Hazard characterisation in food allergen risk assessment: The application of statistical approaches and the use of clinical data

R. W.R. Crevel, D. Briggs, S. L. Hefle, A. C. Knulst, S. L. Taylor

Research output: Contribution to journalArticle

56 Scopus citations

Abstract

A structured approach to assess the risk to allergic individuals from food allergens requires as a first step the experimental measurement of minimum eliciting doses in a population that is as representative as possible of the relevant allergic population, using a standardised protocol. These doses are established in controlled challenge studies, but logistical and statistical constraints mean that a proportion of the allergic population may still be at risk of reacting at doses below those which have been or could feasibly be tested. However, statistical modelling of the dose distribution resulting from such challenges permits inferences to be drawn about the proportion of allergic individuals that are likely to react to specified (low) amounts of residual allergen in food. However, different statistical models, which all provide good fits to the experimental data yield different values outside the experimental range. Consequently, the outputs from these models require a form of validation, which demonstrates how close the predictions are to reality. In addition to characterisation of the hazard, for each allergenic food this validation requires information about exposure to undeclared allergen, the actual number of reactions taking place in the wider allergic population, and the prevalence of allergy to that food.

Original languageEnglish (US)
Pages (from-to)691-701
Number of pages11
JournalFood and Chemical Toxicology
Volume45
Issue number5
DOIs
StatePublished - May 2007

Keywords

  • Minimum eliciting dose
  • Modelling

ASJC Scopus subject areas

  • Food Science
  • Toxicology

Fingerprint Dive into the research topics of 'Hazard characterisation in food allergen risk assessment: The application of statistical approaches and the use of clinical data'. Together they form a unique fingerprint.

  • Cite this