Abstract
Background: Ultra-Processed Foods (UPFs) make up a growing share of older adults' diets and may contribute to frailty through pro-inflammatory pathways. The objective of this study was to examine the association of UPF intake with frailty development and with annual changes in select frailty components. Methods: This prospective cohort study used data from 2547 participants in the Framingham Offspring Cohort. UPF intake was assessed using a food frequency questionnaire and classified according to the NOVA framework, and frailty was defined by the Fried frailty phenotype. We used cumulative and mixed logistic regression models to examine the association between daily servings of UPF and odds of developing frailty, adjusting for baseline age, education, energy intake, multivitamin use, smoking, self-rated health, history of diabetes, cancer, cardiovascular disease, and diet quality. For the frailty component analysis, we used cumulative linear regression models to assess the association between UPF intake and annual changes in grip strength, gait speed, and weight, further adjusting for BMI and physical activity. We also evaluated potential effect modification by sex and baseline age (<60 vs. ≥60 years). Results: The study population was 55.1% female, with a mean age of 60.3 ± 8.9 years. Over an average follow-up of 10.8 years, 233 participants (9.2%) developed frailty. UPF intake was not associated with frailty development in either the cumulative or mixed regression models. UPF intake was not associated with annual weight change but was inversely associated with annual change in gait speed and with annual change in grip strength in men only. Conclusions: Our findings contribute preliminary evidence that, in middle-aged and older adults, increased UPF intake is not associated with frailty but may be related to worsening muscle strength and function. Further research with a more granular approach to UPF classification is required to translate these findings to practical recommendations and to clarify their clinical significance.