TY - JOUR
T1 - Dietary patterns and the risk of diabetes in Korean adults
T2 - A cross-sectional and prospective cohort study
AU - Hwang, Jiyoung
AU - Kim, Hyesook
AU - Kwon, Oran
N1 - Publisher Copyright:
© 2024
PY - 2024/9
Y1 - 2024/9
N2 - Objective: The aim of this study was to identify dietary patterns associated with diabetes in Korean adults and to investigate their association with diabetes risk in both a cross-sectional and prospective study. Methods: Predefined food groups collected by the Korea National Health and Nutrition Examination Survey (KNHANES 2015–2018, n = 19 721) were entered in a reduced rank regression (RRR) model, followed by stepwise linear regression analyses to identify the most predictive dietary patterns. We evaluated the construct validity of dietary patterns in two independent samples from KNHANES 2019 to 2021 (n = 14 223) and the Health Examinees (HEXA) cohort study (n = 30 013). Associations between dietary patterns and diabetes risk were examined using multivariable regression and multivariable-adjusted Cox proportional hazard models, respectively. Results: A dietary pattern was identified with high positive loadings for refined white rice, kimchi and salted vegetables, wheat flour and bread, and seasonings, and high negative loadings for whole grains, legumes with tofu and soymilk, poultry, eggs, and plant oils. The higher pattern scores were significantly associated with diabetes risk in KNHANES 2015 to 2018 (male: odds ratio [OR]: 1.59; 95% confidence interval [CI]: 1.35, 1.88; female: OR: 1.37; 95% CI: 1.18, 1.52), KNHANES 2019 to 2021 (male: OR: 1.47; 95% CI: 1.01, 1.69; female: OR: 1.37; 95% CI: 1.18, 1.54), and HEXA study (male: hazard ratio [HR]: 1.10; 95% CI: 1.01, 1.34; female: HR: 1.24; 95% CI: 1.02, 1.52). Conclusions: Dietary patterns derived by RRR followed by stepwise linear regression analyses were associated with increased risks of diabetes among Korean adults.
AB - Objective: The aim of this study was to identify dietary patterns associated with diabetes in Korean adults and to investigate their association with diabetes risk in both a cross-sectional and prospective study. Methods: Predefined food groups collected by the Korea National Health and Nutrition Examination Survey (KNHANES 2015–2018, n = 19 721) were entered in a reduced rank regression (RRR) model, followed by stepwise linear regression analyses to identify the most predictive dietary patterns. We evaluated the construct validity of dietary patterns in two independent samples from KNHANES 2019 to 2021 (n = 14 223) and the Health Examinees (HEXA) cohort study (n = 30 013). Associations between dietary patterns and diabetes risk were examined using multivariable regression and multivariable-adjusted Cox proportional hazard models, respectively. Results: A dietary pattern was identified with high positive loadings for refined white rice, kimchi and salted vegetables, wheat flour and bread, and seasonings, and high negative loadings for whole grains, legumes with tofu and soymilk, poultry, eggs, and plant oils. The higher pattern scores were significantly associated with diabetes risk in KNHANES 2015 to 2018 (male: odds ratio [OR]: 1.59; 95% confidence interval [CI]: 1.35, 1.88; female: OR: 1.37; 95% CI: 1.18, 1.52), KNHANES 2019 to 2021 (male: OR: 1.47; 95% CI: 1.01, 1.69; female: OR: 1.37; 95% CI: 1.18, 1.54), and HEXA study (male: hazard ratio [HR]: 1.10; 95% CI: 1.01, 1.34; female: HR: 1.24; 95% CI: 1.02, 1.52). Conclusions: Dietary patterns derived by RRR followed by stepwise linear regression analyses were associated with increased risks of diabetes among Korean adults.
KW - Diabetes
KW - Dietary patterns
KW - Hybrid methods
KW - Reduced rank regression
KW - Stepwise linear regression analysis
UR - http://www.scopus.com/inward/record.url?scp=85195563057&partnerID=8YFLogxK
U2 - 10.1016/j.nut.2024.112491
DO - 10.1016/j.nut.2024.112491
M3 - Article
AN - SCOPUS:85195563057
SN - 0899-9007
VL - 125
JO - Nutrition
JF - Nutrition
M1 - 112491
ER -