To the Editor: In their cohort study, Dr Paynter and colleagues1 examined the utility of cardiovascular risk panels of 12 and 101 genetic markers and concluded that they are not useful compared with conventional risk factors alone. However, there are 2 important issues with their interpretation.
First was the choice to examine the relationship between the single-nucleotide polymorphisms (SNPs) tested and the combined end point of coronary artery disease (CAD) or myocardial infarction (MI) and stroke, even considering only the 12 SNPs chosen for the cardiovascular disease (CVD) SNP panel. The authors mixed together markers that have been shown to correlate with different phenotypes. Some are associated with CAD/MI, while others (such as the 4q25 atrial fibrillation marker2) are associated with stroke. However, of all 101 markers used (SNPs previously reported to be associated with CVD or an intermediate phenotype), only 9p21 has been convincingly replicated for both stroke and CAD/MI.3,4 Using markers that are correlated with stroke but not MI will not be useful to predict risk for MI, and vice versa, and pooling different end points and risk variants together only dilutes the predictive power of the genetic risk panel.
Holm H, Thorleifsson G, Stefansson K. Genetic Risk Score and Cardiovascular Events in Women. JAMA. 2010;303(20):2032-2033. doi:10.1001/jama.2010.660