Abstract 235: Type 2 Diabetes and Glycemic Traits in Relation to Colorectal Cancer Risk: A Mendelian Randomization Study
Mingyang Song,Yiwen Lu,Marc Gunter,Neil Murphy,Barbara L. Banbury,Wenjie Ma,Jennifer Prescott,Graham Casey,Stephen B. Gruber,Edward L. Giovannucci,Ulrike Peters,Andrew T. Chan
DOI: https://doi.org/10.1158/1538-7445.am2018-235
IF: 11.2
2018-01-01
Cancer Research
Abstract:Background: Although type 2 diabetes (T2D) has been linked to an increased risk of colorectal cancer (CRC), a causal relationship remains to be established. Also, the influence of hyperglycemia and impaired beta-cell function on CRC remains to be determined. Methods: We performed a Mendelian randomization (MR) study using common single nucleotide polymorphisms (SNPs) identified from prior genome-wide association studies for T2D (78 SNPs), fasting glucose (36 SNPs), glycosylated hemoglobin (HbA1c, 11 SNPs), and homeostatic model assessment for beta-cell function (HOMA-B, 13 SNPs) (all SNPs reached genome-wide significance at 5 x10-8). We first assessed the association of each SNP with CRC risk using logistic regression in 26,357 CRC cases and 20,505 control participants in the Colon Cancer Family Registry, Colorectal Transdisciplinary Study and the Genetics and Epidemiology of Colorectal Cancer Consortium. The MR estimates for the association of each risk factor with CRC were calculated using the inverse-variance weighted method that combines summary data on the association of the SNPs with each trait derived from previous genome-wide association studies. Given the potential sex difference in the relationship of metabolic factors with CRC, all analyses were performed in men and women separately. Results: We did not find any statistically significant association between each of the genetically predicted risk factors and CRC incidence for sex-stratified or combined analyses. The odds ratios (ORs) of CRC associated with genetic risk of T2D were 1.02 (95% confidence interval [CI], 0.95-1.08, P=0.62) in women and 0.96 (95% CI, 0.90-1.03, P=0.25) in men. For fasting glucose, the ORs of CRC associated with one mmol/L increment was 0.91 (95% CI, 0.68-1.20, P=0.49) in women and 1.20 (95% CI, 0.90-1.59, P=0.22) in men, respectively. For HbA1c, the ORs of CRC associated with one percent increment was 0.83 (95% CI, 0.57-1.22, P=0.35) in women and 0.99 (95% CI, 0.67-1.45, P=0.95) in men. For HOMA-B, the ORs of CRC associated with a unit increase in log-transformed measurement was 0.92 (95% CI, 0.53-1.60, P=0.76) in women and 0.77 (95% CI, 0.44-1.35, P=0.36) in men. Conclusion: Our findings do not support a causal association of T2D, impaired beta-cell function, and hyperglycemia with CRC risk. However, the null findings may be due to weak instrument bias since these common SNPs account for only a limited proportion of the inter-individual variation in these glycemic traits. Citation Format: Mingyang Song, Yiwen Lu, Marc Gunter, Neil Murphy, Barbara L. Banbury, Wenjie Ma, Jennifer Prescott, Graham Casey, Stephen B. Gruber, Edward L. Giovannucci, Ulrike Peters, Andrew T. Chan. Type 2 diabetes and glycemic traits in relation to colorectal cancer risk: A Mendelian randomization study [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 235.