A Mendelian randomization study of type 2 diabetes and cancer risk in East Asians

Our research aims to explore genetic correlation between T2D predisposition and risks of several cancers, which have been predominantly focused on populations of European ancestry. In an East Asian population, we leverage two-sample Mendelian Randomization to investigate the complex association between Type 2 Diabetes (T2D) and cancer susceptibility. This investigation utilizes genetic data summarized from three reputable sources: the Japanese ENcyclopedia of GEnetic associations by Riken (JENGER), the Asian Genetic Epidemiology Network (AGEN), and the Meta Analyses of Glucose and Insulin-related traits (MAGIC). We explored the associations between exposure datasets, which included T2D, glycated hemoglobin (HbA1c) and fasting glucose (FG) levels, and the risk of several prevalent cancers for the outcome datasets. By analyzing 174 SNPs associated with T2D, 15 SNPs related to FG, and 74 SNPs linked to HbA1c, we discovered a significant inverse relationship between T2D and the majority of cancers, including gastric (OR = 0.875, 95%CI: 0.825-0.928), breast (OR = 0.907, 95%CI: 0.850-0.967), esophageal (OR = 0.761, 95%CI:0.681-0.851), colorectal (OR = 0.877, 95%CI:0.834-0.923), hematological malignancy (OR = 0.837, 95%CI:0.752-0.933), lung (OR = 0.916, 95%CI:0.858-0.977), hepatocellular (OR = 0.865, 95%CI:0.787-0.951), prostate (OR = 0.862, 95%CI:0.812-0.914), and endometrial cancer (OR = 0.841). The power of gastric cancer, breast cancer, esophageal cancer, colorectal cancer and prostate cancer reached the desired 80%. Conversely, we did not observe a reverse correlation between T2D and specific cancers in the East Asian population. There is limited evidence that genetically predicted FG and HbA1c levels are associated with specific cancer. Moreover, conducting multivariable MR analysis did not alter the correlation between T2D and specific tumors. These findings explain the causal associations of genetic liability to T2D with various types of cancers in East Asian ancestry, and this association is not affected by glycemic biomarkers.