Ultra-Processed Food Consumption and Mortality: Three Cohort Studies in the United States and United Kingdom

Abstract

Given the increase in ultra-processed food (UPF) consumption, their potential health effects have aroused concern. Whether UPF consumption is associated with cancer and cardiovascular disease mortality is debatable. This study evaluates the association of UPF consumption with mortality. A total of 108,714 U.S. adults from the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial (1993-2001), 208,051 UK adults from UK Biobank (2006-2010), and 41,070 U.S. adults from National Health and Nutrition Examination Survey (1999-2018) were included. Dietary data were collected by dietary questionnaire and classified using the NOVA classification. UPF consumption was expressed as the weight proportion of UPFs in total foods consumed. Cox proportional hazard models were used to calculate hazard ratios and 95% CIs. Mediation analysis was used to evaluate whether multiple metabolic pathways mediated the associations in UK Biobank. Analyses were performed in 2022-2023. Combined analyses of the three cohorts showed that those with the highest quartile of UPF consumption had higher risks of all-cause mortality (hazard ratio, 1.16; 95% CI, 1.11-1.20) and cardiovascular disease mortality (hazard ratio, 1.17; 95% CI, 1.06-1.28) compared to the lowest quartile of UPF consumption. UPF consumption was not associated with cancer mortality risk. Biomarkers of liver function have the greatest mediating effects on all-cause mortality (20.3%), and biomarkers of inflammation have the greatest mediating effects on cardiovascular disease mortality (29.2%). Higher UPF consumption was associated with increased all-cause and cardiovascular disease mortality risk, with multiple metabolic pathways playing mediating roles.