Fig. 3

Smoothing curves between dietary metrics and accelerated HD, AL, KDM, and PA. Binomial logistic regression models and RCS were performed with adjusting for age, sex, ethnicity, year, BMI, smoking, exercise, education, income, nutrient supplement use, self-reported cancer, cardiovascular diseases, hypertension, diabetes, and chronic kidney disease. Models for DASH and HEI-2015 were additionally adjusted for drinking. The solid black lines correspond to the central estimates, and the gray-shaded regions indicate the 95% confidence intervals