Recent epidemiological studies have indicated that baseline C-reactive protein (CRP) levels may have value in prediction of cardiovascular risk. Using six tag single-nucleotide polymorphisms (SNPs) selected from our complete list of SNPs on the CRP gene, we investigated the association of CRP genotypes with plasma CRP levels and cardiovascular risk in the National Heart, Lung, and Blood Institute (NHLBI) Family Heart Study cohort (1,296 Caucasians, 48.5% male, 54.7 +/- 12.8 yr old). There was a significant trend toward association of CRP haplotypes with CRP levels (P = 0.045). SNP analysis indicated a highly significant association of SNP -757 (rs3093059, P = 0.0004) and SNP -286 (rs3091244, P = 0.0065) and a borderline association of SNP -7180 (rs1341665, P = 0.06) with CRP levels. Neither CRP haplotypes nor individual SNP genotypes were associated with intima-media thickness of the common carotid or internal carotid artery or the bifurcation of the carotid arteries. These results indicated a strong impact of local SNPs of the CRP gene on plasma CRP levels, but there was no direct evidence that these genetically controlled CRP elevations by local CRP SNPs contributed to cardiovascular disease phenotypes.