Abstract

The correlation between vitamin K-dependent coagulation factors (CF) and lipids has been previously established, with Factor VII notably receiving attention in the investigation of atherosclerosis. Current evidence suggests that plasma phospholipids (PL) may enhance the procoagulant reactions and may increase the availability of fat soluble vitamin K. PL may also affect the clearance of CF. There appears to be a concomitant decrease in CF as the result of lipid lowering therapies. While investigating single nucleotide polymorphisms (SNP) of the Factor IX (fIX) gene as potential quantitative trait loci (QTL) for fIX activity level (fIX:C) in the GAIT study, we observed an apparent relationship between two SNPs, G-793A and C-698T, with plasma cholesterol concentration. As expected, a similar relationship existed with low density lipoprotein (LDL) concentrations. In the current study, we examine these fIX promoter SNPs, which exhibit near perfect linkage disequilibrium (ρ=0.998, p<0.0001), as a possible QTL for cholesterol and LDL. We performed variance components analyses using Solar. We observed significant (p<0.05) marginal associations between LDL and factor VIII activity level (fVIII:C), fIX:C, cholesterol, VLDL, triglycerides, and BMI. We observed significant marginal associations between cholesterol and fVIII:C, fIX:C, LDL, VLDL, triglycerides, lipoprotein A (Lp(a)), and BMI. There was no association for smoking status or contraceptive use. For LDL, the most parsimonious model included age, age2, sex, age2 by sex, fIX:C, and the SNP covariate (coded −1,0,1). C(−698)T was significantly associated with LDL levels (nominal p=0.0262) and these covariates accounted for 33% of the total variance of LDL, with a similar result for G(−793)A. For cholesterol, the most parsimonious model included age, age2, fIX:C, Lp(a), and the SNP covariate. C(−698)T was significantly associated with LDL levels (nominal p=9.54 X 10−3) and these covariates accounted for 40% of the total variance of cholesterol, with a similar result for G(−793)A. Interestingly, these SNPs showed no association with fIX:C. Even though an increase in fIX:C results in an increase in cholesterol in this study, it is the same unit increase regardless of which allele is present, i.e. there is no interaction. The results of a linkage analysis for markers on the X chromosome were insignificant. Though this finding is intriguing and interesting with regard to the correlation between the vitamin K-dependent coagulations factors and lipids, it is entirely possible that it is by chance alone. These findings and the correlation between lipids and the vitamin K-dependent coagulation factors merit verification in an independent study population.

Author notes

Corresponding author