Abstract 2184: Prospective Evaluation of a Comprehensive Genetic Risk Score for Future Cardiovascular Events Among 19,313 Initially Healthy Women: The Women’s Genome Health Study
Background: While multiple genetic markers associated with CVD have been identified by genome wide association studies (GWAS), their aggregate effect on risk prediction beyond traditional factors is uncertain, particularly among women where data are sparse.
Methods and Results: Using the NHGRI catalog of GWAS results published between 2005 and April 2009, we identified 80 uncorrelated SNPs (pairwise r2 less than 0.5) reported to associate with CVD events or at least one traditional CVD risk factor with a published p-value less than 10−7. These 80 SNPs were then used to construct a genetic risk score (GRS) that was prospectively evaluated among 19,313 initially healthy Caucasian women followed over a 12 year period for incident vascular events. In crude analysis, the GRS had a hazard ratio (HR) of 1.02 per affected risk allele (95%CI 1.01–1.03) such that women with an above median GRS (> 79 alleles) were at significantly higher risk of future vascular events than women with a below median GRS (Figure⇓) (HR 1.26, 95%CI 1.10 –1.45). However, after adjustment for factors included in either the Framingham or Reynolds Risk Scores, the GRS was no longer significantly associated with risk, nor did it improve indices of either discrimination or reclassification with a change in the c-index of 0.007 (95% CI −0.001, 0.001) and a Net Reclassification Improvement 0.5% (p = 0.4).
Conclusions: In this large scale prospective evaluation of women, a comprehensive GRS derived from previously published GWAS data was associated with increased vascular event rates, but did not improve global risk prediction models based on currently available traditional factors.