Protecting our food: Can standard food safety analysis detect adulteration of food products with selected chemical agents?

Abstract

The food we eat and water we drink is routinely tested for a range of biological and chemical contaminants, which can be hazardous to human health, as part of food safety legislative requirements. The vulnerability of the food industry to deliberate contamination events, rather than naturally occurring events, was explored as one aspect of the EU FP7 project EDEN (End-User Driven Demo for CBRNe). We wanted to investigate if routine food safety testing could detect deliberate contamination with three chemical contaminants and three matrices (cooked ham, sugar and water). The contaminants selected had to be hazardous to human health at levels in the final food product that could occur with a deliberate contamination event.

Standardised reference panels were developed and homogeneity and stability were tested prior to distribution for food safety chemical testing, as required by EU legislation, in the meat food chain (cooked ham and water) and the sugar food chain (sugar and water). Each reference panel contained 11 samples analysed in triplicate (33 analyses per matrix). The meat food chain panels contained bromadiolone (a rodenticide) in the meat and sodium trifluoroacetate (a simulant for a toxic pesticide) in the water at levels from 0 to 4000 parts per million (ppm). The sugar food chain panels contained mercury chloride in both the sugar and water, at levels from 0 to 12 500 ppm. The food safety standard chemical analysis methods were compared to the following external laboratory methods for the meat food chain panels: liquid chromatography coupled to mass spectrometry for meat and nuclear magnetic resonance spectroscopy for water. Inductively coupled plasma with mass spectrometry was used to analyse the sugar food chain panels containing both sugar and water samples. Neither the meat nor the sugar food safety methods detected contamination in any of the samples whilst the external laboratory correctly identified and quantified the contaminants in all the samples.

The results for these three contaminants (bromadiolone, sodium trifluoroacetate and mercury chloride) are not surprising given that they are not the target of today's food safety testing procedures. These limited results are of note and highlight food chain vulnerability to deliberate contamination events with novel contaminants. The EDEN project is exploring a 2-level approach: screening food with non-specific detection tools which are supplemented by targeted detection tools when an alert is triggered. This approach could lead to increased consumer protection whilst simultaneously reducing the economic burden of testing and product recall for the industry.