Adulteration analysis: modern strategies

Abstract

With the earliest records of food fraud dating back to ancient history, adulteration of food products is not a contemporary issue. However, the ways in which the industry combats the issue have undoubtedly changed considerably, especially over recent decades. From an analytical perspective, we now have an arsenal of techniques at our disposal to aid identification of adulteration issues and can even pinpoint where in the supply chain an ingredient or product has become affected.

By definition, adulteration of food is the addition of an extraneous (or lower grade) substance to a food product which reduces its quality and, in some cases, can have an impact on consumer safety. Where intentional, the primary motivation is usually economic, with the aim of lowering costs or increasing the volume of a high value product. However, adulteration can also arise incidentally, where foreign substances are introduced as a result of ignorance, negligence or through the use of improper manufacturing facilities.

The food industry is experiencing a period of intense economic uncertainty, driven by both immediate factors such as increasing overhead costs, and longer-term factors such as climate change and geopolitical unrest, all of which compromise supply chain security. These pressures mean that some food businesses could be pushed into crisis situations which, without appropriate management, could allow instances of food adulteration to arise. Irrespective of the cause, the inclusion of materials which have not been considered for their toxicological impact, the subsequent mislabelling of the product and the departure from transparent supply chains can all have a serious impact on consumer safety.

The potential severity of these incidents is illustrated best by the reporting of past examples in the media. One such example was the Chinese milk scandal in 2008, where substandard milk intended for infants was adulterated with melamine in order to generate an artificially high nitrogen result. The intention was to fool the tests that checked for any undeclared dilution of the milk by giving the appearance of a higher protein content. However, melamine is toxic at the concentrations added and as a result, a large number of babies fell sick - in some cases fatally so. Once exposed, techniques able to detect the presence of melamine could be added to testing lists or specifications. However, this example helps to illustrate the potentially lethal cycle created within the field of adulteration detection - fraudsters will often show incredible ingenuity by adapting their strategies in response to advancing technology, and traditional methods of targeted adulterant analysis quickly become inadequate. Bearing in mind the ever-increasing ingenuity of fraudsters and increasing economic factors, how does the food industry protect itself from adulteration threats that are yet to reveal themselves?

Traditional methods of testing for adulteration have been based on understanding the potential risks that the testing process is meant to identify and address. Described as ‘targeted’ approaches, testing strategies could encompass spectroscopy, microscopy, chromatography or DNA-based techniques, designed to test for a selected analyte of interest. Prior to the horsemeat scandal of 2013, despite awareness of previous incidents of adventitious contamination between species, focus on meat species testing had been reduced significantly, relying instead on supply chain audits. Specific requirements for targeted horsemeat testing by enforcement officers had ceased in 2003, and so it was fortuitous that an incident of horsemeat substitution was indeed detected. If non-targeted speciation testing (which allows for the screening of multiple species at once) had been readily available, it is likely that the threat could have been detected earlier.

A decade later, non-targeted methods for the detection of adulterated meat are becoming increasingly popular. The expediated development of DNA sequencing-based techniques (Next Generation Sequencing, or NGS) has led to cost and time effective ways to harness non-targeted DNA analysis. Complementary to DNA techniques, analytical tools which provide chemical profiles of food ingredients or products have become increasingly prevalent in adulteration detection. Predominantly spectroscopy-based, these techniques provide a ‘fingerprint’ of the material in question and can be used to create a reference database of what an adulterant-free sample should ‘look’ like, harnessing the power of statistics to predict with confidence whether or not a product has been adulterated.

Multivariate methods of statistical analysis, such as Principal Component Analysis (PCA), are commonly employed for this purpose and can be used to conclude whether a suspect sample is statistically different enough to the reference database to conclude evidence of adulteration. However, serious consideration must be given to ensuring reference databases are reliable and kept up to date, as the quality of the reference dataset can affect the testing outcome dramatically. It's also important to note that the ‘fingerprint’ of a food product is likely to vary, due to seasonal variations in the ingredients, or subtle differences in the mode of preparation such as extractions of powdered bulk materials. Periodic statistical cross-validation should be applied to ensure time variabilities can be tracked across seasons and that a genuine product is not flagged as suspicious purely due to an aged reference database.

Non-targeted methods allow a more holistic picture of a sample to be built, although a single technique rarely provides sufficient evidence to label a product ‘free-from adulteration’. We need to decide whether relying solely on a chemical profiling technique or a DNA-based analysis is enough to demonstrate the absence of adulteration, or if using a combination of both approaches is necessary. For example, in the field of herbs and spices, adulteration could occur through the addition of chemical bulking materials (e.g. inorganic minerals or starches), or substitution or dilution with an alternative, but chemically similar, plant material. Likewise, the ways in which products or ingredients are processed may restrict the effectiveness of a specific strategy. For example, edible oils are extracted from plant materials in such a way that very little cellular material remains, making biological profiling difficult. As a result, it's important to explore methods that specifically examine the chemical profile distinctions between high-value commodity oils and potential adulterants.

Depending on application, sensitivity afforded by NGS methodologies can either be too sensitive in some instances or not sufficiently sensitive in others. When employed as a non-targeted approach, NGS will likely detect DNA to a concentration of less than 1% of the overall food commodity. At this level, the technique will detect adulterants, but will also detect species from cross contact contamination from other organic material without discrimination. This can lead to either misleading results or a result that is not actionable. On the other hand, there are instances where it would be dangerous to assume that the sensitivity of a NGS method would suffice for the detection of contaminants that could pose a food safety risk at extremely low concentrations. An example of this is the use of NGS to detect food allergens, as the technique does not provide the extreme sensitivity required for the detection of levels that could pose a risk to an allergic consumer. As such, PCR remains the industry standard technique for DNA-based allergen adulteration analytical strategies.

The use of analytical techniques to screen for adulterants, whether targeted or not, can be expensive, time-consuming and impractical to apply to every batch of product/ingredient. A more feasible approach involves conducting a comprehensive risk assessment of the supply chain to pinpoint potential weak links. Referred to as Vulnerability Assessment and Critical Control Points (VACCP), this approach focuses on economically motivated adulteration. While it's beneficial for all products, its significance is particularly pronounced when dealing with high-value commodities sourced from regions where geopolitical conditions may pose challenges to supply chain transparency. One potential recommendation stemming from this type of assessment is the implementation of an analytical testing regime.

The exercise will involve input from a multifunctional team that includes personnel from technical and quality functions, manufacturing, public affairs, sales, marketing and legal counsel. Key responsibilities, such as investigation and reporting to the authorities, should be defined and ready access to manufacturing and supply chain information should be available to expedite investigations. They should also have procedures and systems in place, which are designed specifically to reduce the risk of adulteration within the production environment.

Legislation across geographic regions varies, but all governing bodies in the Western world make specific provisions or include chapters requiring food manufacturers to put in place mitigation strategies to protect food against intentional adulteration. In the UK, guidelines are issued by the Food Standards Agency (FSA) or Food Standards Scotland (FSS) to enable manufacturers to both comply with the legislation and protect consumers. There is an ever-increasing list of tools, guidelines and reports to help with mitigation of adulteration. These include databases (such as the service provided by Food Chain ID), intelligence sharing networks (e.g. Food Industry Intelligence Network) and even vulnerability assessment tools, such as the National Food Crime Unit's food fraud resilience tool, which can be completed in as little as 15 minutes.¹ As an indicator of how seriously the UK government now takes the threat of intentional adulteration, dedicated law enforcement units were set up within both the FSA and FSS in 2015 to allow for easy reporting and subsequent investigation of suspected food fraud incidents – the National Food Crime Unit and the Scottish Food Crime and Incidence Unit.

Technical innovation within instrumentation is generating smaller and faster analytical solutions to adulteration testing all the time. Coupled with enhanced computing power required to undertake the necessary statistical modelling, handheld spectrometers, as well as DNA-profiling capabilities, are now becoming available for use within factory settings (or even in a field). Furthermore, the types of statistical methods used to discriminate adulterants can be built upon so that trends can be modelled and potential modes of adulteration within a particular industry can be predicted. Moreover, artificial intelligence can be used to monitor numerous market factors, allowing in-depth pre-learning of each risk area and predictions as to which ingredients/products are likely to be the next topic of interest. However, as with the databases that underpin our statistical evaluations, these predictive models also require a wealth of high-quality data points, collected from reputable sources, which are supplied with meta data information to allow trends to be identified with high accuracy. There is clearly a need for those within the industry to work together to ensure the predictive models employed are fit for purpose.

As we consider the current state of the industry's combative strategy to adulteration issues, it is clear that the high impact cases of the last 20 years have heralded an increased awareness in the field, as analytical technologies, risk assessment strategies and legislative requirements have all developed significantly.

However, when compared to the industry's response to the threat of allergens, for example, there are many who would argue that there is still room for improvement. A potential barrier is the cost and inconvenience of analytical testing, especially when traditional targeted approaches can result in lengthy testing regimes, which add little value when considering the overall risk of adulteration without first completing a VACCP assessment. As quicker non-targeted analytical approaches become more easily accessible, it is likely that where questions arise about the integrity of a food product, an analytical strategy which provides cheap, quick and unequivocal conclusions will become more appealing.