Genome-Wide Association Study of CRP Response to Statin Therapy in Diabetes in the CARDS Trial

Reduction of C reactive protein (CRP) may underlie part of the benefi ts of statins in people with Type 2 diabetes (T2D), but whether this is through different pathways than low-density lipoprotein cholesterol (LDL-C) reduction is unclear. Our aim was to examine the variability in statin induced change in CRP in diabetes and to seek genetic determinants of this change. CRP was measured at baseline and one year in T2D patients in the CARDS trial. ~2.5 million SNPs typed on Perlegen platform and imputed with HAPMAP 2 were used. Multivariate linear regression was used to model change from baseline to 1 year in CRP in 1005 patients allocated to Atorvastatin. Posttreatment (log)CRP level was used as the dependent variable with baseline CRP, age, sex, BMI, smoking, duration of diabetes and HbA1c as covariates. Genome-Complex Trait Analysis (GCTA) was used to identify narrow sense heritability of CRP response. The median (interquartile range) of change in CRP on statin was -0.12(-1.03, 0.69) mg/L with 45.3% of those on statin having no fall in CRP. In contrast LDL-C fell in 96.5% of people on statin. Narrow-sense heritability h2(±SE) for CRP response was 0.19(±0.24) compared to 0.05(±0.12) for LDL-C response. Previously identifi ed variants for LDL-C -response to statin (in genes APOe2, APOe4 and LPA) were not associated with CRP response (p>0.05) but three borderline GWAS signifi cant loci (p