Abstract 1807: DNA Sequence Variants Related to Plasma Triglycerides or High-Density Lipoprotein Cholesterol and Risk of Myocardial Infarction
BACKGROUND: In epidemiologic studies, plasma triglycerides (TG) and high-density lipoprotein cholesterol (HDL-C) are consistently associated with risk for myocardial infarction (MI) but it is unclear if these associations are causal. If plasma TG or HDL-C were causal factors, one might expect genotypes to alter MI risk in a manner predicted by their effect on plasma TG or HDL-C.
METHODS: We studied thirty genetic loci that we recently identified for plasma lipids using genome-wide association mapping. For single nucleotide polymorphisms (SNPs) associated with plasma TG or HDL-C and for combinations of SNPs as genotype scores, we estimated in a population-based cohort:
the change in plasma lipid level due to genotypes; and
the hazard for MI conferred by that degree of difference in plasma lipid level.
This predicted risk was compared with that observed when SNPs were tested for association with MI in up to 15,133 MI cases and 15,133 controls.
RESULTS: We validated the statistical framework using a score of SNPs related to LDL-C, a proven causal risk factor for MI. LDL-C genotype score increased plasma LDL-C by 19 mg/dL, leading to a predicted increase in MI risk of 45%. The observed risk for LDL-C genotype score was a 60% increase (95% CI 43% −78%, P <5 × 10−8). In contrast, whereas TG genotype score increased plasma triglycerides by 23 mg/dL and led to a predicted increase in MI risk of 73%, the observed risk for TG genotype score was only 26% (95% CI 13% - 40%, P<0.001). HDL-C genotype score lowered plasma HDL-C by 9 mg/dL and led to a predicted increase in MI risk of 58%; however, the observed risk for HDL-C genotype score was a 7% increase (95% CI 3% decrease - 19% increase, P > 0.05). In individual SNP analysis, SNPs at 8 of 9 loci predominantly related to plasma LDL-C were associated with MI risk (P < 0.05); however, of 19 loci predominantly related to TG or HDL-C, SNPs at only 6 loci were associated with MI risk (P < 0.05).
CONCLUSIONS: In contrast to LDL-C, the risk for MI conferred by genotypes does not track simply with their effect on plasma TG or HDL-C. Going forward, a key challenge is to identify genes that not only alter plasma TG or HDL-C but also affect risk for clinical events like MI.