News

In this paper (purchase required) the authors make use of a recently developed ambient ionisation source for mass spectrometry.  Self Aspiration Corona Discharge Ionisation (SACDI-MS) enabled the direct measurement of volatile compounds from coffee, and when the authors coupled this with an Air Curtain Sampling Device it meant that, by removing interfering volatiles from neighbouring batches, they could design an in-line sensor suitable for use in a production environment.  The use of Deep Learning Algorithms with an Artificial Neural Network enabled them to compensate for other interfering peaks from environmental volatiles, often a problem in direct ionisation mass spectrometry.  They constructed a chemometric classification model using a reference set of coffees from 6 different geographic origins and proved that they could differentiate between them in a high-throughput, rapid production environment.  They conclude that this makes the approach ideal for in-line screening of coffee authenticity in situations when there is a consistent “expected” origin, used to train a classification model, that needs to be distinguished from substitution by “unexpected” origin coffee.

Photo by Andrew Neel on Unsplash

European legislation permits the inclusion of non-cocoa vegetable fats up to 5% into milk chocolate recipes.  Anything beyond this is adulteration.

In this paper (purchase required) the authors report the use of proton NMR combined with chemometrics to discriminate between milk fats, cocoa fats and non-cocoa vegetable fats (“cocoa butter equivalents”, CBE).  They prepared known mixes (0-100%) of different fats.  They used both a targeted and an untargeted approach.  The targeted approach used the integrals of the signals belonging to ω-3, ω-6, ω-9, and saturated fatty acids.  The untargeted approach used the spectra as fingerprints.

The authors reported that the untargeted partial least-squares discriminant analysis model (PLS-DA) recognized the type of CBE in blends with sensitivities in prediction higher than 75%. The targeted PLS-DA model was capable of recognizing non-adulterated milk chocolate fats with 100% sensitivity and specificity in prediction. Conversely, low percentages in sensitivity were achieved for most of CBEs. Both targeted and untargeted PLS regression models efficiently determined the amount of CBE in blends. Fingerprinting models showed better results both in the classification and quantification of CBEs.  They conclude that this proves the applicability of ¹H NMR in milk chocolate quality control.

Photo by Kaffee Meister on Unsplash

New designated protected names came into force yesterday, under the UK/Japan trade agreement signed in 2021.

Within the UK, the following foods will now have protected descriptors and Geographical Indications

• Daiei Suika 
• Daisen Broccoli 
• Echizen Gani/Echizen Kani 
• Edosaki Kabocha 
• Futago Satoimo/Futago Imonoko 
• Hiba Gyu 
• Higashiizumo no Maruhata Hoshigaki 
• Hiyama Haishen 
• Ibuki Soba/Ibuki Zairaisoba 
• Iburigakko 
• Iwadeyama Koridofu/Iwadeyama Meisan Koridofu 
• Koge Hanagoshogaki 
• Kumamoto Akaushi 
• Matsudate Shibori Daikon 
• Mito no Yawaraka Negi 
• Monobe Yuzu 
• Nango Tomato 
• Okukuji Shamo 
• Ozasa Urui 
• Sayo Mochidaizu 
• Taisyu Soba 
• Tokyo Shamo 
• Toyama Hoshigaki 
• Tsunan no Yukishita Ninjin 
• Tsuruta Steuben 
• Yamadai Kansho 
• Yamagata Celery 
• Yatsushiro Tokusan Banpeiyu 
• Zentsujisan Shikakusuika 
• Hagi 
• Harima 
• Hokkaido 
• Mie 
• Nadagogo 
• Tone Numata 
• Wakayama Umeshu 
• Yamanashi

Within Japan, the following foods will now have protected descriptors and Geographic Indications

• Cornish Clotted Cream 
• Cornish Pasty  
• Anglesey Sea Salt/Halen Môn 
• Arbroath Smokies 
• Conwy Mussels 
• East Kent Goldings 
• London Cure Smoked Salmon 
• Lough Neagh Eel 
• Lough Neagh Pollan 
• Melton Mowbray Pork Pie 
• Orkney Scottish Island Cheddar 
• Pembrokeshire Earlies/Pembrokeshire Early Potatoes 
• Scotch Beef 
• Scotch Lamb 
• Single Gloucester 
• Staffordshire Cheese 
• Stornoway Black Pudding 
• Traditional Ayrshire Dunlop 
• Traditional Cumberland Sausage 
• Traditional Grimsby Smoked Fish 
• Traditional Welsh Caerphilly 
• Welsh Beef 
• Welsh Lamb 
• Welsh Laverbread 
• West Country Beef 
• West Country Lamb 
• Yorkshire Wensleydale 
• English Wine 
• English Regional Wine 
• Herefordshire Cider 
• Herefordshire Perry 
• Irish Poteen 
• Kentish Ale 
• Kentish Strong Ale 
• Somerset Cider Brandy 
• Welsh Wine 
• Welsh Regional Wine 

The UK government press release is here.

Photo by Richard Iwaki on Unsplash

The European Commission has published its first monthly report on EU Agri-Food Fraud. The stated intent of this free list of incident collations is to help both regulatory authorities and food businesses target their fraud defence activities to the highest risk areas.

The report collates all entries from iRASSF that have been categorised as “suspicious”.  It therefore includes official controls, border rejections, whistleblower complaints and media reports (unlike, for example, the JRC monthly food fraud collation which is fed purely by media reports).  It does not include “suspicions” which have not been communicated outside of the originating nation.

In January 2024, 277 “suspicions” are listed.  Read the January report here.

(image from the report)

This paper (purchase required) reports a simple screening method for vegetable oil adulterants (corn, sesame, soy, sunflower, almond, and hazelnut) in olive oil.  It is based on a multispecies DNA sensor that can be read with the naked eye. It is the first report of a DNA sensor for olive oil adulteration detection with other plant oils. The researchers have identified unique nucleotide variations which enable the discrimination of the seven plant species. Following a single PCR step, a 20 minute multiplex plant-discrimination reaction is performed, and the products are applied directly to the sensing device. The plant species are visualized as red spots using functionalized gold nanoparticles as reporters. The spot position reveals the identity of the plant species. The authors report that <5–10% of adulterant was detected with good reproducibility and specificity.

Image from the paper

The results of the FSA's most recent surveillance testing project of food and beverages are published here.  This covers retail samples (including online sales) purchased in October 2022 by various local authorities in England, Wales and Northern Ireland.  It included authenticity testing on 437 samples including meat and meat products, herbs and spices, basmati rice, coffee, cheese and olive oil.   97% were reported as authentic and two reported as inconclusive. The main commodities with authenticity issues identified were oregano with 13% of samples containing other leaf types and basmati rice with 10% of the basmati rice samples reported as having been adulterated with non-basmati rice varieties or in one instance having no approved basmati varieties at all.

FSS have published (here) their 2024-27 strategic plan for food crime prevention.  It is a high-level document and, although written for Scotland, the principles are applicable for regulators and enforcement bodies anywhere.  It describes a model for the identification, analysis and implementation of measures to reduce or prevent the ccurrence or re-occurrence of food crime and the identification and mitigation of related future food crime risks.

(image from the strategic plan)

In this proof-of-concept study (purchase required) the authors used direct-sampling ambient mass spectrometry (REIMS, sometimes called the “ion knife”) to build a classification model for different cuts of beef.  They analysed untargeted lipid profile data from a reference set of 125 authenticated samples purchased directly from an abattoir (25 each of ribeye, sirloin, brisket, shank and foreshank) from 12-month old bulls of the Pirenaica breed.  They used machine learning to select discriminatory features (11 fatty acids and 44 phospholipids).  They report good discrimination between different cuts of meat using the model.

Photo by amirali mirhashemian on Unsplash

The 2023 BSI Supply Chain Risks and Opportunities Report is available for free download.

Food & beverage was the league-leading sector for supply chain theft, accounting for 22% of global reported thefts (up from 17% in 2022). 

Other watchouts include

• The reverberating impact of climate change impacts, a stark example being a trebling of virgin olive oil commodity price over the past 2 years.
• In the US, 48 ransomware attacks in the food and drink sector reported in 2023 by the FBI Crime Compliant Sector.

Image from the report

In this paper (purchase required) the authors build a Raman spectroscopy classification model to differentiate 5 brands of very similar dairy products.  The model was trained on 30 retail samples of each brand.  The authors reported significantly improved feature selection and model performance (particularly in detecting differences in dairy fat content, the key question for adulteration) when they used a data fusion approach of 4 different Machine Learning protocols rather than training using one protocol.  They recommend this fusion approach for constructing classification models.

In this paper (open access) the authors developed a rapid PCR-based test protocol for three species of tuna (Thunnus thynnus (Blue Fin Tuna), Thunnus albacares, and Katsuwonus pelamis) under simulated conditions for canned and flavoured products.  Home-made canned simulants were prepared by mixing each fish tissue of the three tuna species with salt, pepper, paprika, onion, oil, vinegar, and tomato followed by frying and boiling. DNA was then isolated from the home-made canned products. Binary mixtures were prepared using the isolated DNA in various percentages of adulteration that ranged from 1 to 100%.  DNA was extracted, followed by amplification by rapid small-scale PCR using species-specific primers.  The PCR products were hybridized (10 min) to specific probes and applied to the rapid sensing device. The signal was observed visually in 10–15 min using gold nanoparticle reporters.  The authors report that the method was reproducible and specific for each tuna species and 1% of tuna adulteration (in the isolated DNA) could be detected with the naked eye.

photo by Farhad Ibrahimzade on UnsplashP

This review article (purchase required) covers the recent application of CRISPR/Cas technology to food authenticity sensor applications.

The CRISPR/Cas system is a defence mechanism in bacteria and archaea against phage invasion.  It allows them to store fragments of invading DNA and use them as a guide to recognize and destroy future invasions. Bound to a CRISPR RNA (crRNA) that precisely matches the nucleotide sequence of the target (spacer sequences), the CRISPR-associated (Cas) protein is precisely navigated to these sequences. Through enzymatic processes, it cleaves target nucleic acids at specific sites.  This sequence selectivity underpins gene editing technology.  It is less well known that it can also be used to build selective analytical sensors.

The authors describe how, in recent years, CRISPR/Cas-based detection systems have been applied to precise DNA detection for food authentication. Such systems entail designing a guide RNA that complements the target DNA, triggering precise cleavage by the Cas enzyme and activating collateral cleavage. This leads to the cleavage of fluorescent probes, enabling detection. CRISPR/Cas-based detection systems can be a rapid and highly specific tool for identifying target DNA in food. Coupled with DNA amplification strategies, it has the potential to achieve sub-attomolar sensitivity for nucleic acid detection.

CRISPR/Cas assays are qualitative, but can be made quantitative by combining with digital sample partitioning. The sample is divided into multiple individual partitions, each containing a distinct number of biological entities (0, 1, 2, 3, etc.). Each partition reacts independently and the partition containing the target will produce an increased fluorescent signal, allowing absolute quantification using Poisson distribution.

The authors cite and review published examples, and conclude that these digital isothermal amplifications and dCRISPR methods are rapid and offer single copy sensitivity and high accuracy, making them a promising tool for accurate DNA quantification with great potential for application in authenticating food DNA.

This PhD thesis (request a copy here) describes the application of a previously-published LC-MS analytical method for triglycerides in fats to build an authenticity classification model for oils and fats based upon their triglyceride profile.  The author reports good discrimination between different pure oils and also good discrimination when oil or sesame oil were adulterated with lower value oils.  The model was also used to discriminate aged and degraded oils, and those which had been heat-processed.  The author concludes that this fast, simple, robust and reliable method offers significant benefits in authenticating edible oils, evaluating oil degradation, and differentiating meat products from their fats. The method has excellent potential for universal use.

Part II - Food Fraud Prevention and Types of Fraud

Welcome! In support of the Food Authenticity Network, this blog series reviews key topics related to food fraud prevention. Watch here for updates that explore the definitions of food fraud terms and concepts.

This blog post builds on our previous review of the definition and scope of food fraud.

Food fraud was first clearly defined in 1820 by Frederick Accum in ‘A Treatise on Adulteration of Food and Culinary Poisons.’ Over the next two hundred years, the subject continued to be reviewed as a  food science or food safety problem, as by Wiley and other pillars of scholarship. Along the way, ‘someone else’ was relied upon to actually prevent the problem. The ‘someone else’ was never assigned.

Interdisciplinary areas of study converged over time, to enable the shift to focus on prevention. In the 1970s, criminology theory expanded from focusing on the criminal and punishment to prevention. In the 1980s, quality management became a separate area of business theory with a shift to understanding and reducing the root causes of problems. In the 2000s, risk management became more formalized, such as in ISO 31000 Risk Management, which focused on likelihood and consequence as well as risk and vulnerability. In the 2010s, Enterprise Risk Management expanded the resource allocation decision-making to evaluate not only how to mitigate but also to prevent problems. 

This holistic view of vulnerability applied criminology concepts to all criminal acts and all possible targets. For food products, that led to the need to define the ‘types of food fraud’ and the ‘types of products.’ If we are going to prevent food fraud, we need to consider all types of actions and products. This led to the holistic and all-encompassing definitions:

Type of Food Fraud & Definition (From various sources including GFSI and SSAFE):

• Adulterant-Substances (Adulterant/ Adulteration):
  • Dilution: The process of mixing a liquid ingredient with a high value with a liquid of a lower value.
  • Substitution: The process of replacing an ingredient or part of the product of high value with another ingredient or part of the product of lower value.
  • Concealment: The process of hiding the low quality of a food ingredient or product.
  • Unapproved enhancements: The process of adding unknown and undeclared materials to food products in order to enhance their quality attributes.
• Mislabeling or Misbranding: The process of placing false claims on packaging for economic gain.
• Grey market production/ diversion:
  • Gray Market: A market employing irregular but not illegal methods.
  • Theft: Something stolen and then covertly re-entered into commerce.
  • Diversion/ Parallel Trade: The act or an instance of shifting a product from one intended market to another, which is unauthorized but either legal or illegal.
• Counterfeiting (IPR): The process of copying the brand name, packaging concept, recipe, processing method, etc., of food products for economic gain.

The types of food fraud are intentionally broad – holistic and all-encompassing - to frustrate the criminal against action of any kind.

Watch out for the next blog in March, which will review fraud susceptibility of different types of products (e.g., raw materials or ingredients to finished goods in the marketplace).

If you have any questions on this blog, we’d love to hear from you in the comments box below.

References:

• https://publicdomainreview.org/collection/a-treatise-on-adulteration-of-food-and-culinary-poisons-1820
• Spink, J, and Moyer, DC, (2011) Defining the Public Health Threat of Food Fraud, Journal of Food Science, Volume 75 (Number 9), p. 57-63. (ISI 1.791; SJR 0.378), URL: https://ift.onlinelibrary.wiley.com/doi/full/10.1111/j.1750-3841.2011.02417.x
• Spink, John, Bedard, Brian, Bruner, Leon, Keogh, John, Scimeca, Joseph, Goodridge, Lawrence, Moyer, Douglas C, & Vasan, Akila (2019). International Survey of Food Fraud and Related Terminology: Preliminary Results and Discussion, Journal of Food Science,  84(10), 2705-271800, URL: https://ift.onlinelibrary.wiley.com/doi/full/10.1111/1750-3841.14705

The EC Knowledge Centre for Food Fraud and Quality (the Joint Research Centre, “JRC”) have published their monthly collation of global food fraud media reports for January 2024.  Thanks, as always, for FAN member Bruno Sechet for formatting these into this infographic.  If you would like to join the JRCs mailing list to sign up for these monthly summaries then the link is here.  You can follow Bruno's LinkedIn feed here.

The authors of this study (open access) developed a chemometric classification model to distinguish true cinnamon from its potential adulterants, Cassia or Saigon cinnamons.  The model is based on simple and low-cost LC-UV analysis of four marker chemicals: eugenol, cinnamaldehyde, coumarin and cinnamic acid.  Sample pre-treatment was vortexing/sonicating with methanol followed by centrifugation.  Reference samples for the model were purchased from retail outlets rather than fully traceable sources; 25 samples of each type of cinnamon, including both sticks and powder.  The model was first constructed to differentiate pure powders.  Then the authors used an experimental design on a training set of in-house prepared mixtures (down to 1%/99% mixes) and a Partial Least Squares algorithm to model the classification of mixtures.  They found the model was linear and – in the case of true cinnamon mixed with either of the two adulterants – could discriminate adulteration down to 1%.  The model could not discriminate Cassia from Saigon cinnamons but the authors consider this a less important question.

Image from the published study.

A Toolkit to Support Weight of Evidence Approaches for Food Authenticity Investigations has been published by Defra.

Applying a weight of evidence approach for food authenticity investigation is relevant in situations where screening tests which do not give a definitive answer are used, for example with non-targeted fingerprinting approaches for food authenticity testing which rely on probability-based interpretation of the data. In these situations, gathering and assessing the weight of evidence can help in drawing a conclusion on the authenticity of a sample/product.

This document provides a structured outline on how to approach a weight of evidence assessment to verify the authenticity of food and drink samples where there is no single confirmatory test result available.

It has been developed by a sub-group of Defra’s Authenticity Methods Working Group (AMWG), drawing on analytical testing, enforcement, and food industry expertise.                                                                                                                                                                                             

Access the Weight of Evidence Toolkit.                                                                                                                                                                                                                                                                                                                                                                                                                 This report has also been added to the 'Guides' tab of the 'Tools_Guides_reports' part of our Food Fraud Prevention section.

A recent criminal conviction in the UK highlights the fraud risk from rogue employees.  Two Despatch Managers at 2Sisters Food Group, the country’s largest poultry supplier, were supplying another company, Townsend Poultry, with chicken. Townsend Poultry was not a customer of 2 Sisters Food Group and there were no records of any deliveries. The fraud was uncovered during an audit when Townsend Poultry appeared incongruous on the customer records.  Enquiries made with local hauliers used by the 2 Sisters Food Group confirmed there had been 84 deliveries from the 2 Sisters Food Group to Townsend Poultry, worth hundreds of thousands of pounds. The Despatch Managers had destroyed the records of those deliveries.

2 Sisters suffered the theft of £300K of stock over an extended period between 2019 and 2021.  This stock was then fed into the UK market with falsified or non-existent traceability records; a food safety risk.

Read the FSA statement here

Read the story here.

This paper (purchase required, free to IFST members) reports a quick, non-destructive technique to add to the panel of analytical tools needed to detect olive oil adulteration or mis-labelling.  This test is to detect addition of sunflower, rapeseed or corn oils.  It is based on electrochemical examination of the peak of alpha-tocopherol oxidation on a pencil graphite electrode (PGE).  There is no sample pre-treatment needed.  The authors prepared in-house oil mixes and were able to confidently discriminate “adulterated” samples at around 10% added non-olive oil. The method's relative standard deviation (RSD) was 20%, and the α-tocopherol in cold pressed olive oil cut-off value was 30.98 ± 12.57 nA.   The authors believe this is the first publication on utilisation of voltammetric techniques for the detection of olive oil adulteration using a PGE.

The European Parliament and Council agreed to review and strengthen the existing marketing standards applicable to honey, fruit juices, jams and milk. The so-called Breakfast Directives lay down common rules on the composition, sales names, labelling and presentation of these products to ensure their free movement within the internal market and help consumers make informed choices.

The revised Directives agreed upon by the co-legislators will introduce the following changes:

• Mandatory origin labelling for honey:  the countries of origin in honey blends will have to appear on the label in descending order with the percentage share of each origin. Member States will have the flexibility to require percentages for the four largest shares only when they account for more than 50% of the blend. The Commission is empowered by the co-legislators to introduce harmonised methods of analysis to detect honey adulteration with sugar, a uniform methodology to trace the origin of honey and criteria to ascertain that honey is not overheated when sold to the final consumer. A Platform will be set up to advise the Commission on those matters. This will limit fraudulent practices and increase the transparency of the food chain.

• Innovation and market opportunities for fruit juices in line with new consumers demands: Three new categories will become available: ‘reduced-sugar fruit juice‘, ‘reduced-sugar fruit juice from concentrate‘ and ‘concentrated reduced-sugar fruit juice‘. This way consumers can choose a juice with at least 30% less sugars. It will be possible for fruit juices to indicate on their labels that “fruit juices contain only naturally occurring sugars” to clarify that, contrary to fruit nectars, fruit juices cannot by definition contain added sugars – a feature that most of the consumers are not aware of.

• Higher mandatory fruit content in jams: an increase of the minimum fruit content in jams (from 350 to 450 grams per kilo) and in extra-jams (from 450 to 500 grams per kilo) will improve the minimum quality and reduce the sugar content of these products for EU consumers. Member States will be allowed to authorise the term ‘marmalade' as a synonym of ‘jam', to take into account of the name commonly used locally for these products. The term “marmalade” was authorised until now only for citrus jams.

• Simplified labelling for milk: the distinction between ‘evaporated' and ‘condensed' milk will be removed, in line with the Codex Alimentarius standard. Lactose-free dehydrated milk will also be authorised.

The political agreement reached by the European Parliament, Council and Commission is now subject to formal approval by the co-legislators. From entry into force 20 days after publication of the final text, Member States will have 18 months to transpose the new provisions into national law and 6 more months before it applies throughout the European Union.

Read full press release.