Logistic regression models fitting results of the association between tea categories and cognitive impairment 1.

1Binary logistic regression analysis was used to calculate ORs and 95% CIs for tea categories related to cognitive impairment which assessed with CCM, with non-consumption group treated as reference.2 P value were tested by logistic regressions in which tea category was treated as categorical variable.3 Crude model.4 Adjusted for age, sex, race, education, marriage, tea consumption volume and tea concentration.5Adjusted for variables in model 2 plus physical examinations (BMI, WHR, SBP, DBP), family status (family income, have children or not) and disease situation (history of present illness and family history of hypertension, diabetes, CHD, AD, PD).6 Adjusted for variables in model 3 plus behavioral risk factors (cigarette smoking, alcohol consumption, and physical activities), dietary intake (vegetables, fruits, meat, fish, beans, milk).7Adjusted for variables in model 4 plus nutrition supplement, depression and ADL.Logistic regression models fitting results of the association between tea categories and cognitive impairment 1.