Probabilistic modelling techniques in dietary exposure assessment: application on the risk assessment of cadmium for Austrian adults

ABSTRACT

Probabilistic exposure and risk assessment of chemical hazards in the diet have increasingly gained ground in recent years as a pragmatic approach for the approximation of reality. This work presents the outcomes of a project which aimed at applying probabilistic techniques for basic modelling of chronic dietary exposure to food contaminants following EFSA guidance. These techniques, based on Monte Carlo Risk Assessment (MCRA) software and on the programming language R, were employed for the risk assessment of cadmium for Austrian adults, enabling the validation and the critical comparison of the two approaches. Harmonisation and optimisation of procedures, refinement of exposure assessment skills and confidence in the results were the main benefits. Data amount and validity were identified as critical parameters, influencing the precision of the results. Cadmium was selected as a case study due to its toxicological properties, its ubiquitous presence in food and the availability of Austrian occurrence data. Similar exposure and risk estimates were generated through MCRA and R in alternative optimistic and pessimistic exposure scenarios, suggesting low levels of concern, except for vegetarians, whose upper tail exposures are close to the established Tolerable Weekly Intake. However, as occurrence data gaps have been identified as the major element of uncertainty, the estimated exposure and risk levels are characterised as underestimated. Grains and grain-based products, potatoes and leafy vegetables are the main contributors to the intake. The results will contribute to risk management and to a future refinement of the assessment.

KEYWORDS:

• Cadmium
• probabilistic dietary exposure modelling
• Monte Carlo Risk Assessment
• MCRA
• programming language R

Acknowledgments

This work was part of the work programme “Joint venture on the further development of chemical exposure assessment by use of probabilistic modelling” within the 2018-2019 EU-Food Risk Assessment (EU-FORA) Fellowship Programme of the European Food Safety Authority (EFSA), implemented in the Risk Assessment Department of the Austrian Agency for Health and Food Safety (AGES), which was the hosting site for the first author. Food and drinking water samples were collected by the Austrian Food Control Authority with the support of the Austrian Federal Ministry for Social Affairs, Health, Care and Consumer Protection. Consumption data were collected by the University of Vienna with the support of the Austrian Federal Ministry for Social Affairs, Health, Care and Consumer Protection and EFSA.

Disclosure statement

No potential conflict of interest was reported by the author(s).