Abstract 1028: Association of plasma C-reactive protein (CRP) and CRP genetic risk score with cancer risk in the Atherosclerosis Risk in Communities (ARIC) study

Abstract

Abstract Background: Chronic inflammation is related to carcinogenesis. Many studies reported a positive association between plasma CRP - a non-specific marker of systemic inflammation - and colorectal cancer (CRC) risk. In the ARIC study, hazard ratio (HR) for CRC was 1.97 (95% CI,1.13; 3.43) for the highest versus lowest CRP quartile, although it is unclear if the association is causal. The associations of CRP with other cancer types are inconsistent. Our goals were to prospectively 1) examine associations of plasma CRP with other cancer types and total cancer and 2) examine associations of a genetic risk score based on single nucleotide polymorphisms (SNPs), affecting CRP levels, with cancer risk. Methods: The ARIC cohort aged 45-64 y at Visit 1 in 1987-89 (55% women, 28% African-American and 72% Caucasian) was followed until 2006. Among 8,657 Caucasians, we examined the incidence of total (n=1,929), CRC (n=205), breast (n=371), lung (n=274), and prostate (n=395) cancers in relation to a weighted genetic CRP risk score (CRPRS) composed of 20 SNPs located in/near CRP, APOC1, GCKR, HNF1A, IL6R and 15 other genes. These SNPs were associated with CRP levels in a previous meta-analysis of genome-wide association studies. The CRPRS was created as a sum of risk alleles for each person by multiplying each SNP by the beta coefficient from the meta-analysis. Plasma CRP was measured at Visit 4 (1996-98). We examined associations of plasma CRP with the risk of total (n=1,481), breast (n=235), lung (n=185), and prostate (n=352) cancers among 9,857 Caucasians and African-Americans, followed from Visit 4 until 2006. Cox proportional hazards models were used to estimate HRs of cancers in relation to plasma CRP and the genetic score. Results: CRPRS was positively associated with CRC risk HR =1.18 (1.03; 1.36) per one standard deviation of the score but not with total or any other cancer in a multivariate analysis. Additional adjustment for plasma CRP only slightly attenuated the association. The HRs were similar for colon and rectal cancer. After multivariate adjustment, plasma CRP was associated positively with the risk of breast and lung cancers: 2.40 (95%CI, 1.25; 4.61) (p-trend=0.004) and 1.90 (95%CI, 1.12; 3.22)(p-trend=0.03), respectively, for the highest versus lowest quartiles. The association for breast cancer did not markedly change after excluding subjects with < 3 years of follow-up or CRP>10mg/L, whereas for lung cancer both exclusions resulted in weaker associations. Conclusions: This study provides the first evidence of a significant association of CRP polymorphisms with CRC risk suggesting a role of CRP and inflammation in colorectal carcinogenesis. CRP polymorphisms may be linked with CRC risk through plasma CRP or other mechanisms. We also found that not only colorectal but also breast and lung cancers are positively associated with plasma CRP in the ARIC study. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 1028. doi:1538-7445.AM2012-1028