Abstract :
[en] Food safety remains a major public health concern worldwide, particularly in a context of globalised supply chains, diversified agricultural practices, and evolving dietary habits. Among chemical hazards, pesticides are widely used to improve agricultural yields, and their residues may remain in food and be ingested by consumers.
In addition to pesticides, processing contaminants such as polycyclic aromatic hydrocarbons (PAHs) and acrylamide are of particular concern due to their genotoxic and carcinogenic potential. European regulatory frameworks ensure a high level of consumer protection. However, significant uncertainties persist regarding occurrence data for under-investigated foods, the quantitative impact of food processing and domestic preparation, and up-to-date dietary exposure assessments.This doctoral thesis aimed to strengthen the dietary risk assessment of the Belgian population by integrating analytical innovation, updated occurrence data, and the quantification of processing effects, including consumer behaviour, in order to refine dietary exposure and risk assessment. The work focused on three relevant contaminant/matrix combinations: pesticides in tea and herbal infusions, PAHs in spices and dried herbs, and acrylamide in selected food products insufficiently or not documented in previous European assessments.
To achieve these objectives, sensitive, adaptable, and robust analytical methods were developed and optimised to address complex and heterogeneous food matrices. Multi-residue LC–MS/MS and GC–MS/MS methods were implemented for tea and herbal tea leaves as well as their corresponding brews in order to determine pesticides transfer factors (TFs). For PAHs, a modular GC–MS/MS method inspired by the QuEChERS approach, combining a common extraction step with adaptable clean-up procedures, was designed to cope with the diversity and complexity of spice and herb matrices. This strategy enabled reliable quantification of PAH4 across a wide range of products, including smoked spices and complex mixtures. The innovative character of this method led to its use as the basis for an interlaboratory ring trial within the EURL/NRLs network, addressing the lack of harmonised methods capable of dealing with complex spice and herb matrices. For acrylamide, the QuEChERS method initially developed for pesticides was adapted through the implementation of matrix-specific clean-up procedures. This resulted in a sensitive UPLC–MS/MS method validated for heterogeneous and under-investigated food categories such as pastries, confectionaries, dried fruits, nuts and oil seeds, vegetable chips, black olives, and potato-based dough products.
These analytical methods were subsequently applied to generate missing occurrence data necessary for accurate exposure and risk assessment. The occurrence results revealed several important findings. In tea, 38 % of the 55 analysed samples contained at least one pesticide exceeding the EU maximum residue level (MRL) in dry material or contained prohibited pesticides, confirming the relevance of continued monitoring.
Furthemore, up to 15 pesticides has been identified in one tea sample. However, a key innovation of this work lies in the systematic quantification of pesticide transfer from dry leaves to infusion under brewing conditions representative of Western consumption habits. Experimentally determined transfer factors ranged from 0 % to 100 %, revealing clear relationships with physico-chemical properties: highly lipophilic pyrethroids exhibited limited transfer (<10 %), whereas polar neonicotinoids frequently exceeded 50 %.
New occurrence data also revealed relatively high acrylamide concentrations in several less-studied food categories, such as sterilised black olives, potato-dough-based products, and particularly vegetable-based crisps. In spices and dried herbs, PAH4 were detected in several samples, especially in products subjected to smoking, confirming their contribution to dietary exposure.
A distinctive and innovative aspect of this thesis is the integration of consumer behaviour into exposure assessment. For tea, experimentally derived and modelled transfer factors reflecting realistic brewing practices were incorporated. Results showed that brewing parameters such as temperature and infusion time had limited influence on transfer factors under typical preparation conditions, whereas intrinsic characteristics of the tea matrix and the presence of flavouring ingredients played a more significant role. A comprehensive TF database was established, and four predictive models based on physico-chemical parameters were developed, enabling transfer estimation for non-tested compounds and providing a tool for future risk assessment and regulatory use. Regarding acrylamide, participants were recruited to prepare food products according to their usual cooking habits. Questionnaires combined with analytical measurements revealed that acrylamide levels were generally well below current benchmark levels. Interestingly, participants tended to stop cooking when a golden colour was reached, corresponding to their preferred sensory characteristics, which aligns with the colour recommended by EU regulation to limit acrylamide formation. These studies highlight the importance of integrating consumer behaviour into research in order to obtain more realistic occurrence data and exposure assessments. Once occurrence data and processing factors were acquired, risk assessment was performed using approaches adapted to contaminant toxicological profiles. Hazard indices and hazard quotients were calculated for acute and chronic exposure to pesticides, while the margin of exposure (MOE) approach was applied for genotoxic and carcinogenic substances such as PAHs and acrylamide. No significant acute or chronic risk associated with pesticide intake via tea infusion was identified, despite the high proportion of MRL exceedances in dry leaves, demonstrating that MRL exceedance is not necessarily synonymous with consumer health risk. For PAHs in spices and herbs, integration of new occurrence data into national exposure calculations resulted in only a limited increase in total PAH intake, with margins of exposure remaining largely above levels considered of a concern for the consumer.
For acrylamide, new food categories analysed moderately modified overall intake estimates, and updated exposure assessment confirmed that margins of exposure for the Belgian population remain a public health concern, in line with EFSA conclusions of 2015, thereby reinforcing the need for continued mitigation measures.
Overall, this thesis contributes to reducing knowledge gaps in dietary exposure assessment by combining analytical innovation, updated occurrence data, modelling of processing effects, and integration of consumer behaviour. Importantly, the TF database and predictive models developed during this work are already being used by the Belgian Food Safety Agency to support risk evaluation in cases of MRL exceedances in tea. Furthermore, the acrylamide occurrence data generated were transmitted to EFSA to support the revision of the EU regulation and the establishment of benchmark levels for new food categories, thereby strengthening consumer protection and improving risk characterisation at EU level.
