Consumers are regularly exposed to well-known food contaminants (FCs), which are typically assessed for risk on an individual basis. However, there is limited knowledge about the overall levels and combinations of these compounds depending on dietary choices. The goal of this study was to estimate the real-life mixtures of FCs in different dietary models by integrating extensive data from the scientific literature concerning the reliable quantification of FCs in foods. A FAIR database detailing the occurrence of 73 FCs in 16 foods commonly consumed was built. The data were integrated into an omnivorous and a vegetarian dietary model. A weighted estimate of the 25th, 50th, and 75th percentiles of FCs in both dietary models revealed that the omnivorous model presented slightly higher levels of FCs than the vegetarian. At the 25th percentile, the FC levels in both dietary models fall within the European Food Safety Authority (EFSA) reference exposure levels for chemical hazards, except for arsenic, lead, cadmium, fumonisin B1, and OTA. At the 75th percentile, the FC levels exceed the EFSA reference levels for those FCs and additional mycotoxins. Using in vitro models, the 25th percentile can mimic real-life FC exposure, while the 75th percentile simulates a possible worst-case scenario.
Keywords: FAIR database; dietary models; food contaminants; hazard assessment; omnivorous diet; vegetarian diet.