Effect of Genetic Predisposition and Gene-Environment Interaction on the Incidence of Type 2 Diabetes Mellitus in a Community-based Cohort Study

Abstract

Introduction: Contribution of genetic predisposition to risk prediction of type 2 diabetes mellitus (T2DM) was investigated, and genotype variation was tested for gene-environment interaction effect using a prospective study in middle-aged adults in South Korea. Methods: From a community cohort of 6,910 subjects with 8 years follow-up, genetic predisposition score with subsets of 4, 16, 36 selected single nucleotide polymorphisms (SNP) (genetic predisposition score, GPS) in association with T2DM were determined, and their effect was evaluated using risk prediction models. Also, genetic and environmental factors were tested for their interaction effect on T2DM risk. Results: Sixteen SNPs were found to be in significant association with T2DM in the study population, and hazard ratios of GPSs above median risk allele scores were 1.19 (95% confidence intervals: 1.04-1.36), 1.23 (1.08-1.41), 1.42 (1.19-1.70) with GPS-4, GPS-16, GPS-36, respectively. Significant changes in C-statistics for discrimination were observed in both subject groups tested with GPS-16 and GPS-36, and while significant net reclassification index (NRI) in subpopulations weakened upon addition of glycated hemoglobin (HbA1c) in subjects with GPS-4, those with information on GPS-16 and GPS-36 remained robust. Significant positive gene-environment interactions by additional scales were observed for GPS-16 and GPS-36 with environmental variables such as family history of DM, high-density lipoprotein cholesterol and HbA1c, and multiplicative scales were also significant for family history of DM. Conclusions: From a cohort of middle-aged Koreans, subjects with genetic predisposition information on multiple numbers of SNPs showed significant model discrimination and reclassification, even after adjusting for HbA1c, a stronger predictor for T2DM risk. For gene-environment interaction, although some effect is suggested from the results of the current study, further investigations with larger number of subjects or in an independent population is suggested.