Updated Analysis of Gene Variants Associated With Deep Vein Thrombosis FREE

To the Editor: A single-nucleotide polymorphism (SNP) association study¹ assessing variants associated with deep vein thrombosis found that, of 19 682 gene-centric SNPs, 18 were consistently associated with deep vein thrombosis in the Leiden Thrombophilia Study (LETS) and in a subset of the Multiple Environmental and Genetic Assessment of Risk Factors for Venous Thrombosis (MEGA-1). Nine of these SNPs were genotyped in the remaining part of the MEGA study (MEGA-2). Assays for the other 9 SNPs for MEGA-2 were not available at that time; the associations of these 9 SNPs with venous thrombosis in 1314 cases and 2877 controls of MEGA-2 were subsequently assessed.

Details of the study population, DNA collection, SNP selection, and genotyping were described previously.¹ Chromosome 1 associations were stratified by factor V Leiden (rs6025, chr 1:167,785,673) to avoid reporting false-positive associations due to linkage disequilibrium. This analysis was previously performed for rs2227589 (SERPINC1 [NCBI Entrez Gene 462]) and rs670659 (RGS7 [NCBI Entrez Gene 6000]), which appeared to be independently associated.¹ The SNPs rs6016 (chr 1:167,778,744), rs4524 (chr 1:167,778,379), rs3820059 (chr 1:167,657,778), and rs6131 (chr 1:167,847,509) were among the 9 SNPs pending genotyping in MEGA-2. The rs6016 SNP (F5 [NCBI Entrez Gene 2153]) was in linkage disequilibrium with rs4524 (F5, D" = 0.99 and r² = 0.97 in LETS and MEGA-1), which has been reported to be associated with venous thrombosis.² The SNP rs3820059 (C1orf114 [NCBI Entrez Gene 57821]) was in linkage disequilibrium with factor V Leiden (D" = 0.91 and r² = 0.09) and not associated with venous thrombosis after adjustment for factor V Leiden in LETS and MEGA-1. The SNP rs6131 (SELP [NCBI Entrez Gene 6403]) was also not associated with venous thrombosis after adjustment for factor V Leiden, although linkage disequilibrium was not strong (D" = 0.34 and r² = 0.03).

Six additional SNPs in MEGA-2 were genotyped: rs4524 (F5), rs1801690 (APOH [NCBI Entrez Gene 350]), rs3087546 (EPS8L2 [NCBI Entrez Gene 64787]), rs369328 (CASP8AP2 [NCBI Entrez Gene 9994]), rs881 (TACR1 [NCBI Entrez Gene 6869]), and rs2266911 (ODZ1 [NCBI Entrez Gene 10178]). The false discovery rates for the first 9 SNPs in MEGA-2 were previously reported.¹ For the present analysis, the false discovery rate was manually recalculated for all 15 SNPs. Men and women were combined in all analyses including rs6048 (F9 [NCBI Entrez Gene 2158]) because the association with deep vein thrombosis in MEGA-2 was similar in men and women.³ Odds ratios, 95% confidence intervals, and P values were calculated by logistic regression using SPSS for Windows release 16.0.2 (SPSS Inc, Chicago, Illinois).

The associations between SNPs and deep vein thrombosis in MEGA-2 are presented in the Table. Of 6 newly genotyped SNPs, only rs4524 in F5 replicated in MEGA-2. The other SNPs had odds ratios close to 1 and corresponding P values >.20. After adjustment for factor V Leiden, rs4524 remained associated with venous thrombosis (odds ratio, 1.21; 95% confidence interval, 1.10-1.34). Linkage disequilibrium between factor V Leiden and rs4524 was low (r² = 0.02 in MEGA-2).

The previous study¹ identified 7 SNPs with a false discovery rate of less than 20%; 3 had P < .05 and a false discovery rate of less than 10% (rs13146272 in CYP4V2 [NCBI Entrez Gene 285440], rs2227589 in SERPINC1, and rs1613662 in GP6 [NCBI Entrez Gene 51206]). After recalculating the false discovery rate for 15 SNPs, 7 SNPs (including rs4524) had a false discovery rate of less than 20% (Table).

The rs4524 SNP in F5 was consistently associated with deep vein thrombosis in 3 large case-control studies. Five SNPs in APOH, EPS8L2, CASP8AP2, TACR1, and ODZ1 were associated in 2 case-control studies but failed to triplicate in the third study. The associations of 3 SNPs in F5, C1orf114, and SELP were due to linkage disequilibrium with other SNPs.