Julia Stewart:
Hello, this is Julia Stewart, and welcome back to PMA’s audio blog, “Ask Dr. Bob” with PMA’s Chief Science & Technology Officer Dr. Bob Whitaker. We’re continuing to talk about the benefits and challenges in product testing. Bob, there is a tremendous amount of product testing going on today in some commodities. What are some of the things we should consider when looking at all this testing and the resulting data?

Bob:
I think there are three basic questions we need to ask about testing, as well as what this data may mean and what we might do with it:
1. What pathogens should we be testing for?
2. Does a positive test result necessarily mean that the product would cause disease? And,
3. How can a company use the data generated for product testing programs to improve its food safety program?

First, let’s take the question “What pathogens should we be testing for?” Today, you’ll find many companies who test for E. coli O157:H7 and Salmonella species – others add a test for nterohaemorrhagic E. coli or EHECs – meanwhile still others are beginning to look at shigatoxin-producing E. colis or STEC’s – and a few even include Listeria monocytogenes in their product testing protocols.  Meanwhile, there are still some who look for indicator species like generic E. coli or total coliforms, even though the correlation between these and actual pathogen contamination is questionable in produce. 

So, which organisms should we test for?  A producer needs to make the decision on what to test for based on the risk profile of the particular products they grow and produce, the timing window they have between testing and when they need to ship, and the costs involved.  It’s important to avoid a “one size fits all” approach.  Instead, we should employ a science- and risk-based approach to determine commodity-specific and/or pathogen-specific testing strategies. 

For example, let’s say we’re a tomato grower. Then we’d clearly want to prioritize testing for Salmonella species, since this pathogen is most commonly associated with tomatoes.  On the other hand, if we are a leafy greens grower, then we’d want to broaden the pallet of tests to include Salmonella and E. coli O157:H7 and perhaps consider adding EHEC or STEC’s to cover all pathogenic strains historically associated with these foods. 

Julia:
So let’s look at the second question: Does a positive test result mean the product would cause disease? 

Bob:
You can look at this question from a number of perspectives.  Clearly, the most important is public health.  The Food, Drug and Cosmetic Act – that’s the federal food safety law we in the produce industry are currently subject to – considers product that tests positive to be adulterated, and prevents that product from being put into commerce.  The Food and Drug Administration will always act first and foremost to protect public health.  Further, as of September 2009, you have the obligation to report any products that may pose a public health risk via the Reportable Food Registry – so, if you had a positive test result and the product is in the marketplace, you’d need to report that to FDA. 

From a strictly scientific perspective, we know that new strains of E. coli have emerged in the last 30 years, most notably, E. coli O157:H7; and, unlike its harmless brethren, as few as 10 cells of O157:H7 can cause significant health issues in a person, particularly children, the elderly or other at-risk populations. Conversely, we can now identify more than 2,000 strains of Salmonella, whose illness-causing dose rates are much higher than with E. coli O157:H7, if they cause human illness at all.  Also, Salmonella infection is generally not lethal, and we’re learning that Salmonella may be more common in the environment than originally thought. 

So, while the zero-tolerance approach currently mandated by the FD&C Act is appropriate for a pathogen such as E. coli O157:H7, today’s science may or may not justify such a strict standard for other, perhaps less harmful pathogens.  We’re also starting to understand that perhaps a positive PCR test result or even a bacterial culture may not necessarily correlate to that organism’s ability to cause human illness.  Research is showing that our screening for DNA fragments that are associated with the pathogen’s ability to produce toxin, or to attach to the intestines, might not by themselves be proof that the organism can cause disease.  Other factors such as the pathogen’s acid tolerance and the physiological state of the pathogen likely play a critical role. 

Julia:
So, how can the data generated by a product testing program be used?

Bob:
Obviously, enacting a product testing program is an important decision and a significant financial commitment.  So, it seems reasonable that a company that chooses to do product testing should look to use the resulting data to improve their food safety programs.  In order to do that, you must have procedures in place to receive testing data from lab partners, to store and sort that data, and then analyze the data so that trends can be identified and acted upon.  For instance, you may be able to use product testing data in conjunction with additional data from your GAP or HACCP programs to identify and prioritize specific risk factors.  As an example, a preharvest risk assessment that identifies an animal intrusion event has occurred may be used in conjunction with a concentrated in-field product testing program to verify if the intruding animal species carried a pathogen, and perhaps even what buffer zone distances are necessary to manage the potential risk. 

Julia:
Bob, you’ve raised some interesting issues here to consider when setting up product testing programs.  How much data from product testing do you think is out there in the industry?

Bob:
Julia, that’s difficult to say.  Right now, the clear focus has been in leafy greens crops.  Some companies have large, sophisticated product testing programs that have generated thousands of data points.  In the last year there’s been a great deal of discussion around how industry might combine some of this data or “mine” it to see if it’s possible to prioritize risks.  We talked about this with FDA during a meeting after the Center for Produce Safety’s research symposium in June.  Sometime in the near future, we should do a blog series on data mining, as it’s a subject all unto itself with a number of positive and negative aspects.

Julia:
That sounds good, Bob; I’ll hold you to it!  Speaking of holding you to it, I promised our listeners that as we concluded this blog series, you’d sum up the real value of product testing.  So next time in our final blog in this series, let’s summarize some of the discussions we have had on this subject and see what the plusses and minuses are.  Until then, thank you, listeners, and we’ll see you back here next time.

Julia Stewart:
Hello, this is PMA PR Director Julia Stewart, and welcome back to PMA’s audio blog, “Ask Dr. Bob” with PMA’s Chief Science & Technology Officer Dr. Bob Whitaker. Bob and I have been talking about the challenges with product testing in produce. Bob, this time we promised to get into the role of confirmation testing. Why is it important for producers and buyers to understand the benefits and limitations of this type of testing, and what does it mean to supply chain logistics?

Bob:
In previous posts we discussed rapid tests such as PCR tests, and the benefits of those types of tests – mainly speed.  Because of the pressures our industry is under to sample, test and get results back so that loads can be harvested or shipped, speed is king as they say and these DNA-based test methods have become the standard in the industry.  But, we have also discussed some of the issues with these types of tests; mainly they are not always sensitive or accurate enough. 

We have all heard of instances where initial positive samples fail to be confirmed on follow-up testing.  It is extremely important to be sure a sample is in fact positive, because growers and processors are often making decisions not to ship or harvest thousands of dollars worth of products.  And of course, adding another round of testing costs money and time.  You all might recall that as little as a few years ago, some producers recalled product based on initial test results only to find out later that the original result might not really have been positive for a pathogen, but was simply a closely related non-pathogenic bacteria that shared many similar DNA fragments.

So, this is where “confirmation testing” enters the picture. Initial positive samples are often re-tested using another type of test.  To be clear, this isn’t just running the same test again – this is using a completely different test to determine if the first was accurate or not.  Some choose to use FDA recognized microbiological plating methods commonly referred to by many as “BAM testing” – “BAM” stands for FDA’s Bacteriological Analytical Manual.  Others use another round of PCR-based testing focusing on a more diverse set of DNA primers characteristic of a specific pathogen.  Sometimes, depending on the pathogen, there are other types of tests that can be used.  For example, if an initial test indicates E. coli O157:H7, one might opt to use an immunological-based test to determine if shigatoxin, which is characteristic of E. coli O157:H7, is present. 

Julia:
Wow, sounds like a very tricky issue. I see why you stress that anyone who is testing their products understand the benefits and limitations of each approach. 

Bob:
That’s right Julia, there are pluses and minuses for any of the approaches I just described. 

For example, using another round of PCR-based testing has the advantage of being fast.  Once you have the “positive” results from the initial testing, you can have the lab immediately run a second round of testing using a different set of primers designed to the target the pathogen.  You can have results back within another 12 to24 hours using this approach.  The logic is that the more DNA primers you test for, the more likely it is that your original positive test was correct – and, correspondingly, the less likely that the positive test was just a close cousin of the pathogen that shared some of the pathogen’s same gene sequences.  The drawback is that you have to be certain the primers you use really uniquely distinguish pathogens from non-pathogens.  The science is not entirely conclusive on this matter yet, so you have to work with your labs to understand the specificity and selectivity of the probes used in your test and the relative significance of the number of unique primers used in building the confidence that your results are accurate.  

On the other hand, BAM testing is kind of the gold standard for microbiologists.  It’s based on the ability of bacteria to grow on different media “recipes”.  Every microorganism’s DNA genetically codes for specific proteins or enzymes that allow that microorganism to use certain sugars and nitrogen sources to grow.  Different bacterial species, of course, have different DNA and therefore differ in which sugars and nitrogen sources they can grow on.  Over time, microbiologists have developed a complex set of testing media which can distinguish specific pathogens from related non-pathogenic strains.  These media recipes are well recognized and accepted by scientists and can be a very powerful tool to identify specific pathogens.  

So as a confirmation testing tool, BAM testing can be very specific and accurate.  It can distinguish between closely related species of bacteria and at the end of the confirmation test, if the sample tested positive by PCR and BAM methods, you can be reasonably confident that you have a  pathogen-positive sample and a live culture of the pathogen from the lab   However, BAM testing isn’t without issue itself. These tests take time, anywhere from 48 to 96 hours depending on how you handle samples.  It is also important to note, that there is also a degree of human discretion in reading BAM results, and bacterial colony shape and colors can be interpreted incorrectly on occasion. 

Julia:
So, confirmation testing is an important tool to verify initial positive results and give the producer and the buyer confidence that products either are or are not truly contaminated with a pathogen.  Thanks, Bob.

Next time, we’ll talk about a very complex area of concern: product sampling.  So, listeners, please join us for what is sure to be an interesting discussion.  

Remember you can email Bob at AskDrBob@pma.com. In addition to listening to these and other Ask Dr. Bob blog posts, we invite PMA members to visit our new online Food Safety Resource Center on PMA.com and check out the lab testing white paper in the Education section. Thanks!

Julia Stewart:
Hello, this is PMA PR Director Julia Stewart, and welcome back to PMA’s audio blog, “Ask Dr. Bob” with PMA’s Chief Science & Technology Officer Dr. Bob Whitaker. We’re in the middle of a continuing series on the challenges of pathogen testing in produce. In the last post we talked about how a positive is not always a positive and the importance of looking at the selectivity and sensitivity of tests.

Bob, in the last post you also said that a negative test result might not actually be a negative. That seems even more worrisome to me. I mean if you test a product and the result is negative and you ship it out only to find later that it was positive, that can be a big problem.  Can you elaborate on that?

Bob:
Sure, Julia.  There are basically three issues to look at here: enrichment, interference and sampling issues.  Let’s take enrichment first.  Enrichment is simply the method used to permit any bacteria that might be in a sample to grow and multiply so you can get enough DNA to be able to isolate it and start the process of determining if any of it comes from a pathogen.

Since pathogens are typically present only in very low levels, when present at all, this process of enrichment can be critical.  You also have to appreciate that these pathogens thrive in warm, humid, high nutrient environments — in other words our intestinal tracts and the surface of a fruit or vegetable is not necessarily the best environment for them.  This enrichment method permits these bacteria to recover, if you will, and begin growing again.  So, it’s important that conditions and times are maximized to permit pathogen growth.  This typically takes 12 to 18 hours.  Failure to use proper enrichment methods can result in pathogens going undetected even though they may have been present in the original sample.

If that wasn’t enough of a challenge, we’ve got the issue of interference. Fruits and vegetables are extremely complex in terms of their chemistry. They’re made up of thousands of chemicals, and that composition can sometimes change dramatically based on variety or growing conditions.  Some of these chemical constituents can actually interfere with pathogen detection by interfering with DNA isolation, purification or detection.  It seems like it is almost necessary to develop specific isolation protocols for every commodity to be sure the tests are selective and sensitive enough to find pathogens if they are present.

Lastly, we have the sampling issues. If your sampling program is not sufficient, it is quite possible that if there is a low level or non-uniform contamination event, like we typically see in produce, that you could simply miss including contaminated fruits or leaves in your sample.  We’ll talk about this more in a future post.  In fact, I think this is actually a more important consideration in product testing than even the tests themselves so it warrants a more extensive discussion.   But in the end, if you don’t have a dynamic sampling program, you can easily miss contamination and get a “negative” result for a field or batch of finished product that might indeed be contaminated.

Julia:
Thanks Bob.  You’ve raised some good points here.  After hearing this, I can see why you always say that food safety programs should first focus on preventing contamination from ever happening.  Detecting it after the fact is not a simple issue.  I can also see why you repeatedly emphasize how important it is for those doing testing to ask questions and fully understand what their labs are doing, and what the data may or may not mean. 

Next time, let’s dive into this issue of confirmation testing and how that can be used to verify these initial or PCR-based screening test results.  Listeners, thanks for tuning us in today and we look forward to speaking with you further on this issue of product testing.