Abstract
INTRODUCTION: Ultra-processed food (UPF) intake is linked to an increased risk of death, but underlying molecular mechanisms are largely unknown. We examined associations of a previously derived UPF metabolomic signature with all-cause, cancer, and cardiovascular disease (CVD) deaths, and whether the signature potentially mediates these associations.</p>
METHODS: We used baseline data from 63,401 participants in the UK biobank cohort and examined the association of a metabolomic signature of UPF intake with mortality over a maximum of 15 years of follow-up, accruing 2936, 1553, and 420 all-cause, cancer, and CVD mortality events. The metabolomic signature was computed as a weighted sum of previously identified 17 UPF-related serum metabolites. Associations of the metabolomic signature with mortality outcomes were examined using confounder-adjusted flexible parametric survival models. The proportion of the association of UPF with mortality mediated by the metabolomic signature was also computed.</p>
RESULTS: A standard deviation increase in the metabolomic signature was associated with 1.17 (95% confidence interval (CI), 1.13-1.21), 1.14 (1.08-1.20), and 1.15 (1.05-1.27)-times increased risk of all-cause, cancer, and CVD mortality, respectively. Comparing the highest vs. lowest quartiles, HR (95% CI) for all-cause, cancer, and CVD mortality were 1.46 (1.31-1.63), 1.40 (1.2-1.62), and 1.40 (1.05-1.85) respectively. The signature mediated 30%, 28%, and 21% of the association between UPF intake all-cause, cancer, and CVD mortality respectively. Component analyses showed that docosahexaenoic acid was robustly inversely associated with all-cause and CVD mortality, but glycoprotein acetyl was positively associated with all-cause and cancer mortality. Valine was inversely associated with all-cause mortality.</p>
CONCLUSION: The metabolomic signature of UPF is a key prognostic marker for mortality and potentially mediates UPF-mortality associations through systemic inflammatory pathways.</p>