Better together: evaluating single and combined biomarkers of fermented food intake and their associations with cardiometabolic health

Summary

Cardiometabolic diseases (CMDs) represent one of the largest health and socioeconomic burdens to modern society. Fermented foods are commonly consumed in diets worldwide and can be an important part of an effective dietary strategy to help prevent and manage CMDs. However, precise associations between the consumption of fermented foods and cardiometabolic health have not been well-established. This could be partly due to the difficulty of accurately capturing their intake through self-report dietary assessment methods. This thesis set out to identify food intake biomarkers (FIBs) for the habitual intake fermented foods, which can act as more objective measures of intake. In addition, several candidate FIBs for dairy foods that were identified from non-targeted, controlled intervention studies needed to be validated under free-living conditions. These aims were achieved using data from a free-living adult cohort in the Netherlands (NQplus). Alongside, associations between fermented food and dairy (food) intake, biomarkers, and cardiometabolic factors were explored, which can reveal insights about the diverse metabolic effects of these foods.

Approximately 16-18% of foods consumed by Dutch adults consisted of fermented food items. Fermented foods with the highest consumption (in g/day) included coffee, yoghurts, beer, wholegrain bread, wine, and cheese. A relative validation of the FFQ (compared to multiple 24-h recalls) for estimating the intakes of fermented foods revealed acceptable or good validity for commonly consumed foods in The Netherlands (e.g., coffee, wholegrain bread, and cheese), but not for less frequently consumed foods (e.g., quark and buttermilk). A non-targeted analysis of the plasma and urine metabolomes of 246 NQplus participants resulted in the identification of 36 metabolites as candidate FIBs for fermented foods, the majority corresponding to the intakes of coffee, wine, and beer (none were identified for cocoa, bread, cheese, or yoghurt intake). Some metabolites appeared to originate from the food raw material and have been previously linked with a specific food (e.g., trigonelline for coffee), while others overlapped across several fermented foods (e.g., ethyl α-D-glucopyranoside for both wine and beer). In addition, several possible fermentation-dependent metabolites were identified (e.g., erythritol and citramalate). Given the non-specificity of single biomarkers, there is potential in further exploring multi-markers of fermented food intake. Associations between these identified metabolites with CMD risk parameters were weak and inconclusive.

The robustness of several previously-identified FIBs of dairy foods (milk, cheese, and yoghurt) were also evaluated in the NQplus cohort using both single and multi-marker approaches. Multi-marker models that also accounted for common physiological covariates better captured the subtle differences for milk and cheese intake over single-marker models. However, no significant single or multi-marker models were able to properly assess yoghurt intake. In a subcohort of 131 NQplus participants, 14 free fatty acids in plasma were positively associated with dairy intake (and in particular, dairy fat inake). Additionally, several of these fatty acids were also associated with plasma total and low-density lipoprotein cholesterol, blood pressure, and a composite score of cardiovascular risk. These findings suggest that milk-derived fatty acids could act as biomarkers for dairy intake and (cardio)metabolic effect. However, no clear associations were observed between dairy intake and CMD risk factors, and more studies are required in larger populations with a larger gradient of CMD risk.