- Irene D. Bezemer1,
- Andre R. Arellano2,
- Carmen H. Tong2,
- Charles M. Rowland2,
- Helen A. Ireland3,
- Kenneth A. Bauer4,
- Joseph Catanese2,
- Pieter H. Reitsma5,6,
- Carine J.M. Doggen1,
- James J. Devlin2,
- Frits R. Rosendaal1,5,3⇓ and
- Lance A. Bare2
- 1 Department of Clinical Epidemiology, Leiden University Medical Center, Leiden, The Netherlands
- 2 Celera, Alameda, CA, USA
- 3 Department of Cardiovascular Genetics, Division of Medicine, University College London, London, UK
- 4 Beth Israel Deaconess Medical Center, Molecular Medicine Unit, Boston, MA, USA
- 5 Department of Thrombosis and Haemostasis, Leiden University Medical Center, Leiden, The Netherlands
- 6 Einthoven Laboratory for Experimental Vascular Medicine, Leiden University Medical Center, Leiden, the Netherlands
- Correspondence: Frits R. Rosendaal, Department of Clinical Epidemiology, C7-P Leiden University Medical Center P.O. Box 9600 2300 RC Leiden, The Netherlands. E-mail:
Background We recently reported the association between the Malmö sequence variant in F9 (rs6048) and deep vein thrombosis.
Design and Methods We aimed to study whether the association between F9 Malmö and deep vein thrombosis is explained by linkage disequilibrium with nearby single-nucleotide polymorphisms, and whether the association is explained biologically by F9 Malmö affecting factor IX antigen levels or activation of factor IX. We investigated the association of F9 Malmö and 28 nearby single-nucleotide polymorphisms with deep vein thrombosis in men from two case-control studies, LETS (n=380) and MEGA (n=1,469). We assessed the association of F9 Malmö with factor IX antigen level in male control subjects from LETS (n=191) and two subsets of MEGA (n=823 and n=484) and the association with endogenous thrombin potential in LETS control men. We studied the association between F9 Malmö and factor IX activation peptide in 1,199 healthy middle-aged men from the NPHS-II cohort.
Results In the combined LETS and MEGA studies, the odds ratio (95% confidence interval) for the G allele of F9 Malmö, compared with the A allele, was 0.80 (0.69–0.93). One single-nucleotide polymorphism in F9, rs422187, was strongly linked to F9 Malmö (r2=0.94) and was similarly associated with deep vein thrombosis. No other single-nucleotide polymorphism or haplotype tested was more strongly associated. Factor IX antigen level, factor IX activation peptide levels and endogenous thrombin potential did not differ between F9 Malmö genotypes.
Conclusions The F9 Malmö sequence variant was the most strongly associated with deep vein thrombosis among common single-nucleotide polymorphisms in the region. However, the biological mechanism by which F9 Malmö affects risk remains unknown.
The incidence rate of deep vein thrombosis in the general population is estimated to be about 1 per 1,000 person-years leading to 500,000 new cases per year in Europe and 400,000 in the US.1,2 Deep vein thrombosis is a multicausal disease that results from acquired (e.g., age, oral contraceptive use, surgery) and genetic risk factors (e.g., factor V Leiden, prothrombin G20210A). Elevated plasma levels of coagulation factors (e.g., VIII,3 IX,4 X,5 and XI6) have been shown to be important risk factors for deep vein thrombosis. Antigen levels of factor IX above the 90th percentile (129U/dL) have been associated with a 2 to 3-fold increased risk of deep vein thrombosis.4 We recently tested 19,682 potentially functional single-nucleotide polymorphisms (SNPs) for association with deep vein thrombosis.7 One of the SNPs identified was an A>G sequence variant in the gene encoding factor IX (rs6048, F9 Malmö) and the G allele was associated with a 15–43% decrease in deep vein thrombosis risk compared with the A allele, in three case-control studies of deep vein thrombosis. This common variant (minor allele frequency = 0.32) leads to the substitution of alanine for threonine at amino acid position 148 in the portion of the factor IX zymogen that is cleaved from the zymogen to activate factor IX.8 This variant has not been reported to be associated with hemophilia B. The mechanism by which F9 Malmö affects risk of deep vein thrombosis is unclear.
In this study we calculated the pooled odds ratios in the case-control studies in which the association of F9 Malmö with deep vein thrombosis was initially identified. We investigated whether the association between F9 Malmö and deep vein thrombosis could be explained by linkage disequilibrium between F9 Malmö and other F9 variants. We assessed whether the association could be explained biologically by F9 Malmö affecting factor IX antigen levels, activation of factor IX or endogenous thrombin potential.
Design and Methods
Study populations and data collection
The case-control populations (LETS, MEGA-1 and MEGA-2) used to analyze the association of genotypes with deep vein thrombosis are derived from two large population-based case-control studies; the Leiden Thrombophilia Study (LETS) and the Multiple Environmental and Genetic Assessment (MEGA) of risk factors for venous thrombosis. Samples from the Northwick Park Heart Study–II (NPHS-II) were used to evaluate the association of F9 Malmö genotype with factor IX activation peptide.
Collection and ascertainment of events in LETS have been described previously.9 Briefly, patients were recruited between January 1, 1988 and December 30, 1992 from three anticoagulation clinics in the western part of the Netherlands: Leiden, Amsterdam, and Rotterdam. A total of 474 consecutive patients, 70 years or younger, with a diagnosis of a first deep vein thrombosis of the leg and without a known malignancy were included. For each patient, an age- (±5 years) and sex-matched control subject who had no history of deep vein thrombosis was enrolled. In the present analysis of LETS, we excluded 52 participants due to inadequate quantity or quality of DNA. After these exclusions, 443 patients and 453 controls remained.
Collection and ascertainment of events in MEGA have been described previously.10,11 Briefly, MEGA enrolled consecutive patients aged 18–70 years who presented with a first diagnosis of deep vein thrombosis in leg or arm or pulmonary embolism at any of six anticoagulation clinics in the Netherlands (Amsterdam, Amersfoort, The Hague, Leiden, Rotterdam, and Utrecht) between March 1st, 1999 and May 31st, 2004. Control subjects in MEGA included partners of patients and random population control subjects frequency-matched on age and sex to the patient group and were selected so that the distribution of age and sex matched that of the patient group. For the present analyses we excluded patients who had a pulmonary embolism without a documented deep vein thrombosis or a history of malignant disorders to obtain a study population similar to the LETS population. We split the MEGA study to form two case-control studies, based on recruitment date and sample availability (blood or buccal swab). The first subset of MEGA, MEGA-1, included 1,398 patients and 1,757 controls who donated a blood sample. The remaining 1,314 patients and 2,877 controls, who donated either a blood sample or a buccal swab sample, were included in MEGA-2.
We calculated the pooled odds ratio for deep vein thrombosis in men and women from the LETS and MEGA studies. The gene effect in men, who have one X-chromosome, was assumed to be equivalent to the effect in women homozygous for one allele, compared to women homozygous for the other allele. Heterozygous women were assumed to have an in-between gene effect. The analysis of gene variants in F9 and deep vein thrombosis risk was restricted to men. The analysis of F9 Malmö with factor IX antigen levels and endogenous thrombin potential, was restricted to men not using vitamin K antagonists. The LETS study included 188 male patients and 192 control men (191 with factor IX antigen level measurements and 157 with endogenous thrombin potential measurements), the MEGA-1 study included 637 male patients and 832 control men (823 with factor IX antigen level measurements) and the MEGA-2 study included 620 male patients and 1,327 control men (484 with factor IX antigen level measurements).
The Northwick Park Heart Study-II (NPHS-II) has been described previously.12 Briefly, a cohort of 4,600 men aged 50–63 years registered with 9 general medical practices in England and Scotland were screened for eligibility in the NPHS-II. Exclusion criteria for the original study included: a history of unstable angina or acute myocardial infarction (AMI); a major Q wave on the electrocardiogram (ECG); regular aspirin or anticoagulant therapy; cerebrovascular disease; life-threatening malignancy; conditions exposing staff to risk or precluding informed consent. Factor IX activation peptide was measured in the first available sample from individuals who, at the time of sampling, had not had a coronary heart disease event.
Gene variants in the F9 region and risk of deep vein thrombosis
To investigate whether other SNPs in this region might be associated with deep vein thrombosis, we used results from the HapMap Project to identify a region defined by linkage disequilibrium with F9 Malmö (r2≥0.2, chr X: 138,404,951 to 138,494,063). Apart from the F9 gene, this region also contains part of the gene encoding MCF.2 cell line derived transforming sequence (MCF2). This region contained 48 SNPs with allele frequencies >0.02 (HapMap data release #221, phase II April07, on NCBI B36 assembly, dbSNP 126). Allele frequencies and linkage disequilibrium were calculated from the SNP genotypes in the HapMap CEPH population, which includes Utah residents with ancestry from Northern and Western Europe. Eighteen SNPs were selected for genotyping using pairwise tagging in Tagger (implemented in Haploview).13 We were unable to construct assays for 3 (rs4149754, rs438601, rs17340148) of the 18 tag SNPs; the remaining 15 SNPs are reasonable surrogates for 45 of the 48 SNPs in this region because the 45 SNPs are either directly genotyped or in strong linkage disequilibrium (r2>0.8) with at least one of the 15 genotyped SNPs. In addition, we genotyped 14 candidate SNPs.14,15 Thus, in total 29 SNPs were initially investigated in the men of LETS. SNPs that were associated with deep vein thrombosis were subsequently investigated in the men of MEGA-1.
Factor IX antigen level
The levels of factor IX were determined in LETS and MEGA by enzyme-linked immunosorbent assay (ELISA) as previously described.4 This ELISA is highly specific for factor IX and results are not affected by the levels of the other vitamin K–dependent proteins. The intra-assay and interassay coefficients of variation were 7% and 7%, respectively, in LETS and 4 % and 3% in MEGA. Results are expressed in units per deciliter (dL), where 100 U is the amount of factor IX present in 1 dL pooled normal plasma.
Factor IX activation peptide
Factor IX activation peptide was determined by double antibody radioimmunoassay16 as a marker of turnover of factor IX in the NPHS-II samples. The intra-assay coefficient of variation was less than 5% and the interassay coefficient of variation was about 10% for factor IX activation peptide.
F9 Malmö and endogenous thrombin potential
Endogenous thrombin potential is an activate protein C (APC) sensitivity test that quantifies the time integral of thrombin generated in plasma in which coagulation is initiated via the extrinsic pathway.17,18 The sensitivity of the plasma endogenous thrombin potential to APC was measured in LETS under conditions where the test is insensitive for small amounts of phospholipid present in plasma as previously described.19,20
Allele frequency and genotype determination
DNA concentrations were standardized to 10 ng/μL using PicoGreen (Molecular Probes, Invitrogen Corp, Carlsbad, CA, USA) fluorescent dye. Genotyping of individual DNA samples was performed as previously described7 using 0.3ng of DNA in kPCR assays21 or using multiplexed oligo ligation assays (OLA).22 Genotyping accuracy of the multiplex methodology and kPCR has been assessed in three previous studies by comparing genotype calls from multiplex OLA assays with those from real time kPCR assays for the same SNPs, and the overall concordance of the genotype calls from these two methods was >99% in each of these studies.23–25 Primer sequences are available upon request.
The combined analysis of F9 Malmö in LETS, MEGA-1 and MEGA-2 was performed using the meta package version 0.8–2 available in R software language version 2.4.1 (http://www.r-project.org) by treating the individual studies as fixed effects and using the inverse variance method to estimate the pooled effect of the SNP.26
Deviations from Hardy–Weinberg expectations were assessed using an exact test in controls. Odds ratios (OR) and 95% confidence intervals (95% CI) were computed as an estimate of risk of deep vein thrombosis associated with each SNP using logistic regression.
The association between haplotype and deep vein thrombosis was assessed using the R language package (http://cran.us.r-project.org/) of haplo.stats,27 which uses the expectation maximization algorithm to estimate haplotype frequencies followed by testing for association between haplotype and disease using a score test. A sliding window was used to select and test haplotypes consisting of 3, 5, and 7 adjacent SNPs from the set of SNPs including F9 Malmö and the 28 other SNPs.
Differences in the factor IX antigen level and endogenous thrombin potential between F9 Malmö genotypes were assessed in control subjects of LETS, MEGA-1 and MEGA-2 using a t-test assuming equal variance among the groups. Differences in factor IX activation peptide between F9 Malmö genotypes in NPHS-II were also assessed using a t-test. t-tests were performed with SAS version 9. Power to detect a difference in mean factor IX antigen levels, activation peptide or thrombin potential between F9 Malmö genotypes was calculated using nQuery Advisor version 4.028 for a two sample t-test at a 0.05 two-sided significance level and assuming equal variance among the groups.
Combined analysis of F9 Malmö and deep vein thrombosis
In the combined analysis of LETS, MEGA-1 and MEGA-2, we found that F9 Malmö was associated with deep vein thrombosis in men: the pooled odds ratio was 0.80 (95% CI, 0.69–0.93) for the G (n=1,108) compared with the A genotype (n=2,688). The pooled corresponding odds ratio in women was 0.89 (95% CI, 0.74–1.08) for the GG (n=405) compared with the AA genotypes (n=2,190), assuming an in-between effect in heterozygous women (n=1,750) (Figure 1).
Factor IX antigen level and risk of deep vein thrombosis
Factor IX antigen levels in LETS were previously reported to be positively associated with risk of deep vein thrombosis. Individuals with factor IX levels above the 90th percentile of the control group levels had a 2- to 3-fold higher risk of deep vein thrombosis.4 In the MEGA study, we confirmed the results published for LETS. We found that individuals with factor IX antigen levels above the 90th percentile had a 1.7-fold (95% CI 1.4–2.1) increased risk of deep vein thrombosis, compared with those with levels below the 90th percentile. After adjustment for age, sex, oral contraceptive use and vitamin K-dependent coagulation factors II, VII and X (all coagulation factors dichotomized at the 90th percentile) the odds ratio was 1.8 (95% CI 1.4–2.2). Among men, the unadjusted odds ratio in the MEGA study was 1.8 (95%CI 1.3–2.4).
Gene variants in F9 region
To study the region surrounding F9 Malmö we investigated 15 tag SNPs that served as surrogates for other SNPs in the region and 14 additional candidate SNPs in LETS men (Figure 2). We found that 7 of the 29 SNPs, including F9 Malmö, were associated (p≤0.05) with deep vein thrombosis (Table 1). We then investigated the association between these 7 SNPs and deep vein thrombosis in MEGA-1 men and found that 2 were significantly associated with deep vein thrombosis in MEGA-1: F9 Malmö and rs422187, a SNP that is located 421 bp from F9 Malmö in intron 5 (Table 1). The minor allele frequency and risk estimate for the association between rs422187 and deep vein thrombosis was similar to that of F9 Malmö as LD between rs422187 and the F9 Malmö SNP was high (r2=0.94). We did not find any haplotype of adjacent SNPs (data not shown) that was more strongly associated with deep vein thrombosis in LETS or MEGA-1 than F9 Malmö or rs422187.
F9 Malmö and factor IX antigen level
We investigated whether the association between F9 Malmö and deep vein thrombosis could be explained by an association between F9 Malmö and factor IX antigen levels. Among the male control subjects from LETS, MEGA-1 and MEGA-2 for whom factor IX antigen levels were measured we did not find a difference in factor IX level between the A and the G genotype (Table 2).
F9 Malmö and factor IX activation
We investigated whether the association between F9 Malmö and deep vein thrombosis could be explained by differential activation of factor IX. Among men of NPHS-II for whom factor IX activation peptide levels were available we compared the mean activation peptide levels between the A and G genotypes. There was no evidence of lower levels of factor IX activation peptide in men with the F9 Malmö genotype (Table 2).
F9 Malmö and endogenous thrombin potential
We also investigated the association between F9 Malmö and deep vein thrombosis through measuring the endogenous thrombin potential. In the test performed in LETS, coagulation is initiated via the extrinsic pathway. Among the male control subjects of LETS for whom endogenous thrombin potential was measured, we found no difference between F9 Malmö genotypes (Table 2).
The LETS, MEGA and NPHS-II studies had sufficient power to detect small differences in factor IX or factor IX activation peptide levels between F9 Malmö genotypes A and G. The combined LETS and MEGA factor IX antigen level assays had a standard deviation of 17 U/dL and the combined study had 90% power to detect mean factor IX antigen level differences of 0.18 standard deviations (3.1 U/dL) or greater. The factor IX activation peptide assay had a standard deviation of 75 pmol/L and the study had 90% power to detect mean factor IX activation peptide level differences of 0.25 standard deviations (19 pmol/L) or greater. The endogenous thrombin potential assay had a standard deviation of 348 Nm.min and the study had 90% power to detect mean thrombin potential differences of 0.55 standard deviations (190 Nm.min) or greater.
We investigated whether the previously reported association between F9 Malmö and deep vein thrombosis could be explained by linkage disequilibrium with SNPs that had a stronger association with deep vein thrombosis than F9 Malmö. We also studied whether factor IX antigen levels, factor IX activation peptide or endogenous thrombin potential differed by F9 Malmö genotype, which would provide a biological explanation for the association with deep vein thrombosis. However, the present results did not support any of these hypotheses.
The OR for the association of F9 Malmö with deep vein thrombosis was slightly lower in men than in women of all three studies, and broader confidence intervals were observed in women.
We did not identify any SNP that was more strongly associated with deep vein thrombosis than F9 Malmö in LETS and MEGA-1 among 29 SNPs we genotyped in the F9 region. For 3 SNPs initially selected for genotyping in LETS we were unable to construct genotyping assays. These 3 SNPs are unlikely to be responsible for the observed association between F9 Malmö and deep vein thrombosis since they were not in strong LD (r2<0.2) with F9 Malmö in the HapMap CEPH population. A recent study found that rs4149755 in the F9 gene was associated with deep vein thrombosis in postmenopausal women.15 We included rs4149755 in our analysis in LETS, but rs4149755 was not associated with deep vein thrombosis in men (Table 1) or in women (data not shown). We also did not find any haplotype that was more strongly associated with deep vein thrombosis in both LETS and MEGA-1 than F9 Malmö.
The lack of an association between F9 Malmö and factor IX antigen levels is consistent with the results of a previous study, where none of 27 SNPs including F9 Malmö were associated with factor IX levels.14 In a recent analysis in LETS, no association was observed either;29 the present study also included the LETS and confirmed the findings in the MEGA study, which has a much larger number of samples. The combined analysis of LETS and MEGA had sufficient power to detect small factor IX antigen level differences between genotypes; the observed levels were almost identical. In addition to factor IX antigen levels we studied factor IX activation. This analysis did not yield an explanation for the observed association between F9 Malmö and risk of deep vein thrombosis either. The power calculations suggest that we can likely (90% power) exclude differences of 19 pmol/L or greater, but we cannot rule out smaller differences. We used an indirect assay to estimate activation of factor IX. The plasma level of the factor IX activation peptide is dependent on the release of activation peptide that occurs when factor IX is activated to factor IXa and the steady clearance rate of the activation peptide in the kidneys.30 Direct measurements of factor IX activation by tissue factor (TF):factor VIIa or factor XIa might reveal effects on the activation of factor IX caused by F9 Malmö.
The endogenous thrombin potential is the end product of the TF activation pathway and is therefore an alternative measure of the activity of factor IX or any other factor in the coagulation cascade. We found no evidence of an association between F9 Malmö and thrombin potential, although we cannot rule out a small difference between the genotypes.
If the associations of F9 Malmö with deep vein thrombosis observed in this study are replicated in future studies, the question remains by which mechanism F9 Malmö reduces risk of deep vein thrombosis. Our study included SNPs with minor allele frequencies of at least <2% and was not designed to detect functional variants that are less frequent. It is also possible that one or more of the observed associations is the result of linkage disequilibrium with a low frequency variation.
In conclusion, the Malmö SNP in F9, rs6048, was associated with a 20% reduction in risk of deep vein thrombosis in a combined analysis of men from LETS, MEGA-1 and MEGA-2. The association was not explained by linkage disequilibrium with other SNPs in the F9 region. We were not able to provide a biological explanation for the association of F9 Malmö with deep vein thrombosis, as we found no evidence for an association of F9 Malmö with either factor IX antigen levels or activation of factor IX, indirectly measured by factor IX activation peptide plasma levels and by the endogenous thrombin potential. Additional studies should focus on investigating direct measurements of factor IX activation and on linkage disequilibrium between F9 Malmö and a rare (minor allele frequency 0.02) sequence variant.
Acknowledgments: the authors acknowledge the contributions of the Celera High Throughput Laboratory and Computational Biology group and thank J. Sninsky and T. White for helpful comments on this manuscript. We thank R. van Eck, I.K. van der Linden, J. van der Meijden and P.J. Noordijk who performed laboratory measurements.
Authorship and Disclosures
IDB, HAI, KAB, PHR, CJMD, FRR, LAB: study design; IDB, ARA, HAI, JC, CJMD, FRR: data acquisition; IDB, ARA, CT, CMR, HAI, KAB, PHR, CJMD, JJD, FRR, LAB: analysis and interpretation; IDB, LAB: drafting of the manuscript; critical revision of the manuscript: all authors. HAI, KAB, PHR, CJMD, JJD, FRR, LAB: study supervision.
Preliminary results were presented at the International Society on Thrombosis and Haemostasis (ISTH) Congress; July 10, 2007; Geneva, Switzerland.
The remaining authors have no disclosures.
Funding: the Leiden Thrombophilia study was supported by the Netherlands Heart Foundation (grant 89.063). The Multiple Environmental and Genetic Assessment of risk factors for venous thrombosis was supported by the Netherlands Heart Foundation (grant NHS 98.113), the Dutch Cancer Foundation (grant RUL 99/1992) and the Netherlands Organisation for Scientific Research (grant 912-03-033| 2003). Ms. Bezemer was supported by a grant from the Leducq Foundation, Paris, France for the development of Transatlantic Networks of Excellence in Cardiovascular Research (grant 04 CVD 02).
The funding organizations did not play a role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; or the preparation, review, or approval of the manuscript. LAB, ARA, CHT, CMR, JC, and JJD have employment and ownership interests.
- Received November 7, 2008.
- Revision received December 5, 2008.
- Accepted December 15, 2008.
- Copyright© Ferrata Storti Foundation