Lessons Learned from Intentional Adulteration Vulnerability Assessments (Part I)
Three Practices for Supply Chain Management in the Food Industry
Changes in the Food Safety Industry: Face Them or Ignore Them?
How Technology is Elevating Food Safety Practices & Protocols
Five Tips to Add Food Fraud Prevention To Your Food Defense Program
E. Coli Update: FDA Advises Consumers to Avoid All Romaine Lettuce Harvested in Salinas, California
For more than three years, more than 100,000 pounds of giant squid from Peru was imported into the United States by a father-son duo who owned two Long Island food processing and distribution companies, and then marketed the squid as the more expensive octopus. The mislabeled seafood was worth over $1 million, and 10 grocery stores were defrauded during this time period. This kind of fraud carries steep fines and a possible five-year prison sentence.
Honey is a popular item for adulteration, and honey with a specific botanical source is seen as a more valuable product. The Czech Agriculture and Food Inspection Authority took samples of organic Spanish lavender honey in a Czech supermarket, and analyzed the pollen. The analysis showed that the honey was from alternative botanical sources and certainly not lavender.
The nose knows: In case fish smells “fishy”, it is no longer fit for human consumption. A Canadian fish importing company pleaded guilty to the import of 9,000 pounds of rotten and partially decomposed fish into the United States. The potentially adulterated fish was sampled by the FDA, who declared it to be too spoiled to be sold in the country, hence refused its entry into the United States—but the fish was imported via a wrong shipment declaration anyway. The crime of importing refused food carries a prison sentence of up to a year.
In a large study of nearly 6000 products, more than a quarter (27%) of herbal medicines and foods sold in 37 countries on six continents was found to be deliberately or accidentally adulterated. In this study, the products, which came in a variety of forms such as softgels, tea and more, were analyzed with high throughput DNA sequencing and showed mislabeling, added fillers, substituted ingredients or contaminants. Such fraud can be a harmful to consumer health and safety, and must be monitored and tracked closely.
Producers of food-based products are faced with challenges of maintaining the safety and quality of their products, while also managing rapid screening of raw materials and ingredients. Failure to adequately address both challenges can be costly, with estimated recall costs alone starting around $10 million, in addition to any litigation costs.1 Long-term costs can accumulate further as a result of damage to brand reputation. A vast array of methods has been employed to meet these challenges, and adoption continues to increase as technology becomes smaller, cheaper and more user friendly. One such technique is Fourier transform infrared (FTIR) spectroscopy, an analytical technique that is widely used for quick (typically 20–60 seconds per measurement) and non-destructive testing of both man-made and natural materials in food products. The uniformity and physical state of the sample (solid vs. liquid) will dictate the specifics of the hardware used to perform such analyses, and the algorithm applied to the identification task will depend, in part, on the expected variability of the ingredient.
Infrared spectral measurements provide a “compositional snapshot”— capturing information related to the chemical bonds present in the material. Figure 1 shows an example of a mid-infrared spectrum of peppermint oil. Typically, the position of a peak along the x-axis (wavenumber) is indicative of the type of chemical bond, while the peak height is related either to the identity of the material, or to the concentration of the material in a mixture. In the case of peppermint oil, a complex set of spectral peaks is observed due to multiple individual naturally occurring molecular species in the oil.
Once the infrared spectrum of an ingredient is measured, it is then compared to a reference set of known good ingredients. It is important that the reference spectrum or spectra are measured with ingredients or materials that are known to be good (or pure)—otherwise the measurements will only represent lot-to-lot variation. The comparative analysis can assist lab personnel in gaining valuable information—such as whether the correct ingredient was received, whether the ingredient was adulterated or replaced for dishonest gain, or whether the product is of acceptable quality for use. The use of comparative algorithms for ingredient identification also decreases subjectivity by reducing the need for visual inspection and interpretation of the measured spectrum.
Correlation is perhaps the most widely used algorithm for material identification with infrared spectroscopy and has been utilized with infrared spectra for identification purposes at least as early as the 1970s.2 When using this approach, the correlation coefficient is calculated between the spectrum of the test sample and each spectrum of the known good set. Calculated values will range from 0, which represents absolutely no match (wrong or unexpected material), to 1, representing a perfect match. These values are typically sorted from highest to lowest, and the material is accepted or rejected based on whether the calculated correlation lies above or below an identified threshold. Due to the one-to-one nature of this comparison, it is best suited to identification of materials that have little or no expected variability. For example, Figure 2 shows an overlay of a mid-infrared spectrum of an ingredient compared to a spectrum of sucrose. The correlation calculated between the two spectra is 0.998, so the incoming ingredient is determined to be sucrose. Figure 3 shows an overlay of the same mid-infrared spectrum of sucrose with a spectrum of citric acid. Notable differences are observed between the two spectra, and a significant change in the correlation is observed, with a coefficient of 0.040 calculated between the two spectra. The citric acid sample would not pass as sucrose with the measurement and algorithm settings used in this example.
When testing samples with modest or high natural variability, acceptable materials can produce a wider range of infrared spectral features, which result in a correspondingly broad range of calculated correlation values. The spread in correlation values could be of concern as it may lead to modification of algorithm parameters or procedures to “work around” this variation. Resulting compromises can increase the potential for false positives, meaning the incorrect ingredient or adulterated material might be judged as passing. Multivariate algorithms provide a robust means for evaluating ingredient identity for samples with high natural variability.
Click below to continue to page 2.
In the United States, the FDA is responsible for regulations and recommendations to protect public health, which includes the prevention of any type of food adulteration (unintentional contamination, intentional adulteration, and food fraud – or “economically motivated adulteration”). FSMA (the Food Safety Modernization Act) resulted in new regulations and guidance with strategies to reduce all types of risks in food facilities. It was the most comprehensive reform of FDA’s food safety regulations in more than 70 years.
Governments are responsible for the regulatory framework and consumer food protection to keep their citizens safe. In Germany, the Federal Office of Consumer Protection and Food Safety is responsible for national food safety as well as cross-border trade and international information exchange. The German BLV is also the contact point for the EU’s Rapid Alert System for Food and Feed, RASFF. To ensure consumer safety, the Max Rubner-Institut employs some 200 scientists who research food safety, nutrition and food fraud.
Pet food is a highly profitable business. Global pet food sales hit a record $90 billion in 2018, and adulterated or mislabeled feed is not uncommon. In the United States, the FDA ensures correct labeling and adherence to quality standards in pet food. Over the course of six years, a processing facility in Texas shipped low quality, mislabeled ingredients such as feathers and by-products, labeled as premium single ingredients, to pet food manufacturers and distributors. The guilty party had to pay $4.5 millions in restitution to the fraud victims, and the defendant is on a five year probation.
In 1995, a honey processing company was indicted on charges of adulterating industrial honey labeled “USDA Grade A” with corn syrup to increase profits. Ultimately, the jury found in favor of the honey processor, in part because there “weren’t enough regulations governing honey to make the charge stick.”
Honey is defined as “the natural sweet substance produced by honey bees” from the nectar of plants. However, there is not currently an FDA standard of identity for honey in the United States, which would further define and specify the allowed methods of producing, manufacturing and labeling honey (there is, however, a nonbinding guidance document for honey). Some of the details of honey production that a standard of identity might address include allowable timing and levels of supplemental feeding of bees with sugar syrups and the appropriate use of antibiotics for disease treatment.
In circumstances where strict regulatory standards for foods are not available, they may be created by other organizations.
A food standard is “a set of criteria that a food must meet if it is to be suitable for human consumption, such as source, composition, appearance, freshness, permissible additives, and maximum bacterial content.”1
To ensure quality, facilitate trade, and reduce fraud, everyone in the supply chain must have a shared expectation of what each food or ingredient should be. Public standards set those expectations and allow them to be shared. They help ensure that stakeholders have a common definition of quality and purity, as well as the test methods and specifications used to demonstrate that quality and purity. Public standards help ensure fair trade, quality and integrity in food supply chains.
A method is generally an analytical technique to assess a particular property of the content or safety of a food or food ingredient. For example, methods for detection of nitrates in meat products or baby food, coliforms in nut products, or high fructose syrups in honey. Methods are an important component of food standards.
A food standard goes a step further and provides an integrated set of components to define a substance and enable verification of that substance. Standards generally include a description of the substance and its function, one or more identification tests and assays (along with acceptance criteria) to appropriately characterize the substance and ensure its quality, a description of possible impurities and limits for those impurities (if applicable), and other information as needed (see Figure 1).
Figure 1. The Anatomy of an FCC Standard (Source: Food Science Program, Food Chemicals Codex, USP)
A standard defines both what a food or food ingredient should be and documents how to demonstrate compliance with that definition.
Many of the foods prone to fraud are those that are not simple food ingredients, but agricultural products that can be more complex to characterize and identify (such as honey, extra virgin olive oil, spices, etc.). Milk products are an example of a commodity that is prone to fraud with a wide range of adulterants (for example, fluid cow’s milk is associated with 155 adulterants in the Food Fraud Database). Ensuring the quality and purity of a product link milk requires implementation of multiple analytical techniques or the development of non-targeted methods.
The creation of effective public standards with input by a range of stakeholders will be particularly important for ensuring the quality, safety and accurate labeling of these high value commodities in the future.
