• Decrease font size
  • Return font size to normal
  • Increase font size
U.S. Department of Health and Human Services

Food

  • Print
  • Share
  • E-mail

Transcript of Proceedings of Public Meeting on Sprout Safety, May 17, 2005

Back to Food Safety Information Concerning Sprouts


 

2005 Sprout Safety Public Meeting

Transcript of Proceedings


 

Aside from minor spelling and style corrections, the wording of this document is exactly as received from the commercial transcribing service. Accordingly, FDA makes no representation as to its accuracy.


Contents

Opening Remarks

Robert Brackett, Ph.D.

Overview

Nega Beru, Ph.D.

Government Perspective:

Foodborne Illness Outbreaks
--Amy Dechet, M.D., Epidemic Intelligence Officer, Foodborne and Diarrheal Diseases Branch, Centers for Disease Control (CDC)
(presentation available upon request)

Sprout Guidance, Background and Next Steps
--Michelle Smith, Ph.D., Interdisciplinary Scientist, Office of Plant and Dairy Foods, CFSAN, FDA
(presentation available in PowerPoint)

Questions and Answers

Industry Perspective:

State of the Sprout Industry
--Bob Sanderson, International Sprout Growers Association (ISGA)

Seed Handling and Distribution Systems
--Bob Rust, International Specialty Supply

Consumer Perspective:

--Caroline Smith DeWaal, Center for Science in the Public Interest
(presentation available upon request)

Questions and Answers

Scientific Perspective:

Microbiological Findings on Sprout Operations Following FDA Guidance
--Jed Fahey, D.Sc., Johns Hopkins School of Medicine
(presentation available in PowerPoint)

Interventions
--William Fett, Ph.D., Eastern Regional Research Center (ERRC), ARS, USDA
--Kathleen Rajkowski, Ph.D., Eastern Regional Research Center (ERRC), ARS, USDA
(presentation available in PowerPoint)

Testing Methodologies and Sampling
--Mary Lou Tortorello, Ph.D., Research Microbiologist, Office of Plant and Dairy Food, CFSAN, FDA
--Tong-Jen Fu, Ph.D., Research Chemical Engineering, Office of Plant and Dairy Foods, CFSAN, FDA
(presentation available upon request)

Alternative Seed Sanitation Methods and the Results of Practical Field Applications
--Kean Ashurst, Caudill Seed Company
(presentation available upon request)

Risk Analysis
--Don Schaffner, Ph.D., Rutgers
(presentation available upon request)

Questions and Answers

Public Comments

Summary of Meeting and Adjourn

Nega Beru, Ph.D.


Proceedings

DR. BERU: Good morning, everyone. I guess we'll get started. We said the program would start at 8:30. We gave a little time for some of the stragglers, traffic and so on but good morning and I would like to welcome you to this public meeting on sprouts.

My name is Nega Beru and I'm the Associate Director of the Office of Plant and Dairy Foods in CFSAN. I also am the moderator for this meeting.

As you know, this meeting is intended to elicit information on the current science related to foodborne illness associated with the consumption of sprouts.

Before I introduce the agenda and the panelists for our meeting, I would like to invite Dr. Robert Brackett, Director of the Center for Food Safety and Applied Nutrition, to make his opening remarks.

Opening Remarks

DR. BRACKETT: Thank you, Nega.

Good morning to all of you and welcome to this public meeting on sprout safety. It really is a pleasure to be with you here today and I had hoped to be here the entire day with you but I had a hearing that was changed to today so I will stay for a little while and head downtown right away again.

Microbial food safety of fresh produce is a priority for CFSAN and so is the role of fresh produce and a healthy diet, which is also important to us.

In October of 2004, CFSAN finalized an action plan for fresh produce. This meeting continues our dialog with stakeholders on how to ensure that foodborne illnesses associated with fresh produce, and that includes sprouts, is minimized to the greatest extent possible.

During the past decade, over 20 percent of all produce related foodborne illnesses were associated with the consumption of raw or lightly cooked sprouts. CDC first brought sprouts to our attention as a vehicle for foodborne illness back in 1995. In a 1998 white paper on fresh produce, the National Advisory Committee on Microbiological Criteria for Foods identified raw sprouts as a special food safety problem. In 1999, the National Advisory Committee on Microbiological Criteria for Foods issued a report entitled "Microbiological Safety Evaluations and Recommendations on Sprouted Seeds." Since then, FDA has issued several consumer advisory warnings about health risks associated with the consumption of raw sprouts. FDA also released two guidance documents, one on seed disinfection and the other on testing irrigation water, concerning practices to minimize microbial contamination of sprouts. We have worked collaboratively with other agencies, groups, and the industry in a variety of different ways. Although the guidance documents were well received, and their reissuance and adoption appears to have resulted in some improvement, outbreaks implicating sprouts continue to occur.

As a way to address the public health risk associated with the consumption of raw and lightly cooked sprouts, FDA is considering the development of a regulation. The purpose of this meeting is to elicit information on the state of the science with respect to sprout safety and intervention strategies and also to engage you, the stakeholders, early on in the rulemaking process.

We believe that the most effective strategy for reducing foodborne illness from raw and lightly cooked sprouts is likely to be one that approaches the problem from several different angles. For example, we need to look at the seed producers and distributors as well as the sprout producers to understand the practices that contribute to the contamination of sprouts with human pathogens. It's important that we consider the views and ideas of all of our food-safety partners. By working together, we think we can achieve this goal. Consequently, I am looking forward to hearing your views and seeing them later on, and your comments, and I hope that you find that the discussion is worth your while as well.

So with that I would like to again welcome you to CFSAN and wish that I could stay for the whole day. Thanks.

Overview

DR. BERU: Thank you, Bob.

As you will note from the agenda you picked up on your way in, we have a pretty fully day but before I go over the agenda with you let me just say a few words about housekeeping.

The all important rest rooms--this is a full day meeting so I should point that out first--are upstairs where you registered as you came in that hallway. There are also rest rooms down the hall out this door.

We will provide refreshments during the breaks but you are on your own for lunch. There is the Wiley Cafe, which you probably saw as you walked into the building. And if you drove, there are a number of places to eat up and down Route 1, as well as on Kenilworth Avenue.

We have a number of people who have registered to provide comments at the end of the formal presentations during the comment portion of this meeting. If you have not registered but still wish to make a comment, please see me some time during the meeting and we'll add you to the list.

And when you ask questions or make comments, please wait to get the microphones or step to one of the two microphones on the sides of the auditorium. We are making a transcript of the meeting and we want to be sure we capture everything, including your name and affiliation. The transcript is going to be made available on our website in about four weeks time.

Now let me go over the agenda. It is in your packets if you want to follow along. We've had the opening remarks by Bob Brackett, our Center Director.

We'll first have a session which we've termed "Government Perspective". Dr. Amy Dechet from the CDC, who is an Epidemic Intelligence Service Officer in the Foodborne and Diarrheal Diseases Branch of the CDC, will first present on foodborne illness outbreaks.

That will be followed by Dr. Michelle Smith, Interdisciplinary Scientist in the Office of Plant and Dairy Foods, CFSAN. Dr. Smith will present on the sprout guidance, background and next steps.

I'll ask you to hold questions until both speakers have gone through their presentations and they will come to the front and then we'll have a question and answer session.

We'll take a break at that point and we'll go into what we term the "Industry and Consumer Perspectives." First we'll have a presentation by Mr. Bob Sanderson, International Sprout Growers Association, on the state of the sprout industry.

And then Mr. Bob Rust from the International Specialty Supply will present on seed handling and distribution systems.

And we'll have a consumer perspective by Ms. Caroline Smith DeWaal from the Center for Science in the Public Interest.

When all three have done their presentations, again we'll have a question and answer session of those three panelists.

Then we'll go into what we call the "Scientific Perspective" section of the meeting. First microbiological findings on sprout operations following FDA guidance by Dr. Jed Fahey from Johns Hopkins School of Medicine.

And we'll take a few questions but we'll hold most of the questions for Dr. Fahey until after the other scientific presentations have--we've gone through the other presentations. We'll have lunch at that point. As I say, lunch is on your own.

And then after lunch the first presentation will be by Dr. William Fett, Eastern Regional Research Center, on interventions, and Dr. Kathleen Rajkowski, Eastern Regional Research Center again from USDA, ARS.

That will be followed by "Testing Methodologies and Sampling" by two of our own scientists at the Moffitt Center, Office of Plant and Dairy Foods, CFSAN. Dr. Mary Lou Tortorello, who is a research microbiologist at the Moffitt Center and Dr. T-J Fu, who is a research chemical engineer, also of the Moffitt Center.

That will be followed by "Alternative Seed Sanitation Methods and the Results of Practical Field Applications" by Kean Ashurst--I hope I'm saying that--pronouncing the name correctly--Caudill Seed Company.

And the last of the scientific presentations will be one on "Risk Analysis" by Dr. Don Schaffner from Rutgers.

There will be a question and answer session again of the scientific perspective presentations. We will take a break and we'll go into the public comment section of the meeting. And then I will summarize the meeting and we will end that.

With that, I would like to invite Dr. Amy Dechet from the CDC to present on "Foodborne Illness Outbreaks".

As I said, please hold until both Drs. Dechet and Smith have made their presentations.

Government Perspective

Foodborne Illness Outbreaks

[See presentation slides for Dr. Dechet]

DR. DECHET: Good morning and thank you for the opportunity to speak with you today.

I thought to begin this meeting it would be helpful to give you an overview of some of the characteristics of sprouts and their outbreaks, and put it in the context of the produce-associated outbreaks as well as foodborne outbreaks in general.

Each year in the United States we estimate that there are over 76 million foodborne illnesses, over 300,000 hospitalizations and over 5,000 deaths. Now these numbers sound large and they are large but what does this really mean?

This means that one in four Americans will become ill. So if you look in the room around you and look at three other people, one of you is probably going to become ill from some sort of foodborne illness this year. One in 1,000 Americans are expected to be hospitalized and this will result in over $6.5 billion in medical and other costs such as time lost from work.

The Centers for Disease Control and Prevention and, in particular, the Foodborne and Diarrheal Diseases Branch, where I work, maintains a Foodborne Outbreak Surveillance System. The Local and state health departments are a primary contributor to this system as they are the ones responsible for detecting, investigating and controlling outbreaks. At CDC we will participate in these investigations when we are specifically requested by the states to do so and we do some field investigations every year with states and provide over 100 phone consultations as well.

Once information is gathered from the states it's usually sent to the CDC where we collect data on the number of cases, the implicated food and the etiology. To be entered into the system, we define an outbreak as two or more cases of a similar illness resulting from the ingestion of a common food.

I want to point out that this system is completely voluntary. States are not required to report, although most do participate and some of the data we get is incomplete. It could be because at the state level they were unable to get the information or it simply did not make it on to the form. I'll also point out that some of the information that we have may not match exactly the same numbers that FDA has and this is because again our systems are slightly different.

If you look at now at our Foodborne Outbreak Surveillance System over the past decade or so, you'll see on the X axis the year and on the Y axis the number of outbreaks per year. You notice that in 1998 there is a significant jump in the number of outbreaks. This is probably an artificial increase because we actually enhanced surveillance quite a bit during that year and, as you can see, since that time we've averaged around 13 to 1,400 outbreaks per year.

Looking specifically at produce associated outbreaks from 1998 to 2002 where we have the most complete data, we know of 249 outbreaks, and this represents six percent of outbreaks with a reported food source and 13 percent of outbreak associated cases.

If then we turn to the implicated produce, many of these produce outbreaks are sort of a generic or multiple produce item outbreak where people mention salad or salad was the implicated item but then we also do have some information on produce specific outbreaks. And as you can see the top six there are highlighted in yellow, which includes lettuce, sprouts, juice, melon, tomato and berries.

Taking those same top six now I've again listed here the number of outbreaks. Again this is up to 2002. And then the percent of population eating that food in the week prior to interview. And this is based on our FoodNet population survey where people are randomly called and asked, "Did you eat this food item in the seven days prior to interview on the phone?" And as you can see lettuce, which is the top of our list of outbreaks, over 70 percent of people reported eating lettuce in the week prior to interview. The rest of them, excluding sprouts, range anywhere from 21 to 68 percent but, if you look at sprouts, only eight percent of people recall eating sprouts in the week prior to the interview on the phone and yet it's the second leading cause of produce associated outbreaks.

Now we fondly refer to sprouts as our stealth vehicle in foodborne outbreaks because many people don't realize they're eating sprouts. They're often in salads or in sandwiches and people don't remember that they ate sprouts. In fact, in many of the outbreak investigations only 30-40 percent of people will outwardly recall eating sprouts but then when we look at other implicated food items and break it down by ingredients we often do find sprouts then in those food items.

So what is it about sprouts that makes them so unique? Many of you, I think, probably already know this information. I'll just briefly review it. There are, of course, multiple opportunities for contamination along the spectrum from the farm to the table. At the field level there are birds flying overhead, there are animals passing through, and even if animals are grazed on the field months beforehand we have some evidence to show that actually the seeds can be contaminated when they are grown months later.

During the harvesting process sometimes a few combines are used for many different fields and they are also processed in one large central processing facility and so, therefore, if you have a contaminated batch it can actually then mix with other batches and contaminate a larger amount of sprouts or seeds.

During the scarification process where the seeds are rubbed between hard surfaces in order to crack the seed it's a perfect time for bacteria to enter into that seed and then sit there and actually they can survive there for a several months period of time.

Then, finally, there's multiple times where during transport it could become contaminated. From the field to the processing facility to distribution by the sprouters to the grocery stores and back home. So you've got many people handling the sprouts and again bacteria can amplify throughout the process.

And then, of course, there's the sprouting process as well. This is a lovely microbiological environment for bacteria to grow. It's warm. It's moist and, depending on the literature that you read, anywhere from a two to four log increase in colony forming units per gram can occur.

Unfortunately, it's difficult to detect pathogens. The contamination is usually non-homogenous so if you sample one part of the seed or the sprout you may not actually be getting a representative sample of levels of contamination elsewhere. Also, there can be low levels of contamination which are difficult to pick up on culture methods.

Finally, sprouts are rarely cooked or washed by the consumer and this is often our final safety net for many food items that at least you can hope that the consumer will cook things well and that will kill a lot of the bacteria but this usually does not occur with sprouts.

There are a couple of other things that make sprouts unique. Sprouts is very democratic. It likes all sorts of different organisms. We have many different serotypes of Salmonella that have been implicated in sprout outbreaks as well as E. coli . We have some of the serotypes a little bit unusual here in the United States such as bovismorbificans, mbandaka and we also have the E. coli 0157:non-motile. These organisms are more common in animals overseas and again sort of suggests that there may be some connection with the field and contamination at the seed level.

There also are multiple kinds of sprouts as you very well know. I have just listed a few of them here. While they have similar characteristics, they may not all actually be identical in terms of their likelihood of transmitting pathogens the way in which they're eaten.

And then, finally, as I've alluded to already, sprouts--we have international partners with sprouts in the seeds and the whole process, and while it's challenging enough to control what happens here in the United States, it's even more challenging to control things overseas and on the individual farms.

So I'd like now to turn to sprout outbreaks by year and I've done this in a graphical form. We have on the X axis again the year and on the Y axis the number of outbreaks. The peach colored color is the alfalfa. The dark grey the mung bean, the orange clover and then green is mixed, and that generally is clover and alfalfa combined. As you can see, when we first learned about sprout associated outbreaks in 1995 with Salmonella , we had a pretty strong increase up to 1999 which then decreased and again rose in 2003.

I'm going to take away all but alfalfa simply to clarify a point here. We know that alfalfa outbreaks through 2004--actually to the current day as far as we're aware from 1995 is 26 outbreaks. As you can see there was--and I think it was mentioned this morning--there was growing concern about the cause of alfalfa sprouts and other sprouts leading to foodborne illness. So there was an interim advisory and then guidelines passed in 1998 and '99 by the FDA to advise chlorination of the seed with 20,000 parts per million and then also testing of the irrigation wash water, which I think we'll hear a little bit more about in the next talk.

This looked very hopeful because in 2000, 2001 and 2002 the numbers of outbreaks decreased but then we had this explosion of outbreaks in 2003 and this raised concern about what might be happening in the industry.

The next logical question is, well, perhaps why we have all these outbreaks is that people weren't following the guidelines and that certainly is a very reasonable question to ask. We do have some information on who was following the FDA guidelines. The ones in peach here we don't know whether they were or not. The orange, we have evidence to suggest that they complied with the guidelines. And the green suggest they did not comply. As you can see, we have both orange and green after the guidelines were passed. So again what this represents is that perhaps the current practices--current regulations are not sufficient enough to prevent outbreaks.

I'd like to turn from that time period of 1998 onwards and just look at the average size of outbreaks by year. I have again listed the number of outbreaks, the total number of cases that we're aware of as reported by the states, and then the mean number of cases per outbreak.

As you can see, particularly looking at the year 1999 and then looking towards the later years, 2003-2004, it suggests that perhaps the outbreaks are getting smaller and this could be a very positive thing and could be a result of the guidelines being followed by the sprouters. I caution, though, in too much interpretation of this data because we are every year getting better at detecting outbreaks. We now have a good system of DNA fingerprinting across the nation and actually even internationally. So I think we are picking up outbreaks quicker before they become larger. We sometimes are picking up outbreaks of two to three people because states--they have a lot of people working in the state to follow through on outbreaks to look at some of the DNA fingerprinting measures and we're doing more and more biological testing.

Of course, a remaining question and one I'll end with is, well, perhaps outbreaks are not the best way to measure the effectiveness of an intervention. I want to just return very quickly to a different produce item, to juice, number three on our list that I showed earlier.

Here we have the number of juice outbreaks from 1995 to 2005 similar to what I've shown you with sprouts. Each box represents an outbreak and the various colors are suggestive of different kinds of juice. Around the same time period there was a juice labeling regulation that was passed in 1998-99. As you can see, it certainly rose in '99 and then actually has decreased and remained low. There have been various HACCP regulations throughout.

The last four in red there from 2002 to 2005 actually represent outbreaks that would not have been included in the HACCP regulations due to specific quality and size, for instance, of the juice facility. So this graph suggests that actually you can use outbreaks to show that an intervention is effective or perhaps that it's not as maybe is suggested by the sprout outbreaks.

So to summarize and conclude, sprout associated outbreaks represent a small proportion of the foodborne outbreaks in general but they are one of the most common vehicles identified in produce associated outbreaks and this is particularly in the setting of a produce that actually is not eaten as much as other produce items in those top six items.

I think we're seeing actually with the organic movement as well as the sort of health related movements that more and more people are eating sprouts and again the numbers I suggested maybe perhaps are not totally accurate because probably more people are eating sprouts and aren't even aware.

As suggested by the number of outbreaks most recently, the current practices and regulations are not adequate to prevent disease from sprouts and I think that is why we're all here and looking forward to a good discussion throughout the day.

Finally, outbreak surveillance does offer opportunities for tracking effectiveness of interventions and we certainly at CDC will continue to collaborate with all parties involved to try to prevent disease in humans from sprouts.

Thank you very much.

DR. BERU: Thank you, Dr. Dechet.

Our next presenter will be Dr. Michelle Smith, Interdisciplinary Scientist in the Office of Plant and Dairy Foods.

By the way, we will have bios copy made available either during the break or during the lunch break.

Dr. Smith will present on "Sprout Guidance Background and Next Steps."

Sprout Guidance, Background and Next Steps

[See presentation slides for Dr. Smith]

DR. SMITH: It's a pleasure to be here this morning and I'd like to thank everyone in the room for coming also and I'm looking forward to what comes out of this program and, hopefully, we can make some good progress together.

This is an outline of some of the things that I intend to cover in my presentation. I don't have a lot of time because mostly we want to hear from you but I would like to talk a little bit more about outbreaks, collaborative efforts, consumer advisories and some of the things that have happened since we started working together on sprouts just as food for thought as this effort goes forward.

Dr. Brackett mentioned the Produce Safety Action Plan. I'll do that also. And Dr. Dechet brought up some of the questions surrounding the adequacy of FDA's guidance and I'll talk a little bit about them along with next steps. Where do we go from here and how do we get there?

As has been mentioned, sprouts were first described as a special problem in the NACMCF Produce White Paper that became available in 1998. NACMCF in that paper talked about opportunities for contamination for fresh produce items, including sprouts, and some of these opportunities have been covered. What makes sprouts special, as everybody here probably knows, is that the conditions that foster the germination of the sprouts also promote the growth of pathogens if they're present.

An additional challenge for this group and for anyone involved in coming up with interventions is that many of the treatments that have been looked at to inactivate pathogens may also decrease germination of the seed, decrease yield or affect the appearance of sprouts. And treatments that result in any of these make that treatment option much less viable.

Now this is my simplified chart for purposes of this meeting of outbreaks. Our numbers are just a little bit different from CDC numbers. According to my total, we have been involved in 27 sprout associated outbreaks since 1996. The first column is the year, the second column is the total number of sprout associated outbreaks in parentheses, followed by columns of different sprout types and how that particular outbreak was attributed to a certain type of sprouts. In a few cases the numbers don't quite add up. Some of that may be due to the fact that it was a mix of sprout types and numbers have fallen in more than one column. For example, in both alfalfa and clover because we weren't able to tease it out. Then finally the last column is the number of reported cases of illness associated with the outbreak and the actual number of illnesses is very likely much larger than the number of illnesses reported.

Now just a few milestones to put some context around the different activities. CDC brought sprouts to our attention in 1995. The second bullet could run two pages. There have been many, many, many work groups, collaborative efforts, other projects started individually and by agencies and industry in cooperation to try and address the issue of sprout safety. This is a very complex issue and we're still working on it and that's why we're here today.

FDA asked NACMCF to look at sprouts in 1997 and they published the Sprout White Paper in 1999. This paper played a very large role in the guidance that FDA released later in that year.

If you haven't seen the Sprout White Paper there is still a lot of very good information in that paper looking from soup to nuts at sprouts. It's available at this website and also published in the International Journal of Food Microbiology.

Some of the findings in the Sprout White Paper, and it was an extensive paper but some of the findings I'd like to highlight here are that seeds are the most likely source of microbial contamination. Things can get worse in the sprouting facility if appropriate practices are not followed but most, if not all, outbreaks it appears that the seed brought the contamination in, in the first place. If the seed is damaged or scarified that can increase the chances of pathogens becoming internalized in the seed and it also makes it much more difficult for any kind of treatment to disinfect the seed prior to sprouting.

The NACMCF Sprout White Paper specifically recommended a five log seed disinfection treatment to be applied before the initiation of sprouting. This five log standard was not formally included in FDA's guidance, and I'll get to what we said in a couple of minutes but it has been very widely used as the target by many of the people looking at interventions and comparing different interventions.

We made available two guidance documents in October of 1999. The first guidance "Reducing Microbial Food Safety Hazards for Sprouted Seeds" is a very simple, broad guidance document containing five steps that we feel are important to minimizing food safety hazards for sprouted seed. The second guidance document "Sampling and Microbial Testing of Spent Irrigation Water During Sprout Production" provides additional guidance on how to implement one of the first five recommendations. Specifically, the microbial testing of spent irrigation water. The testing process is very complex and scientific, and I brought in our microbiologist to write that. I wasn't able to write it and if I tried to rewrite it I would have to share it with them to make sure I got it right but this issue warranted its own guidance document just to help people understand the procedures for the microbiological testing. Both of these documents are available on our website which is noted there.

The objectives of the guidance were to provide recommendations to seed suppliers and sprout producers about reducing microbial food safety hazards and sprout associated illnesses and to ensure that all parties comply with the food safety provisions of the Food, Drug and Cosmetic Act.

Our guidance documents are guidance. The specific recommendations that we feel are important for microbial food safety for sprouts are just that. They are recommendations. They are guidance. The individual recommendations are not requirements. At the same time the Food, Drug and Cosmetic Act requires that food be safe and wholesome and not produced or held under insanitary conditions. That act contains requirements for all foods.

In the Federal Register notice announcing availability of the guidance there were several statements, including "failure to adopt effective preventive controls can be considered insanitary conditions." And a notice that FDA will consider enforcement actions against any party who does not have effective preventive controls in place. And we put particular emphasis on microbial testing because that's the last hurdle before the product is released into the marketplace and gets to the consumer.

Now in the broad sprout guide there are five points. Everyone has a responsibility for this guidance to be effective or as effective as it can be. Everyone along the way needs to contribute to improving and maintaining the food safety of that component over which they have control and it starts with seed production. We recommended good agricultural practices for the production environment for the seed, that seed be conditioned, stored and transported under conditions that would minimize opportunities for contamination. When you get into the sprouting facility we are recommending that producers follow good manufacturing practices which are set out in 21 CFR Code of Federal Regulations Part 110 that these GMPs should be standard operating procedure.

Also, we recommend seed treatment. This is the FDA language and our guidance. We recommend applying one or more approved treatments shown to reduce pathogens prior to sprouting. And as an example we give the 20,000 parts per million calcium hypochlorite. This does not say that the seed treatment has to be 20,000 parts per million calcium hypochlorite and when we drafted the guidance back in 1999 we were very hopeful that treatment alternatives would be discovered and available because we know there are downsides to this particular treatment but it at least up until now has remained pretty much the gold standard. The one that people are most familiar with and the most data exists for.

And then finally because there is no single treatment that we're aware of that can be counted on to completely eliminate pathogens on seed, we recommended microbial testing of the spent irrigation water, the water that flows through the sprouts during the production practice and we recommended that this testing be done on each batch of sprouts before the sprouts enter the food supply. Salmonella and E. coli O157:H7 are the pathogens of primary concern in sprouts and these are the pathogens that we recommend testing for.

Now any guidance, including the sprout guidance, identifies the most important steps that we were able to identify at the time the guidance was released to improve food safety and in this case we believe that those recommendations should be implemented immediately to reduce the risk of sprouts as a vehicle for foodborne illness.

The broad sprout guide did not provide detailed information on all individual steps that should be followed to produce seeds and sprouts. The broad sprout guide just set out the five items. The idea was to have a simple document that could be pasted to the wall. It was the microbial testing where we specifically went into details on how to. However, both documents contain additional resources and references and we have done additional things since the guidance to provide more information.

One of the things that was put together was an educational video "Safer Processing of Sprouts." This was a joint California Department of Health Services-FDA video produced in cooperation with industry and universities. It was a fairly large workgroup that got together and came up with the script. Many sprouting facilities volunteered their facilities to be filmed as part of this video.

One of the things I think that is particularly good about this video is that these are actual sprout production facilities and if you look at the video and you listen to what it has to say it puts a sprout face on many of the very basic issues like worker health and hygiene and good manufacturing practices. More information about the video is available both on the California website and on FDA CFSAN's website.

There has been mention of the consumer advisories that FDA has put out. Our first advisory issued August 31st, 1998, and that advisory focused on alfalfa sprouts. When sprouts first came to our attention alfalfa was the one that we knew about and so that was the subject of the advisory.

In 1999, we updated this advisory to cover all raw sprouts because we were beginning to get more evidence that at least some of the outbreaks were not just alfalfa sprouts but maybe alfalfa, clover mixtures, and at least one, we think, the clover seed was the source.

The 1999 consumer advisory advised all persons to be aware of the risks associated with eating all raw sprouts. We noted, as we do often in consumer advisories, that people in high risk categories should not eat sprouts, should not eat raw sprouts, and there was a new kind of statement in this advisory. Persons wishing to reduce the risk of foodborne illness from sprouts should not eat raw sprouts.

There were several driving forces behind including a statement this strong in the advisory. One of the driving forces was that the majority of people experiencing foodborne illness associated with sprouts were not in the high risk categories. They were just normal healthy large population people. The other driving force you can see in this slide in the red numbers. In 1999, mostly clustered in the summer time, there were six outbreaks associated with sprouts. That summer it felt like sprout outbreaks would never end and I don't think anybody here wants to go back to that summer again.

Before we released the advisory we did call in some members of industry and share it with them and discussed the situation but it was decided that that's the wording that we would go with.

So 1999 was a tough year with the updated consumer advisory, the six outbreaks. We issued our sprout guidance as a direct final. Usually we go through two stages, proposed first, ask for comments, and then go to final incorporating those comments. Because of the large number of outbreaks we went straight to final guidance but we did offer an opportunity for comment on that guidance and we've collected those comments and considered them.

We also initiated an assignment to inspect sprout facilities and I'll get into that in a little bit.

Now this was a lot of activity. It wasn't all pleasant but in 2000 there were no outbreaks associated with alfalfa or clover sprouts so as painful as this might have been it looked like we were having a positive effect.

There was one problem. As the alfalfa and clover situation appeared to be getting better, a new trend started to emerge and that was outbreaks involving mung bean sprouts and Salmonella enteritidis. A couple of things about mung bean sprout outbreaks. Even though our '99 advisory was meant to cover all raw sprouts, we were hearing from some people that maybe mung bean sprout producers did not see themselves in that guidance. There are differences in the production practices. You might look at an alfalfa seed compared to a mung bean seed and notice that the alfalfa seed has a much rougher surface and mung bean is very smooth. There is less opportunity for pathogen attachment, et cetera. There are still natural structures on mung bean seeds that pathogens could very easily attach to so there may have been somewhat of a false sense of security.

There also was a general assumption at least on our part that mung bean sprouts were cooked before consumption so it would have a lower risk associated with them. When we started investigating the outbreaks associated with mung bean seeds and identified the food vehicles involved, we saw that those food items contained as an ingredient either raw mung bean sprouts or mung bean sprouts that had only been very lightly cooked. Not cooked long enough or at a higher enough temperature to eliminate any pathogens that might have been present. So the assumption that mung beans were cooked to the point where microbial food safety wouldn't be a concern didn't necessarily hold true.

So in 2002 we took our previous advisory and we updated it again to advise all persons to be aware of the risks associated with eating raw and lightly cooked sprouts, and we specifically included mung bean sprouts by name so that there could no longer be any confusion or misconception about whether or not bean sprouts were included in our concerns.

I mentioned that we've done a couple of field assignments. The first one was in 1998. This preceded the NACMCF recommendations and it preceded FDA's guidance. So what we looked for in 1998 wasn't based on the guidance itself. We looked at facilities in terms of the current good manufacturing practice regulations in Part 110. We targeted 100 firms. We usually draft an assignment with a goal in mind. We ended up doing inspections at 83 firms. There was a very long questionnaire. Again largely based on the GMPs and also to gather demographic information, information about the size of facilities, products produced, number of employees, things like that. Eighty firms that were visited answered questions on this survey.

Finally, we collected samples for microbial analysis. Samples were collected at different points in the production process starting with raw, dry seed and going all the way to finished product. Samples were tested for the pathogens, Salmonella , E. coli 0167:H7 and a number of other non-pathogenic microorganisms were enumerated. Samples were collected at 78 firms.

Now this is just for context. I'm not going to dwell a lot on the specifics back in '98 because it's not 1998 anymore and I think things have changed but at that point in time out of the 83 facilities that were inspected 47 of the firms or a little bit more than half of the firms received what is called an FDA 483. This is a report of observations that's usually issued when insanitary practices or conditions are observed. And in some of the more recent assignments I'll give some examples of the things that we see as insanitary practices or conditions.

In the microbial analysis we found positives for Salmonella at three firms visited. At firm C the samples were positive in the pre-sprouted or pre-soaked seed but prior to sprouting. Midway between the sprouting process samples were collected and they were positive. When firms washed sprouts after harvest, wash water was collected and this was the case in the third firm. That was also positive. And finished product was positive at all three firms.

In no instance in this assignment were any of the samples of raw seeds that were collected found to be positive for pathogens and this may indicate at least using the methodology of the time some of the difficulties with microbial testing of seed.

We did a follow-up assignment in 2000. This time we targeted 150 firms, did a more limited inspection focusing on the recommendations and FDA's guidance, and also used a much shorter questionnaire. Again looking at the recommendations in FDA's guidance and trying to get more detail. First of all, were they being applied and, if so, how were they being applied? We collected and tested spent irrigation water from firms that reported that they were doing testing themselves.

The good news is that we did not find any samples to be positive for pathogens in the 2000 assignment. The bad news is that FDA 483s or notice of observations were issued to 99 firms or 72 percent. None of these assignments were designed in a way where you can do any kind of quantitative comparison between the numbers. To a large extent the individual investigators may be very subjective in what they decide is an insanitary condition and whether or not they issue a 483 but there is some qualitative information that can be gained from this, particularly with respect to the kinds of observations that were made.

Another thing that we did in 2000 was we issued a fairly significant number of warning letters. This is a step above and beyond issuing a 483. We issued warning letters to 65 firms and we did this when we found a combination of one or more insanitary conditions and failure to implement effective pathogen preventive controls. Again our emphasis was on microbial testing. There was some misconception since we did not give the field authority to issue warning letters based on the absence of seed disinfection that FDA no longer thought seed disinfection was important. The reason we didn't give that direct authority was that seed disinfection has many complexities and it would have been too difficult, and we still haven't come up with a tool to fully describe to the field investigators what would or would not constitute an effective treatment, the variation between tolerances of different seed types and all those issues. So when it came to seed treatment, this is still important, we gathered as much information as we could but we did not make that a criteria for the warning letter.

Okay. We found that at 54 firms there were no significant deficiencies noted. We also found that firms that were inspected in 1998 did better in 2000 than firms that were being visited for the first time. So maybe there was some advantage either to FDA presence or to increased awareness from having been visited before.

Some of the categories where we saw deficiencies include personal cleanliness, things like hand washing facilities, rest room facilities, their condition, accessibility, even their existence, insanitary food contact surfaces, how they were cleaned, what they were made out of, the presence of pests, including insects and rodents, and water quality issues. Things like maybe having a private well as a water source and the private well had never been tested to ensure water quality.

Now, we said earlier that it looked like the guidance was doing some good for a little while. What we have seen in 2001, 2002, a single outbreak associated with alfalfa sprouts, which may or may not have been mixtures of alfalfa and clover because we're starting to become more aware of that practice. In 2003 there were five outbreaks. This is one short of the summer of the '99. In 2002 or in 2004--two outbreaks by mid 2004. So this raises the question that was mentioned just a little while ago about the adequacy of the guidance.

Some of the initial reports from the investigators investigating the outbreak when they have been able to trace back product to an individual facility were starting to write in their report that the sprouter appears to be following FDA sprout guidance.

Now again our broad guidance is the equivalent of five bullets so our first question was does the investigator know enough to assess accurately whether or not the guidance is being followed. And we drafted questions here in house that we've shared with investigators whenever we've become involved to kind of help them ask the second and third level of questions to get at whether or not the guidance is being consistently and appropriately applied by those firms that say they're following the recommended practices.

In all fairness, this also raises the question about the adequacy of the current guidance. Is there something different we should do? Is there more or less or some other way that the guidance should be revised to be more effective?

The most recent information we have about practices within the sprouting industry comes from a joint California Department of Health Services Food and Drug Branch and FDA Inspection that was initiated in February 2004. This was an inspection of all registered sprout facilities in California so again the data that comes from this you need to be careful about extrapolating it to national numbers but I think that some of the observations that they made in this assignment are or should be serious food for thought in what we do next as we go forward. Again they used the standardized questionnaire that was largely based on FDA recommendations.

This is the inspection checklist that was part of the inspection so this is a list of everything that they looked at.

In this particular assignment 50 percent of the firms were described as having deficiencies in the same four areas that were noted before. Numbers are somewhat different but again this was not designed for direct comparisons.

When looking at seed treatment the majority of facilities visited employed some type of seed disinfection treatment. Seventy percent of the facilities were using calcium hypochlorite, 25 percent of facilities were using sodium hypochlorite, ozone and peroxyacetic acid were also used. However, only two facilities, if you look at all the steps that are recommended for the seed disinfection treatment with 20,000 parts per million calcium hypochlorite, only two facilities used the correct concentration, the recommended time and the best method, including agitation and consideration of the seed solution ratio. So this indicates to me that there is at least a strong possibility that the folks that are doing seed disinfection treatments may need a little bit more help in the steps to follow to do that the most effective way.

Many firms, 71 percent of firms, collected spent irrigation water for microbial testing. Almost all firms tested for Salmonella and E. coli O157:H7 but the testing method varied greatly and some of our colleagues here will get into testing in a lot more detail this afternoon.

When it says "unapproved test" we cited several tests in our guidance document. That doesn't mean that you can't use alternative tests but if you're using a different test than the one that we cited you should have some confidence and some experience that it really does work on spent irrigation water.

California also noted that they had some issues with who does the tests and where the tests are done, qualified people in a controlled environment, and also pooling has the potential to dilute pathogens if they're present, and holding samples for an extended period of time. If pathogens are present they could die off in that sample and give you a false negative which is a false sense of confidence, a waste of money, and a public health concern.

Sixty-five percent were conducting confirmatory tests. Many firms said that they have never had an initial positive but would do confirmatory tests if they saw one. A little more than a third of the firms were using the enrichment media for confirmatory testing and this is what microbiologists would recommend.

Almost half the firms wait to ship until the results have been received, and again this is from the confirmatory testing or the follow-up testing to an initial positive most likely with the rapid test kit.

One area certainly where more could be done is the issue of record keeping. One of the firms visited maintains a record of sampling for microbial testing but no records of the outcome of that sampling and this would also help us to have information like this as we look at the incidence of positives.

Less than 20 percent that conduct confirmatory testing maintain records of positive results from that so there's even less record keeping further along the line and most firms do not contain or do not maintain records of what they do when they find positives, how was the product disposed of. Now this may be kind of the philosophy of don't ask, don't tell or we don't want to put anything on paper that's not good news but this really would be to the benefit of all people involved to maintain records of their due diligence, their appropriate actions, and also would help us in investigations to know what happened, and to have documentation to support that knowledge.

Traceback is another area. A larger proportion of firms report having the ability to identify their seed source than firms that reported having some kind of lot number or other identifying data on finished product. Absence of this information makes traceback very difficult. Traceback is even harder when different seedlots are mixed and the California survey found that for every production lot of sprouts that they looked at from one seedlot up to as many as ten were used to make that lot of sprouts.

Dr. Brackett mentioned the Produce Safety Action Plan that was finalized in October 2004. This is the website for that action plan. With some other commodity groups we've made the suggestion that they go visit the action plan and look at it and see if there are ideas about things that could be done. It covers risk assessment, education outreach, research components, just a lot of things that could have application here.

It's on our program priority list for this year to hold a public meeting. We'll be able to check that off after today so thank you all.

As individual tasks are identified they also make it on to our to do list and one of the tasks that we have on our priorities list is to evaluate testing protocol for the recovery of Salmonella in sprout seeds. Now the actions in the Produce Safety Plan and anything that comes out of this collaboration might be individual group actions, they might be collaborative efforts but to make true progress we really need the work on the part of all parties.

The Food Code has defined sprouts as a high risk food. Cut cantaloupe is a high risk food. There are a number of things on that list. I think by your presence here and your interest that's an indication that there's appreciation that a rigorous risk reduction strategy is needed. The question that we're working on now with your help is what should the strategy include to help us answer the who's, how's, when's and best applied to get the most appropriate effect of the best possible solutions.

The Federal Register document announcing this meeting set out a number of questions to kind of help focus comment on the areas that we're interested in getting more information on. This is my short list. My personal favorite is how can we better reach seed producers and distributors. We've said all along that this is a collaborative effort. It involves all parties at every step along the supply chain. Our Produce Safety Action Plan no longer focuses just on the farm and packing house. We've extended that to cover the entire supply chain and that would be an approach that we would most like to use with sprouts also.

In addition to this public meeting, we have a comment period that runs until July 18th and the instructions for submitting additional comments are in the Federal Register notice for the meeting.

I thank you for your time and attention.

Questions and Answers

DR. BERU: Thank you, Michelle. I would like now to invite you to sit up here together with Dr. Dechet.

I know we're running a little behind time but we'll have a short question and answer session.

As I said before, please use the mikes. We are recording this and we'll be having a transcript made of the meeting which will be posted on our website so when you have questions either step to one of the microphones or they can be handed to you.

Yes, please?

MR. SANDERSON: My name is Bob Sanderson. I am president of the International Sprout Growers Association and I would like to point out what--I may be hearing things that aren't there but something of an inconsistency in Dr. Dechet's presentation and Dr. Smith's.

Dr. Dechet spoke about the chlorine recommendation of the guidance and discussed it at some length and then it seemed kind of as an aside and said, "And also there is the testing." And Dr. Smith pointed out that when they were evaluating compliance--when the FDA was doing this they said, "In particular, seed testing or testing procedures."

And I have a feeling that there has been a mixed message on this emphasis so I think that's a serious problem because growers really are counting every penny and they're going to put those pennies where they really think they'll have to and the power of the inspector is huge. Okay. Thank you.

DR. DECHET: In terms of my reference to the chlorination process, certainly the sprouters and the growers would have much more information about that. I really use that more as a sort of time table to talk about some of the interventions that--or the guidelines that were issued. Probably Michelle can speak a little bit more about that but we would, you know, really defer.

I mean, our job at CDC is really to detect the outbreaks and report that information, and I think sort of actual guidelines and who is following those guidelines and what the emphasis is of those guidelines really is best left to the FDA.

DR. SMITH: And I think maybe another reason CDC focused on the seed disinfection is that for a long time it looked as if the firms that were disinfecting seed with 20,000 parts per million calcium hypochlorite were not involved in the outbreaks, whereas firms using the same seedlot that weren't consistently or appropriately applying 20,000 were the ones that were having the outbreaks so that put kind of a spotlight on the seed disinfection treatment. At the same time there are issues with the 20,000 parts per million. Not just whether it's being used appropriately but the impact that it has on seed, whether or not it's consistent with organic production, the fact that some types of sprouts, not necessarily the major ones but some of the other types can't tolerate that kind of seed treatment. It's very complex.

And then finally there is research about pathogens surviving that treatment and recovering during sprouting to the point where they can have a foodborne--cause foodborne illness. So our last hurdle, the final thing that people can do to make sure that contaminated product is not shipped is the microbiological testing.

We are looking at a multiple hurdle approach. I understand people want to do the one thing that can make the most difference and our guidance does provide for alternative approaches if they do not violate laws or regulations and if, hopefully, they get you to the same food safety place. That second part is really a hard call. I tend to take calls from the field that observe something that is just woefully inadequate and I can say that's not an appropriate alternative but there's a huge gray area that I think we need more information to map out better.

DR. BERU: Yes, please?

MS. SANDERSON: My name is Barbara Sanderson and I'm from Jonathan Sprouts also. Michelle, I really appreciated your reiterating the guidance where it said seed disinfection such as...and I have--I would like to point out a situation that happened in the industry. It hasn't to do with your recommendation but how it was used and I believe it began with an outbreak where a grower was using a combination of a heat treatment and 2,000 parts per million calcium hypochlorite and still had an outbreak. The lawyers used this information to turn your guidance into a rule that we had to use 20,000 parts per million calcium hypochlorite and there was no way--that kind of put a damper on getting any other kind of treatment into use or even into the visibility of how effective it might or might not be with sprouts so that we really got paralyzed by the legal profession.

DR. NORTON: I'm Dick Norton. I'm a retired microbiologist from the Food and Drug and working with different companies and asepsis for the last 17 years now. Currently I'm working with a small company with asepsis and we've sent the results of Dr. Reiser's tests on alfalfa sprouts for this. It's something new to this field. We have been working off in a little corner somewhere else and I don't think it's generally recognized here but we've been working with this fatty acid combinations for 14 years now with Tyson's, with various meat, types of meat and fish, and we've been marketing products for hand anti-asepsis in hospitals. So we're unknown in this field but we have background and your question as to how can we all put this together got me interested in finding out how we can help with this and our interest is in treating seeds. It turns out there's no effect on germination rate. There are safe compounds that are cheap so if there's any interest in this we would like to work with these people and unfortunately I have to go to the EPA on this very same subject at 1:00 o'clock so I can't stay around for the whole day so if anybody has any questions.

DR. BERU: Thank you. We are running a little behind time so let me suggest that we will have Dr. Dechet and Dr. Smith also join the next set of panelists during the question and answer session so if you have any additional questions let's hold them until then and let's take a break until 10 past 10:00 and, as I said, we have refreshments in the area where we registered.

Thank you.

DR. SMITH: And as Dr. Beru mentioned, you're all on your own for lunch but I have a few copies of the menu for the cafe that's right out the front door and I'll tape the last copy to the table and put a stack out there when I get up.

[Whereupon, a break was taken.]

Industry Perspective

DR. BERU: Thank you. Our next presenter will be Mr. Bob Sanderson and he will present--he is president of the International Sprout Growers Association and he will present on the status of the sprouting industry.

State of the Sprout Industry

MR. SANDERSON: Thank you, Dr. Beru.

On behalf of the ISGA I would like to submit what is our current set of sanitary guidelines, an industry model HACCP and a beginning to attempt to define good agricultural practices for seed.

These things weren't built for my--can people hear me okay if I stand fairly straight?

Dr. Smith suggested that I provide an overview of the sprout industry's, as she put it, strengths, weaknesses, needs and desires.

My talk is not going to be the ISGA's comment per se. That will be submitted prior to the deadline in July but I'm going to try to talk in more general terms about the sprout producer and his relationship with product safety issues and their requirements.

In discussing the sprout industry's strengths and weaknesses, there is the product itself to consider and there are people who are involved in producing sprouts, and I think any discussion of the people has got to factor in the kinds of rules and regulations under which we're operating, whether they're clear, consistent and effective in achieving their intended purpose.

The product itself is remarkable in some obvious ways. Sprout producers may plant from one to several crops a week, 52 weeks a year, with an interval of from four to ten days between beginning the process and shipping the product to the market. What happens during this four to ten day period is a combination of very rapid chemical and biological processes which transform a mostly unpalatable seed into a nutritious food.

In addition to being nutritious, it appears that many, if not all, varieties of sprouts contain high levels of one or more phytochemicals, which may have significance to human health. The understanding of these properties is still in the early stages.

The very rapid growth of sprouts and the conditions in which they're grown present unique challenges in terms of food safety as well as some unusual opportunities, which I'll come back to in a moment.

Regarding the people who grow sprouts, it's a very diverse group. The largest in terms of numbers of producers and amount of sprouts produced in the U.S. and certainly in the world as well are those involved in the growing of mung bean sprouts, which have been a staple food in Asia for thousands of years. The other main group is companies, which beginning about 40 years ago, began to introduce a variety of sprouts, which historically have not been widely used as a human food. Increasingly, more and more sprout producers are combining the growing of the traditional sprouts and the more recently developed varieties so the industry is a mix of people from very different cultures. Some carrying on with the family business which may have been started generations earlier and others who are relative newcomers.

The increase in concern with sprout safety seems to have coincided with the appearance on the market of the newer types of sprouts so, as Dr. Smith mentioned, there may be some feeling on the part of the producers of the more traditional sprouts that they've been caught up in somebody else's problem. After all, these types of sprouts have been grown for many generations and wouldn't we have known about safety problems long ago?

Although it may be argued that different types of sprouts, which are used in different ways, may require different risk reduction strategies, it's not accurate to say that any particular type of sprout is traditional because the nature of the market, consumer preferences, production methods continue to change very rapidly. This is true both for the producers of bean sprouts and the newer green leaf sprouts since both are functioning in the context of increasing food safety concerns and both have to deal with continually increasing regulatory and third party requirements as well as changes in the nature of the market and consumer preferences.

There are some other groups with their priorities within the sprouting community. The ISGA being an international organization has members from many different countries who have quite different perspectives but since the ISGA core membership is made up of U.S. growers the association tends to have a focus on FDA regulations and food safety which some growers operating outside the U.S. seem to view as a kind of negative fixation.

After all, the ISGA was started to promote the positive aspects of sprouts, not to be talking about safety all the time and, furthermore, sprouts seem not to be considered such a high risk in many other countries. But the issue isn't simple because of differences in types of sprouts, how they're usually consumed, total volume sold, as well as differences in available public health monitoring and epidemiology.

The FDA's recommended 20,000 ppm chlorine seed treatment is generally considered very excessive by growers from other countries where in some cases the use of this treatment wouldn't be allowed. Even among U.S. growers it would be high on the wish list to have an approved treatment which did not create such an irritating and hazardous working environment.

It would also be a good thing for the ISGA as an international association to have a single internationally acceptable set of safety practices or at least an agreement on underlying principles, which isn't to say that sprout growers all around the world would apply these equally but at least we would have a common frame of reference.

Another point of debate in the sprout industry which exists inside the U.S. as well as between the U.S. and the rest of the world is the status of organic sprout production and, as I think someone mentioned, sprouts kind of is associated with organics, which is kind of ironic in a way nowadays, again primarily relative to the 20,000 ppm chlorine treatment recommendation. Although some organic certifiers have allowed the use of this 20,000 ppm treatment, the basis for this allowance is felt by many to be quite a stretch with organic production standards and is based on the need for organic standards to take second place to FDA safety requirements in any situation where there may be a perceived conflict.

This is entirely appropriate and necessary but unfortunately there isn't presently a clear way to evaluate the effectiveness of risk reduction strategies which might be more consistent with organic production methods.

So the ISGA's wish list might also include, if possible, the allowance of safety protocols which were more consistent with organic standards. This isn't to suggest that safety criteria should be diluted in order to be acceptable to different groups but that, if possible, safety requirements should be based on end product criteria rather than on any specific method of getting there.

One aspect of sprouts which many consider to be their greatest drawback is that they have proved to be extremely resistant to treatment based approaches to assuring their safety. This is because, and this is very much in a nutshell, sprouts--with sprouts any sanitizing step is followed by four or more days of growth at room temperature. However, the same characteristics which make sprouts difficult to sanitize may also make them accessible to more thorough sampling and testing than is possible with most foods.

The sampling and testing of seedlots prior to use in sprouting with seed which test positive for pathogens being diverted to non-food uses is a common sense preliminary step. The fact that in several instances sampling and testing of isolated pathogens from implicated seed following outbreaks clearly suggests that if the same sampling had been doing prior to the use of this seedlot for sprouting these outbreaks wouldn't have occurred.

Bob Rust is going to go into more detail on ways to carry out this initial screening of seedlots.

Since sprouts emerged as a significant health concern in the late '90s there have been dozens of attempts by leading food safety researchers to develop effective sanitizing interventions. These investigations continue into the present and it may be that in the near future a truly effective seed sanitizing procedure will be developed which is affordable and which doesn't entail significant risks to sprout growers or have other serious drawbacks or possibly such a procedure already exists and, if so, then one can only hope that it will get a speedy regulatory approval.

In the event of the allowance of such a treatment, one question will remain and it will need to be addressed. It's whether the use of such a treatment would eliminate or significantly reduce the present need for every batch spent irrigation water testing.

I don't know of anyone who presently is proposing that this should be expected at least in the near future and this is a very important question because testing is very, very expensive and if you introduce the idea of intermittent or periodic testing in a very competitive market there will be a significant economic incentive to minimize it. This may lead to even more confusion about what constitutes compliance, who is in compliance and who isn't, which will not help the overall level of confidence in the market.

So while we're waiting for improved treatment interventions or even if we find them, it seems worthwhile to consider how effective the recommended testing procedures are. The 1999 FDA guidance recommends that the spent irrigation water from all production batches be tested in duplicate for the two pathogens of concern and that negative test results be back in house before any product is shipped from the facility.

In preparing for this meeting I have tended to assume that my company's testing program could be duplicated by other sprout producers but we are fortunate in being an hour away from a qualified testing lab which is very responsive to us and which charges us very reasonable rates for its services. For other growers I've heard that lab costs per test can be much greater and some may have to overnight ship their samples which will add a full day before they can get their lab results back. This will add significant problems to inventory control as well as hold and release programs which is that you don't let anything out the door until you've got your results back.

So if this every batch testing is going to be a necessity until some significantly better treatment is developed, things like cost per test, how responsive a lab is, how long it takes to get the sample to the lab, how long it takes to get the sample results back are very important.

To put this in perspective, if my competitor and I are both doing all the required tests but he's doing duplicate testing and I am only doing single samples for each test, I will save enough money on testing alone to offer a price which will allow me to take my competitor's business and at the same time make more money at the lower price than I'm making now. This testing is very expensive and one thing on many growers" wish list is that the required testing could be significantly reduced.

One bean sprout grower from outside the U.S. told me that if he tried to do all the kinds of testing recommended in the FDA guidance he couldn't possibly remain in business.

But I feel it's important to come back to the question do we need to do this testing. It seems that until some treatment comes along which clearly indicates that we can cut back on our testing we need to do it and so we have to focus on how to make it more effective and more consistent.

High testing costs are frequently mentioned as a burden to sprout producers but I think this question needs to be looked at in context. Are they a burden because the retail price of sprouts is as high as it can be and consumers will just stop buying them if the price goes up? Or are they a burden because every grower is trying to survive in a very competitive market? And a difference of pennies per package can be a determining factor in obtaining or keeping an account with a supermarket customer.

In talking about the costs of the testing program, I'm talking about something in the range of ten cents a package wholesale cost. If you ask a sprout producer what his bottom line would look like with an extra dime per package coming in, I think you'd see quite a reaction. In my own company this--and this is, I think, literally true--would finance an entirely new facility and give everybody a raise at the same time.

So an important question about the burden of these testing costs is the price sensitivity of sprouts inherent in the product or is it dependent on perceived value so that if this value goes up the customer will readily pay the extra amount. And I think there is good evidence at least from some markets that customers will readily pay more for sprouts if they have a reason to believe they're getting a better or safer product.

When the FDA guidance were issued in 1999 sprout producers or sprout products were redefined. They still looked and tasted the same as they always had but the costs of producing them to an adequate standard had increased substantially. At the same time it was left up to the industry or the individual producer to convince the customer that the new improved product which looked and tasted just like the old one was worth the increase in price.

Since the supermarket buyer is under considerable pressure to buy the least expensive product, it's the producer's job to justify his higher production and safety costs but neither the supermarket buyer nor the consumer has any way to evaluate the safety claims of the producer. This can encourage marketing strategies which play upon customers" insecurities.

In the present guidelines the recommended treatment and testing protocols are preceded with the qualifier "e.g." which means "for example," which suggests that what is recommended is only one option out of several but in practice these guidance recommendations are very strictly interpreted and have acquired the force of regulations.

Deciding what, if anything, might be considered "e.g." is a difficult thing to do in part because it isn't clear what the treatment referred to actually accomplishes. The sprout industry would like treatment and testing options to be based on the best available science and we want, if possible, for them to be flexible so that "e.g." can allow for a range of approaches provided they achieve an acceptable end result.

There was an attempt several years ago for the ISGA to take responsibility for setting safety standards by way of a seal program. Two problems were encountered. One was that at the time the seal program was being developed there had not been very thorough inquiry into the effectiveness of the safety protocols which were then available. For example, there was no consideration at that time that growers could do this every batch sampling. The other problem was that the ISGA was not in a position to effectively police compliance with the requirements of the seal program as it wouldn't be today.

Whether we have a guidance recommendation or a regulation or whatever the structure, it needs to be responsive to developments in treatment and testing technologies and as flexible to the needs of individual companies as possible. Although the ISGA can't police this, clearly the FDA can't immediately review every new idea that comes along.

And a question, I think, underlying this is, is it worthwhile to evaluate safety methods for a very small industry which might be quite different from what is used with most foods? Of course, it's worth it to the industry to solve its problems and to prosper but are the sprout industry's challenges relevant to other areas of food safety or research generally? If we consider the testing challenge in its simplest form to be the challenge of quickly and accurately telling friend from foe in the microbial world, this would seem to connect the sprouting industry's needs with many important research areas. Therefore, the industry needs to continue to work with people in government and academic research who have qualifications and review capabilities necessary to develop effective sprout safety standards and procedures. The ISGA technical review board can provide a framework for this effort.

Allowing for flexibility and safety protocols has got a number of interrelated parts. There is the question of how well something works in a research setting and how well it would translate into a production setting. This has repeatedly come up in the case of the 20,000 part per million chlorine treatment recommendation with some arguing that lab results using inoculated seed suggests a very high level of effectiveness in the real world with others arguing just the opposite, and I think there's good arguments on both sides.

Another challenge is how can compliance with a number of different methods be evaluated by food safety and third party inspectors. It's certainly simpler for an inspector to evaluate a single treatment and testing protocol for all producers in a given category than to evaluate several which might differ from company to company.

And it becomes even more difficult when, as in most cases, the sample leaves the building and is delivered or overnight mailed to a lab. Is there any way the grower and his food safety auditor can be assured that the lab is doing the testing in a way which will provide the best likelihood of detection of pathogens if they're present in the sample? This may require skills which are not within the safety inspector's usual job description.

The challenges of developing good sampling protocols, assuring the best testing methods and providing adequate inspection criteria is another area where the ISGA technical review board can provide a significant contribution by working with growers to develop appropriate inspection checklists for treating and testing options.

So as long as we have the need for this testing, we might as well look at it as an opportunity and make sure it's being done as well as possible.

If the playing field is as level as possible and the methods are as good as possible, the added costs may actually benefit the entire industry by raising the standard of acceptable of the product in the whole market. In the meantime we can hope that treatment interventions are developed which will make the sprout producer's life much simpler. Perhaps some day requiring only a minimal amount of microbiological monitoring.

Most sprout growers are acutely aware of the microbial hazards which are possible with their products, not through any choice. One of the greatest strengths of the sprouting industry is our collected awareness of microbial risk and our growing knowledge of food safety. This may not only be our best resource for our industry's future but may also provide opportunities for learning which would be relevant to many other areas in the food industry as well.

I believe in the past calendar year there haven't been any reported outbreaks connected with sprouts. I don't know if the last 12 months--I'm not sure if that is the calendar year but May to May, which is a very good indicator of significant improvements in the industry. We need to focus our energies on how to increase our margins of safety, restore confidence in sprouts, and make the sprouting industry a model of food safety.

Thank you.

DR. BERU: Thank you, Mr. Sanderson.

The next presentation will be on Seed Handling and Distribution Systems and will be made by Bob Rust, International Specialty Supply.

Seed Handling and Distribution Systems

MR. RUST: I just wanted to thank Dr. Michelle Smith and the FDA for inviting me and giving me an opportunity to discuss a subject that I'm passionate about.

In order to stay within the allotted time I recorded the presentation and it will be available after this meeting in the Seed Safety Section of our website, www.sproutnet.com.

[Video presentation.]

Hello, I'm Bob Rust, owner of International Specialty Supply, with divisions of Prime Seed Company, Prime Packaging, Centrix Equipment Company and Sun Garden Sprout Company.

We also test sanitation products and processes for other companies and we have a pathogen testing laboratory. We do consulting on such things as GMPs and HACCP. We write a newsletter called the SproutNet and have an informational website, www.sproutnet.com, with most of the research related to sprout safety.

I appreciate the honor of being invited to speak regarding seed handling and distribution systems. The world has seen 26 sprout related outbreaks since 1998, of which 25 involved sprout growers in the United States and Canada. The FDA stated that they believed that the source of the contamination in most sprout related outbreaks is the seed. It's impractical for sprout growers to line up small quantities of seed from farmers all over the world so seed companies who specialize in sprouting seed contract the production with farmers and work with seed processors near the farmer to process the seed.

It's difficult to predict how much seed to contract and crop failures are common occurrence. When this happens, seed companies will purchase seed from other seed companies who may or may not specialize in sprouting seed. Although there may be thousands of seed companies throughout the world, there's only a handful that specialize in sprouting seed.

There are three ways the seed industry can improve its record of shipping contaminated product to sprout growers. It can grow, harvest, process, store and ship seed as though the seed were food product. It can decontaminate seed destined for sprouting and it can screen the seed for pathogens prior to shipping it out.

As the owner of a seed company and as a commercial sprout grower, we've put all three of these methods to the test. There are so many farmers and processors involved in so many parts of the world that our attempts at controlling the production was unrealistic. There are some good decontamination possibilities on the horizon but it may take several years of actual field use to determine if they are as effective as they appear.

The last one, seed screening, is something our company has been developing since 2000. There has been 13 sprout related outbreaks since we started our sampling protocol and we're convinced that it's our seed screening process that has kept us from being involved in any of these outbreaks.

Seed screening is a simple approach to risk reduction. The idea is to determine if the seed is contaminated prior to selling it for sprouting purposes. The Camden Research Group in England in their 2004 report titled "Review of Microbial Risks Associated with Sprouted Seed" concluded that the absence of pathogens in seeds is critical and, consequently, microbial testing of seeds prior to use for production of sprouts is essential. Seed sampling is also suggested in the Codes of Practice of Food Safety in Ireland and the Ontario Ministry of Agriculture and Food is adopting its recommendations for seed companies and sprouters.

In early 2000, Bob Sanderson came up with the idea that if the source of sprout related outbreaks is often the seed maybe we should check to see if the seed is contaminated before we use it. Hmm. Check to see if the seed is contaminated before we use it? We thought this was a great idea so in conjunction with Jonathan Sprouts we developed a process for screening seed before it sprouted. When seed arrives it's placed in a quarantined area and inspected for evidence of mouse droppings on the bags, holes in the bags where mice or insects may have entered, insect larvae, bird droppings, that sort of thing. The bags are black lighted for traces of urine.

In order to find contamination in seed prior to sprouting it, a pathogen must be captured for identification. Alternatively, one may capture and identify evidence of pathogen contamination. Twenty-five grams of sample from each and every bag. If the sample comes to less than three kilograms, enough bags are re-sampled in order to get three kilograms.

The composite sample of seed is very carefully inspected for indicators of contamination with both a magnifying glass and a microscope to determine its fitness for human consumption. The entire sample is sprouted using commercial sprout reduction methods, the seed is not sanitized prior to sprouting. A sample of the runoff water is collected using FDA procedures recommended for commercial sprout producers. The water is then enriched and tested for Salmonella and 0157:H7. Everything is documented and signed by the person doing each procedure, their times, such as drop shipments directly from the country of origin where sampling and testing isn't practical, in these cases no documentation is provided and the sprout grower needs to perform the sampling and testing procedures prior to using the seed.

Neither this certificate nor the procedures we perform is a guarantee that the seed is not contaminated. It's merely a risk reduction step.

We have an extra step as well. Before we receive a shipment we bring in a sample and we inspect it, sprout it out and test it. We stop at the point of rejection. That is if the seed doesn't pass visual inspection, we reject it without taking the time to sprout it out.

Recently when I mentioned to a researcher that we found pathogens four times, she was shocked that it was only four times. There have been 13 sprout related outbreaks since the year 2000. If these four lots had not been screened out there would have been 17 outbreaks instead of the 13. That's 31 percent more outbreaks. I think we would find more except seed with visible contamination never makes it past our inspection process and by doing what we're doing we learn whom the safe seed processors are and who to avoid so it's also a supplier screening process.

For over 100 years seed has been sampled and tested for germination, purity, hard seed, plant pathogens, and other properties that relate to the quality of the seed. The statistical probabilities used to determine these percentages is well documented and extremely accurate.

Testing seed for human pathogens can be even far more accurate. When testing for human pathogens we're not looking for percentage but for any at all. If a single pathogen is detected the lot is contaminated and can't be used for sprouting purposes.

When it comes to seed sampling for human pathogens, sample size and total contamination per lot are the two factors that determine the probability of capturing a pathogen.

There is a possibility that two or more pathogens could be lodged on one seed. This makes our probability charts less reliable so we base probabilities on contaminated seeds rather than on CFU. Seed while being processed does get mixed very well, well enough for very accurate sampling estimates.

For ease of explanation throughout the rest of this presentation, all samples will be 25 grams and all bags will be 25 kilogram bags so the sample size is always 1/1,000th of a bag or 1 seed per 1,000.

Some question the reliability of sampling because there is a possibility that pathogens can be in a corner of the bag or in a clump and never detected. Anything is possible but this scenario is extremely unlikely. The seed is harvested, transported and dumped into a silo or bins. It is then poured or agered into the seed cleaning equipment, processed and poured into a bag. The cleaning and grading process doesn't allow even two seeds to clump together or they won't fit through the screens. Seed with pathogens are not going to stay next to each other throughout this process. They will be somewhat, if not thoroughly, distributed.

When trying to detect plant pathogens, similar sample sizes are used and the probabilities are extremely high as well. When looking for plant pathogens you are looking for frequency rather than any at all. In order to determine the percentage of pathogens in wheat, eight kilograms is sampled for each 100 tons of seed. Then 300 seeds are pulled from that eight kilograms for testing. Distribution of the plant pathogens is good enough that this method is very accurate. In a same lot size we would sample and inspect the lot, then 25 million seeds would be tested in order to find a single human pathogen.

Larger samples increase the probability of capture at a given contamination rate. Considering that we sample at least 1/1,000th of a seed, the larger the lot, the larger the sample. If you have seed that is contaminated at one seed per kilogram, pulling one sample from one bag would only get you a 2.5 percent chance of finding it but, if you didn't find it in the first bag, you have another shot with identical odds in the second bag. The more bags there are the more 2.5 percent shot you got at it. You get enough 2.5 shots, say 400, and you have a 99.99 percent chance of capturing at least one of those contaminated seeds.

But what if the pathogens are in just a few bags? This is a problem if you're trying to find a particular level of contamination but it is not a problem if you are trying to find any at all. What matters is the total number of pathogens in the lot, not how many bags they're in.

This chart shows what happens if you sample various size lots in which all lots have 2,000 contaminated seeds. Notice if you have one contaminated bag, you have 2,000 contaminated seeds and one pull will capture one of those seeds 86.5 percent of the time. If you divided those 2,000 seeds among two bags, now you have 1,000 contaminated seeds per bag but you have doubled the number of pulls and the odds even out.

Our charts are actually based on the odds of capturing clean seed, then we reverse the numbers in order to predict the number of contaminated seed. If you tried this with small numbers such as a few marbles in a cup, the odds will change as you move the marbles from cup to cup but there are about 12.5 million seeds in a bag. In this example 1,000 seeds is moved into another bag with another 12.5 million seeds also. When you move 1/100,000th of the seed from one bag to another the number of clean seeds is virtually unchanged so the probabilities for all practical purposes remain the same.

This next chart shows a large lot in which--except in the bottom row--is unevenly distributed among 800 bags. The first row has all 2,000 seeds in one bag. Your odds as in the previous slide are 86.5 percent but it doesn't matter that you pulled seeds from the 799 bags that are clean. You pulled one pull from the one bag that was so contaminated that it gave you an 86.5 percent chance of finding one contaminated seed.

Some say the blending makes seed sampling unreliable. Blending dilutes contaminated seed per kilogram but it has no effect on the total number of contaminated seeds and, therefore, no effect on the probability of capture.

As an example, two bags of seed contaminated at the rate of four contaminated seed per kilogram has 200 contaminated seeds. If it's blended in with 800 bags of non-contaminated seed the new lot still has 200 contaminated seeds. The probability of detection by pulling 802 samples is 18.5 percent. These are the same odds as if you had pulled a sample from each of the two contaminated bags before they were blended in.

However, using our seed sampling protocol on the two bags, three kilograms would have been pulled for testing. This increases your sample size from 1/1,000th to 1/16th. The odds of capture go up to 99.9994 percent. So blended lots are okay as long as the seed is screened properly before the lots are blended.

So why is it hit and miss when health officials try to find pathogens in seed that they are certain caused an outbreak? It's the total number of contaminated seeds in the lot that determine the probability of capture. Seed companies can screen the seed when the entire lot is intact. This is when the total number of contaminated seeds is its highest. If just a portion of that lot is contaminated, the seed in the contaminated bags are included in the population that's sampled. By the time the epidemiologists determine that sprouts are the most probable cause of an outbreak, a good portion of that lot is gone and a portion, if not all, of the contaminated seeds was used up in the outbreak.

There's something else that should be pointed out. The effectiveness of seed sampling is inversely related to the effectiveness of chlorine. The more contaminated the seed is the less effective seed sanitizing is. Yet the more contaminated the seed is the easier it is to detect a pathogen using seed sampling and testing. Seed screening along with chlorine or other forms of decontamination complement each other very well to reduce the risk that Salmonella or E. coli O157:H7 is in sprouted seeds.

Is seed sampling too much of a burden on the seed industry? Well, first let's look at what sprout growers have to go through in order to ensure that their sprouts produced from the seed are safe. An ISS Rototech Rotary Drum uses 50 pounds of seed to produce one batch of sprouts. The FDA recommends each batch of sprouts be tested separately and that 12 pounds and 13 ounces of chlorine be used to sanitize 50 pounds of seed.

At a $1.55 per pound the chlorine costs about $20.00 to sanitize the 50 pound batch. Runoff water is collected and sent to a lab to be tested in duplicate for Salmonella and E. coli O157:H7. The best price we could find shopping around was $25.00 for O157:H7 and $33.00 for Salmonella . Overnight delivery of the refrigerated package is $36.00. Chlorine reduces yields 10 to 30 percent. I just plugged in 10 percent into these figures. Collection, ice pack, packaging, shipping, labor and record keeping are at least

$15.00. I didn't put in any time for management, training, hold and release or false positives. So the minimum it costs a grower who follows the FDA guidelines is $195.50 to sanitize and test one 50 pound batch of seed. This is $3.91 per pound of seed.

The seed screening by the seed company consolidates the efforts and the finances. An experienced sampler can sample 44,000 pounds of alfalfa in six hours and mung in eight hours. This is about $120.00 in labor. The seed and bag inspection takes a trained technician about three hours. This is about $45.00. The seed is about $60.00. Growing and disposing of the sprouts is a couple hours and sending off the sample is $36.00. The FDA recommends the tests be done in duplicate so this costs about $66.00 for Salmonella and

$50.00 for O157:H7 tests. Collecting the sample, packaging it for refrigerated shipping and send it off is about an hour. This totals $407.00, which in 44,000 pounds of seed comes to about a penny a pound.

So the seed company pays $407.00 per 44,000 pounds which is less than a penny pound and the sprout growers pay $3.91 a pound, which in 44,000 pounds comes to about $172,000.00. That's 422 times what it would have cost the seed industry to screen the seeds. With $172,000.00 the sprout grower should have at least got the seed, maybe even the semi that brought it to him.

Seed screening doesn't do the following things: It doesn't cost much. It does not produce hazardous waste. It doesn't put production workers at risk. It doesn't affect germination, vigor, yield or the quality of the sprouts and it doesn't reduce background flora. It doesn't introduce the possibility of resistance. It doesn't disenfranchise organic sprout growers and it doesn't negatively affect or take away in any way from other food safety procedures such as decontamination.

But this is what seed sampling can do: It can help the seed industry identify practices that introduce pathogens into foods. It can prevent contaminated blended lots from entering the market if sampling is done prior to blending. It can prevent seed with visible contamination from entering the market, and this includes glass and other things that don't show up on a pathogen test. It's most effective when sanitizers are least effective. And it can help identify farms and seed processors lacking good agricultural processes. It can be used identically to all types of seed and its application will affect commercial sprout growers of all sizes with all levels of sophistication and using all types of equipment. It can be applied evenly throughout the sprout industry. It can be used on seed destined for home sprouters and it may be the only realistic defense that home sprouters have. Chlorine is not a great option for home sprouters. It would shift some of the responsibility of providing safe seed to the seed industry and it would substantially reduce the number of foodborne illnesses from sprouts.

In conclusion, I'd just like to say that seed needs to be screened for pathogens before it's sold as sprouting seed. It just makes sense. Thanks for listening.

DR. BERU: Our next speaker is Caroline Smith DeWaal, Center for Science in the Public Interest.

 

Consumer Perspective

[See presentation slides for Ms. Smith DeWaal]

MS. DeWAAL: Good morning. My name is Caroline Smith DeWaal and I don't know how many of you were here--was it 1998 they had one of these meetings or '99--FDA had another meeting on sprouts. Probably Bob was here but I was here as well and I was very interested to realize during lunch that I was eating with one of the sprout growers from Vermont, which is my home state. At that point I think it really hit the sprout industry by surprise because a lot of the people who had entered the industry were people who were really working on producing clean safe food and suddenly to be hit with a major food safety problem with their product was something that I know took the industry by surprise.

The Center for Science in the Public Interest is a non-profit consumer advocacy organization and we focus largely on food safety and nutrition issues. We represent consumers both in Canada and the United States and we're funded by about 900,000 subscribers to our Nutrition Action Health Letter. As I've just noted, we've really been involved with the issue of sprout safety for many years now.

CSPI applauds the Food and Drug Administration for moving towards the creation of a regulation to address foodborne illnesses linked to the consumption of sprouts. The fact that the number of foodborne illness outbreaks associated with alfalfa and clover sprouts decreased immediately after the two sprout guidance documents were issued by FDA in October 1999 signalled that those were steps in the right direction. Even at that time, however, CSPI believed that the guidance recommendations needed to be mandatory and advocated for regulatory rulemaking to commence. The continuing sprout outbreaks over the last five years are evidence that further FDA action is needed to ensure the sustained adoption of effective preventive controls by the seed and sprout industry.

So FDA laid out a series of questions in preparation of this meeting. I only answer questions I want to so these are the three that I decided to focus on: Should the sprout guidance be expanded or revised? Is the regulation likely to be effective in reducing foodborne illnesses linked to sprouts? And how can progress be measured?

Although routine microbial testing is an essential element in sprout safety, FDA should move towards the long-term strategy of implementing a hazard analysis and critical control point system for sprouts.

However, until an effective pathogen reduction step is developed and mandated, sprouts should be labeled as a high risk food.

A regulation is a more effective tool towards reducing foodborne illnesses associated with sprout consumption because it will ensure that all sprout producers are legally required to institute preventive control measures. It really puts all sprout growers on a level playing field.

One of the sprout guidance documents that was issued in 1999, "Sampling and Microbial Testing of Spent Irrigation Water During Sprout Production," recommended specific techniques for the routine testing of every production lot or batch of sprouts for Salmonella and E. coli O157:H7.

FDA should examine how successful this microbial testing has been in the past five years, both in terms of the proportion of the industry following the recommendations and what their results were, and I know we're hearing some of those results at this meeting today.

I'm going to talk briefly about what CSPI brings to meetings like this. I don't just come as representing consumer advocates. I think Bob Buchanan at FDA challenged me a number of years ago to actually bring something to the table so what we started doing is tracking foodborne illness outbreaks with our Outbreak Alert Database. It right now contains about 4,500 outbreaks with both known food and known etiology which have occurred in the U.S. between 1990 and 2003. It is compiled from multiple sources, largely especially in recent years from the Centers for Disease Control and Prevention, which publishes an outbreak line listing but they don't organize it by food groups so we take their data and reorganize it so we can actually track it by the food groups but we also take data provided by state health departments and scientific or medical journals.

By linking the outbreaks to specific foods, CSPI is able to alert consumers of food safety hazards and provide better information for food safety resource allocation.

So this is what our data--I'm going to go from the broad dataset to the more narrow issue that we're dealing with today. Between 1990 and 2003 FDA regulated foods were linked to nearly 3,000 outbreaks and over 83,000 cases of illnesses. These FDA regulated foods constituted 66 percent of the outbreaks listed in the Outbreak Alert Database, which again you only get an outbreak alert if we have known food and known hazard, and it's 60 percent of the cases.

And this shows FDA has the lion's share of the outbreaks and the USDA the meat and poultry products, which we normally associate with foodborne illnesses, actually are smaller contributors.

So here is our trends in outbreak reporting on produce. Produce was the food category associated with the most foodborne illnesses linked to outbreaks. This was a surprise when we first started to find this. We started publishing our data in 1999 and it took--we had to actually FOIA the information from the CDC because they didn't want to give it to us. As you see, the early years of the database is much less robust because CDC actually wasn't--they started electronic reporting of outbreaks to CDC in 1998, I believe, and so we started getting a much better dataset. I know you've heard from CDC this morning but what you see here is produce. Both the outbreaks and the illnesses from produce are quite significant.

And here we've got the top causes of produce outbreaks from 1990 to 2003 with norovirus at the top and the two of greatest concern to the sprout industry is Salmonella and E. coli .

Okay. This data reflects probably what you've already seen this morning from CDC that the principal pathogens of concern here are E. coli , not just O157:H7 though, there are other types of pathogenic E. coli strains, and then a large variety of Salmonella strains.

Okay. I know we've talked a little bit this morning about this outbreak. This was a 2001 outbreak. In our last meeting actually we looked at a number of outbreaks and did these kind of profiles and they were quite interesting. This one is a little bit different. It was 32 cases of illness, which is relatively small. Still significant. It was focused in California, Arizona, Colorado and New Mexico. And three of the cases involved or required hospitalization so we had some high risk individuals who ended up getting very sick in this outbreak.

What was interesting is that it found that there was a single seed source, a single sprouter and a single seedlot were associated with the illnesses but the seeds had been imported from Australia and had undergone, as we've discussed earlier, multiple decontamination steps, both heat and low dose hypochlorite solution.

Now I use this outbreak to illustrate a couple of points that sprout producers often are not the source of contamination. Contamination frequently appears to occur earlier in the production chain. That seeds intended for human consumption must be protected from exposure to manure, animals and contaminated water; that contamination decontamination treatments need to be routinely and systematically monitored for effectiveness; and finally that production conditions in other countries need to be evaluated prior to seed import to ensure that the growing conditions are suitable for seeds intended for human consumption.

Well, consumers always have an important role to play when it comes to preventing food safety problems. In this area we can't rely on consumers to make unsafe sprouts safe to eat. We can't ask them to cook their sprouts. I mean sprouts are natural and fresh and you want to eat them raw. They were intended to be eaten raw. It's unreasonable to ask people to soak them in chlorine, for example, once they get them home and probably wouldn't be very good for them anyway so the onus is on you to really deliver to us a safe product.

If sprout safety cannot be ensured then the industry should use labels to alert high risk consumers to avoid the product.

We recommend that alfalfa seeds should not be used unless they've been produced under conditions which are suitable for human consumption; that FDA should ban the use of mixed batches to aid in traceback; the FDA should encourage the development of safe and natural decontamination methods; the government should provide greater oversight to the sprout industry; and that consumer warning labels should be required until effective controls are identified and fully implemented.

I know this isn't what you want to hear or at least some of it.

So let's talk about it. Let's talk about a couple of these recommendations. The practice of using seeds that have been grown for agricultural use should stop. I know that this is a radical suggestion. I know that implementing this suggestion would have wide implications for the entire industry. However, the outbreak data clearly illustrate the contaminated seed are the primary cause of sprout outbreaks. While farmers can safely use manure on alfalfa grown for agricultural purposes it should be strictly banned in the growth of seeds intended for human consumption. Farms that supply seeds to sprout growers should observe strict guidelines for the growth of these seeds and should exclusively root their seeds towards the production of human food.

In addition, the practice of growing sprouts from seeds imported from other regions of the world should stop unless it has been demonstrated that the seeds have been produced under similar strict guidelines.

The practice of mixed batches of sprout seeds makes traceback nearly impossible. Alfalfa seeds intended for human consumption should be maintained in intact batches that are carried through the food production chain from the farm to the table. The batches of seeds and packages of sprouts should be clearly labeled or tagged for ready identification during a recall or traceback. This will help assure that problem seeds and sprouts are readily identified and may be quickly removed from the market.

The methods currently in use, calcium hypochlorite and chlorine, are primarily reduction steps and not elimination steps. While these methods may eliminate some of the harmful bacterias, others may survive treatment. We recommend that all decontamination methods should be challenge tested with seeds positive for E. coli 0157:H7 which may be more resistant to some treatments than other pathogens. Both FDA and the seed and sprout industry should provide clear and accurate information to consumers on the effectiveness of each of the various decontamination treatment methods.

We think the government should require that all sprout growers be classified as food handlers; that unsanitary conditions in the processing plants can lead to contamination of the seeds or sprouts, particularly because sprouts are grown in a warm moist environment which facilitates bacteria growth along with sprout production.

CSPI recommends regular inspections of sprout processors by both state and federal food safety inspectors. In addition, CSPI believes that implementing a HACCP system for the sprout industry would be an effective tool and should be considered.

Although there is no pasteurization step currently available for sprout processors that would leave the product fit to eat in its natural form, there are possible barriers or multiple hurdles that could be incorporated into a HACCP system.

In response to the foodborne illness outbreaks associated with sprouts, FDA has issued a number of consumer advisories over the past few years so let me tell you where that leaves consumers. We know that's not what--when it happens your sales go down dramatically but where it leaves consumers is this: This relies on consumers' memories of what they might have read or heard in the news to protect themselves. This is a far less effective technique for transmitting consumer information than a label on the container. We believe that labels on sprout containers should alert consumers that the product may not be safe to serve to consumers, immunocompromised or elderly consumers or children.

In previous years FDA has implemented mandatory labels for several high risk foods, including unpasteurized apple cider, that pose a comparable risk to sprouts.

So, in conclusion, CSPI urges FDA to move quickly and expeditiously to start the notice and comment rulemaking process for sprout safety regulations. There have been over two dozen outbreaks linked to sprouts in the past ten years, representing thousands of consumers who became ill. Until effective controls are identified and fully implemented, we believe sprouts should be prominently labeled to alert consumers to the risk of eating sprouts.

I want to make one final note. I was very impressed with the last presentation on actually putting the burden on the people who distribute the seeds to actually test them for the safety and it reminded me of practices in the ground beef industry. When I was working on the Jack in the Box outbreak back in the mid 1990s the beef industry said, "Oh, we can't test the ground beef. We can't test every lot for E. coli 0157:H7. That would be a terrible burden and it wouldn't be effective."

Today many companies in the ground beef industry are testing every lot. They are using test and hold programs and they are accepting that responsibility. I thought the presentation that showed that the people distributing the seeds to you would be willing to accept that responsibility by actually sampling the lots is a very important statement about where we need to be going on this.

Thank you.

Questions and Answers

DR. BERU: Thank you, Ms. DeWaal.

I would like now to invite the panelists who have spoken thus far to the table for a question and answer session.

If you do have questions, please step up to the microphones as we are recording this meeting. The transcript, as I said, will be available in about four weeks time.

MR. WARRINER: My name is Keith Warriner from the University of Guelph. This is a question for Bob R. You mentioned in your talk about preventing four cases through spent irrigation water testing. Well, there was two cases where they did do all the testing but they did deliver contaminated product to the marketplace. So is this application of no sanitation and just depend on testing really going to be effective?

MR. RUST: I'm sorry. I didn't understand the question.

MR. WARRINER: All right. So we're talking about just testing rather than sanitation and trying to put greater reliance on the testing of seed batches to determine contamination of that batch and you said that saved about four cases of outbreak cases but there was two, wasn't there, that they did all the testing but they still delivered contaminated product to the marketplace.

MR. RUST: Yes.

MR. WARRINER: So do you think your application of just relying on testing will really be adequate?

MR. RUST: Relying on seed sampling and testing, no, it won't be adequate. It's just one step in seed safety, a reduction step.

DR. SMITH: If I can say something, I think the point that you just made is that screening the seed is an additional step above and beyond what we specifically laid out in our guidance and in the instances where contamination is heavy enough that you can catch it at that point, you keep that seed out of the food supply and you avoid some of the outbreaks.

Something else I was thinking of just before we came to the table, the question earlier this morning about seemingly mixed messages from CDC and FDA about the relative importance of seed disinfection treatment versus testing, I think there are a lot of examples where the outbreak investigation shows that seed disinfection had an impact, either on implicated lots used at facilities where seed disinfection programs were not causing illnesses, things like that, or lower numbers of illnesses at least from firms.

But when it comes to testing, the testing of the spent irrigation water at the facility, we have fewer anecdotes from our investigations possibly because the testing caught contaminated product and avoided the outbreak.

On the other hand, even if you're testing--if you're testing the wrong thing or doing it the wrong way, you may have a false negative that lets product through and results in an outbreak. It may be well intentioned but still an error in the application of the test itself or an error in the way that the results from the test are interpreted, and we have a couple of examples of situations where that has happened.

MR. WARRINER: Yes. I'm pretty amazed that with all this testing going on nobody can actually give a figure about how effective testing is and I think people would buy into it more if they knew it was working.

DR. SMITH: It would be nice to know who is testing, how often and how often they find a positive and what they do with product. People are probably afraid to tell us but we'd love to know.

MR. WARRINER: Definitely. Thank you.

DR. BERU: Are there any other questions?

Yes, please?

MR. LALLEY: Good morning. I'm Mike Lalley from Living Foods and a couple of quick questions. One for Ms. DeWaal. You mentioned in your presentation that there were 4,500 cases identified specifically with various food products. Of which I see a whole 27 have been related to the various sprout products.

By my poor mathematics, I come out to some place just less than a half of one percent of all of these particular issues that you've raised which are of concern to all of us are related to sprouts. That seems to me to be a fairly low number and not one to be ignored obviously but, that being said, we're all well aware of the green onion problem in the Pennsylvania and Georgia, I believe, where we had dead bodies littering the streets.

Are we labeling--is your organization proposing that we label each package of green onions with a warning label? It has been shown that apples which have been defecated upon by birds are absolutely impossible to decontaminate. There's just no way.

Now maybe we should suggest to the public let us quit consuming raw apples, cook your apples prior to eating. And these are just the items that have been identified.

Like I say, if the bird defecates on the apple, we don't have a multi-state situation. He may go over on to another tree and his cousin and his brothers, they may be defecating on other apples and so we may have numerous people ill but not able to traceback.

So it just seems to me that we're a little over anxious to attribute these terrible situations to something that is--to an industry that is in a position that's very difficult to defend ourselves.

DR. BERU: Let's give Ms. DeWaal a chance to respond.

MR. LALLEY: A quick response. I'd like to have one other question.

MS. DeWAAL: Thank you very much. I understand the challenge for the industry of addressing the problem just because of the production issues but by my calculation it's--at least in our database it's about 29 outbreaks. And the bottom line is there--it's serious.

If people get E. coli 0157:H7 it can result in hemolytic uremic syndrome. People exposed to Salmonella have other serious complications. So these problems can be addressed and other industries have faced them.

Now unpasteurized apple cider--I mean FDA held meetings just like that to address that issue.

MR. LALLEY: And that contamination came from--the unpasteurized apple cider, the contamination came from apples?

MS. DeWAAL: It's apples.

MR. LALLEY: Maybe we should--are we warning on apples? I mean, if, in fact, if we drew some multitude of apples and are able to put E. coli into the finished product--well, then it's a no brainer that the apples are contaminated. If the apples are contaminated at a level to which we can no longer consume fresh apple cider without having a warning in our faces, where is the--maybe then if I juice my sprouts--I mean, we're--

MS. DeWAAL: This is a process--

MR. LALLEY: I mean, there's just no--yes, but you're ignoring--I mean, just--I mean, is there any--well, just answer the question then.

Is there any specific desire on your organization or FDA or anybody else to warn consumers of the consumption of raw produce, fruits, vegetables, et cetera, which we all know are in a position--are grown in a warm--you know, we have talked about the warm, moist environment in the sprout operations.

I also do a little dirt production myself of culinary earth, which just so happens that it's a warm, moist environment that produces all of the raw fruits and vegetables that we consume. It's not only in our sprout environments so it just seems to me that we're focusing an inadequate or an over abundant amount of attention on this particular group.

MS. DeWAAL: Well, as you saw in my presentation, I did give you the data on all produce because we do focus our efforts there but more importantly, I think to your question, is that you are producing a processed food product and the processing may be something simple and natural that can be done in my home kitchen but nonetheless if you bring your products to the market consumers have an expectation of safety that you need to address and these outbreaks demonstrate that the processing needs to be improved.

MR. LALLEY: Absolutely and we're in total agreement but the question is, is there any desire on your group or the FDA to take this warning label and apply it to all items that may fall into this type of a situation? Items that are consumed raw primarily, fruits and vegetables. Do you have the--does your organization have any of that in process?

MS. DeWAAL: We supported the use of labels also on another raw process food.

MR. LALLEY: Green onions?

MS. DeWAAL: Excuse me. No. Raw--unpasteurized apple cider.

MR. LALLEY: How about green onions? We had dead bodies littering the streets. I mean, no--I mean, it's--like I say, by virtue of the fact that I'm strong and my position in no way reflects the fact that I disbelieve that there's problems, potential problems on any food product, with that being said why the inordinate amount of attention in the complete looking the other way on these other products? I mean, it just seems absolutely ridiculous.

DR. BERU: I think that question has been asked and--

MR. LALLEY: And the answer was we are or we are not considering warning labels on other fruits and vegetables? Yes or no?

DR. BERU: I can tell you the--

MR. LALLEY: It's a yes or no question.

DR. BERU: --FDA is not considering labeling of all produce as you're laying that out.

Are there any questions?

While we wait for questions, I had a question for Dr. Dechet actually. In your presentation you mentioned that the number of illnesses per outbreak average over the years seems to have gone down. Can you tell us what you attribute that to?

DR. DECHET: Certainly one possibility might be that the guidelines are being followed and that has had some effect. I think we probably think it's more due to the fact that awareness has been raised about outbreaks throughout the states. The Department of Health and Human Services now has field epidemiologists in all states, EIS officers in almost all states, so foodborne outbreaks are very rapidly investigated.

We also have improved our gel electrophoresis, our DNA fingerprinting program. It's now every--and this actually touches on one of the questions that was raised just a minute ago, you know, are we testing other food products and how do we know that when Salmonella is due to one product or another. Depending on the state, for instance, every tenth sample of Salmonella gets DNA fingerprinted and those results are sent to the CDC, and it goes into our national database and then our database staff will compare those isolates all across the country.

So I think now we're able to see that something in California, oh, look, it's the same exact same DNA fingerprinting as something in Arkansas, as something in Massachusetts, and then those states are brought together and then they actually do an investigation.

We've also included some of the high risk food items, green onions, unpasteurized cider, sprouts, on our general--we call our shotgun questionnaire when we can't find a food item. It involves over 400 food items and we'll start looking for some of those high risk food items. So I think it's greater awareness. I think it's earlier detection and rapid investigation that probably has actually reduced the numbers per outbreak.

DR. BERU: Panelists, feel free to ask each other questions as well.

Yes?

MR. SANDERSON: This is a question for Caroline. I think all of us agree with you completely that good agricultural practices for seed production is absolutely common sense but it's really kind of difficult when you try to figure out how to bring it about. I wonder if you know of any industry where good agricultural practices went from being a good idea to being a regulation that was consistently followed and how that was done because maybe we could take some pointers from that.

MS. DeWAAL: Bob, that's a great question and I'd be happy if FDA also looked at that issue. One of the challenges here, and I do work on food safety issues across both FDA and USDA issues really from farm to table, but one of the challenges here is that we don't really regulate farm production and so we are relying on good agricultural practices at the seed production level as we do for green onions, as we do for cattle, as we do for many other items so those are voluntary standards. So what it means for the sprout growers as an industry is that you're really relying on your source--on your source material and that's the seed.

I think that we don't regulate--we don't have an agency you can go to that regulates on the farm. You don't have anyone you can go to and say could you help ensure that the seeds that I'm being sold are adequate.

So I think you have to rely on some of the programs we've heard about today that give you some level of assurance but, no, I have not seen an on farm program that--where good agricultural practices work. Now, that said, one--or has moved to a voluntary program that has been highly functioning.

Now the exception there--there is always an exception--is eggs. And in Pennsylvania they developed--they had a major problem with Salmonella enteritidis in the flocks and they did develop on farm practices that showed themselves to be highly effective at controlling SE in eggs and that is slowly moving towards a regulation now that will provide a level playing field for the egg industry.

So again it's harnessing kind of the kind of programs that are being field-tested and then harnessing those into good regulations. That's really where we are. I mean, the sprout issue isn't new. We've been looking at it since the mid 1990s and this is at least the second major meeting I've attended on this so this is not a new issue. FDA has not moved quickly to consider regulations but with the outbreak data today it's clear we need more standards in place.

DR. SMITH: If I can add to that just briefly. You're right that good agricultural practices are guidance and we don't have the authority to enforce guidance, and that's not what guidance was about but we do have authority over all FDA regulated foods, including the farm. So we do--we are able to get there but we don't have the implementing regulations to lay out what's right and wrong. It has to be an obvious call.

MS. DeWAAL: You can get there after an outbreak. You cannot implement regulations on the farm before an outbreak. I mean, you get to the farm after you--

DR. SMITH: Well, we don't have the implementing regulations and with the exception of a couple of invited friendly visits we don't tend to go to the farm very often unless there is cause that brings us.

DR. BERU: Are there any questions?

Yes, please? Again, name and affiliation, please.

MR. BOYD: My name is Steve Boyd and I'm with a company called Hydrose. We do molecular diagnostics of diseases and plants.

I submit relative to Bob's question, Bob Sanderson, that one model that you might take a look at is the international seed testing program that has been slowly developing over the years. In this case their focus has been primarily on diseases of plants relative to the plants. They have no human interest per se but the process and how they've gone about pulling together other nations as well as coming up with standardized test procedures. One of the bigger problems with this whole group is standardized test procedures and there are a number of different people. In fact, you just attended one of the meetings recently in Thailand.

I mean, this is a group that has been doing this sort of testing. They're well aware of the statistical issues. Their focus is on the quality of seed both sold into the United States and globally. The question is what the pathogen is relative to us so I submit that there is one organization that if you look at a pathway it might be of some use.

Scientific Perspective

DR. BERU: Thank you. We are really running behind time and so I will stop the questions and answers at this time and move on to the next presentation. Thank you, panelists.

We'll have one of the science presentations before we break for lunch and I would like to invite Dr. Jed Fahey from the Johns Hopkins School of Medicine to make his presentation on Microbiological Findings on Sprout Operations Following FDA Guidance.

Microbiological Findings on Sprout Operations Following FDA Guidance

[See presentation slides for Dr. Fahey]

DR. FAHEY: Good morning. Can you hear me?

I'd like to spend the next 15 minutes or so getting you ready for lunch and talking about food safety problems of course.

I've got an asterisks next to my name because I've got my e-mail address there. If there are slides that you have--that you'd like copies of or if you have questions, I'd be happy to provide them to you later.

I didn't hear the first couple of presentations. I apologize for that. I hope I'm not re-hashing anything that was already said but I'd like to present some data that we developed actually four years now on hold and release testing of sprouts, a variety of green sprouts. And we, as with most of you, were too busy and too overworked and we just haven't had time to try to get this--make a serious effort to get this published. We should do that and will do that, especially if there's interest in the data here. So the data has been around for a few years. We haven't been holding out. We've just been--just haven't gotten to it.

So by way of overview, I'd like to tell you about a one year program of hold and release testing conducted in concert with strict seed and facility cleaning procedures by 13 U.S. broccoli sprout growers. These tests were not all done on broccoli sprouts but my colleagues and I developed broccoli sprouts and, in fact, started a company to promote chemoprotective food products and the company sells broccoli sprouts. I'll show you a formal disclosure or disclaimer in a second so we obviously had an interest in working with those folks that did grow broccoli sprouts.

The microbial contamination tests were performed on almost 7,000 drums of sprouts, equivalent to about 5 million consumer packages of fresh green sprouts, and less than one percent of those sprout samples gave an initial positive test for E. coli 0157:H7 or for Salmonella . And, when retested, only three drums again tested positive and obviously they were destroyed and did not make it to the marketplace. I'll go through some of the details in a minute.

We also looked at composite testing and although this is not a commercial message for composite drum testing, it seemed to be every bit as good as testing individual drums.

By way of disclosure, we started a company to sell broccoli sprouts, Braccisca Protection Products. I'm an unpaid consultant to the company now and my stock is under lock down by Johns Hopkins, who by the way also owns stock in the company. But clearly we've been working with broccoli sprouts, we've been working with broccoli and cruciferous vegetables and have great interest in their thriving as well as the interest of the sprout--all of the green sprouts.

So I present to you just a couple of slides specifically about broccoli sprouts and then I won't say anything more about them just to remind you that there is considerable or maybe I should just say significant antibacterial activity in seeds and sprouts of cruciferous vegetables, which may indicate that there is some advantage to growing them in terms of microbial safety. I don't know. But after only a short amount of time when we measure the antibacterial activity against a lab strain of E. coli , not the nastier one, we find a marked impact on bacterial growth from that leachate or I've called it leachate but it's essentially the rinse water.

And also, and this is published--the previous slide wasn't but we published a couple of years ago in the Proceedings of the National Academy of Sciences there is tremendously potent antibacterial activity of sulforaphane, one of the major phytochemical components of broccoli and some other crucifers against helicobacter pylori, which is the causative agent of many, if not most, ulcers and is a risk factor for stomach cancer.

The contrast there between the top four bars and the bottom one is just to show that there's actually two or three logs of difference in the potency of this particular phytochemical from broccoli seeds against helicobacter pylori, a human pathogen, compared to a number of other environmental bacteria and fungi so I just mention that in passing. I'm not going to present any more data on that but there may be some additional oomph or punch to be attributed to the cruciferous seeds and sprouts from them.

So again this may be--I've heard warm moist environment a couple of times since I walked in already. Sprouts, as you know, are seeds that have been placed in environmentally controlled hydroponic conditions. They're incubated in a warm moist nutrient rich environment made nutrient rich by the leachate from the seed itself and they are ideal environments for microbial growth so that if E. coli or Salmonella is present it's going to multiply. It's almost a guarantee.

So to date I'm not aware of, and we may hear data later but I'm not aware that practical methods have been developed to check the growth of these contaminants during germination and sprout growth or processing so you've got to prevent contaminants from entering the process or if you find that the sprouts are contaminated obviously they've got to be destroyed but first you've got to identify them.

It's essential in my mind and our mind in my group at Johns Hopkins that seed that's going to be used for sprouting undergo some sort of surface disinfestation or disinfection treatment using a biocide of some sort. The efficacy of these agents--in particular, calcium hypochlorite--has been, I think, very well documented by the laboratories of Dr. Beuchat and others, many others. When these agents are used correctly the resulting sprouts are safe to eat and so as you know and I've seen in earlier presentations there is a guidance that the FDA issued six years ago on how to accomplish this surface disinfestation.

So we--well, this has been said before but the guidance then indicates that testing the spent irrigation water is the way to ascertain whether seedlots are contaminated or sprout lots.

The FDA received considerable guidance or considerable--not guidance but input from the sprouting industry and it is continuing to do so today obviously designed to prevent contaminated sprouts from ever reaching the public. The FDA, as far as I'm aware, followed up the issuance of this guidance with an inspection of 150 or so sprouters and determined that only about half of them were complying with the guidance.

So what we've done is examine a subset of these sprouters in whose facility compliance with the FDA regulations was verified based upon inspections by two third party auditors.

This hold and release testing obviously was done--obviously based on my disclaimer or disclosure--with co-packers who agreed to participate. Co-packers for the broccosprouts product line and Brassica Protection Products. It was done during the calendar year, the entirety of the calendar year 2001, and there were 13 producers. These companies were all established companies. They hadn't just sprung up to grow broccoli sprouts. They grew and continue to grow many different kinds of green sprouts and they had agreed to comply with Brassica Protection Products' rigid standards of sanitation. We selected the plants so that they were geographically distributed across the U.S.

All of the surveyed growers followed all of the steps specified in 1999 FDA guidance. They surface disinfested or disinfected--I apologize. I use surface disfestation [sic] and disinfection interchangeably. I'm sure there are subtle differences but they are unknown to me. We--they rather performed microbial testing of the spent irrigation water from each batch of green sprouts and they were subject to, and those of them that are still working with BPP are still, of course, subject to both announced and unannounced third party audits as well as state and federal inspections of their procedures.

The routine was exposure to 20 parts per million calcium hypochlorite for 15 minutes followed by extensive rinsing to remove residual chlorine. They were grown in either trays or drums and, of course, the only three inputs are light, heat or constant temperature at least and clean water. And in the case of the 13 sprouters we examined they used one of three sources of water. It was either filtered. It was well water or it was chlorinated municipal water.

So a sample of the irrigation water, typically one liter, was collected after two days of growth of the sprouts. Since they are typically grown for 72 to 120 hours, the rapid microbial test, therefore, permits the sprouters to abort contaminated batches either before packaging or certainly before shipping and distribution. And in the cases of these sprouters all of them were held until the test results were obtained.

The data was not forwarded to me. I'm not particularly good at handling large masses of other people's paperwork. I have enough problems with my own. They were forwarded to a company who specializes in quality assurance measurement, Quality Associates, Incorporated, in Columbia. And they also reported some other data, including the types, the other types of green sprouts that they were growing, which is listed in the second bullet there. Some of them actually included wash water or irrigation water from some non-green sprouts, some mung or soy bean lots. And to the extent that that data was combined--in pooled drum samples--to the extent that that data was combined with green sprout data, we included it. To the extent that it was exclusively bean sprout, we did not include it.

All but one of the growers used the services of an external micro-testing company to which samples could be delivered on the day of collection so these were obviously in their geographic area.

One of the growers had set up an in-house laboratory and we insisted that if we were going to use their data that we receive all the laboratory notes from their in-house micro lab.

All of the samples were screened for E. coli O157:H7 and for Salmonella species. Most of the E. coli testing--you know, you can't force people to use your favorite micro test certainly in this kind of a voluntary data submission so most of those testing for E. coli used a test kit that had been identified in the guidance. Not most of them or only a small number of them used such test kits for Salmonella . They used other tests in other words. Probably motivated by cost of the tests.

Most of the tests were AOAC official methods. Most of the kits rather. And I'm not even familiar--I'm not a microbiologist by training. I'm not even familiar with some of the subtle distinctions here. One was an AOAC performance tested method. One was an FDA bacterial analytical method for Salmonella testing and a couple of them didn't specify the types of tests that they used.

All of the results, however, were entered in the database and the data entries were independently verified by an audit process at Quality--at QAI.

So I've got a list here which is somewhat hard to read but again I'll be happy to e-mail this table to anybody who wants it. This is a list of the different test methods that were used along with an indication of their reported sensitivity and specificity. Some of them, of course, or I guess all of them were developed for other products or the sensitivity and specificity rather was assessed on other products like milk, apple cider, orange juice and then some vegetables and poultry.

So the distribution of sampling then included a majority of single drum samples but as you can see about a third of the tests were composite tests on two to four drums done because of the cost. That was, I heard, mentioned a couple of times earlier. It is an expensive process and a number of growers are concerned that they be able to use pooled drum samples and then my understanding is if there's a positive or a presumptive positive in a pooled drum sample they would go back and retest individual drums.

So the results that we found were that there was again less than one percent presumptive positives, a total of 24 presumptive positives across all of these tests, no apparent skewing towards multiple drum tests or single drum tests. The total number of drums tested was 6,839 and of those 24 presumptive positives, as I mentioned earlier, when retested only three gave positive tests on retest and those lots of sprouts were destroyed and not sold.

So in most but not all cases the presumptive positive--the water sample from the presumptive positive was reanalyzed. In two of the cases the follow-up analysis was again positive for Salmonella and the sprouts were destroyed and in a single case there was a presumptive detection of E. coli . It was assumed by the grower to be real and he or she destroyed the sprouts without a follow-up test to confirm whether this was, in fact, E. coli 0157:H7.

None of the samples included broccoli sprouts as a parenthetic statement and all of the sprouts had been held at growing facilities until the microbial results were back so that there were no instances in which contaminated sprouts were released for distribution.

So we feel--as far as I'm aware this is the largest collection of such data that's at least been shared publicly. We're--we at Hopkins since we developed sprouts and we're concerned about and continue to be concerned about food safety, we were very eager to do this test and happy to have the cooperation and collaboration of the sprouters who did join us.

I think this is a kind of testing program to be proud of and not to hide. My only apology is that we have had this data for a while and just haven't sort of gotten it out there. It has just been a case of entropy, I guess, but I'm glad to share it with you now.

This work was supported by a philanthropic donor to our lab, Lewis B. and Dorothy Cullman Foundation. And I want to acknowledge the contributions of a number of colleagues. In particular, Phil Ourisson at QAI, the Quality Associates, and Kitty Stevenson, who works with me. Again thanks to the sprout companies. I don't know if any of you are here who agreed to be audited and share their records and sort of show us their petticoats.

That's the end of my presentation.

DR. BERU: Thank you, Dr. Fahey.

We are going to hold questions until all of the science presentations have been done. It's now 12:00 o'clock. We'll take a lunch break and we'll start promptly at 1:00 o'clock.

Thank you.

[Whereupon, at 12:00 p.m., a lunch break was taken.]

- - -

Afternoon Proceedings

[1:00 p.m.]

DR. BERU: Good afternoon and welcome back from lunch.

We will continue the science presentations and our next speaker is going to be Dr. William Fett. He is from the Eastern Regional Research Center at the USDA ARS and his presentation will be on interventions.

Interventions

[See presentation slides for Dr. Fett]

DR. FETT: Good afternoon. I know it's good snooze time right now so, hopefully, we can keep you awake throughout the afternoon with all the presentations.

My presentation is actually joint with Dr. Rajkowski, who is going to cut me off after the first 20 minutes.

What I'm going to do is give you a very brief overview in the 20 minutes on the interventions for use on sprout seed that have been published. I'll only be talking about publishing information in my part of the talk and Kathleen will be talking a little bit about unpublished information with her work.

So I'll start out on interventions for use on sprout seed and then if time permits talk a little bit about interventions during sprouting and postharvest. There is actually less information on those two areas and then Kathleen will finish up with a brief presentation of her ongoing collaborative study with industry.

So we already heard about the current guidance and that seed should be treated with one or more approved antimicrobial treatments such as the 20,000 ppm calcium hypochlorite and that at least one treatment should be applied immediately before sprouting.

So the challenge is to obtain a five log reduction of human pathogens on sprout seed while maintaining seed germination, vigor, sprout appearance and yield at acceptable commercial levels.

Really what we ultimately would like to achieve is elimination of all pathogens on the seed due to their potential for outgrowth during the sprouting process that you've already heard about in a couple of presentations this morning.

So I thought this was an interesting graphic. I just pulled out the papers that I had in my file on interventions for application to seeds or sprouts or during the sprouting process. As you can see, after the outbreaks began in '95 and sprout safety became a much more visible arena in food safety, you can see that initially '96-98 there was little in the way of published information but during that time I think it was obvious that people were thinking about the problem, submitting proposals, proposals were becoming funded and the research was getting underway. You can see the explosion of papers until 2003 that have been published in this area.

There was a little bit of a drop off in 2004 and I expect for this year there has been two papers published so far that I'm aware of and we'll probably end up with around seven or eight papers. And, you know, research is continuing in this area and you obtained a couple handouts this morning concerning unpublished information on potential interventions so it's still an active area of research. We probably won't see the number of papers that we saw in 2003 again but in the produce arena it's probably the most active area of research for a single produce item.

So there has been a lot of work done on chemical interventions on seed primarily with aqueous sanitizers and they are listed here. They have been tested at room temperature or at elevated temperatures, 42 to 50 degrees Centigrade, and with or without the presence of surfactants up to about one percent.

If you look over the literature the elevated temperature during treatment was usually beneficial but the problem was that, even though you got a little bit higher log kill, the seed germination was reduced often to an unacceptable level.

The effect of the addition of surfactants has been variable. Even with one percent Tween 80, the most increase that is in the literature is about one log and often less than that.

There has also been a variety of gas phase treatments looked at for sprout seed and those are listed here from acetic acid vapor down to ammonia.

In addition there has been physical methods tested for interventions on sprout seed, including dry heat, hot water, irradiation, both gamma and pulsed UV, hydrostatic pressure and radio frequency dielectric heating.

And, also, there has been biological intervention work done on a more limited scale than chemical or physical interventions. There has been a little bit of work on the use of lactic acid bacteria as an antagonist to Salmonella and E. coli 0157. And, also, there has been some work at our center on the use of fluorescent pseudomonades as biological agents or competitive exclusion agents. And, also, some work on whole bacterial communities, isolated both from laboratory grown sprouts and also market purchased sprouts.

And, finally, there was one paper out there and at least two groups interested in the application of bacteriophage for controlling human pathogens or eliminating human pathogens on seed.

There has also been a limited amount of work on combination treatments for sanitizing seed, both combination of aqueous chemicals such as lactic acid followed by 2,000 parts per million of chlorine, combination of physical treatments with aqueous treatments such as ultrasound and heat in combination with calcium hydroxide Tween 80, Tsunami 200 and Fit. The last two are commercial products. I'm not sure if Fit is still available in the marketplace. Finally, a combination of physical treatments such as a dry heat step followed by gamma irradiation.

So just briefly to examine what has been found with chlorine and high levels of chlorine. Usually 16 to 20,000 ppms has been studied in several laboratories. There are many papers on this subject. First of all, the data that has been obtained has been highly variable as far as actual log reductions that have been reported ranging from two logs to about seven log reduction. This is probably due in great part to differences in methodology that have been used in different laboratories for this kind of research and some are listed here but there are several others.

I think this makes it difficult to assess alternative treatments unless the alternative treatment is done in the same laboratory with the same set of conditions and along with the high level of chlorine treatment so you can directly compare.

The one consistent finding is that this treatment does not consistently eliminate pathogens from inoculated seed in the laboratory. As I mentioned, if we're going to compare alternative chemical treatments it would be best to run 20,000 ppms chlorine in those experiments to get a direct comparison using the methodology that you're using in your particular laboratory.

Little has been done with chlorine treatments on naturally contaminated seed. Trevor Suslow's group and my group have both reported that the use of high levels of chlorine with or without buffering to a neutral pH did eliminate Salmonella from two naturally contaminated seedlots that we looked at.

In Dr. Tortorello's lab in Chicago they found that similar treatment with actually one of the same lots of naturally contaminated seed was not effective in eliminating Salmonella . There were some differences in methodology that the three groups used that could lead to such a difference in the results.

It also could be just that the particular group of seed that were used, even though it was from the same lot, might have had contaminates located in different locations on the seed where she might have seed with contaminates that were more difficult to reach with a chemical aqueous sanitizer than the seed that we looked at.

So what are some potential alternatives to high levels of chlorine for treating alfalfa seed that had been mentioned in the literature? And there are really not that many that have been put forth as alternatives. Larry Beuchat's lab, he has proposed calcium hydroxide at one percent, ten minute treatment at room temperature as an alternative and also the commercial product Fit used 15 to 30 minutes at room temperature. In our lab we have looked at a couple of commercial products, Citrex and Pangermex at 20,000 ppms, 10 minutes at room temperature, and found similar log reductions to chlorine. We're not usually seeing a five log reduction on alfalfa seed and we're certainly not eliminating the pathogen from the seed but the log reductions in these direct comparisons were similar.

One of the problems, if you will, with the research as a whole here is that they were tested only in a single laboratory by a single group and really it would be nice to see some independent confirmation of efficacy by other laboratories and also scale up studies. Usually these were done in very small scale with small amounts of seed. Also there is again safety issues primarily with calcium hydroxide and also possibly cost issues with some of these other commercial products.

As far as biological interventions go, the most successful intervention so far was the addition of whole microbial communities isolated from market sprouts. These are added at the seed soak stage and at day seven the sprouts were harvested. Actually days one, three and seven the sprouts were harvested and the populations of Salmonella were determined and compared to the control at day seven the sprouts grown from seed treated with these whole microbial communities showed a five log reduction in population of Salmonella . This was some work done at our center in collaboration with J. Garland down at the Kennedy Space Center.

One interesting thing was that when they tried whole microbial communities from laboratory grown sprouts they weren't as active as from market sprouts.

What has been published as far as chlorine treatments, high levels of chlorine for mung bean seed? We did a little bit of work in my lab a few years ago on this and we did find that it was pretty effective for reducing Salmonella on mung bean seed. We got about a five log reduction but not in elimination and with E. coli 0157:H7 about a four log reduction but again not elimination. There was no effect on percent seed germination.

So what are the potential alternatives to chlorine, high levels of chlorine that have been proposed for mung bean seed? And actually they are more numerous than for alfalfa seed. As Michelle mentioned, I think one of the reasons is that I think mung bean seed is a little bit easier to sanitize than alfalfa seed because the seed coat is often intact, fairly smooth, where as alfalfa seed often has cracks in the seed coat and also you can get wrinkled seed because of insect damage, which makes it much more difficult to sanitize than mung bean seed.

Bob Rust was telling me last night that often mung bean seed, which usually comes from China, is actually hand harvested versus alfalfa seed, which would be combine, which would lead to probably damage to the seed coat.

Anyway, in 1999 a group from Canada, Pasqual Delaquis, published this paper on gaseous acetic acid treatment of mung bean seed and they were very successful in eliminating Salmonella and E. coli 0157:H7 from inoculated seed. They did not eliminate Listeria monocytogenes but did report a five log reduction. There was a slight loss in germinability of the seed but at least from a statistical standpoint it wasn't significant. Unfortunately, the same treatment applied to alfalfa seed led to much greater loss in seed germination.

More recently Randy Warbol's group at Cornell reported that a prolonged dry heat treatment of mung bean seed, four to five days, was very effective in eliminating Salmonella and E. coli 0157 without affecting seed germination. Again, the similar treatment of alfalfa seed was found to be detrimental to seed germination.

A group in Europe reported that hot water treatment of inoculated mung bean seed were effective in achieving a five log reduction of Salmonella in these various temperature time combinations for this and non of the treatments were reported to reduce seed germination below 95 percent.

So a very--apparently a very effective method for mung bean seed with little effect on seed germination. Most likely due to the fact that mung bean has a much thicker seed coat than does alfalfa seed and that protects it from the detrimental effects--thermal effects against the embryo.

Dean Glyver's group published this work several years ago on use of ammonia gas treatments for mung bean seed and reported very good activity, although not elimination of Salmonella or E. coli 0157:H7 and no effect of germination. Again the similar treatment to alfalfa seed was not as effective but apparently did not affect germination.

As far as combination treatments go, Dr. Bari at the National Food Research Center over in Japan reported that if you expose the seed to dry heat 50 degrees Centigrade for one hour followed by hot acetic electrolyzed water along with sonication you would get between around a four or five, 4.6 log reduction of E. coli 0157:H7. You did have survivors but no effect on germination and sprout growth rate.

Alternatively, they tried dry heat treatment followed by gamma irradiation at two kilogray and two kilogray has been shown to reduce populations of pathogens on seed without reducing the subsequent germination or yield so it's probably an acceptable exposure of the seed. And they got about a 4.6 log reduction of E. coli 0157:H7 with no survivors and no effect on germination but they did see a reduced sprout growth rate.

As far as other seed types there is not much information out there. One group published on the use of high pressure for reducing populations of pathogens on garden cress, sesame, radish and mustard seed. They had most success with garden cress where they obtained a six log reduction of Salmonella , generic E. coli and Listeria innocua. There were survivors after treatment and germination was delayed by about one day but after that the sprouts grew just fine.

The other seed types tested in this work, sesame, radish and mustard seed, were more pressure sensitive and showed a greater loss in germination and sprout growth rates.

For radish seed Dr. Bari in Japan again published that a dry heat treatment, 50 degrees C. for one hour, followed by gamma irradiation at two kilogray again gave about a five log reduction of E. coli 0157:H7. Again there were survivors.

So, in summary for the seed, there is no method currently available which eliminates pathogens from artificially inoculated alfalfa seed while maintaining seed germination and vigor. As far as effectiveness of high levels of chlorine to eliminate Salmonella from naturally contaminated seed the data is variable at this point. I think the published investigations of outbreaks indicate that the use of high level of chlorine may not always be 100 percent effective under common commercial practice and that may be due to the way it is being applied and, as we heard earlier, I think there is a lot of variability in how these treatments are actually being applied by commercial growers but I think it was also evident that it can reduce risk.

For mung bean seed especially there are several potential alternative treatments to high levels of chlorine. Acetic acid vapor treatments, ammonia gas, dry heat and wet heat. High pressure might be applicable to some seed types such as garden cress.

There is another report out with alfalfa seed and high pressure where they didn't have such good successes with garden cress. They got a large reduction in germination of the seed by high pressure treatment.

Finally biological control with whole communities appears promising but it's a very long term research and development project which I think would take--you know, I think it's worth doing but I think it's going to take years to come up with a commercial product.

And then most of these results have been done--you know, from investigations in a single laboratory and they really need independent confirmation. They need scale up studies and regulatory approvals may be needed for some of these alternative treatments.

As far as sprouts go, the challenge is to significantly reduce or eliminate pathogens on sprout surfaces during growth or postharvest without compromising sprout appearance. Not yield, that shouldn't be in here but shelf life and nutrient content.

And there is substantial laboratory evidence for a variety of sprout types that Salmonella and E. coli 0157:H7 can actually get internalized into the sprout that are grown from inoculated seed or if you inoculate the roots at an early stage. Whether this happens under commercial practice I don't think we really know but the potential is definitely there.

So during sprouting as far as addition of antimicrobials to the irrigation water there has been quite a few chemicals looked at from chlorine to acetified sodium chlorite but no intervention was highly effective.

And then I'm going to quickly go through this because Kathleen is giving me the cut off sign here. As far as postharvest interventions there has actually been a fair amount of work and I'm going to skip over that.

And one of the most effective postharvest interventions has been irradiation, either gamma or electron beam, chlorine but at 2,000 ppms rather than 200 ppms, and chlorous acid at 268 ppms but regulatory approvals to use these chemicals, I think, are required.

So, in summary, on sprouts really during the sprouting process there's no effective way of eliminating Salmonella or E. coli 0157:H7. And addition of the antimicrobials to the irrigation water would interfere with the spent irrigation water testing. The test may come back negative but you still might have viable pathogen cells on or in sprouts.

And after harvest probably only gamma or beta irradiation would most likely eliminate both surface borne and internalized pathogens and neither is approved right now for use on sprouts postharvest.

So I'm going to turn it over to Kathleen to give you just a brief overview of her collaborative study that's ongoing.

Interventions

DR. RAJKOWSKI: Good afternoon.

As a quick means of introduction, I was asked to revisit two issues. One particular one is a liquid sanitizer, Tsunami, which was reformulated by the commercial company. And to investigate what would be considered a chambered ozone treatment.

The first table are the results obtained when gaseous ozone with a 65 percent relative humidity was applied to the seeds. There were six batches or six separate inoculums done at six different times. The consistency was that they were all done for 24 hours with the same ozone rate or ppm which was 30,000 ppm. The average reduction using this ozone treatment is approximately one log.

But in many of the studies using ozone what we did find is that the recovery on a selective ager showed injury. The results presented here are on a non-selective ager. Based on the fact that there was indication of an injury we took these particular treated alfalfa seeds and gave them treatment with the Tsunami. Now being a research group I started using the Tsunami at many different percent levels. I went from the one percent, which was the approved, all the way up to five percent to see what would happen and I found no difference between three and five percent so the upper level that I used was three percent.

What we did do is we did verify before I used any of the liquid sanitizer that the ozone treatment did not damage the seeds whatsoever so they were verified through independent laboratories. Two separate sprouters were given the seeds to germinate them and determine the yield ratio within their sprouting facility. Unfortunately, these people could not use the sprouts for sale but they were able to say, yes, they were good, we would buy these particular seeds. And one grower supposedly has been using the ozone treated seeds.

Using the peracetic hydroperoxide sanitizer to reduce Salmonella using percentage 0 through 3 percent, the non-treated is the control. The ozone treated is what I retrieved after using the sanitizer. At the two percent there was a two log reduction but at the three percent there was no growth after recovery using plating media, which was the tryptic soy ager which is a non-selective, and in order to prove that there was no growth at the three percent I did what would be considered the traditional MPN, most probable number, recovery and it confirmed the no growth scenario.

As Bill mentioned, when I did use the peracetic hydrogen peroxide sanitizer I also used the 20,000 ppm on the same seeds. After treatment you notice there it's blank because I got the same value as the non-treated. I could not find a calcium hypochlorite reduction and as a result we are going to be revisiting different methods of drying the seeds after the wet artificial inoculum. As Bill mentioned, several laboratories have been using different treatments. No one has been using consistently the same method of inoculation and drying the seeds.

I did my initial work using irradiation. In order to do that work I dried my seeds down to almost the same level as before inoculation and it is possible that that might be the reason why I am not seeing the reduction with the calcium hypochlorite because I would have seeds very much similar to a natural inoculated seed and the fact that it has the same or comparable percent moisture.

This is all unpublished data. We are going to be continuing this in collaboration with the FDA, Dr. Wu, and we will try to do, what Bill suggested, a split sample where I would take my inoculated seeds, send them to other laboratories and have them confirm the reduction using their technique but the same protocol.

Thank you.

DR. BERU: Thank you, Drs. Fett and Rajkowski.

Our next set of presentations is going to be on testing methodologies and sampling, and that will be by Dr. Mary Lou Tortorello and Dr. T-J Fu from OPDF, Moffitt Center, Chicago.

Testing Methodologies

[See presentation slides for Dr. Tortorello and Dr. Fu]

DR. TORTORELLO: Well, as Nega mentioned, T-J and I are sharing this overview on testing methodologies and sampling.

My part of this talk is going to be primarily a historical perspective providing you with the background, the rationale for how the FDA guidance in 1999 came to be for testing recommendations, that is what was known at the time based on the best available science that we had.

T-J, however, has done a lot more recent work on sprout research and is going to be giving you more of a looking forward perspective, that is what have we learned since that guidance was published and what are some of the new issues that should be considered.

So I'll begin this overview with a question. Why are we testing when a lot of the rest of the food industry has decided that endproduct testing of foods does not really ensure their safety for a couple of reasons? First of all, in a lot of foods there are low levels of microorganisms present and the type of contamination that is present could be very sporadic. So endproduct testing for the most part has been moved away from the rest of the food industry and this is why HACCP has come to be such a focus of food safety for the food industry.

However, sprouts are a unique food. In sprouts there is a natural microbial growth to high levels during the sprouting process and depending on the way the sprouts are produced those microorganisms can be quite uniformly distributed throughout the batch and so for sprouts product testing could very well make a lot of sense.

And so the questions that we asked before the guidance was issued was: What do we want to test and how and when should that testing be done?

And Caroline--are you still here--you asked if anybody remembered that first FDA public meeting on sprouts and I remember it. It was in 1998 and I remember it quite distinctly because it was at that meeting that I first heard of the idea of testing spent irrigation water to indicate the microbiological condition of the sprouts and that idea came from an industry representative. At that time it was decided that we at the National Center for Food Safety and Technology at the Moffitt Center in Chicago would try to verify this idea and so we went out and bought the big rotary drum sprouters and we bought a bunch of materials that would allow us to produce sprouts in our pilot plant and we attempted to verify this idea of testing the spent irrigation water as opposed to the sprouts.

Spent irrigation water, of course, being that water that is used to mist and hydrate the sprouts during their growth. It trickles through the batch of sprouts and as it trickles through it sort of collects the microorganisms that are present in the sprouts. Therefore, making up a quite uniform representative sampling of that batch and this is kind of a unique situation in food production.

So we tested this first looking at the total microbial counts in the batch looking at the difference in total counts between the sprouts and the water and we found that there is a very good correlation. The mean difference being between about half and one log in counts between the sprouts and the spent irrigation water.

We did the same thing on inoculated sprouts, that is that we inoculated with E. coli 0157 and found again that there was a pretty good match between counts of E. coli in the sprouts versus counts in the water.

But the real test of this idea came from using seeds that were naturally contaminated with Salmonella that we had obtained from outbreaks and we were fortunate to be able to get three different lots of seeds that we tested this idea on.

What we found was that every time we found Salmonella in the sprouts we also found it in the water for these naturally contaminated seeds and so this was a real good proof of concept for us that we could, in fact, use that spent irrigation water as an indicator of the contamination status of the sprouts.

Okay. So that answered the question what should be tested. We then tried to determine how and when should testing be done.

As you all know, the guidance has some very specific recommendations on how to test the spent irrigation water and what the FDA microbiologists were charged with at the time was to come up with some kind of test that would be practical to use by the sprouting industry. That is we wanted the testing results to be available within 48 hours of sampling and we also wanted to be able to recommend a test that had been extensively validated.

At that time there were two AOAC official methods for Salmonella . One was the Assurance Gold Enzyme Immunoassay and the other one was the VIP, the Visual Immunoprecipitate Assay. These had undergone the extensive collaborative testing that is prescribed by AOAC International to become an official method and at the time for E. coli 0157:H7 the VIP was the only official method. However, we had had some additional experience within the agency in testing sprouts with the Reveal test and we were pretty confident that that would also be a good test to include in our recommendation.

So the requirements for testing that were in the guidance as a result of all of this consideration was to test each sprouting batch for both Salmonella and E. coli 0157:H7. Because of the methodology restrictions on Salmonella there were a couple more enrichment steps included. For E. coli things were a little bit simpler before getting to that rapid test kit. These were obviously the screening tests that were going to--with results available within 48 hours.

And then the guidance recommended what to do if you got positive results. Well, you had two alternatives and that was first either to discard the sprouts and seed and disinfect the sprouting equipment or you could go through the confirmatory series of tests to prove whether that initial screening result was a true positive or not. The confirmatory tests are quite extensive. They involve isolation of the organism as well as biochemical characterization to prove whether that isolate is, in fact, the pathogen.

A couple of comments that I wanted to make here, although the guidance did not explicitly say so, how to run those confirmatory tests should be done from the original enrichment. I guess there was at least one sprout outbreak that occurred because of this incorrect confirmation. The confirmatory test, I guess, was run not from the original enrichment but from the original sample that had been stored in the refrigerator. This was not explicit but should be made very explicit now.

The other comment that I would like to make is about the disinfection. If you get a positive result how do you disinfect your equipment and how do you validate whether that equipment has been properly disinfected? This was kind of a personal consideration of mine because it happened to me in our own laboratories and I wondered what was the procedure in the sprouting industry when a positive happened. How do you validate your disinfection?

We tested the sensitivity of these rapid test kits for Salmonella and E. coli and found that they were pretty sensitive. Both the sprouts and water could be detected down to an inoculation level of 1 CFU/gram. However, they are not as sensitive to be able to get away with not doing an enrichment and this chart shows that for E. coli 0157 tests, the VIP and Reveal, that the detection limits are pretty high. That is for the VIP you need seven log CFU/ml sprouts or water to be able to detect those organisms. Reveal is a little bit more sensitive. You need five to six log CFU/ml to detect them and so these tests can never be run without doing an enrichment because the pathogens don't--they grow to high levels but not that high during sprouting.

As is shown on this slide and this is the growth of Salmonella during sprouting of the naturally contaminated seeds associated with outbreaks. And this slide actually helps me to make two points. The first point is about the maximum population level that is reached during sprouting at least for these two lots of seeds. In lot A populations of Salmonella got up to maybe 10 MPN per gram and in lot B perhaps between 104th-5th MPN per gram. So again to reiterate the test kits must be used in conjunction with an enrichment because they're just not sensitive enough to be able to detect these levels.

And the second point that I wanted to make from these slides is to show you that the maximum population levels are reached at about the second day of sprouting, that is 48 hours after initiation of sprouting. So that was the rationale for including in the guidance the recommendation that the sample be taken at 48 hours of sprouting because this has been shown not just for the Salmonella , as seen here in the naturally contaminated seeds, but also in other types of microbial populations that we had tested. We saw the same trends for the total microorganisms, for the cola forms and for inoculated E. coli as well.

That said, I will also show you this table which shows the Salmonella rapid test kit results for each day of sprouting of naturally contaminated seeds. We are looking at two seedlots here showing that on day zero seedlot A was detected as positive. That must have been a very hot batch. Lot B--I see lot B was not detected on day zero. However, on day one both seedlots showed positive testing by the EIA so this table would imply that even though the maximum population levels are shown at 48 hours after beginning of sprouting it's possible that by day one even though those populations are not at maximum levels they may be high enough for the test kits to detect the presence of the contaminate after enrichment.

I also wanted to talk a little bit about the validation of rapid tests. Although we were working with AOAC official methods which have, as I said, undergone very extensive collaborative testing, I believe there are a minimum of eight to ten laboratories that simultaneously test a kit in order for AOAC official method recognition as well as a lot of other testing that must be performed, including sensitivity, specificity, false positive rate, false negative rate. It is very, very extensive testing.

Even though we were working with known AOAC official methods we wanted to do our own in-house verification that these methods were tested--were okay to use for sprouts and sprout water because although the AOAC official method allows use for various categories of foods, sprouts may not have been included in that testing. So we wanted to make sure that these tests were okay for use with sprout--for sprouts.

And after consulting with the FDA statisticians who gave us their blessing on the number of reps to use and number of lots to test and number of serotypes of Salmonella to include in the testing, we did a total of four seedlots, a total of 66 replications for each rapid test kit, comparing them to the BAM, which is the FDA reference method for Salmonella . We decided that, in fact, these tests were good to go for including in the guidance.

And this type of testing was also done for E. coli 0157:H7 by Steve Wiegant, who is an FDA colleagues of our's out near Seattle.

I just also wanted to show you some results that we did on rapid test kits that did not wind up in the guidance. The guidance had specified rapid test kits that were all immunoassays. We wondered if there were any nucleic acid based rapid test kits that could also be used and at that time the Gene Trak test for Salmonella was an AOAC performance tested method, which is a little less stringent testing regime than an official method. I believe the performance tested method involves only a minimum of one independent laboratory confirmation of the test as opposed to the eight to ten labs that are required in official method.

Anyway, we looked at the BAX, which is a PCR based method, compared to the Gene Trak, which is a nucleic acid hybridization method, and compared that to the Assurance Gold EIA, which was included in the guidance and which is an immunoassay. These are the detection limits of the three assays and we can see right off the bat that the immunoassay, the Gold, is the least sensitive when it comes to detection limit requiring 106 CFUs for positive result to occur. The Gene Trak was next and requiring about 105 cells and the BAX, which is a PCR based method which amplifies the genetic target, is very sensitive requiring only 100 CFUs for detection. At that time the BAX was neither an official method nor a performance tested method.

So that was pretty impressive to us until--and these were done for Salmonella cells in buffer, and that was pretty impressive until we looked at the performance of these tests in sprouts. In sprouts and water from the naturally contaminated seeds.

And here we saw that the BAX did not perform that well at all and that was largely because its procedure recommended a non-selective enrichment of the organism and so it detected very poorly. Only 13 positives out of 60 samples. The Gene Trak was very good, detecting 36 out of the 60 samples. And the Assurance Gold EIA detected 41. Whereas, the BAM detected 30.

Now these were not inoculated samples. These were naturally contaminated samples so some of them might not have been positive at all and that's why you see the low percentages of positives among all of the tests.

My point here is that we need to consider in considering other tests that could be run, we need to consider the entire method, that is from the enrichment all the way through the rapid test, and whether or not the test has been validated for sprouts.

I've put up here in this table just a few AOAC performance tested methods and I just wanted to show you if you look at the column all the way on your right the matrix for which these various tests have been approved. The Gene Trak Salmonella , which is the one that we included in that last slide, has been performance tested for very general categories for food, animal feed and environmental samples. Whereas, some of these tests have been tested for very specific applications. Look at the Pathogen E. in the very bottom row has been performance tested just for raw ground beef. So my point here is that in considering other tests we need to look at what the matrix has been that was included in the testing.

This is an interesting slide because it is the same slide that Bill Fett put up for chemical interventions but I did the same thing for pathogen detection methods and publications in those areas.

It's interesting that there are a whole lot of publications on different detection methods for sprouts. The question, however, should be not whether the test has been published but whether it has been validated with independent laboratories testing the method and not just a single publication.

So, in summary, in considering what methods to use, we want to ask has the method been tested specifically for sprouts and for sprout irrigation water and has the entire method been tested against a standard method, including the enrichment, the detection and the confirmation. In short, has the method been extensively evaluated or has it been validated?

And that's it. Thank you.

Testing Methodologies and Sampling

DR. FU: Okay. Good afternoon.

In this part of the talk what I will do is to review some of the issues that have been raised since the publication of the last guidance in 1999 and then provide some of the research findings that have been published or has been accumulating in our lab that may help to address some of those issues.

Okay. I will divide the presentation into three areas. First is to look at the issue relating to the sampling and testing during sprouting and then I will also look at the sampling and testing method that has been developed and used for testing seeds and, finally, if time allows, I will look at the detection methodologies and share with you the experience we have in validating and evaluating commercially available methods.

There are several methods and issues raised regarding the sampling and testing of spent irrigation water such as when is the best time to sample, 24 hours versus 48 hours, and how to sample. The guidance suggests sampling one liter per production batch. Some people would like to pool and there are some considerations for mung bean so I will go through them.

Regarding the sampling and testing of spent irrigation water: As I said, people would like to know if they can test the irrigation water at 24 hours instead of 48 hours as suggested by the guidance. Testing earlier gives the sprouters some advantages. For example, they can ship the sample a day earlier while the sprouts are at the freshest and also they can free up some of their storage while putting their sprouts in storage waiting for the testing results.

But is it appropriate to test irrigation water at 24 hours? To answer that question we need to know what level of pathogens are present during sprouting at 24 hours and since 1999 a lot of growth study has been published in the literature and the data provided from the study may help us to answer the question that I just asked.

So what we have done is to collect those published studies and to summarize the data and put out this table. What wee have here is we list the level of pathogens that was measured at the beginning, at 24 hours and then 48 hours, during sprouting of the naturally contaminated seeds, as well as the artificially inoculated seeds among different growth studies.

What we can see here is almost all of the studies shows that pathogens proliferate during sprouting. However, the level of the test population are different. In general, for the naturally contaminated seeds the level were lower and the highest level they reach at 48 hours were lower as opposed to those you find in artificially inoculated seeds.

Okay. The other thing I want to point out here is Lou has pointed out those test kits recommended by the guidance gives you a detection limit about one CFU/gram, which is about log zero--log value is about zero. So you can see most of the studies, except one, will give a positive result using the test kit provided by the guidance either at 24 or at 48 hours.

Now just to give you a better idea of the relative log increase between the 24 hours and 48 hours what we have done is to just further analyze this data and put out this table. What I have here is to list the range of the log increase that was calculated from the previous study that was at 24 hours and 48 hours for the naturally contaminated seeds and artificially inoculated seeds.

And if you will divide the value of log increase at 24 hours by the log increase at 48 hours for each of the studies, what we have found is average ratio obtained is about .85 for the naturally contaminated seeds and about .88 for the artificially inoculated seeds. That is on average 90 percent of increase of population during sprouting occur at the first 24 hours of sprouting. So I would say that if you have a test kit that detection limit is way above the test level you can reach at 48 hours most likely that test kit will be able to pick up the pathogen at 24 hours on average but I want to stress the words "on average" because individually the studies show that the level of pathogen at 24 hours is about maybe--on a log scale it's only less than 60 percent compared to that you will see at 48 hours.

Now the other issue that has been raised is the issue of pooling. In the guidance, as I said, the guidance recommend the sample from each production be tested but we all know that testing requires a lot of cost and so some sprouters are trying to reduce the costs by pooling water from multiple production lots and then take a representative sample from that lot and sending it out for testing. That, in effect, you dilute the concentration of the sample in the sample that you take for analysis and because of this concern the FDA actually advised against sample pooling.

Now let's--an alternative approach that may help to resolve this problem is to allow sample concentration so that you can concentrate all the sample you pool and then test by a single test. So the sample pre-concentration, we have work for several years and I will show you the data in later slides but I want to also mention that sample pre-concentration may also be helpful to address another issue in sampling and testing of spent irrigation water for mung bean sprouts.

It is known that mung bean sprout irrigation water cycle and the procedure they use is different. Typically for mung bean sprout they dump a lot of water during each irrigation cycle. The water can be as much as 100 fold greater than the water that you use during alfalfa sprouting and that creates some concern that the microbial count that you have in the mung bean spent irrigation water may be much lower than those of bean sprouts. Therefore, the testing spent irrigation water may not indicate the positive status of the mung bean sprouts.

So the sample pre-concentration may again help to concentrate the level of microbial population before you go for testing.

So what I would do in the next several slides is to share with you a system--a concentrated system that we have developed in our lab and some validation work that we have done to evaluate the potential for this concentrator to be used as a part of sample preparation for detection of Salmonella and E. coli 0157:H7 in large volume of sprout water.

Okay. This is a picture of the concentrator. What it has is--it has a pre-filter that is about ten micron--with a pour size of ten micron. It also consists of a tangential flow filtration cartridge. The idea of a tangential flow filtration is commonly used in the fermentation industry and I won't have time to go through the details but anyway this process concentrator will allow processing up to ten liters of irrigation water within two hours. It can concentrate the water by 100 fold to 100 mls and those concentrated samples can then be used--analyzed by either culture method, dipstick test, ELISA or even biosensor.

And to validate the efficacy of this concentrator as a part of testing there are two aspects we look at. First is to determine whether the increase in background microflora in the concentrated sample affects the enrichment and subsequent detection of pathogens by conventional and/or commercially available rapid test. We also would like to know the lowest recovery rate of pathogens that can be recovered from the system.

To evaluate the first point what we have done here is to concentrate a large volume of spent irrigation water by 100 fold and aliquot out in 10 ml aliquots. And for each aliquot we add different level of pathogens at about one CFU, 10 and 100. And then we perform 20 replicates for each inoculation level. Again this is, just like Lu said, a blessing from a statistician from FDA. That is their testing recommendation.

So what we have here is for the sample that were inoculated at 10 CFU and 100 CFU the test kit are 100 percent efficient to detect the presence of Salmonella and they were also confirmed by the culture method.

When the samples were inoculated at one CFU the test kit was still very good and picked up the presence of Salmonella . Okay. So that is the test kit was not really affected by the concentration. Here the 10 ml concentrated sample actually correspond to one liter so from here we can say that the test kit, although it will be able with the help of the sample pre-concentration we were able to detect the presence of one CFU or Salmonella in one liter of spent irrigation water.

The same evaluation was done on the irrigation water collected from sprouting of seedlot and a result very similar. That is the Assurance Gold EIA test still performed nicely with the concentrated water.

Now with respect to recovery what we would like to know is if we have one cell in this 10 liter of water will we be able to see it in the concentrated stream. So what we have done here is to inoculate a 10 liter of water with one cell, 10 cells and 100 cells, and then test the presence of Salmonella in the concentrated stream and to see the rate of positive reaction. And what we have shown here is for the sample that inoculated at 10 cells per 10 liter or higher, the system will be able to give you a positive result.

The same experiments were performed also for E. coli 0157:H7. This is a busy slide but what I want to show you is in the guidance the enrichment media is special enrichment media developed by FDA scientists using the modified BPW plus some antibiotics. When we used that enrichment media the positive rate were not as good as we would like and not as good or close as what you see in Salmonella . Okay. So what we have done is to try a different kind of enrichment media that was provided by the manufacturer of the test and we were able to boost the positive rate using that enrichment media.

But while we are we trying to validate the result of a device, what we have observed is sometimes the culture methods fails to give you a positive reaction when the device indicates really positive results. So it has turned out that when you are trying to confirm your culture result the background microflora in sprouts was so high that actually it prevents good identification of your typical colonies. So with the help of immunomagnetic separation we were able to increase the positive rate of the culture method so overall for the Salmonella -- E. coli 0157:H7 we have shown that the test kit will be able to detect at a level of 10 CFU or higher per liter of water.

And this is again the recovery study that I talked to you about with Salmonella but for E. coli 0157:H7 the result was better. We were able to detect the presence of E. coli 0157:H7 one CFU/10 liter.

Okay. Let's move to the seed sampling and testing. As several speakers have mentioned that this is recognized that seeds are the major source of contamination in sprouts. So to prevent the use of contaminated seeds for sprout production is critical to ensure sprout safety and I think most importantly is to protect the home sprout grower who grow their sprouts at home. Normally they will not do any microbial testing or seek treatment so clean seeds to them is really critical.

As far as the sampling and testing method that's out there, what I can find is most of the methods were used for either of two ways. For the outbreak investigation and also for screening seeds that are used for sprout production. What I will cover in this section is the method that has been used for determining the presence of pathogens in seed and also identify some factors that may affect the efficacy of seed testing.

I also will just briefly mention the protocol that have been developed by the industry on screening seeds that are destined for sprout production. Okay.

Inami, et al., which is a group with the California Department of Health, they have done many, many outbreak studies and they actually developed a sampling plan, and it's I guess the only one--a very comprehensive sampling plan that you can find in the literature.

What they did is they take 500 grams of sample from different bags of lots and then they take 100 grams of sample unit from their 100 sample for testing and for this paper that they published what they did is they take actually two 100 gram samples from each sample. Two 100 gram sample units from each sample so they can compare the different testing methodologies. They want to compare whether testing the seeds is better or testing the sprouts that are grown from the seeds is better. And as far as the detection goes they were following the BAM method so I will show their result here.

You are invited to look at their paper to look at their specific protocol but what I want to show there is to compare the positive rate testing the seed versus the positive rate observed by testing the sprout and they have done three seedlots and on average their success rate for testing seeds is about only 90 percent than if they were testing the sprouts.

For us we also did a similar study to look at testing seeds versus testing sprouts and how efficacious they are. What we have is three seedlots and they are associated with the outbreak indicated here.

What we have done instead of testing the seeds or sprout itself, we test the water that was collected from the soak water from the sprouts that has been grown for two days. What we have found out is for some seeds the testing for day zero gives you a pretty good prediction of the status of the sprouts but for some seeds you will miss the positive status of the seeds if you only test the soak water. So for our experience testing sprouts will give you better prediction of the contamination status of the sprouts.

Other factors that may effect the testing efficacy is the sampling size. What I have shown here is for sampling for 100 gram seeds--for some lots the success rate is pretty good but for other lots you hardly will be able to determine the positive status of the contamination.

But if you will increase the sampling from 100 grams to one kilograms, what do our experience--

[Audio difficulties.]

--recognized screening seeds is an important step in multi-approach to prevent the occurrence of sprout associated outbreak and Bob Rust and Bob Sanderson has jointly developed this six step procedure for seed screening. I think Bob has mentioned this in more detail earlier so I won't discuss this more here but what I want to point out is that what they have indicated that is this seed screening has prevented at least one potential outbreak of 0157:H7 and then prevent the shipment of contaminated seeds to sprouters and that further protects the consumer who might be using and might be consuming those sprouts. So it provides additional protections.

Okay. The last part is the detection methodology. As Lou mentioned the validation performed specifically on sprouts is very critical. What I would do here is to provide you two examples from our experience in evaluating the performance of two commercially available methods.

This is an automatic immunomagnetic separation system. I will not go into the detail operation because I don't have time. It's called a Pathatrix System but what I want to point out is it has been AOAC performance certified for detection of E. coli 0157:H7 and Salmonella at a detection limit of 1 CFU/25 gram of sample.

So we tried to use this for our sprout irrigation water using their protocol, enrichment protocol, as certified by AOAC. We found out that the system failed to detect the presence of the O157:H7 at the level of 100 grams per sample and also failed on a concentrated sample. In this case the concentrated sample corresponds to 2.5 liter of unconcentrated sprout water.

Okay. So what we have done also is to put the protocol recommended by the guidance test to see if that would give us a better detection. Yes, this did for unconcentrated sprout water and we were able to detect as low as one cell per 25 gram but this protocol still is not good if we use the concentrated water.

What we also tried is to look at the media provided and in this case we were able to increase the detection substantially and so that we can actually reach one cell per 25 ml. For the tangential flow concentrated sample that is actually one CFU/2.5 liters.

The same thing is done for the Salmonella . A gain the protocol provided by the AOAC approved method really is not as good as they would like to--as they expected from other food. It's only with some modification we were able to reach the detection limit that was approved by the AOAC. So again trying to indicate that the validation specifically done on sprout system is critical.

Biosensor will provide some opportunity for real time pathogen monitoring and this fiber optic biosensor that is commercially available in one format. This format actually would allow flow through type of operation and this is more of sample by sample called RAPTOR.

This biosensor has been shown to be able to detect E. coli 0157:H7 in ground beef at 3-30 CFU/ml levels. It is really, really sensitive without enrichment. And so when I saw that I was really, really excited so I decided to purchase one and test it in sprouts.

More recently variation of that biosensor are commercially available and has been shown to detect the presence of Salmonella in spent irrigation water collected at 67 hours of sprouting from seeds inoculated with 50 CFU/gram of pathogen.

But I want to caution this statement because if you go to the paper what happened is the detection limit of this Salmonella is about 10 CFU/ml but the way it is written here indicates in the 50 CFU/gram will give a false sense that's very sensitive but you want to really put the growth study in mind that at 67 hours post sprouting the level can be very high. Okay.

So anyway what we have done here is to evaluate the efficacy of the Analyte 2000 Fiber-Optic Biosensor with multiple replicates and what we have shown here is, yes, at the level--the low level some tests--some trials will give you positive status but until you reach a very high inoculation level you won't get a consistent positive result.

So again a lot of times when people have a new method they test three and everything is positive and they say, wow, this test is great for something but unless we give you a larger number of trial then you won't be able to really say more confidently what is the efficacy of the test.

So, in conclusion, I want to say there are a lot of information out there since 1999 and some of the information will help as we engage in discussion and try to improve the guidance.

And then sample concentration we have developed that can help to improve the detection if--will provide alternative for the pooling of the sample.

And seed screening does provide additional control to prevent the sprout associated outbreak and again I want to emphasize it will provide protection for consumers who produce sprout at home.

And then it appears that testing sprouts to provide--will provide better detection than testing seeds.

And, finally, I want to emphasize that validation of any method should be done in sprouts so that--because the sprout is so high a lot of times those test kits may fail.

Before I leave I want to thank all these people. Actually we. Lou and me want to thank all these people who helped us to collect all the data that we presented.

Thank you.

[Applause.]

DR. BERU: Thank you, T-J and Mary Lou.

Our next presenter is Kean Ashurst from Caudill Seed Company and the title of his presentation is Alternative Seed Sanitation Methods and the Results of Practical Field Application.

Alternative Seed Sanitation Methods and the Results of Practical Field Application

[See presentation slides for Mr. Ashurst]

MR. ASHURST: Thank you. Hello. My name is Kean and good afternoon.

I have been preceded by some very excellent presentations today and I'm proud to be here as a spokesman for this industry and for Caudill Seed.

Caudill Seed Company is dedicated to making sprout eating a safe and nutritious pursuit for all in the industry. CSC provides seed and sprouting equipment to sprout growers and is attempting to positively impact the sprout industry in several ways, both directly through the seed that it sells and indirectly through recommendations and technology development assistance to sprout growers. We want to provide a science-based approach for alternate solutions to all segments of the sprouting industry.

Review methods or the purpose today is review methods tested to improve seed cleanliness and recommend one path; review sprouting procedures and sanitation; and identify risk points and make recommendations; support the use of existing sanitation methods, chlorine, and propose an alternate that may be as or more effective than chlorine; alternate seed sanitation methods and results of field applications.

The sprouting industry seed must be considered a raw agricultural product and must be dealt with accordingly. I cannot over stress that. The 10 to 12 years that I spent with Libby, Nestle and Hunt-Wesson, we washed our pickles, we washed our tomatoes, we washed all those vegetables before we processed them. Okay. You have to make that assumption. All right.

There are other sources of contamination to be considered and controlled for a total quality management of sprout production. Seed sanitation includes not only sanitizers but also techniques and I think what we're going to see as we go through some of our sprouting facilities we're going to see Company A doing this, Company B doing that, Company C doing this and so on and so forth, and that is not standardization. I'm not saying that what A is doing is incorrect but we need to have a total agreement and move forward on a level playing field and everybody is playing the same ball game.

Other sources of contamination: It is essential to recognize that not all risks come from the seed and other areas of contamination need to be addressed. Proper application of sanitizing protocol. Equipment cleaning and sanitation.

Some of you have seen this slide as long as two or three months ago. I want to take some brief time and go through it because I think it's an important slide. It's not all encompassing but it represents my visits to sprout growers, not all but most.

I want to start out with seed contamination. Really no firm regulations on it.

Field growing conditions, we need to have firm regulations.

Procedures for cleaning seed from the field impacted.

Storage and transportation impacted.

Let's go to sprout production. Air quality in the growing chamber. Water quality for irrigation. Again no hard fast regulations in this area. Harvesting procedures. Seed not washed and sanitized. Completely avoiding the guidances. Meetings like today need to be more frequent. Industry and regulatory cooperation. Washing, packaging and shipping.

This is one area, I guess, that I get really excited about when I see water temperature at 70-80 degrees. It should be below 40. Pour or no sanitizer used in the washing system. No real HACCP program in the facility. Temperature abuse of the product in shipping. I have seen a number of sprouters sell product on vans, 110 degrees and the van has got a two hour drive to its destination. Employee practices, again no regulations. No standards of identity. No QA/QC programs. I'm not saying this applies to everybody. There are good sprouters and there are marginal ones.

The other area is facility conditions. About three weeks ago I had an opportunity to go into a sprout grower in an unknown state, unknown name, entered the facility and walked in and within 50 feet I counted 18 critical violations and I didn't move. Equipment condition and design. Product and employee traffic. Cross traffic. Seed coming in, sprouts going out, and they're going just like this. Sign wave sanitation, and everybody in this room knows what it is. You're going to have an inspection tomorrow, let's get it right. Construction of the facility. Okay.

Working--a seed sanitation program. Working with Dr. Kathleen Rajkowski since the summer of 2004, we have collectively developed an evaluation program of current and available processes for reducing seed contamination. Caudill Seed provided the seed and the USDA would provide controlled inoculated seed packets for testing. The initial findings were presented in October 2004 meeting with the FDA.

Evaluation program. We revisited irradiation, E-beam, heat treatment of seed and ozone gas. Irradiation and E-beam technology resulted in poor germination and in some cases as low as 40 percent. Unacceptable endproduct quality. The sprouts were rubbery. They were not crisp in texture. Consumer acceptance issues with regards to irradiation and they ignored the needs of a big segment of the industry, the organic processors.

Heat treatment rapidly aged the seeds and temperatures between 145 and 147 can only be achieved consistently in a fluid bed dryer.

Ozone gas treatment showed some promising results and did not affect the overall flavor or texture of the sprouts but it failed as a standalone process. Ozone gas treatment was able to achieve significant reduction. One log kill with up to two log injury in controlled humidity conditions.

Dr. Rajkowski has reviewed these slides with you.

Tsunami 100 was tested in combination with the ozone gas treatments and the results were significant. You've seen these slides earlier.

We're indicating here that we're getting anywhere from a minimum of four and as high as six log reduction.

When was the last time, and this is going to be a sore subject for some of you--when was the last time you got 50 pounds of clothes cleaned stuffed in a nylon bag and submerged for 25 minutes in water with an astringent detergent dip? Well, it doesn't work on seeds either. I'm sure that the above was not the intent of the guidances.

Effective washing implementation. Effective agitation of the seed is required for adequate wash. We're using a 250 CFM blower here, an oil-less unit that is used in the food industry for transporting powdered sugar. Water temperature is ambient.

I want you to take a look at the slide on the left. It is hard to document this but what we're trying to achieve is an internal rolling action in that nylon bag and we succeeded in doing that. You have to follow the procedure though. You can't overload the bag. You've got to place the bag in the right part of the washing unit. And if you look to the right this represents only a small portion of the debris and the material that came off of this seed. And I say a very small portion. This is with Tsunami 100.

One sprout processor is currently using gaseous ozone treated seed in conjunction with Tsunami 100 with this washing system on a test basis. Dr. Rajkowski will be receiving and evaluating a full scale up system in 14 days at the research facility in Wyndmoor, Pennsylvania.

Caudill Seed will seek an expansion of this program. We will be requesting FDA to consider usage of Tsunami 100 at three percent application. Caudill Seed will be recommending a science-based continuous improvement program on seed washing and sanitation.

To combat other sources of contamination we have made available automated CIP systems for growing chambers. This was tied in with a touch screen unit. As soon as you finish your growing process and you wash it out you hit the wash button and put the doors back on and hit the wash button and the unit is sanitized internally. And that has been a result of some modifications that we've made to the growing chambers.

In addition to that we've made available a point of use water purification system for each growing chamber. This additional feature incorporates a 10 micron spiral wound filter and a .5 micron carbon block with UV light at 254 nanometers.

I have been through a number of facilities. I haven't seen the back flow preventers. I have seen a hose attached to the water line and they're using the hose to flush out the drain. What kind of problems do you think that we're going to have?

The next one is probably very important. It impacts the organic industry and it helps the industry overall. Air intake systems for the growing chamber are equipped with dual wave length UV lights that react and create hydroxyl radicals. These radicals serve to sanitize the air and increase phytochemical production and enhance the immune system of the emerging sprout.

To explain this in detail I've got a lot of data on this information that's true and factual. We can do it in the Q&A program and I can also give you a slide. It would take about four hours if we discussed it today.

There are some of the original references, J.R. Vig, U.S. Army Electronics Laboratory, Institute of Molecular Plant Sciences in the Netherlands, USDA out of New Orleans, and the Department of Plant Pathology at the University of California, Davis.

It is our goal at Caudill Seed to provide the industry with the best available technology and to assist the processor with their food safety needs. In addition, there are numerous valid ideas brought forth by other responsible parties in this forum that need serious consideration.

That's the end of the show. Thank you very much.

DR. BERU: Thank you, Mr. Ashurst.

Our last speaker in this session, this part of the public meeting, is Dr. Schaffner from Rutgers, and his presentation is on Risk Analysis.

Risk Analysis

[See presentation slides for Dr. Schaffner]

DR. SCHAFFNER: Thanks very much to my colleagues at the FDA for the invitation to be here today.

Being the last formal speaker on the program it is always tough. One possibility might be that there is nothing left to say. Another possibility might be that there is no one here to listen and another possibility might be that you're all just already so overloaded with information there's nothing else I can possibly pack into your brains.

Well, I'm going to, hopefully, be able to stand up to that challenge and the opportunity for me really again as the last formal speaker is to try to integrate some of the information that you've heard already or at least present a framework for how that information can be integrated. I am told to do that is risk assessment.

Why are we interested in sprouts at all in my laboratory? Well, we had an initial discussion with Bob Sanderson and Bob was very passionate about his belief in using irrigation water testing. And I don't know how many of you have ever heard the expression "if you give a small boy a hammer he will find that everything he encounters requires pounding." Well, imagine me as that small boy and instead of a hammer I've been given the tool of Monte Carlo simulation. Whenever I'm presented with a microbiological problem, my initial--first initial reaction is, "Well, let's use Monte Carlo simulation to look at this." However, in this particular case, all kidding aside, I think it's a tool that's particularly suited for looking at the trade offs and the complementation of some of these different issues.

I also want to acknowledge my collaborator in all this, Rebecca Montville. Rebecca is a scientist that is currently working at Yale University and she did a lot of the grunt work to get this work done and also contributed a lot of the ideas.

If you'd like a copy of my presentation I would be happy to provide it to you. These are two of--our two publications on this particular topic. They're also both available on my website. If you just Google my last name in conjunction with Rutgers University and food science I'm sure it will pop right up. That's probably easier than giving you the URL to our web page. But again most of what I'm going to say here, at least the meat of what I'm going to say is contained in these two publications so you can read those at your leisure.

The first--I should say the first publication--our first objective before we looked at building a Monte Carlo simulation of the entire process was to look at an analysis of the published studies on seed treatments. And you've already heard from a number of my colleagues about the research that has been done in this area. So what we tried to do was to look at how much had been done, to look in particular at the sorts of treatments, and then try to compile them into some sort of sensible quantitative framework.

This is one of our slides from that or one of our figures from that publication. What you can see is that if you look at 20,000 parts per million chlorine the log reductions are all over the map. The frequency, which is the Y axis there, represents the number of times we were able to see a particular log reduction from a particular study so we looked at all of these studies. We broke down the tables and figures and tried to tease this data out of all of them.

What you can see is that this highly recommended or at least it's in people's minds this is a highly recommended procedure, this procedure gives, if you look at these studies in the literature, gives most commonly about a 2.5 log reduction. In some cases in some people's hands it can give upwards of a five log reduction and even greater than a five log reduction if you look at that Lang 2000 study.

Now, I should mention that those authors listed there are the first authors on those studies. It's not the entire author list. So, for example, the Taormina 1999 study Larry Beuchat is a co-author on that study as well. I just had to be able to quickly reference them and fit them all on this slide.

So what you can see is that there's a great deal of variation in terms of the effectiveness of the 20,000 parts per million chlorine disinfection. Now that's not a problem to a risk assessor or to a modeler because what we can do is to say, well, okay, these data represent the true variability of the system so we'll take that into account when we do our calculations. I'll give you a little bit more on that later but before we leave this topic of log reduction I want to just share this figure with you.

If you take those studies that I just presented and you look at the relationship between the log reduction that they reported and the inoculum size, that is the concentration of organisms that they put on the seeds to begin with, you can see that there's a bit of a relationship. Okay. And if you look at that Lang 2000 study which showed the greatest log reduction, what they did was they took seeds, they immersed them in an aqueous environment that had a lot of pathogens and they vacuum dried that system to boost the concentration.

Now when I look at data like this and I see a correlation it makes me immediately suspicious. What these data appear to show is that--and this is, you know, not really that spectacular a finding, it makes sense. The more organisms you put on the seeds, the more you can take off, the greater log reduction that you can document. What this analysis says to me is that we need standardized methods across laboratories for the way we inoculate seeds and I think I heard at least two other presenters this morning say that same thing.

So it's not just enough that scientist Jones publishes a study that shows a certain log reduction but it needs to be by methods that we in the scientific community and in the sprouting community agree are the appropriate methods for demonstrating the effectiveness of these technologies so that's a really important take home message.

Now if you take--if you look at this slide, what I did here was I've got all this data but then I've sort of summarized that data with that bright yellow line and that's what risk assessors would call a triangular distribution. That shows that you see at least a one log reduction, most commonly a 2.5 log reduction, and at most about a 6.5 log reduction.

Another figure from our first study, the one that was published in JFP in 2004, is this particular one. What this figure does is it relates that triangular distribution from all the different datasets that we analyzed to give us a quick visual way to compare the effectiveness of 20,000 parts per million calcium hypochlorite, 2,000 ppm calcium hypochlorite, 2,000 ppm sodium hypochlorite, and 200 ppm sodium hypochlorite. Now what you can see is that there's not a heck of a lot of difference when you go from 2,000 ppm to 20,000 ppm, at least if you're looking at calcium hypochlorite. In fact, the mode, that is the peak of the triangular distribution, is actually higher based on our analysis of the literature data for 2,000 ppm calcium hypochlorite.

Now granted the maximum you see is less and the minimum that you see is less but the most common number, the one that comes up in more studies than any other, is actually higher so we need to be really careful about recommending one particular log reduction over the other. Again you can use a risk assessment framework, a Monte Carlo simulation model to compare the effectiveness of these various disinfection technologies.

This slide represents a screen snapshot from a piece of software that we developed. This computer model was developed using a tool called Analytica. Analytica is basically--it's like an Excel spreadsheet except that it represents the data visually instead of in rows and columns. It's a tool that we have used in my lab and my colleagues at other universities and federal agencies and international groups have used to do some of these microbial risk assessments.

And what--and again I don't have time to go into the details of it but this particular computer model is available on my website. I have made some changes as I've been sitting here in the audience and listening to people and getting inspired. This represents a relatively recent picture and once I get back to Rutgers I'll be posting a new version on my website.

If you want--the great thing about this particular piece of software is it allows you to hide complexity. So if this were a live version and I were to click on that button in the middle that says "model" that particular node would explode and you could look at all the different factors that go in and the ways that I've combined them mathematically. It uses a technique called--an approach known as "influence diagram" so you can see which inputs influence which outputs.

And you can see on this particular screen snapshot that a lot of the information that my colleagues have presented to you today is contained here. We have the sample size. We have the batch size. We have the detection limit that you heard so eloquently discussed just a few minutes ago, the number of samples taken, the prevalence and the concentration of the pathogen when it is present on the seeds. The computer simulation also takes into account if you want to use a particular type of disinfection and what the range of effectiveness of that disinfection technology is. Also if you want to do sampling of seeds prior to using them for sprouting or whether you want to evaluate irrigation water during the sprouting process.

So what this piece of software is, is a mathematical representation to the best of our ability of the information that's out there in the scientific literature. Since the scientific literature is not static this shouldn't be static either and so it can be updated and changed as the science changes.

I should mention that it was kind of tailored towards alfalfa because there's a lot of the published literature on alfalfa but if you knew the particulars of a--it's Salmonella on alfalfa in particular. If you knew the particulars of another type of seed or another type of pathogen you could customize this to aid you, and again think of it not as the magic solution to all of your problems but as an aid to guiding your thought process in evaluating one approach over another or the interaction of different approaches.

So the way the computer model would work again if this was live I could just click on any of those "Calc" buttons and it would calculate based on Monte Carlo simulation the estimated fraction of contaminated batches, for example, given the assumptions that are input into the computer model.

In the interest of time I'm not going to go into a great deal of detail on this particular table. There is a more complete table that's in the published study that's available on my website but I really just want to focus on the three points that are on the right-hand side of the screen there. This actually echoes something that Bob Rust said earlier in the day.

If the prevalence of the pathogen is high, which means that when you go and you pull a sample the odds of finding it are high, then using testing is good. It's going to give you the most bang for your buck.

If the prevalence is low then disinfection is going to be more effective and that's why an approach of testing, accepting, rejecting, followed by disinfection is really a great combination and it gives you again the most bang for your buck in terms of trying to reduce the risk.

Now another key point here is the assumptions that you make matter and again I don't have time to go into all the assumptions that underlie this particular table. Okay. You can look at the article that we published in Applied and Environmental Micro in 2005 to see some of those assumptions but if you change the assumptions--in other words if you change assumptions about prevalence, if you change the assumptions about the concentration of the pathogen when it is present, if you change your assumptions about the detection limit for the pathogen of concern, these numbers may change.

The good news is that a tool like the one that we've proposed can help you quickly evaluate changes in those assumptions without going back to the laboratory, without doing--spending another

$200,000.00 and two years of your life to come back with the answers. Again the answers from the computer model may not be the gospel truth but they are a guide to future studies and a guide to decision making.

Another key point, the number of samples you take matters, and this just shows the relationship between the percentage of the time that you're able to detect the presence of pathogens and the number of sub-samples you take. Again Bob Rust said this much more eloquently than I can say and I just wanted to echo some of those thoughts but remember that the particular shape of this curve and the particular size of those columns depend upon the assumptions you make about the prevalence of the pathogens and also the number of computer simulations you run.

If you're dealing with very, very low prevalence on the order of one out of every 1,000 25 gram samples being positive for the organism and you only run the computer simulation 1,000 times then the chances of actually finding that pathogen in all of those computer iterations is quite low so you need to be very aware of those particular sensitivities.

Okay. So what are some lessons learned at least from the research that we've published. Seed disinfection is a highly variable process even in the laboratory and you heard a number of other people allude to this. If it's highly variable in the laboratory or between laboratories, what happens when we take it out into the actual production environment? Well, it's certainly not going to be less variable. So seed disinfection is a highly variable process even in the laboratory.

Irrigation water sampling can be highly effective. Again I wanted to call your attention to something that Dr. Tortorello said is that conventional wisdom in food microbiology is that testing doesn't work. Well, the good news about sprouts is that they grow pathogens like gang busters if they happen to be there and you have a wonderful environment that is the environment of the irrigation water from which to take samples. So in the particular case of sprouts testing actually can be effective and again it will take a long time before some food microbiologists accept this concept but testing really does work, I think, quite well in the particular case of sprouts.

But to do that if the prevalence is very low you have to test many samples, and by many I mean hundreds of samples or sub-samples prior to using that seed in production. You also again have a wonderful opportunity during the production of sprouts to sample the irrigation water and to test that irrigation water. And again you heard from a number of my colleagues today about how important that is.

And then, finally, remember that disinfection and sampling at least based upon the research that we've done and the data that I've presented here, disinfection in sampling are complementary tools.

I just want to leave you with a couple of my favorite quotes. These are from George Cox and from John Tukey, a couple of statisticians.

George Cox said, "All models are wrong so don't believe anything a model tells you but realize that some computer models can be useful." Hopefully, today we've shared some of our models that are useful to you.

And then another quote from John Tukey who said that, "An approximate answer..." and these risk assessment models only give approximate answers "...to the right problem is worth a good deal more than an exact answer to an approximate problem." So again, hopefully, these approximate answers that we've given will be useful and shed some light on the subject at hand.

And for those of you who would rather process pictures I want to leave you with this slide from Gary Larson which has been amended slightly. The Sprouts logo didn't appear in the original cartoon. These are clearly microbial risk assessors who are studying sprouts because they've got to be microbiologists, you can see there is the prerequisite microscope. I'm sure they're modelers and statisticians because they've got a blackboard full of equations and finally they must really love sprouts because they're all excited now that the Sprout truck is here.

Thank you all for your attention.

Questions and Answers

DR. BERU: Thank you, Dr. Schaffner.

I would like now to invite our speakers to the table for our question and answer session.

I would like to encourage questions not only from the audience but, panel members, feel free to ask each other questions as well. I will take the moderator's prerogative and ask the first question of Dr. Fett.

Of the methods--you did an extensive review of the various treatment methods and I wanted to ask you in terms of applicability to the widest varieties of seeds in terms of efficacy, which one holds the most promise do you think?

DR. FETT: Well, I don't know if I have a good answer for that. The problem is that most of the methods have been looked at only for alfalfa seed. There has been very little work on other types of seed. I would say that as I discussed in the presentation that I think we do have some viable alternatives for mung bean that look very promising but the work needs to be extended a bit. There could be cultivar differences, various differences in methodologies that could be used that need to be examined but I think with mung bean there are several potential alternatives that look very good.

I think with alfalfa seed I think it's a different story. I don't see too much out there right now that has really--that has been published I should say--that is exciting and that I would consider a viable alternative to the 20,000 ppms or chlorine right now. There could be work that will be published soon but that looks more promising for alfalfa seed but I think alfalfa is a real problem and I think that relates to the nature of the seed coat and the cracks and the wrinkles and everywhere that the pathogens can hide. I don't think--I don't see one method right now that is too exciting for all the different seed types.

With chlorine we do know that there are some problems with effects on germination and yield for some of the seed types so it's not universally acceptable either so I think there's not one universal method that can be really applied for all seed types right now.

DR. BERU: Thank you.

Go ahead, please.

DR. WARRINER: We always talk about a five log reduction in testing efficacy. I've got doubts about that for sprout seed studies for the obvious reason. If you have a low number of survivors in 20 or 38 hours time you can have 103 again so I just wanted to know Professor Fett's thoughts on do you think that needs to be revised as in the log five criteria?

DR. FETT: Well, I don't know if it needs to be revised. I mean the rationale for that was in the back of the white paper, I believe, in '99 that was published by the committee. I think what we're ideally looking for is elimination. I would agree with you there. Definitely elimination is preferable to showing a five log reduction because, as you just said, if we have some survivors there there's a good potential for them to grow out.

One thing that I didn't talk about that I'm really interested in is trying to combine chemical and biological methods where we would treat the seed with a chemical method to knock down the pathogens as low as we can get them, probably not eliminate them, and then come in with a competitive exclusion product which would not allow whatever survivors are there from the chemical or physical treatment to grow out to high enough numbers to cause problems.

I think that's a possibility for the future. It's not something that's going to happen very soon unfortunately but I think it's an exciting possibility because I think to totally eliminate pathogens from this whole gamut of seed types is probably unrealistic. Eliminating the pathogens and keeping the seed viable and vigorous and getting a good yield so I think elimination is ideal but I'm not sure we're going to get there but elimination would be a lot better than a five log reduction.

MR. WARRINER: I agree.

DR. BERU: Is the speaker on? For the benefit of the transcriber, please state your name.

MR. RUST: Now it looks like it's on.

DR. BERU: Yes.

MR. RUST: The reason that we are all here is for risk reduction in sprouts and one of the things to increase the risk reduction or the risk that there are pathogens in sprouts is to increase compliance among sprout growers. And I heard some really good things here today and a lot of it came from questions that people from industry sent to the FDA and the FDA responded, and I think very well.

One of them is talking about a 24 hour sample or testing after 24 hours instead of 48 hours, and that does wonders for the sprout grower because of their hold and release. What can happen in 24 hours is you test it for 24 hours and then you have a presumptive positive and then you have to still have that--at any rate the quicker that it can come back the easier it is to comply with because that is quite a complication when you get a false positive which does happen and then it may or may not be out there and then you have to make a decision of whether you want to contact your customers and tell them that you may or may not have bad product. At any rate, the quicker that you can get those samples back the better.

Also, the concentrator, that would be fantastic. That would save sprout growers a tremendous amount of money.

Don Schaffner was talking about the 2,000 parts per million. That would make a world of difference in compliance because right now 20,000 parts per million there's a lot of sprouts that won't even grow under 20,000 parts per million and then so sprout growers, "Well, this whole thing is just--doesn't really work." And so what is real?

Anyhow--and of course I like seed sampling or I wouldn't have brought it up but I think those four points would make a tremendous difference in the sprouting industry and also I appreciate all the help that you folks have done to do this and I think the industry does too.

DR. BERU: Go ahead, please.

MS. SANDERSON: I have a question for you, T-J. When you use the concentrator and then you put up all those tests that you ran for trying to detect pathogens, I had some trouble understanding all your charts but I got the feeling that the quick tests were pretty good at finding it but the confirmation tests might not have been quite as good. I was concerned if we were to use a concentrator and we got a presumptive positive that when we went to a confirmation test we might get a false negative.

DR. FU: That's why I wanted to bring up the improved confirmation tests with the immunomagnetic separation. With usually the culture method after enrichment you went straight to a selected place and with so high--the background in sprout water is so high that when you streak it there's a lot of things growing on the selected plate so your ability to identify the typical colony is affected by the background. But if you would be able to use the immunomagnetic separation to capture your pathogen before you streak on a plate and that helped improve the confirmative test. So that's why I wanted to point it out that immunomagnetic separation is important for the culture method.

DR. BERU: Yes, Mr. Sanderson?

MR. SANDERSON: This question is also for T-J. It's very good to acknowledge the problems of bean sprout growers in doing spent irrigation sampling and you mentioned something like 100:1 difference in water use maybe. Now did you also do counts? Okay. I point that out because I have been doing just aerobic counts and coliform counts and they're consistently higher by at least a log on bean sprouts from alfalfa sprouts after 48 hours and the bean sprouts are grown at about 90 degrees whereas our alfalfa sprouts are grown at 70. This is Fahrenheit. So I'm mentioning that because in looking around at ways to do this like tangential flow concentration it seems to me there's a huge difference in 10:1 reductions from 100:1 in terms of costs and kind of availability for sprouters.

DR. FU: Right. I want to reiterate that, yes, if you grow the sprout at a higher temperature, say 30 degree versus 20 degree, you do see a log increase in the background. But we didn't see--if we used immunomagnetic separation that you can clear up the background and we didn't see too much problem in terms of detection. Does that answer your question?

MR. SANDERSON: It may be more important if you understood mine because there's a lot of things that I may not understand really clearly but just that the levels--I'm not sure you would get a one percent organism count in bean sprout water after 48 hours. You might get ten percent of the count for the green sprouts after the same length of growing. In which case you might not need as much concentration and the concentrator that you've shown is--for me as a person in business--a very complicated thing and it's expensive. Okay.

DR. FU: Okay. I think now I understand your question more. You are saying if you grow the sprout, mung bean sprout, at a higher temperature, the rate is higher and so the dilution you see with water being irrigated on mung bean may not be too much a problem because the level was higher. That's good but we haven't done really experiment on mung bean yet.

DR. BERU: I had a question for Dr. Schaffner on the issue of 2,000 for this 20,000 ppm calcium hypochlorite. You do show the distribution of high and low reductions in the 2,000 but you also I think said in your presentation on the need to standardize how these measurements are done. In view of that, what can one really say about those distributions?

DR. SCHAFFNER: Right. That's a really good question and you've got to remember, too, that the selection of the people who studied 20,000 probably overlapped with the people that studied 2,000 but it might be from different laboratories. So you're absolutely right that you have to be very, very careful about making those comparisons. My only point in showing those two particular distributions is that people somehow seem to have fixated on 20,000 parts per million chlorine and I just wanted to sort of try to open the door again to alternatives to 20,000 and to say, hey, maybe 2,000 is almost as good to at least plant that seed in your minds but not to say that definitively. Again maybe to encourage my colleagues who are looking at disinfection techniques to really sit down and look at this rigorously to see if maybe 2,000 might not be almost just as good.

MR. LALLEY: I have got just a question for anybody in the audience or the panel. Ms. Tortorello referred to the uniformity throughout the crop of the pathogen in the event of its existence in a crop. I would have a question which is if someone could explain to me how a 50 pound bag of seed which produces perhaps eight to ten to one pounds of sprouts is giving us an equivalent of 1,600 packages per drum or 10,000 servings per commercial drum, how, in fact, such small numbers of people can be affected in the outbreak and in such small geographical areas as some of the outbreaks which have been identified or supposedly identified?

I would further like to comment on the elimination versus germination. Elimination is the key. Germination is a secondary issue. In my mind if I get 50 percent germination and I'm assured of elimination of the pathogen, I couldn't care less what it cost me for that seed as far as additional costs to eliminate the pathogen.

And one final question is the inoculation, the artificial inoculation, of seed for scientific testing. It is very similar in my mind to taking a hypodermic needle and infecting cows thoroughly with E. coli and/or Salmonella and then pointing to the cow and saying, "Kill the Salmonella and E. coli but don't touch the cow."

DR. BERU: Did anyone--yes, please.

DR. TORTORELLO: I'd like to try answering the first part. That is the relationship between uniformity of contamination in a batch and the number of people who become ill. Is that--was that your correlation?

MR. LALLEY: [Not using microphone.]

DR. TORTORELLO: Okay. Well, I guess I'll try answering that with the consideration that illness results because of at least two factors. One, it's a quantitative issue. That is you have to--and perhaps Dr. Dechet from the CDC can also chime in here. It depends on the number of organisms that are ingested and it also depends on who ingests them. So basically it's--you may have uniform contamination throughout the batch but not necessarily uniform numbers throughout the batch so a person who gets part of that batch may not actually be ingesting the same amount of viable microorganisms as another person and that person also may have a higher resistance or better able to withstand a pathogen being consumed.

Dr. Dechet, would you like to say something?

DR. DECHET: Yes, it is actually a very common question that comes up that if there's so much product out there how is it that so few people are getting sick. We often refer to the disease pyramid. If you think about like the part of the--at the top of the pyramid is the number of people reported to the CDC of getting ill. But if you think the bottom, the base of the pyramid, first you have all the people who eat the product, then you have those who actually let's say get diarrhea, and if you think about when you've gotten diarrhea how many of you have actually gone to the doctor. Probably not that many. Most of people sort of ride it out at home. Then you have the group that actually goes to the doctor and how many of those actually get cultured. As a physician myself, I know that very rarely do physicians choose to culture diarrhea because usually by the time the results get back the patient is better. It's really the most severe cases that get cultured. And then of those that get cultured, how many actually make it to the lab, get confirmatory testing and so forth. How many of those then get reported and how many--if at the CDC we get every tenth Salmonella sample, how many actually make the CDC? So there's actually a multiplier effect down the road where we may only being seeing six. We expect that the number of people that actually got sick from a product is really much higher.

MR. LALLEY: How many--

DR. BERU: Excuse me. Please get to the microphone because this needs to be captured by the transcriber.

DR. DECHET: I'll just repeat the question. I believe what it is what is about the geographical distribution? It's a good question. I think, you know, some of the points that were just mentioned here are ones that might address some of those issues. You also have to realize that the degree of reporting behaviors are different from state to state and so again while some states DNA fingerprint every sample that comes in, some only do every tenth, so that may have something to do with it. I also am not that familiar, and probably someone in the industry can help me with this, about are they sent to different areas or the sort of concentration that some areas tend to get seed from one place and not from another place. I don't know but also I think the behaviors of the sprouters are probably quite different from area to area and I would welcome other people to comment on that.

DR. BERU: Any other? Yes?

DR. SCHAFFNER: Just a follow up on that. Another explanation that I think we need to carefully think about is that it may only be a seed or two in a batch and we are doing some research that hasn't been published yet but that shows that the organism actually from a single seed or a very small handful of seeds will spread rather slowly across a tray during the four day irrigation so it may well be that in this entire growing chamber full of sprouts not all of them are positive for the pathogen. So it may well be that these outbreaks, the smaller outbreaks, are isolated because there's not that much product implicated.

And then just a quick reaction if I could to the concept of elimination. I think one of the things that we need to think about very carefully is that elimination of a pathogen--you can't talk about elimination of a pathogen until you know the starting number. If I have 104 and I have a five log reduction there's a probability that I'll have pathogens left at the end but if I start with 106 and I have a five log reduction I'll have one log of organisms there so I think it's very dangerous to start talking about elimination. We really need to be focused on a log reduction.

Now I'm not here to say whether five log reduction is right or wrong but we need to move away from the concept of elimination because elimination assumes a starting concentration. If we don't make that explicit we end up chasing our tails for a long, long time.

DR. BERU: Dr. Smith has a question or comment.

DR. SMITH: Along the five log reduction line, as Dr. Fett said, NACMCF went through a calculation and came up with five log. I think they've explained that in the white paper. Five log is a pretty common performance standard. We used it in HACCP. FDA did not explicitly adopt five log in our guidance for a number of reasons. One was that we, ourselves, had not evaluated the adequacy. If there are six log pathogens present and you only get rid of five log you've still got a problem. And the other thing is before we were to require five log we'd like to have a way to get there.

So we're between a rock and a hard spot and as this process goes forward I can almost guarantee that the people working on the solution or the regulation or whatever it ends up being will be asked to come up with a performance standard, a description of the endpoint that needs to be achieved. Whenever we can we describe that endpoint and then give as much flexibility as possible on how to get there so I want all the great minds in the room together to start thinking about that and put it in your comments.

Also, Dr. Fett, you mentioned internalization. This is an issue certainly with a lot of fresh produce items where we know if you put a warm tomato in cold water a suction effect will pull in water and pathogens if they're present. I saw slides of pathogens getting in the root hairs of sprouts during the production process during sprouting but I'm starting to hear more about the potential internalization of pathogens in sound, intact seed, not scarified and not damaged. So if anybody here can comment on their opinion, the likelihood of that happening.

DR. FETT: Well, I can just comment that seeds normally aren't sterile. You can get internalization of native bacteria into seeds and also plant pathogenic bacteria can be in the seed coat. Unfortunately, as far as I know with bacterial human pathogens on sprout seed, and somebody correct me if I'm wrong, I don't think we have any good data on where exactly they're located. If some are internal and some are external or they're all external, I don't think we have that information.

But I think it's very clear from several lab studies that the potential for bacterial human pathogens entering through the root system primarily where the secondary root hairs emerge where you have natural wounding occurring is a distinct possibility and I think the potential is there under commercial conditions that that could happen.

DR. BERU: This gentleman had a question. You may need to come down here.

MR. MEYEROWITZ: I'm Steve Meyerowitz from Sproutman Publications and I guess I'm here representing the home sprouter, the kitchen gardener as it were, of which there is a large contingent. I have a comment and a question. My comment is that if we were to make our way down to 2,000 parts per million of chlorine, the public--it would solve a lot of public relations problems for the industry because right now there is a perception that the 20,000 parts per million--this is a public perception--is overkill as it was and it actually discredits any other guidance that the FDA might come out with because of this kind of heavy handedness so that's the perception.

My question is for you, Dr. Fett. When you talked about whole microbial communities that really interested me because I really think that it is at the--the competitive microbial community is very interesting because essentially that's what's going on inside our systems is that we have a competitive microbial community which is part of what keeps our immune systems functioning and protects us from most of the pathogens--low levels of pathogens that are commonly found in food--in all foods. And, in fact, in our whole environment--I think--don't you think--I didn't even think about that after I washed my hands in the rest room and I had to walk out and you have to touch the door that hundreds of other people touch. So we really need that internal friendly microbial community helping us.

So my question to you about this competitive microbial community is how can we get more research on the schedule in the queue as it was and is there any downside to using competitive microbes here? Can it cause other problems?

DR. FETT: As far as more research, talk to your congressman. Have him send some money to the FDA and the USDA for that kind of work. I mean we're going to--we're in the process of writing up our next five year project plan right now in ARS for all of our food safety related projects and in my group, which is one of really probably the--there's three produce safety related groups in the U.S. and ARS. One is my group in Wyndmoor. There's one in Beltsville and there's one in California.

As far as interventions go, it's primarily our group and some work in Beltsville.

As far as competitive exclusion goes it's our group in Wyndmoor and it's essentially one scientist doing the work and it's a very complicated issue. We have been looking at single strains of fluorescent pseudomonades as well as whole communities and a little work on mixtures and biocontrol is a little bit frustrating in that it seems to work well and quite often sometimes but then you have an experiment or two where you have a failure. I'm not exactly sure why that is.

Biocontrol has been worked on for years and years and years and years in the plant pathology field. Since the 1920s there has been a lot of work on that. There continues to be a lot of work on that and there's really not too many products out there. That's why I'm saying it's a long term research and development program.

And if you're talking about one scientist being responsible for doing that work it's a really long-term project so, you know, the more emphasis, I think, on that the better but right now we really only have one scientist working in that area.

It's interesting to me why Salmonella and E. coli can compete so well with the native microflora to begin with during the sprouting process. I wouldn't have expected that really to tell you the truth but there is evidence in the literature if you inoculate Salmonella on to a sprout that has been grown for 24-48 hours the pathogen won't grow or grow very, very little.

So the key is to have enough competitive microflora there at the time of germination of a contaminated seed that you get that competitive exclusion happening but I wish we had more money and more people to do the work but right now as far as I know in ARS there's only one person that will be assigned to that kind of study.

DR. BERU: Thank you, Dr. Fett.

We have to save time for the public comment portion of the meeting so I hate to end the question and answer but we'll end it here, and I thank the panelists.

We'll take a 15 minute break and be back here about five minutes to 4:00. Thank you.

[Whereupon, a break was taken.]

DR. BERU: Welcome back. I think if you will take your seats we'll continue with the last phase of this meeting.

Before we go into the public comment period I have a list of four people who are pre-registered to give comments. Are there any others that want to be added to the list?

Okay. Then I'll just call on them in the order they are written here.

Mr. Lincoln Neal is our first commenter. No. Okay. So we have three public commenters.

Mr. Keith Warriner is our next. You can use the microphone.

Public Comments

MR. WARRINER: Thank you. Really I'm just going to speak five minutes on some of the research we've been doing up in the University of Guelph. I think if you picked up the handout that really describes the overheads and the work. Essentially what it is, is a seed decontamination method based on a sanitizer which we termed SDH. It has gone through various names but essentially SDH has been used in lots of different products, be it toothpaste, contact lens cleaners, chemotherapy drug, which has also been considered so essentially it is a very safe sanitizing product, if you like, which has got a long history of safety.

The reason why we chose SDH to decontaminate the seeds was really based on its phytocompatibility because it doesn't damage human tissue or eukaryotic tissue it obviously preserves the actual seed integrity and viability but what it does do is inactivates pathogens fairly readily.

We've just done some studies as you can see in the handout where we have inoculated seeds with E. coli or Salmonella , a treatment of these seeds was something in the range of 200 ppm, which is very low. It eliminates the pathogen. There is no pathogens on the subsequent sprouts and that's a key point because we don't just test the seeds to see if we get the five log reduction as was referred to earlier. We actually cultivate the sprouts for four days afterwards just to ensure that there is no pathogens there.

So what's really different about this approach is that whereas other approaches tend to try these strong oxidants before soaking the seed, we actually put in the sanitizer and join the seed soaking stage so we've got a higher contact time with the seed itself and also as the seed starts germinating obviously the protective sites within the seed coat start to decrease and the pathogens emerge where you can inactivate them.

So we have tried this on a little different seed types. It does depend on the seed but the ones we have successfully decontaminated is alfalfa, mung beans, clover, flax. The ones we haven't been able to decontaminate is radish and a few others. So it is seed dependent. We're not claiming to be able to decontaminate every seed but the simple fact is that it's very cost effective. It's a very simple solution. There's no part where you couldn't do it domestically large scale.

Another part we've done is actually produce naturally contaminated seed. We did this by inoculating the flowers of mung bean plants with a cocktail of Salmonella and E. coli . When we actually harvested the subsequent beans seven weeks later we found all were contaminated to Salmonella and E. coli existed on a few of them. Significantly, with our treatment with SDH we could actually effectively decontaminate the seeds.

We've also done some work with microbial populations through DGGE work, which essentially is a molecular technique for seeing if there's any different population changes due to SDH and we found only really two main bacterial types were missing from SDH treated seeds but there were no new bands which means that no new populations have been added.

I would just like to finish up with some work we've also been doing with looking at the distribution of contamination within sprouting mung bean beds. As you know, sprouting mung bean beds are basically containers--performed in containers, usually 25 kilograms, unlike alfalfa which is obviously a tray. What we did to look at the distribution of contamination was to look for generic E. coli , fecal coliforms and mesophilic aeromonas because it is better than looking for pathogens themselves but what we did find at the end of the day is that contamination was heterogeneously distributed. So if you try to collect one liter sample of spent irrigation water the chances of that representing actual microbial status of the mung bean bed are very low indeed and this is why I was interested in your concentration method.

We also did some lab trials using colonies of sprouting mung beans with E. coli 0157 and Salmonella . What we did here is we had essentially a 15 centimeter colony of sprouts and we put one gram of inoculated seed at different locations to see how the contamination spread and if we could detect it in spent irrigation water.

The reason why we took these small concentrations of inoculated seed and put them in the sprouting colony is because with seed decontamination it's likely that's what you would have. You would have only a proportion of the seed perhaps which is going to be contaminated where the rest has been successfully decontaminated.

What we actually found was that contamination introduced in this way with localized--within specific sites, a specific--more significantly over 50 percent of the spent irrigation water samples gave false negative results. Whether that's actually representative of what is happening in commercial practice is questionable but it just illustrates the need to validate testing protocols and identify if it does--if spent irrigation water testing does represent what's found in the mung bean bed.

So really that's just in a nutshell about the research we have done and, hopefully, we're going to take the SDH treatment further. It is not going to be easy getting regulatory approval, of course, and then commercial trial but I was very interested in the standardization procedures because I think that's what's needed. A standardized procedure to validate and verify seed decontamination methods so the industry itself isn't confused by it and we have a level playing field.

Thanks.

DR. BERU: Thank you, Mr. Warriner.

I'll next call on Mr. Richard Norton.

MR. NORTON: Hello. We have a small company that is interested in getting started in this field. We don't know anybody in the field hardly but we have had some success with it now. It's an interesting new material because it's so nontoxic. We would like to supply small samples to anybody who is treating seeds and different kind of seeds and who has the ability to sample and measure the reductions that they get. So if anybody is interested, let me know and we'll see that you get some advice. We've worked with these chemicals now for a long time and we have found for different uses we have to have different mixtures but that a lot depends on the surface that is being treated and the ones that we worked out for this was for the waxy coated alfalfa seeds. So other uses may have some other mixtures that we are familiar with.

DR. BERU: Thank you, Mr. Norton.

I would next like to call Mr. Steve Meyerowitz.

MR. MEYEROWITZ: Thank you. I just have some brief comments that I want to remind everyone that we have a wonderful product that is very unique. Sprouts are the only form of agriculture where you can grow it in any season of the year in any locality, whether it's Atlanta or Alaska, and we're going to need to protect this form of agriculture. We have an economic system which is seeing fewer and fewer farms, family size farms, in our local communities. They're being sold. They're being turned into strip malls. They're being turned into condominiums. So our local economies are losing the advantage of local agriculture.

Sprouts offer something in return for local agriculture. If we don't have enough land sprouts can be grown in factories. They can be grown in factories that are reclaimed from old industry or paper mills that no longer exist.

So it's an industry we need to support for our local economy and for the fact that it's a wonderful source of food. As our population or the world population doubles every 20 years, we have less space, more people, less space to grow food, our food is becoming internationally distributed long distances, quality and issues of pathogens are becoming more and more complex. Sprouts are a form of agriculture that can return us to local sources of food, whether it's a local sprout grower that's in your county or whether it's you growing it in your own kitchen. That's what is special about it.

But in addition to that these baby plants offer many wonderful advantages in terms of personal health and public health. We've discovered not only that their nutrition is many times multiplied in the young plant, such as things like protein in sprouts is approximately four percent and in iceberg lettuce, the most common competitor in the salad bar, it's .8 percent. So less than one percent. So more than four times the protein. Things like the vitamin A of radishes, of mature radishes that come from your backyard garden, is only about 10 international units and in radish sprouts it's 391. It's 39 times greater.

But more so than just a question of vitamins, there is also the minerals and trace minerals because it's one of the only forms of agriculture where we can eat the whole plant, including the roots and the roots are where we find a concentration of minerals and trace minerals.

Then we have things like the glucosinolates in the whole brassica category of sprouts where they can actually interrupt the development of malignant cells. We have things like sapponins which are very rich in alfalfa and clover and the sapponins actually have advantages in reducing osteoporosis and reducing the symptoms of menopause and hot flashes. And we have isoflavones. Soy bean sprouts are wonderful sources of isoflavones. Also for advantages in the cardiovascular system. You can go on and on. The antioxidants and bio-flavinoids. It's a wonderful source of nutrition that's just really being discovered.

And it's one of the finest sources of raw food nutrition that we have available to us. The National Institutes of Health and the National Cancer Institutes are constantly telling us to increase the number of portions of raw fruits and vegetables in our diet. It's harder and harder for Americans to do that. Sprouts are a source of raw food nutrition that is available at an affordable price that is locally grown. I think those are some of the reasons why we have to work harder to protect this industry to find solutions that are acceptable to the public.

I mentioned earlier that the public has a certain--that the FDA has some credibility issues with the public. Not all of the--not all consumers trust the--what their government--what our government regulations and rules are, especially if they are perceived as being heavy handed.

We have many issues of food safety in our food supply. It's not just a question of pathogens. We have issues such as pesticide levels and what levels of pesticides are acceptable. We have many dangerous things that exist in our food supply and we need to find ways to improve the quality of our food supply and sprouts are one important area that I think can mean a lot for us as we go forward in the future with prices of oil increasing, the cost of transporting food increasing and the quality of our food and the need for more fresh fruits and vegetables. Sprouts as an industry needs to be protected so that we can keep the quality of our food available to the public, and that's my reminder for all of us here.

Thank you very much.

DR. BERU: Thank you, Mr. Meyerowitz.

That brings us to the end of the public comment portion of the meeting.

By way of concluding remarks I would like to say that--

MR. LALLEY: Is there time for one more?

DR. BERU: Yes, by all means.

MR. LALLEY: I am Mike Lalley, Living Foods. I will try to take just a few seconds or minutes of your time. First of all, I would like to recognize and make it crystal clear that I recognize the risk in any food product. Specifically to quote the good folks from labs, they say there is absolutely no such thing as zero risk and that can be evidenced through the green onion outbreaks, the tomato, the strawberries, the apples, the apple cider, et cetera, et cetera. Since 1996 there has been 160 or 1,600 cases and in the last three years there has actually been about an average of 35 cases a year.

I'm holding an FDA document here that talks about an estimated 120,000 illnesses per year are caused by the consumption of eggs. Relatively that number is obviously of greater significance. However, we're not here to talk about eggs.

So what I am here to ask is that FDA and CDC embark on a thorough and critical review of the epidemiological evidence which is the prime evidence against sprouts. There have been--of the 27 cases in the last ten years there has only been a very small handful, perhaps two or three, where they have been able to culture pathogens from either the seed or the finished products. And, as such, like I say, that is the prime evidence against our industry.

In 1995 there was a Salmonella outbreak in Michigan and a few other states and the respondents in the epidemiological survey, 25 percent of the respondents who had the genetic fingerprint, the supposed genetic fingerprint, only 25 percent of those respondents said that they had actually consumed sprouts.

In 1999, when there was an E. coli 0157:H7 outbreak in Michigan and Virginia, when the good folks from CDC, Michigan Department of Agriculture and Michigan Department of Health knocked on my door, they had not completed their epidemiological study but at that point they said, "The reason we're here is vegetarians are ill and it makes sense that we should come and talk with you."

At that moment in time I thought maybe they had a valid point. Upon further discussion with other CDC scientists they tell me that if there's an outbreak amongst a particular ethnic group, say Middle Easterners, and we ask them what they've consumed and they tell us that they have consumed humus, babaganosh and other food items peculiar to their ethnic origin, and then we compare them with a control group off of the street of John Blow and his cousin Freddy, we see that every time John and Freddy consumed no humus and no babaganosh and as a result there is a statistical number that arises that implicates the items which are peculiar to that ethnic group.

In this case the vegetarians obviously were the ethnic group that I'm referring to and the control group in no way, shape or form took that into any consideration whatsoever.

As a matter of fact, Dr. Brewer expressed a great deal of satisfaction in his written report on that study where he was showed a great deal of satisfaction and I would say joy in the fact that he was able to get a 46 percent positive response rate from those afflicted consumers that had the genetic fingerprint. He did this and admits such in his report by first of all asking multiple questions regarding sprouts and, secondly, by preceding his questioning with a leading description of the potential culprit. As in, for instance, a small caucasian man, 5'7", blue eyes, balding, was he the guy? And then, in fact, he was thrilled when 46 percent said, "Well, it's possible that could have been," because he did say that they could have been hiding underneath the sliced ham or baked or who knows what. But he got 46 percent positive rate. 54 percent of the people that were sick in the State of Michigan, never consumed the sprouts, swore to god. I've got newspaper articles--pardon me?

DR. BERU: We are out of time.

MR. MEYEROWITZ: Okay. I will go really quickly. Anyway, so these vegetarians--okay. So let's just see here. I'll try to skip through some of this. Anyway, the epidemiology study in that '99 case showed a prophylactic effect was--occurred by the consumption of ground beef. Well, now I don't know what you know about ground beef, and I know very little being a vegetarian since 1969, but I'm here to tell you that ground beef will not save you from human pathogens.

When the recall was issued--a recall was subsequently issued in Michigan and Virginia and the outbreak continued for over six weeks in Virginia. And it just so happens the reason that happened is that they had secondary contamination. One person--secondary transmission from one person to another because the slobs in Virginia don't wash their hands where thank god in Michigan they did.

Well, I'm here to tell you that is absolutely ridiculous. It didn't--that isn't the way it happened.

Anyway, to conclude, risk/benefit. Obviously there's a risk in all food items and we don't care what we call them, what they are, what your preferences are. Risk/benefit, we had Vioxx and Celebrex. The risk was 25,000 dead bodies littering the streets of America. The benefit, however, great though it may have been, that at least their back didn't hurt for a couple of weeks prior to passing on.

In the case of sprouts, granted we've got a situation here where human pathogens can be transmitted to any food form. However, the benefits are such that the broccoli sprouts, for instance, are better--shown to have been more effective than the prescription drugs in treating ulcers. They have been shown to prevent and in some cases perhaps cure cancer.

So what is the benefit? What is the risk? If I take a 50 pound bag and they bring me a 50 pound bag of lettuce, and I take it into my facility, could I have GMPs, SOPs, HACCP, and I take 25 pounds of this bag of seed and I grow it in my sprouting facility, god forbid that anybody--a consumer of any sort who wants to avoid contamination should consume it. Yet if I take that same 25 pounds, take it into the warm moist environment of my field out back behind the top--my barn, and where animals have defecated for years, where rodents, insects are flying in from the livestock farm across the street, then there is no risk whatsoever there. Eat up, don't worry.

And the final conclusion is what they're trying to do is they're trying to turn you all into little Howard Hughes. You'll never have his money but if you have his fear you will be living in his world.

Summary of Meeting and Adjourn

DR. BERU: Thank you. Thank you very much.

That now truly brings us to the end of the public comment portion of this meeting.

By way of concluding remarks, I would like to say that we do believe that a good first step to improving the safety of sprouts is to engage in solicited views of other government agencies at the federal, state and local levels, industry, consumer groups, academia and the public generally about the current science relating to preventing or minimizing foodborne illness associated with the consumption of sprouts.

In this respect this has been a very useful meeting. You have heard from the panelists as well as the commenters on various aspects of the issue. From the merits of screening and sampling seeds prior to sprouting to the various studies conducted on the various seed and sprout treatment methods.

We hope to also get additional comments and the comment period is open through July 18th.

With that, I would like to end the meeting by thanking all involved in making this truly important meeting happen.

A special thanks to our panelists and thanks to you all very much for taking time from your busy schedules to participate in this meeting.

Thank you.

[Applause.]

[Whereupon, at 4:31 p.m., the proceedings were adjourned.]

- - 

Presentations available upon request.