Abstract
OBJECTIVES: Numerous studies have demonstrated that fasting serum glucose (FSG) levels and certain single-nucleotide polymorphisms (SNPs) are related to an increased risk of colorectal cancer (CRC); however, their combined effects are still unclear. METHODS: Of a total of 144,527 men and women free of cancer at baseline, 317 developed CRC during 5.3 years of follow-up. A case-cohort study (n=1,691) was used, consisting of participants with a DNA sample available. Three well-known SNPs (rs3802842, rs6983267, rs10795668) were genotyped. Hazard ratios (HR) and 95% confidence intervals (CI) of CRC, colon and rectal cancer were calculated, with the Cox proportional hazard models. RESULTS: The crude incidence rates per 100,000 person-years were 41.1 overall, 48.4 for men, and 29.3 for women. Among participants with dysglycemia, SNPs rs3802842 and rs6983267 were both associated with an increased risk of CRC (HR, 3.2; 95% CI, 1.9 to 5.5 and HR, 1.8; 95% CI, 1.1 to 3.1, respectively) and rectal cancer (HR, 3.4; 95% CI, 1.8 to 6.6 and HR, 3.3; 95% CI, 1.6 to 7.1, respectively). The interaction effect of dysglycemia and SNPs was positive, that is, resulted in an elevated risk of CRC, but was not statistically significant. CONCLUSIONS: This study demonstrates that both high FSG and certain SNPs are major risk factors for CRC and rectal cancer but that they did not interact synergistically. The difference in effect size of the SNPs according to CRC subtype (i.e., colon or rectal cancer) and presence of dysglycemia merits further research.