Fig. 1.

Associations of ultra-processed food (UPF) intake with percent body fat defined adiposity and low skeletal muscle mass. Multinomial logistic regression models were used to estimate odds ratios (ORs) and their corresponding 95% confidence intervals (CIs) for the tertile (T) 3 of percent body fat and the T1 of percent appendicular skeletal muscle mass (ASM) comparing quartile (Q) 2, 3, and 4 to Q1 of UPF intake as the exposure variables (T3 of percent body fat: ≥24.5% for male, ≥36.0% for female; T1 of percent ASM: <30.7% for male, <24.1% for female). P for trends was determined by treating the median value of UPF intake as a continuous variable using multinomial logistic regression models. A 10% increase in UPF intake was used to estimate ORs for higher adiposity or lower ASM. A multivariable-adjusted model was adjusted for age, sex, residential area, education level, monthly household income level, marital status, current smoking, current drinking, walking exercise, weight training, and total energy intake.