IARC 60th Anniversary - 19-21 May 2026
Session : Rapid Fire
Ultra-Processed Foods, Food Additives & Biomarkers of Inflammation, Oxidative Stress, & Metabolic Profile: Cross-sectional Analyses in NutriNet-Santé
HASENBÖHLER A. 1, PAYEN DE LA GARANDERIE M. 1, DECHAMP N. 1, RINALDI S. 2, LANDRIER J. 3, DIAZ C. 4, SZABO DE EDELENYI F. 1, AGAËSSE C. 1, DE SA A. 1, HUYBRECHTS I. 2, HERCBERG S. 1, GALAN P. 1, DESCHASAUX-TANGUY M. 1, SROUR B. 1, TOUVIER M. 1
1 Université Sorbonne Paris Nord and Université Paris Cité, INSERM, INRAE, CNAM, Centre of Research in Epidemiology and StatisticS (CRESS), Nutritional Epidemiology Research Team (EREN), Bobigny, France; 2 International Agency for Research on Cancer, World Health Organization, Lyon, France; 3 Centre de recherche en cardiovasculaire et nutrition (C2VN), Aix Marseille Université, INSERM 1263 – INRAE 1260, Marseille, France; 4 Fundación MEDINA, Centro de Excelencia en Investigación de Medicamentos Innovadores en Andalucía. Parque Tecnológico de Ciencias de la Salud, Granada, Spain
Background
Epidemiological and experimental studies on ultra-processed foods (UPFs) and certain food additives highlighted associations between the consumption of these foods/compounds and cancer risk. Mechanistic epidemiology can help identify the underlying mechanisms and gather evidence toward causality.
Objectives
We aimed to investigate the cross-sectional associations between, on the one hand, UPFs and, for the first time, a wide range of food additives (single substances and mixtures), on the other hand, blood biomarkers of inflammation, oxidative stress, and metabolic profile.
Methods
This study included 4,200 adults from the NutriNet-Santé cohort who participated in a clinical-biological component with blood sampling between 2011 and 2014. Exposure to UPFs and food additives was estimated using 24-hour dietary records (including the brand of processed foods consumed) collected within the two years before blood draw. These consumption data were linked to comprehensive food composition databases and ad-hoc laboratory analyses in food matrices. Non-Negative Matrix Factorisation identified food additive mixtures. To assess non-parametric partial correlations, residuals of biomarker levels and exposures to UPF and additive —each adjusted for known or suspected confounders—were first calculated. Spearman correlations were then calculated between these residuals.
Results
168 food additives were consumed by at least one participant of our study population. Among those, 56 were consumed by at least 10% of participants and were then individually studied. On average, participants consumed 30.6% (SD=12.1) UPF (% energy intake).
Higher exposures to UPFs were associated with higher levels of CRP (inflammation), triglycerides, and secondary bile acids (deoxycholic acid and glycoursodeoxycholic acid), and lower levels of LysoPC(15:0) and LysoPC(17:0). These results suggest a consistent metabolic signature of UPF consumption, with increased inflammation/higher levels of pro-inflammatory bile acids and lower levels of LysoPC, which may reflect the replacement of healthier foods with UPFs and metabolic disruption. Numerous associations were also detected regarding the studied food additive exposures. Notably, the analysis of food additive mixtures revealed distinct metabolic and inflammatory profiles, characterised by positive associations with interleukin-10, acylcarnitines and lysophospholipids, alongside both inverse associations with malonylcarnitine, cortisone, tryptophan-betaine, and glycaemia.
Conclusions/Implications
This new study allows us to identify/suspect certain mechanisms of action that may underlie the associations identified in large cohort studies. These results also allow us to generate hypotheses and guide future studies exploring the causal relationships between exposure to UPFs, to specific food additives and food additive mixtures, and cancer incidence.