Gene-environment interaction of coffee with body mass index in multiple populations

Manu Shivakumar1, Dokyoon Kim1§, Eun Kyung Choe1,2§ 1 Department of Biostatistics, Epidemiology & Informatics, Perelman School of Medicine, University of Pennsylvania, Philadelphia, USA 2 Department of Surgery, Seoul National University Hospital Healthcare System Gangnam Center, Seoul, Korea

Obesity is a global pandemic disease, and its prevalence has tripled in the last four decades. A flurry of genome-wide association studies (GWAS) has found many variants associated with body mass index (BMI) and subsequently, polygenic risk scores (PRS-BMI) derived from them have shown the genetic basis and risk of obesity. Additionally, interactions between genetic factors and environmental risk factors like lifestyle and diet, such as coffee consumption, also play an important role in obesity. In this study, we sought to investigate interactions between PRS-BMI and coffee consumption in the European population (UK Biobank, N = 218,777) and the Korean population (GENIE, N = 5,501). GAINT consortium and Biobank Japan GWAS summary statistics were used to calculate BMI PRS for UKBB and GENIE populations, respectively. Subjects who consumed less than or equal to 1 cup of coffee per day were considered as control and greater than 1 cup of consumption as cases. Log-rank test was used to check the significance of the interaction between PRS-BMI and coffee consumption, adjusting for age, sex, IPAQ physical activity score, and principal components. Further, the subjects were stratified into BMI categories - BMI <= 18.5 (underweight), 18.5 < BMI <= 25 (normal) and BMI > 25 (overweight & obese) the PRS BMI-coffee interaction was tested in the subgroups. The interaction between coffee and PRS BMI was significant (UKBB P = 2.65x10-7 and GENIE P = 0.03) in both populations. The interaction was significant even after adjusting for coffee metabolism score. In the subgroup analysis, the interaction was significant in BMI > 25 (P = 0.003) and BMI > 30 (P = 0.027) groups in UKBB. The findings in this study suggest that there is a significant interaction between BMI-PRS and coffee in different populations. In the case of the European population, there is a significant interaction in overweight and obese subjects. The results from this study may be applicable for personalized medicine and preventive treatment through risk assessments and interventions.