi: 10.1056/NEJMra044440.
6.
Lowe G, Rumley A. The relevance of coagulation in cardiovascular disease: what do the biomarkers tell us? Thromb Haemost. 2014;112:860–867. doi: 10.1160/TH14-03-0199.
7.
Song C, Burgess S, Eicher JD, O’Donnell CJ, Johnson AD. Causal effect of plasminogen activator inhibitor type 1 on coronary heart disease. J Am Heart Assoc. 2017;6:e004918. doi: 10.1161/JAHA.116.004918.
8.
Willeit P, Thompson A, Aspelund T, Rumley A, Eiriksdottir G, Lowe G, et al. Hemostatic factors and risk of coronary heart disease in general populations: new prospective study and updated meta-analyses. PLoS One. 2013;8:e55175. doi: 10.1371/journal.pone.0055175.
9.
Lawlor DA, Harbord RM, Sterne JA, Timpson N, Davey Smith G. Mendelian randomization: using genes as instruments for making causal inferences in epidemiology. Stat Med. 2008;27:1133–1163. doi: 10.1002/sim.3034.
10.
Hohlfeld T, Schrör K. Antiinflammatory effects of aspirin in ACS: relevant to its cardiocoronary actions? Thromb Haemost. 2015;114:469–477. doi: 10.1160/TH15-03-0191.
11.
Li W, Wang K, Zhao M, Yang X, Chen M, Lan X. Development of aptamer oligonucleotides as anticoagulants and antithrombotics for cardiovascular diseases: current status. Thromb Res. 2014;134:769–773. doi: 10.1016/j.thromres.2014.05.021.
12.
Montoro-García S, Shantsila E, Lip GY. Potential value of targeting von Willebrand factor in atherosclerotic cardiovascular disease. Expert Opin Ther Targets. 2014;18:43–53. doi: 10.1517/14728222.2013.840585.
13.
Deloukas P, Kanoni S, Willenborg C, Farrall M, Assimes TL, Thompson JR, et al.; CARDIoGRAMplusC4D Consortium; DIAGRAM Consortium; CARDIOGENICS Consortium; MuTHER Consortium; Wellcome Trust Case Control Consortium. Large-scale association analysis identifies new risk loci for coronary artery disease. Nat Genet. 2013;45:25–33. doi: 10.1038/ng.2480.
14.
Schunkert H, König IR, Kathiresan S, Reilly MP, Assimes TL, Holm H, et al.; Cardiogenics; CARDIoGRAM Consortium. Large-scale association analysis identifies 13 new susceptibility loci for coronary artery disease. Nat Genet. 2011;43:333–338. doi: 10.1038/ng.784.
15.
Nikpay M, Goel A, Won HH, Hall LM, Willenborg C, Kanoni S, et al. A comprehensive 1,000 Genomes-based genome-wide association meta-analysis of coronary artery disease. Nat Genet. 2015;47:1121–1130. doi: 10.1038/ng.3396.
16.
Willer CJ, Schmidt EM, Sengupta S, Peloso GM, Gustafsson S, et al.; Global Lipids Genetics Consortium. Discovery and refinement of loci associated with lipid levels. Nat Genet. 2013;45:1274–1283. doi: 10.1038/ng.2797.
17.
Staley JR, Blackshaw J, Kamat MA, Ellis S, Surendran P, Sun BB, et al. PhenoScanner: a database of human genotype-phenotype associations. Bioinformatics. 2016;32:3207–3209. doi: 10.1093/bioinformatics/btw373.
18.
Fieller EC. Some problems in interval estimation. J R Stat Soc Series B Stat Methodol. 1954;16:175–185.
19.
Burgess S, Scott RA, Timpson NJ, Davey Smith G, Thompson SG; EPIC-InterAct Consortium. Using published data in Mendelian randomization: a blueprint for efficient identification of causal risk factors. Eur J Epidemiol. 2015;30:543–552. doi: 10.1007/s10654-015-0011-z.
20.
Burgess S, Thompson SG. Interpreting findings from Mendelian randomization using the MR-Egger method. Eur J Epidemiol. 2017;32:377–389. doi: 10.1007/s10654-017-0255-x.
21.
Li MX, Kwan JS, Sham PC. HYST: a hybrid set-based test for genome-wide association studies, with application to protein-protein interaction-based association analysis. Am J Hum Genet. 2012;91:478–488. doi: 10.1016/j.ajhg.2012.08.004.
22.
Bacanu SA. On optimal gene-based analysis of genome scans. Genet Epidemiol. 2012;36:333–339. doi: 10.1002/gepi.21625.
23.
Smith NL, Chen MH, Dehghan A, Strachan DP, Basu S, Soranzo N, et al.; Wellcome Trust Case Control Consortium;. Novel associations of multiple genetic loci with plasma levels of factor VII, factor VIII, and von Willebrand factor: the CHARGE (Cohorts for Heart and Aging Research in Genome Epidemiology) consortium. Circulation. 2010;121:1382–1392. doi: 10.1161/CIRCULATIONAHA.109.869156.
24.
Antoni G, Oudot-Mellakh T, Dimitromanolakis A, Germain M, Cohen W, Wells P, et al. Combined analysis of three genome-wide association studies on vWF and FVIII plasma levels. BMC Med Genet. 2011;12:102. doi: 10.1186/1471-2350-12-102.
25.
Tang W, Cushman M, Green D, Rich SS, Lange LA, Yang Q, et al. Gene-centric approach identifies new and known loci for FVIII activity and VWF antigen levels in European Americans and African Americans. Am J Hematol. 2015;90:534–540. doi: 10.1002/ajh.24005.
26.
Rocanin-Arjo A, Cohen W, Carcaillon L, Frère C, Saut N, Letenneur L, et al.; CardioGenics Consortium. A meta-analysis of genome-wide association studies identifies ORM1 as a novel gene controlling thrombin generation potential. Blood. 2014;123:777–785. doi: 10.1182/blood-2013-10-529628.
27.
Smith NL, Huffman JE, Strachan DP, Huang J, Dehghan A, Trompet S, et al. Genetic predictors of fibrin D-dimer levels in healthy adults. Circulation. 2011;123:1864–1872. doi: 10.1161/CIRCULATIONAHA.110.009480.
28.
Huang J, Huffman JE, Yamakuchi M, Yamkauchi M, Trompet S, Asselbergs FW, et al.; Cohorts for Heart and Aging Research in Genome Epidemiology (CHARGE) Consortium Neurology Working Group; CARDIoGRAM Consortium; CHARGE Consortium Hemostatic Factor Working Group. Genome-wide association study for circulating tissue plasminogen activator levels and functional follow-up implicates endothelial STXBP5 and STX2. Arterioscler Thromb Vasc Biol. 2014;34:1093–1101. doi: 10.1161/ATVBAHA.113.302088.
29.
Huang J, Sabater-Lleal M, Asselbergs FW, Tregouet D, Shin SY, Ding J, et al.; DIAGRAM Consortium; CARDIoGRAM Consortium; C4D Consortium; CARDIOGENICS Consortium. Genome-wide association study for circulating levels of PAI-1 provides novel insights into its regulation. Blood. 2012;120:4873–4881. doi: 10.1182/blood-2012-06-436188.
30.
Naitza S, Porcu E, Steri M, Taub DD, Mulas A, Xiao X, et al. A genome-wide association scan on the levels of markers of inflammation in Sardinians reveals associations that underpin its complex regulation. PLoS Genet. 2012;8:e1002480. doi: 10.1371/journal.pgen.1002480.
31.
Franchini M, Crestani S, Frattini F, Sissa C, Bonfanti C. ABO blood group and von Willebrand factor: biological implications. Clin Chem Lab Med. 2014;52:1273–1276. doi: 10.1515/cclm-2014-0564.
32.
Desch KC, Ozel AB, Siemieniak D, Kalish Y, Shavit JA, Thornburg CD, et al. Linkage analysis identifies a locus for plasma von Willebrand factor undetected by genome-wide association. Proc Natl Acad Sci USA. 2013;110:588–593. doi: 10.1073/pnas.1219885110.
33.
Dehghan A, Dupuis J, Barbalic M, Bis JC, Eiriksdottir G, Lu C, et al. Meta-analysis of genome-wide association studies in >80 000 subjects identifies multiple loci for C-reactive protein levels. Circulation. 2011;123:731–738. doi: 10.1161/CIRCULATIONAHA.110.948570.
34.
Holmes MV, Ala-Korpela M, Smith GD. Mendelian randomization in cardiometabolic disease: challenges in evaluating causality. Nat Rev Cardiol. 2017;14:577–590. doi: 10.1038/nrcardio.2017.78.
35.
Ye Z, Liu EH, Higgins JP, Keavney BD, Lowe GD, Collins R, et al. Seven haemostatic gene polymorphisms in coronary disease: meta-analysis of 66,155 cases and 91,307 controls. Lancet. 2006;367:651–658. doi: 10.1016/S0140-6736(06)68263-9.
36.
Taylor AE, Davies NM, Ware JJ, VanderWeele T, Smith GD, Munafò MR. Mendelian randomization in health research: using appropriate genetic variants and avoiding biased estimates. Econ Hum Biol. 2014;13:99–106. doi: 10.1016/j.ehb.2013.12.002.
37.
Johnson AD, Yanek LR, Chen MH, Faraday N, Larson MG, Tofler G, et al. Genome-wide meta-analyses identifies seven loci associated with platelet aggregation in response to agonists. Nat Genet. 2010;42:608–613. doi: 10.1038/ng.604.
38.
Palmer TM, Lawlor DA, Harbord RM, Sheehan NA, Tobias JH, Timpson NJ, et al. Using multiple genetic variants as instrumental variables for modifiable risk factors. Stat Methods Med Res. 2012;21:223–242. doi: 10.1177/0962280210394459.
39.
Didelez V, Sheehan N. Mendelian randomization as an instrumental variable approach to causal inference. Stat Methods Med Res. 2007;16:309–330. doi: 10.1177/0962280206077743.
40.
Rothman KJ, Gallacher JE, Hatch EE. Why representativeness should be avoided. Int J Epidemiol. 2013;42:1012–1014. doi: 10.1093/ije/dys223.
41.
Tchetgen Tchetgen EJ, Walter S, Glymour MM. Commentary: building an evidence base for Mendelian randomization studies: assessing the validity and strength of proposed genetic instrumental variables. Int J Epidemiol. 2013;42:328–331. doi: 10.1093/ije/dyt023.
42.
Schooling CM, Au Yeung SL, Freeman G. Mendelian randomization estimates may be inflated. J Am Coll Cardiol. 2013;61:1931. doi: 10.1016/j.jacc.2012.12.049.
43.
Rose G. Sick individuals and sick populations. Int J Epidemiol. 2001;30:427–432; discussion 433.
44.
Freeman G, Cowling BJ, Schooling CM. Power and sample size calculations for Mendelian randomization studies using one genetic instrument. Int J Epidemiol. 2013;42:1157–1163. doi: 10.1093/ije/dyt110.
45.
Ten Cate H, Hemker HC. Thrombin generation and atherothrombosis: what does the evidence indicate? J Am Heart Assoc. 2016;5:e003553.
46.
Sabater-Lleal M, Huang J, Chasman D, Naitza S, Dehghan A, Johnson AD, et al.; VTE Consortium; STROKE Consortium; Wellcome Trust Case Control Consortium 2 (WTCCC2); C4D Consortium; CARDIoGRAM Consortium. Multiethnic meta-analysis of genome-wide association studies in >100 000 subjects identifies 23 fibrinogen-associated loci but no strong evidence of a causal association between circulating fibrinogen and cardiovascular disease. Circulation. 2013;128:1310–1324. doi: 10.1161/CIRCULATIONAHA.113.002251.
47.
Pawlak K, Mysliwiec M, Pawlak D. oxLDL - the molecule linking hypercoagulability with the presence of cardiovascular disease in hemodialyzed uraemic patients. Thromb Res. 2014;134:711–716. doi: 10.1016/j.thromres.2014.07.007.
48.
Stenflo J, Fernlund P, Egan W, Roepstorff P. Vitamin K dependent modifications of glutamic acid residues in prothrombin. Proc Natl Acad Sci USA. 1974;71:2730–2733.
49.
Brocal I, Marco P, Lucas J, Verdú J, Tarín F. Thrombin generation test in patients under anticoagulant therapy with vitamin K antagonists. Thromb Haemost. 2009;101:594–595.
50.
Schooling CM. Plasma levels of vitamin K and the risk of ischemic heart disease: a Mendelian randomization study. J Thromb Haemost. 2016;14:1211–1215. doi: 10.1111/jth.13332.
51.
Boman KO, Jansson JH, Nyhlén KA, Nilsson TK. Improved fibrinolysis after one year of treatment with enalapril in men and women with uncomplicated myocardial infarction. Thromb Haemost. 2002;87:311–316.
52.
Ong HT. Are angiotensin-converting enzyme inhibitors and angiotensin receptor blockers especially useful for cardiovascular protection? J Am Board Fam Med. 2009;22:686–697. doi: 10.3122/jabfm.2009.06.090094.
53.
The RAPT investigators.Randomized trial of ridogrel, a combined thromboxane a2 synthase inhibitor and thromboxane a2/prostaglandin endoperoxide receptor antagonist, versus aspirin as adjunct to thrombolysis in patients with acute myocardial infarction. The Ridogrel Versus Aspirin Patency Trial (RAPT). Circulation. 1994;89:588–595.
Information & Authors
Information
Published In
Copyright
© 2018 American Heart Association, Inc.
History
Received: 17 July 2017
Accepted: 31 October 2017
Published in print: January 2018
Published online: 12 January 2018
Keywords
Subjects
Authors
Disclosures
None.
Metrics & Citations
Metrics
Citations
Download Citations
If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Select your manager software from the list below and click Download.
- Peripheral Coagulation Parameters and Prostate Cancer Association: A Retrospective Study and Mendelian Randomization, Clinical Medicine Insights: Oncology, 18, (2024).https://doi.org/10.1177/11795549241263950
- SCAR-6 elncRNA locus epigenetically regulates PROZ and modulates coagulation and vascular function, EMBO Reports, 25, 11, (4950-4978), (2024).https://doi.org/10.1038/s44319-024-00272-w
- Rivaroxaban, a direct inhibitor of coagulation factor Xa, attenuates adverse cardiac remodeling in rats by regulating the PAR-2 and TGF-β1 signaling pathways, PeerJ, 11, (e16097), (2023).https://doi.org/10.7717/peerj.16097
- Methylomic, Proteomic, and Metabolomic Correlates of Traffic-Related Air Pollution in the Context of Cardiorespiratory Health: A Systematic Review, Pathway Analysis, and Network Analysis, Toxics, 11, 12, (1014), (2023).https://doi.org/10.3390/toxics11121014
- Appraising the Causal Role of Risk Factors in Coronary Artery Disease and Stroke: A Systematic Review of Mendelian Randomization Studies, Journal of the American Heart Association, 12, 20, (2023)./doi/10.1161/JAHA.122.029040
- Dysregulation of metalloproteins in ischemic heart disease patients with systolic dysfunction, International Journal of Biological Macromolecules, 232, (123435), (2023).https://doi.org/10.1016/j.ijbiomac.2023.123435
- Insights into Causal Cardiovascular Risk Factors from Mendelian Randomization, Current Cardiology Reports, 25, 2, (67-76), (2023).https://doi.org/10.1007/s11886-022-01829-8
- What is the role of Von Willebrand factor in chronic hepatitis B virus infection to hepatocellular carcinoma: a review article, Therapeutic Advances in Chronic Disease, 13, (2022).https://doi.org/10.1177/20406223221125683
- Elevated plasma levels of plasminogen activator inhibitor‐1 are associated with risk of future incident venous thromboembolism, Journal of Thrombosis and Haemostasis, 20, 7, (1618-1626), (2022).https://doi.org/10.1111/jth.15701
- Cardiovascular Risk Stratification Based on Oxidative Stress for Early Detection of Pathology, Antioxidants & Redox Signaling, 35, 8, (602-617), (2021).https://doi.org/10.1089/ars.2020.8254
- See more
Loading...
View Options
Login options
Check if you have access through your login credentials or your institution to get full access on this article.
Personal login Institutional LoginPurchase Options
Purchase this article to access the full text.
eLetters(0)
eLetters should relate to an article recently published in the journal and are not a forum for providing unpublished data. Comments are reviewed for appropriate use of tone and language. Comments are not peer-reviewed. Acceptable comments are posted to the journal website only. Comments are not published in an issue and are not indexed in PubMed. Comments should be no longer than 500 words and will only be posted online. References are limited to 10. Authors of the article cited in the comment will be invited to reply, as appropriate.
Comments and feedback on AHA/ASA Scientific Statements and Guidelines should be directed to the AHA/ASA Manuscript Oversight Committee via its Correspondence page.