Refinement of flavour substance intake estimates using probability of addition: A proposed method based on sales data

Most dietary intake estimates used in the regulatory risk assessment of flavouring substances assume that the flavouring substance is present in all foods at a given concentration. They do not account for the use of alternate flavouring substances for the same taste modality, nor do they consider the use of differing concentrations within a particular food category. This highly conservative assumption leads to an over-estimation of exposure. To refine this assumption in exposure models, the probability of addition can be utilised. In this study, a methodology to estimate the probability of addition was developed, using the volume of a flavouring substance sold for flavours created for a particular food category against its total volume sold for all flavouring use. The method was trialled, as a proof of principle, using historical data sets of two test flavouring substances, benzaldehyde (a large volume/use material) and bornyl acetate (a low volume/use material), and the impact on dietary exposure assessment was assessed. The resulting exposure estimates for both flavouring substances were shown to be significantly lower than those without the incorporation of probability of addition, demonstrating that current methods are overly conservative and that the use of probability of addition provides a more realistic estimate of exposure of flavouring substances present in foods.