The U.S. Pharmacopeial Convention (USP) has developed a guidance document for using so-called non-targeted adulterant testing, a new approach to screening food that can help producers to spot adulteration in the supply chain.
An alternative approach to adulterant detection
Non-Targeted Adulterant Testing (NT testing) has developed over recent years. The technique requires considerable data to function, and advances in data-processing capabilities and commercially available analytical techniques have made the approach much more feasible.
The NT method uses a library of known unadulterated samples to create a model which represents all the variability inherent in these typical samples, and a surrounding “boundary” is statistically determined from this information. Unknown samples are then compared to this model and anything falling outside the boundary is determined to be Atypical, anything falling inside it is Typical. While easy enough to describe, it is not as straightforward as one might think and involves dedicated planning and processing.
Guidance document review
The U.S. Pharmacopeial Convention (USP) has developed a guidance document to assist with exploring this relatively uncharted territory for food-related analytical applications. The draft draws on contributions from scientists from multiple disciplines from around the world and provides some much-needed insights and suggested standardization in the area. The guidance has been in development since 2015 and it is currently open for public comment.
The document covers all aspects of NT testing, from the collection and analysis of representative samples, through development of the NT models, and on to monitoring and maintenance, as well as advice on how to proceed when anomalies are detected.
More about non-targeted screening for adulteration with routine analytical equipment
Detecting food fraud (economically-motivated adulteration, or EMA) is a challenging analytical task because for any food or food ingredient at risk of adulteration there may be numerous potential adulterants, many of which are unknown.
A non-targeted method consists of an analytical measurement that is sensitive to multiple potential adulterants coupled with a statistical model that recognizes deviations from the signal associated with the nominal material: it is not calibrated for any specific adulterants. The method is therefore not a complete answer in that the user only gets a warning if something appears abnormal with a sample. They still need to take further action to determine the exact nature of the adulteration, for example with targeted screening or classical analytical methods. However, the speed and ease of use of the screening provides an effective first line of defence.