Fifteen years of statin pharmacogenetics: A systematic review

Purpose: Statins lower LDL cholesterol (LDLc) and thereby reduce the risk of cardiovascular events. There has been a considerable amount of research into genetic variation influencing statin response. We evaluated the evidence of these pharmacogenetic associations. Methods: PubMed was systematically searched to retrieve publications reporting on the pharmacogenetics of statins. Two separate outcomes were considered of interest: modification of LDLc response and modification of risk for cardiovascular events; publications investigating other outcomes were excluded. Results: 597 publications were identified in the initial search, of which 139 remained after applying the exclusion criteria. The remaining publications varied widely on design, statistical procedures, selection of participants, and reporting of results. SNPs in CETP, APOE, HMGCR, SLCO1B1, ABCB1 and LDLR were investigated most often in the candidate gene studies, with 15 or more publications each. In these candidate gene studies, more than 150 loci were claimed to be associated with statin response, but less than 10% of these associations were positively replicated and none of these showed conclusive evidence. For example, a SNP in KIF6 (rs20455) was shown to affect clinical statin benefit, and successfully replicated once, but 8 subsequent studies did not find this association. Five genome-wide association (GWAS) studies were performed, all investigating modification of the lipid-lowering effect. SNPs in APOE, LPA and ABCG2 reached genome-wide significance and were replicated. Effect sizes found in GWAS were modest, approximately 2-3% of the total LDLc-response to statins. Conclusions: While many papers have been published on statin pharmacogenetics, there are only three associations of SNPs with LDLc-response to statins with strong evidence. None of the investigated SNPs consistently affect clinical outcome. A more standardized way of reporting results will facilitate efforts to combine evidence. Larger samples and meta-analysis of genome-wide association studies might increase the number of associated loci. However, as effect sizes found thus far are generally small, at this moment there is no role for genetic testing in clinical practice to guide statin treatment.