Abstract 3137: Failure of the Chromosome 9p21 Locus to Predict Myocardial Infarction when Non-Event Patients’ Coronary Disease Phenotype is Angiographically-Characterized
A chromosome 9p21 locus is associated with coronary heart disease (CHD) phenotypes in at least 25 independent populations, but multiple clinically-distinct phenotypes have been evaluated. Utilizing angiographic coronary artery disease (CAD) phenotyping, this study evaluated whether 9p21 single nucleotide polymorphisms (SNPs) predict ischemic events (e.g., myocardial infarction [MI]) among CAD patients. Two populations of patients undergoing coronary angiography during 1994 –2007 (population 1: N=1,748; population 2: N=1,014) were evaluated for association of a 9p21 tagging SNP (rs2383206, A–>G) with a composite MI/death endpoint vs. no event among patients with angiographically-significant CAD. Secondary analyses evaluated rs2383206 in two additional angiographic populations that included both CAD and non-CAD patients (population 3: N=2,122; population 4: N=1,466) for the phenotypes of MI vs. CAD/no MI, MI vs. no CAD, and CAD/no MI vs. no CAD. Post hoc analysis evaluated three other 9p21 SNPs. No association of rs2383206 was found with MI/death in populations 1 (OR=0.95 per G allele, p-trend=0.48) and 2 (OR=0.91 per G allele, p-trend=0.28), or with MI vs. CAD/no MI in populations 3 (OR=0.96 per G allele, p-trend=0.57) and 4 (OR=0.89 per G allele, p-trend=0.21). Among CAD patients, no association of rs2383206 was found with Duke CAD Index, number of significantly diseased vessels, left main disease, ostial disease, or proximal disease in the four populations. In contrast, rs2383206 was associated with CAD/no MI compared with no CAD (population 3: p-trend=0.0001; population 4: p-trend=0.0008). These findings were confirmed for rs2383207, rs10757274, and rs10757278. The chromosome 9p21 locus was not associated with MI or a composite of MI/death, or with extent or location of CAD, although it did predict CAD presence. This suggests that 9p21 is involved in the initiation or facilitation of CAD, not CAD progression or the precipitation of acute events. These findings contradict prior reports of a 9p21 association with MI, likely due to differences in study design and phenotype quality. High-quality angiographic phenotyping would be a useful standard in genetic association studies of CHD.