Can we define a tolerable level of risk in food allergy? Report from a EuroPrevall/UK Food Standards Agency workshop

Authors


Correspondence: Charlotte B. Madsen, Research Leader, Department of Toxicology and Risk Assessment, National Food Institute, Technical University of Denmark, 19, Mørkhøj Bygade, DK-2860 Søborg, Denmark. E-mail charm@food.dtu.dk

Summary

Background

There is an emerging consensus that, as with other risks in society, zero risk for food-allergic people is not a realistic or attainable option. Food allergy challenge data and new risk assessment methods offer the opportunity to develop quantitative limits for unintended allergenic ingredients which can be used in risk-based approaches. However, a prerequisite to their application is defining a tolerable level of risk. This requires a value judgement and is ultimately a ‘societal’ decision that has to involve all relevant stakeholders.

Objective

The aim of the workshop was to bring together key representatives from the stakeholders (regulators, food industry, clinical researchers and patients), and for the first time ever discuss the definition of a tolerable level of risk with regard to allergic reactions to food.

Results

The discussions revealed a consensus that zero risk was not a realistic option and that it is essential to address the current lack of agreed action levels for cross-contamination with allergens if food allergen management practice is to be improved. The discussions also indicated that it was difficult to define and quantify a tolerable level of risk, although both the clinical and the industry groups tried to do so. A consensus emerged that doing nothing was not a viable option, and there was a strong desire to take action to improve the current situation.

Conclusions and Clinical Relevance

Two concrete actions were suggested: (1) Action levels should be derived from the data currently available. Different scenarios should be examined and further developed in an iterative process. On the basis of this work, a tolerable level of risk should be proposed. (2) ‘One-dose’ clinical trial with a low challenge dose should be performed in multiple centres to provide additional information about the general applicability of dose-distribution models and help validate the threshold levels derived.

Ancillary