Dietary Patterns and Prevalence Odds Ratio in Middle-aged Adults of Rural and Mid-size City in Korean Genome Epidemiology Study

ABSTRACT

Recently, dietary pattern analysis was emerged as an approach to examine the relationships between diet and risk of chronic diseases. This study was to identify groups with population who report similar dietary pattern in Korean ge-nome epidemiology study (KoGES )and association with several chronic diseases. The cohort participants living in Ansung and Ansan (Gyeonggi province )were totally 10,038. Among those, 6,873 subjects with no missing values in food frequency questionnaire were included in this analysis. After combining 103 food items into 17 food groups, 4 dietary factors were obtained by factor analysis based on their weights. Factor 1 showed high factor loadings in vege-tables, mushrooms, meats, fish, beverages, and oriental-cereals. Factor 2 had high factor loadings in vegetables, fruits, fish, and factor 3 had high factor loadings in cereal-oriental, cerial-western and snacks. Factor 4 showed positive high factor loadings in rice and Kimchi and negative factor loadings in mushrooms and milk and dairy products. Using factor scores of four factors, subjects were classified into 3 clusters by K-means clustering. We named those 'Rice and Kimchi eating' group, 'Contented eating' group, and 'Healthy and light eating' group depending on their eating cha-racteristics. 'Rice and Kimchi eating' group showed high prevalence in men, farmers and 60s. 'Contented eating' group and 'Healthy and light eating' group had high prevalence in women, people living in urban area (Ansan Citizen ), with high-school education and above, and a monthly income of one million won and more. 'Contented eating' group appeared lower distribution proportion in the sixties and 'Healthy and light eating' group does higher in the fifties. 'Contented eating' versus 'Rice and Kimchi eating', odds ratio for hypertension, diabetes, metabolic syndrome and obesity significantly decreased after adjusting age and sex (OR =0.64, 0.73, and 0.85 respectively, 95% CI ). Although our results were from a cross-sectional study, these imply that the dietary patterns were related to diseases.