Effets sur la santé de la diversité alimentaire : exploration du risque de cancer, signatures métabolomiques et diversité du microbiome

In an era of increasing dietary homogenization, this comprehensive research investigates the role of food biodiversity in human health. Dietary risks accounted for 178 million disability-adjusted life years and 7.94 million deaths among adults aged. Dietary diversity represents an important dimension of diet quality, measured at different levels of food classification, from food groups to individual items. Food biodiversity, an emerging concept, specifically examines the variety of biological species consumed, capturing both the nutritional benefits and environmental impacts of food systems. Today, human diets have shifted towards more limited diversity; however, dietary species richness (DSR), referring to the unique number of biological species in an individual’s diet (rather than the number of food groups) within a specific reference period, has shown an inverse association with both total mortality and cause-specific mortality. Despite this, evidence remains scarce regarding food biodiversity’s health benefits, particularly for cancer risk and physiological pathways. By investigating the relationships between dietary species richness, biological systems, and health outcomes, this thesis provides insights into how biodiverse diets influence physiological processes and cancer prevention. Additionally, living within complex ecosystems, while food biodiversity may influence human health systems (metabolome and microbiome), the gut microbiome could in turn also shape eating behaviours, creating a bidirectional relationship. The first objective of this thesis uncovered an inverse relationship between DSR and gastrointestinal cancer in the European population using the large European prospective investigation into Cancer and Nutrition (EPIC) cohort. Here, higher DSR was associated with lower overall gastrointestinal cancer risk, and specifically for esophageal squamous cell carcinoma and colorectal and proximal colon cancer, after accounting for socio-demographic, lifestyle, and other known dietary risk factors. In the second objective, we investigated the metabolic profile associated with greater DSR and its relationship with inflammatory and insulin resistance markers in cancer-free control participants of the EPIC cohort. Results showed a favourable metabolic profile characterized by lower levels of compounds associated with inflammation and metabolic stress, such as specific short-chain sphingomyelins, lysophospholipids, trans-fatty acids, and certain non-essential amino acids. Furthermore, higher DSR corresponded with increased concentrations of polyphenols, essential amino acids, omega-3 and omega-6 fatty acids, and long-chain phospholipids, known for their roles in reducing inflammation, supporting immune function, and promoting metabolic health. Additionally, C-reactive protein concentration was observed to have an inverse relationship with DSR, and no significant relationship was observed for the insulin resistance marker, C-peptide. In our last objective, we conducted a preliminary investigation on the relationship between DSR and microbial diversity in the French prospective cohort NutriNet-Santé. DSR was associated with increased gut microbiome diversity and enrichment of potentially beneficial bacterial taxa. A complementary systematic review showed inconsistent effects of probiotic and synbiotics on eating behaviour outcomes and their modulators, while prebiotic supplementation consistently reduced appetite and increased satiety. These results suggest a bidirectional relationship: food biodiversity may shape microbial communities that, in turn, could reinforce or modify dietary choices, creating a feedback loop between diet, microbiome, and eating behaviour. Overall, this thesis provides baseline evidence justifying the health benefits of food biodiversity. These results should be confirmed in other large-scale observational cohorts for generalizability and robustness of this evidence.