C-reactive protein and the risk of cancer: a mendelian randomization study

TY - JOUR

T1 - C-reactive protein and the risk of cancer: a mendelian randomization study

AU - Allin, Kristine H

AU - Nordestgaard, Børge G

AU - Zacho, Jeppe

AU - Tybjaerg-Hansen, Anne

AU - Bojesen, Stig E

N1 - Keywords: Adult; Aged; C-Reactive Protein; Cross-Sectional Studies; Denmark; Female; Genotype; Humans; Inflammation; Male; Mendelian Randomization Analysis; Middle Aged; Neoplasms; Odds Ratio; Polymorphism, Single Nucleotide; Predictive Value of Tests; Proportional Hazards Models; Prospective Studies; Risk Assessment; Risk Factors; Tumor Markers, Biological

PY - 2010/2

Y1 - 2010/2

N2 - Elevated plasma levels of C-reactive protein (CRP), a marker of inflammation, are associated with an increased risk of cancer, but it is unclear whether this association is causal. We examined whether four common single-nucleotide polymorphisms (SNPs) in the CRP gene that are associated with altered plasma CRP levels are causally associated with an increased risk of cancer. The study population included participants in a prospective study (n = 10215) and a cross-sectional study (n = 36403) of the adult general population in Denmark, all of whom were genotyped for the CRP SNPs. The association between plasma CRP levels measured by a high-sensitivity turbidimetry assay and the risk of cancer was examined for 8224 participants in the prospective study. The hazard ratio of cancer for a doubling of the plasma CRP level was 1.09 (95% confidence interval [CI] = 1.03 to 1.14). The nine most common genotype combinations of the four CRP SNPs were associated with up to a 72% increase (95% CI = 58% to 87%) in CRP levels but not with an increased risk of cancer. The estimated causal odds ratio for cancer associated with a genetically induced doubling in CRP level was 0.94 (95% CI = 0.81 to 1.08). This finding suggests that elevated CRP levels do not cause cancer.

AB - Elevated plasma levels of C-reactive protein (CRP), a marker of inflammation, are associated with an increased risk of cancer, but it is unclear whether this association is causal. We examined whether four common single-nucleotide polymorphisms (SNPs) in the CRP gene that are associated with altered plasma CRP levels are causally associated with an increased risk of cancer. The study population included participants in a prospective study (n = 10215) and a cross-sectional study (n = 36403) of the adult general population in Denmark, all of whom were genotyped for the CRP SNPs. The association between plasma CRP levels measured by a high-sensitivity turbidimetry assay and the risk of cancer was examined for 8224 participants in the prospective study. The hazard ratio of cancer for a doubling of the plasma CRP level was 1.09 (95% confidence interval [CI] = 1.03 to 1.14). The nine most common genotype combinations of the four CRP SNPs were associated with up to a 72% increase (95% CI = 58% to 87%) in CRP levels but not with an increased risk of cancer. The estimated causal odds ratio for cancer associated with a genetically induced doubling in CRP level was 0.94 (95% CI = 0.81 to 1.08). This finding suggests that elevated CRP levels do not cause cancer.

U2 - 10.1093/jnci/djp459

DO - 10.1093/jnci/djp459

M3 - Journal article

SN - 0027-8874

VL - 102

SP - 202

EP - 206

JO - National Cancer Institute. Journal (Print)

JF - National Cancer Institute. Journal (Print)

IS - 3

ER -