While Paul Simon used “One man’s ceiling is another man’s floor,” to refer to apartment house rules of courtesy, you might find a parallel to this separate-but-connected concept in the food supply chain. That is, one producer’s output is another processor’s input; just as you test and verify your finished product before it goes out the door, you expect no less from your supplier. This is especially true if you have no “kill” stage in your process. But there’s a major difference between apartment house rules and food safety: The latter is the law of the land under FSMA.

“For products that do not use a cooking or ‘kill’ step, a robust, reliable testing program is particularly important to help minimize the risk of contaminated ingredients moving forward in the production chain and potentially reaching the market,” says Jun Li, senior research investigator, diagnostics, DuPont Nutrition and Health. “Under FSMA, the initial farmers/growers/producers are first responsible for taking the appropriate measures to minimize or eliminate the risk of illness-causing contamination.”

“The FDA FSMA is very clear in conveying that food manufacturers are ultimately the party responsible for ensuring the safety of the food they sell," says Kevin Habas, 3M Food Safety global communications, scientific marketing and education manager. “The onus rests on food brands to reach back to their suppliers to make sure that those partners have proper food safety plans in place and are taking all the necessary precautions to mitigate and minimize health risks. This is true when working with both domestic and foreign suppliers.”

In fact, in its Sec. 103. Hazard Analysis And Risk-Based Preventive Controls, FSMA states, “(4) the preventive controls implemented under subsection (c) are effectively and significantly minimizing or preventing the occurrence of identified hazards, including through the use of environmental and product testing programs and other appropriate means; and (5) there is documented, periodic reanalysis of the plan under subsection (i) to ensure that the plan is still relevant to the raw materials, conditions and processes in the facility, and new and emerging threats.”

Since FSMA is a risk-based program, microbiological hazards must be reduced or eliminated along the supply chain. This means it’s necessary to determine a preventive control point and establish parameters to reduce risk. Monitoring and corrective actions also must be established, as well as the verification of the effectiveness of the preventive controls.

“A key aspect of FSMA is establishing a supply chain verification program,” says James Cook, SGS global food inspection technical manager. “If the verification program establishes the supplier COA [certificate of analysis] contains the appropriate verification information, and the preventive control parameters are in compliance, the process of acceptance has been confirmed through the verification of the effectiveness.”

But can you trust your supplier’s COA? “If you can’t trust your supplier’s COA, it shouldn’t be your supplier,” says Cook. “President Ronald Regan frequently quoted the old Russian proverb, ‘Trust, but verify.’ Trusting a supplier doesn’t mean you shouldn’t verify its information is correct.”

Some may complain that existing FSMA rules are already too tough and exhaustive, but there is an aspect of common sense that could go a long way in preventing major headaches and potential injuries. “Although not spelled out in the FSMA rules, it is extremely important to the food industry that testing for any adulterants occur in a manner that allows for the product to be held until the test results indicate that the product has tested negative for the adulterant,” says Timothy Freier, division vice president, scientific affairs and microbiology, North America, Mérieux NutriSciences.

What can happen when a product is not held? Freier explains, “Recently, roasted sunflower kernels were tested for Listeria monocytogenes months after the product had been produced and widely distributed. Dozens of downstream customers had to recall millions of pounds of product because the sunflower seeds were used as an ingredient, then distributed before there was any knowledge of the adulteration. If a customer needs a supplier to test an ingredient for an adulterant, arrangements should be made for testing to be done by the supplier, or a ‘pre-shipment’ sample be tested to allow the product to remain under hold until the test results indicate that it is safe to be released.”

As a follow-up note, the original producer’s first recall was issued on May 3, 2016, covering products made between the dates of February 1-19, 2016. An expanded recall added products made from February 20-April 21, 2016—the date on which the facility ceased production. Further recalls from other downstream brands using the same sunflower seeds have progressed through June 20, 2016.

Sampling programs

As the Peanut Corporation of America (PCA) recalls demonstrated, the science—not art—of knowing where to take samples is crucial in discovering the presence of bacteria. “Sampling programs must be tailored to the materials and subsequent processes,” notes Stuart Ray, technical director at Seward Laboratory Systems Inc. “A producer must design a credible sampling program based on knowledge of its products. Batch numbers, dates of manufacture and sources may all be a good basis, but there is a possibility very large batches may spoil in a variety of places.”

Although the quantity, location and frequency of sampling are widely dependent on factors such as the facility size, production volume and history of microbial contamination, a robust testing program ensures the number of samples tested is statistically significant, says DuPont’s Li. “According to the specifications in the US FDA Bacteriological Analytical Manual [BAM] for food sampling, a product being tested for Salmonella contamination should include at least 15 separate, randomly selected samples for products that undergo a ‘kill’ step—or at least 60 samples for products that do not.”

“The decision of how often or where to sample needs to be based on a hazard analysis of the product and process,” adds Freier. “If the risk for a particular pathogen is high, every lot produced might be tested [with product held until the test results are known]. If the risk is low, occasional verification samples may be analyzed. Sampling should be done where it will have the greatest impact on reducing risk.”

“Unfortunately, there is no magic rule of thumb for the number of samples that should be taken,” observes Cook. For this reason, there are many sample plans and sampling procedures, such as ANSI/ASQ, US Military, Codex, US FDA, USDA FSIS, USDA AMS, USDC, CFIA, ISO and AOAC. If the product is under FDA jurisdiction, the number of samples taken and the procedure to take them should be those specified by that agency. Sampling frequency should be based on the need to verify any hazard is being reduced and eliminated.

Mycotoxins, the chemicals created by molds and some fungi, are dangerous because they are heat-stable and often survive food production, processing and cooking steps. While it doesn’t take many mycotoxins to spoil a batch, sampling for them can be a real challenge, according to Marjorie Radlo-Zandi, managing director VICAM, a Waters Business. “Sampling is the source of greatest variability when testing raw materials for mycotoxins. Because the majority of contamination in a single lot or load of product may be contained in just one (or a few) nuts or kernels of grain, samples should be taken from a number of locations in the shipment, bag or storage container, including low-lying sections where broken or damaged pieces may collect during movement or transfer.”

Traditional and rapid tests

Standard test methods for pathogen detection (i.e., conventional or reference methods) generally involve one or two enrichment steps in a generic medium, streaking the enriched sample onto selective agar plates, and confirming any typical colonies that grow with a series of biochemical and serological identification techniques, says DuPont’s Li. “These plating methods are still considered the ‘gold standard’ in terms of their accuracy and are generally less expensive than rapid or proprietary methods. However, standard methods are the most labor-intensive, requiring an expertly trained lab technician to perform the procedure and interpret the results. Plus, these methods are the most time-consuming; in some extreme cases, results may not be available for up to two weeks.”

Four main technology options are used to test for pathogens. 3M’s Habas lists them below, from oldest to newest.

• Traditional culture methods—Although Petri dishes and poured plates like agar have been around for a long time and can seem primitive, they are still used worldwide for growing and evaluating organisms.
• Enzyme-linked immunoassay (ELISA) and lateral-flow test methods—These tests require samples to be enriched and added to a well, before extending up a wick that reveals lines indicating if the specific organism is present.
• Polymerase chain reaction (PCR) methods—These tests allow food processors to take multiple steps to prepare reagents and extract, or reproduce strands of DNA, to signal whether the DNA of a certain type of pathogen appears to be present.
• Molecular-based methods—These technologies amplify DNA at the molecular level, improving accuracy and minimizing time and technician steps. For example, the 3M molecular detection system targets a gene with a sample using many more primers than PCR can and employs a special enzyme to help continuously read amplified DNA results without interruption.

“The key issue with rapid tests is how ‘rapid’ they are,” offers Alan Traylor, business manager, microbial detection, MOCON, Inc. “Most official methods are rooted in manual processes from a past era. They are provable with science but slow.” Traylor lists four important considerations in selecting a rapid method:

• How does it compare with older methods?
• Is there an unseen overhead cost in sample prep that negates the speed gain?
• Does the cost of the test create value in the time-to-result and preparation savings?
• Does the rapid or new method outperform the old so much that comparisons are hard to achieve? This is the case with total plate count, where the rapid methods have greater sensitivity than agar plates and, therefore, can measure much lower values (smaller number of organisms).

According to Mérieux’s Freier, rapid detection technologies are shortening the turnaround time of tests in a number of ways:

1. The enrichment can be optimized to allow faster growth.
2. Immunomagnetic concentration, filtration and centrifugation may be employed to concentrate the target or reduce the background.
3. Detection systems are being developed that can “see” smaller and smaller numbers of target organisms. In fact, the detection capabilities of some rapid methods can be so sensitive that dead cells can trigger a reaction.
4. Turnaround times for many foodborne pathogens have decreased from three to seven days for cultural tests to one to two days for rapid tests.
5. A few same-shift assays (less than eight hours) have been validated, but for only a very few pathogen/food matrix combinations.

“Rapid methods must be chosen with care as they are not viable for all products, and an authorization body should have verified they perform as indicated,” says SGS’s Cook. “The USDA FSIS, AOAC, CFIA and others publish this information.”

Viruses and mycotoxins

If you think the transmission of viruses is limited to restaurants and foodservice workers, think again. About half of the foodborne illnesses in the US are associated with norovirus, according to CDC. Norovirus isn’t the only culprit. Cook points to September 20, 2013, when CDC confirmed 162 people became ill from products containing pomegranate seeds contaminated with the Hepatitis virus genotype 1B.

“In 2012, frozen oysters were found to be contaminated with norovirus. While it is not necessary to test all products for viruses, for some products, such as mollusks, raw fruits, leafy greens and raw seeds, this may be a cost-effective measure. Failure to detect viruses is expensive and risks the prospect of spending millions of dollars in recalls and more money settling lawsuits,” adds Cook. (Currently, PCR and DNA methodologies are the general methods used to test for viruses.)

“One of the limitations of testing for viruses is they cannot be enriched in the same way as bacteria, yeast and molds,” notes Freier. “This greatly limits the sensitivity of virus assays.” Some manufacturers are beginning to test routinely for viruses such as norovirus, and Mérieux NutriSciences provides this service.

Mycotoxins and aflatoxin (a mycotoxin produced by Aspergillus molds) are highly toxic and can be carcinogenic. Mycotoxins tend to occur in a very non-homogeneous way throughout a crop field, storage container or shipment and may be localized within a single kernel of grain or one small section of a farm field or orchard, says VICAM’s Radlo-Zanti. “A single kernel of corn may contain several thousand parts per billion [ppb] of aflatoxin, and success in detecting the contaminated product depends on good sampling practices, followed by onsite test methods or laboratory analysis by liquid chromatography [LC] or LC/MS/MS. While COAs provide a snapshot of raw material quality and safety, processors benefit greatly from the ongoing verification of raw materials prior to acceptance, storage or processing.”

Raw material producers can monitor for the presence of mycotoxins using VICAM field-ready, lateral flow strip tests, which offer qualitative screening (yes/no) and quantitative (numerical) results within five to seven minutes. The advantage of strip tests is that they can be performed almost anywhere, with little or no training or expertise required. Rapid mycotoxin methods can detect and quantify at very low concentrations—as low as two ppb up to 300 ppb for total aflatoxins in grain, tree nuts or peanuts, says Radlo-Zanti.

Extending the reach of tests

Over the last few years, more testing options have become available to processors. “Now, it is easy for a small processor with access to aseptic sampling and minimal training to achieve rapid-test data for plate count, coliforms, yeasts and molds—and other organisms like Pseudomonas genera,” states Traylor.

“Most facilities, regardless of their size, are well-equipped to conduct indicator testing, ATP hygiene monitoring and sample collection,” says 3M’s Habas. “When it comes to rapid pathogen—and to some extent, toxin—testing, affordable options, which can be performed by individuals of varying skill levels, are available.”

“The decision as to where testing is performed may depend more on the preferences of the company than on the details of the testing method,” adds DuPont’s Li. “For example, a company that values speed may elect to perform a rapid test method in-house, regardless of cost. On the other hand, a company with a limited budget for product testing may outsource its testing and request the most cost-effective method to avoid the larger capital expenditure of establishing an internal testing laboratory.”

“The purpose of the laboratory, whether a commercial or in-house facility, is to produce correct results,” explains SGS’s Cook. “However, humans are involved in the process, so errors do happen. Quality assurance is necessary, so when errors occur, a root-cause analysis can be performed to determine the cause, and the necessary corrective and preventive actions can be taken to avoid recurrence.”

Cook also notes QC procedures must be in place for negative samples, positive samples and percentage duplication for pathogen testing, as well as the verification of standard curves and blanks. Report files must be reviewed, and documentation must be done to ensure written procedures are being followed. Finally, test methods must be validated; equipment must be calibrated and kept in good working condition; personnel must be trained in the methodology; and their ability to perform the methodology as written must be verified. Good laboratory practices must be followed—whether in-house or at a commercial lab.

For more information:

Jun Li, DuPont Nutrition & Health, 800-863-6842,
nicolette.kerr@dupont.com, www.fooddiagnostics.dupont.com

Kevin Habas, 3M Food Safety, 800-328-1678,
kahabas@mmm.com, www.3m.com/foodsafety

James Cook, SGS/Agriculture, Food and Life, 973-575-5252,
james.cook@sgs.com, www.sgs.com

Timothy Freier, Mérieux NutriSciences, 312-324-3411,
tim.freier@mxns.com, www.merieuxnutrisciences.com

Stuart Ray, Seward Laboratory Systems Inc., 954-862-1421,
stuartray@seward.co.uk, www.seward.co.uk

Marjorie Radlo-Zandi, VICAM, 800-338-4381,
vicam@vicam.com, www.vicam.com

Alan Traylor, MOCON Inc., 763-493-6370,
atraylor@mocon.com, www.mocon.com