U.S. flag An official website of the United States government
  1. Home
  2. Animal & Veterinary
  3. Safety & Health
  4. Antimicrobial Resistance
  5. The National Antimicrobial Resistance Monitoring System
  6. Animal Arm
  1. The National Antimicrobial Resistance Monitoring System

by Dr. Paula Fedorka-Cray

DR. CRAY: I actually did these slides so they were done yesterday morning, as per Dave. I called him. I said I was sending them, and my father has a dial-up connection. I dropped my son off at my parents’ house, and I tried to download a 10 megabyte file over a dial-up connection. It was a rather painful experience for all concerned.

I thank Bob Walker, Linda English, Dave and Pat for staying later in the evening to do that. I could have driven here faster it turned out and xeroxed them and did what I did.


Nonetheless, this is the animal arm of NARMS. Dave went over some of this, but I will reiterate and give you a little bit of a different perspective on some aspects. It began as a research project, the animal arm actually did, with Dr. Leanne Thomas and myself. Dr. Leanne Thomas was ta the National Veterinary Services Laboratories and I was a scientist at the USDA National Animal Disease Center in Iowa.

We got together and we decided that at the time, in 1994 and 1995, there was an increased interest in antimicrobial resistance. So we actually bought panels that just had breakpoints on them, and we were doing the testing when we received a call from Dr. Tollefson, who had heard what we were doing, and asking us if we would be interested in joining NARMS.

And we were certainly interested in doing that and NARMS officially launched in 1996, and we used up our first 1,500 plates of just breakpoint and then we went to the MIC format. Although we didn’t have comparable plates until 1997/1998.


The evolution is that NARMS is really a passive system and the original goals and objectives were to provide descriptive data on the extent and temporal trends of antimicrobial susceptibility in Salmonella and zoonotic enteric organisms from human and animal populations. To identify resistance that arise. To provide timely information. To prolong the life span of approved drugs. To identify more areas for more detailed investigation and guide responsibility.

It is actually a passive system and it relies primarily on the submission of isolates. And really, all three arms either relies on the submission of isolates or the submission of some type of samples. This isn’t an active system that we have gone out and actually designed ourselves.

We do believe it does provide a comprehensive snapshot, and that really is what it is. It is a one time point snapshot of food animal production through the U.S. We have all animal species that are representative in some way, shape or form.

We look at small, medium and large slaughter plants now, and Dr. Anandaraman will talk about it this afternoon. This did not start until a later date in time. We had the plants as they were coming onboard. So the system was not as robust as it is now until 2000.

We also have isolates from ill animals. There has been a lot of discussion about the necessity of having diagnostic isolates, but we believe that these can provide information on isolates that may be emerging or that we may see in the food chain over time, because these animals are presumed to have undergone some type of treatment over time, and while they should not be part of the actual food that is being consumed, at least in the raw way, shape or form, they may, in fact, be representative of what is to come.

We can see in that some respect with Salmonella Uganda, which is first emerged as a diagnostic isolate and now we are trying to see more prevalence in some of our other samples.

Our isolates from on-farm samples are limited. They are limited to the NAHMS studies, the National Animal Health Monitoring System studies that occur, and they are also limited to epidemiologic studies that we might conduct or our collaborators might conduct over time.

However, at some point in time, and for the most part, during the year we can receive isolates representative of all geographic regions in the United States. So we have these three sources of isolates, on-farm, diagnostic and slaughter, and we looking for Salmonella and Salmonella is our sentinel organism, E. coli, Campylobacter and Enterococci.


This just illustrates that and shows the differences between the human and the retail. We have tested Listeria. The question was asked earlier this morning. We did a snapshot a couple of years ago on about 600 hot dog isolates. We occasionally have some coming in on various other -- from other sources now. We actually have one student who we acquired through a reorganization who is looking at Listeria for part of his project, which is not part of NARMS.


Again, our isolates come from diagnostic and non-diagnostic sources. These diagnostic isolates are presumed to be associated with clinical illness. The animals are not likely to enter the slaughter facility. Right now we have 12 veterinary diagnostic labs. California, Colorado, Iowa, Indiana, Kentucky, Michigan, Nebraska, New York, Oklahoma, Texas, Washington and Wisconsin who have submitted isolates over time.

We have approached Minnesota about submitting isolates, and Pennsylvania will actually join us, the University of Penn, later on this year.

Then we also go up twice a year and we collect isolates from the National Veterinary Services Laboratories, and what we are doing is we are collecting the diagnostic isolates from there that were associated with either a primary or a secondary etiologic outcome. So, Salmonella has to be identified as the main cause of disease.

However, we exclude selection from the sentinel sites so we are not duplicating isolates. So this way we then can saturate the rest of the United States, as far as the isolates coming from other states and the various regions.

For our non-diagnostic isolates we presume that all of these are coming from healthy animals. Our on-farm isolates are part of the NAHMS studies, as I said; however, there is a five-year rotation between commodity. So, for instance, there was a swine study in 1995. There was a swine study in 2000. There is scheduled to be a swine study coming up within this next year.

But when we look at these, again, they are snapshots of time, although they do provide a statistical representation of greater than 95 percent of animal production on-farm in the United States for that particular time period. So, they tend to be very invaluable as far as resources because they also collect management information at the same time, which gives us an idea of what antimicrobials may have been used in practice that may be affecting outcome of the resistance patterns that we may see.

I regret that our APHIS partners were not invited to the meeting, because I think that they could have provided some important additional information.

The slaughter isolates then are rinsates, carcass swabs and ground product. Now we are receiving ready-to-eat products and isolates from eggs. This is a comprehensive snapshot based on the FSIS compliance. It is related to retail. It is not statistically significant, but Dr. Anandaraman will address that this afternoon.


I think it is important for you to realize for you that for the diagnostic isolates what we are just getting are NVSL isolates in the sentinel sites, and we receive anywhere from about 75 to 500 isolates from these various sentinel sites. So, the larger the sentinel sites, like Michigan and New York, we tend to get about 500 of their diagnostic isolates because that is about what they get per year. So we get all of them.

There are some smaller sites, and we actually get all of those too, like Washington State. For actual diagnostic submissions we tend to get about 75 to 125 different isolates.


All of this information is collected and reports are generated and sent back to the sentinel sites, and we give them free range to publish their own information, either including us as co-authors or citing the NARMS system as providing the information to begin with.

The compilation of data that you see on the reports is usually all of the sentinel sites together.

E. coli right now only comes from our sentinel sites, as far as diagnostic lab goes. We get very few Enterococci, too little to really count, and we do get a few Campylobacter and they tend to be sporadic. Those are available too in ancillary reports.

For our non-diagnostic isolates then all of the E. coli, Salmonella, Campy and Enterococci come from the raw product, federally inspected slaughter plants and from focus studies, but there is a caveat. Only the Salmonella comes from the actual FSIS compliance testing.

For E. coli, Campylobacter and enterococci then we actually take the spent carcass rinsates that are sent to the eastern laboratory for FSIS. Once they take their aliquot for Salmonella testing, anything that is left over that is greater than 10 mils is sent to our laboratory, and from that then we can isolate E. coli, Campylobacter and Enterococci.

Now, the FSIS laboratories, the eastern laboratory, doesn’t just receive samples from the eastern part of the United States. The distribution is randomized so they are receiving them from all different -- from through the U.S. on a rotational basis. But it is important to realize that these carcass rinsates are also only coming from chickens.

So, the information we get for Campy, E. coli and Enterococci through non-diagnostic slaughter isolates are limited to chicken carcass rinsates from spent Salmonella samples.


This is what I see as one of the weaknesses of NARMS when we look at critiquing our own program and how we can improve the robustness of the program. But farm sampling and diagnostic isolates are independent of the slaughter program. So, we don’t have a flow through.

We have farm isolates that are coming because of NAHMS related studies, we have diagnostic isolates coming from the diagnostic labs and then we have the compliance testing out here. And we really don’t have an association of what is happening between the farm then and the plant.

We believe that while you can say that, yes, there is some association -- and, of course, what we see on the farm is what we see, in part, of what is coming in the plant. But we can’t say a lot of times what the use is associated at a plant, particularly with antimicrobials, and what is occurring then in the slaughter plant.


If we move to looking at the individual organisms, Salmonella is our sentinel organism. We have all sources. We receive the isolates primarily. We are not doing a lot of the isolation.

These serotypes will vary over time and vary by source, and we believe it is absolutely critical to take the information and break it down by serotype and by source because we see distinctly different patterns associated with these various sources and serotypes.


When we look at our sampling scheme, what we were doing up until just recently took us several months to complete when we would receive an isolate, either slaughter isolates, which are the actual isolates themselves received weekly, or diagnostic isolate, which is the actual isolate.

On farm feces we have always been the testing laboratory for NAHMS studies. So we actually receive the feces or farm sample, and we have to go through a seven to 10-day confirmation and isolation procedure. For this we have used our published methods, which have been compared and contrasted and now are considered to be one of the gold standards in the industry.

Once we had the isolate then it went for serotyping, if we were doing the isolation, or it came with the serotype if they were slaughter and diagnostic isolates. They were susceptibility tested. Then we freeze and store all of our isolates. So we have every isolate that we have tested since 1995 in triplicate frozen in the building.

And right now, we are lyophilizing everything, and one isolate will be sent to the National Germ Plasma Repository in Ft. Collins for, hopefully, lifetime storage if anything happens to the Russell Building.

If we don’t have a phage type, then we have to send it back out to NVSL for phagetyping, and because of our involvement now with USDA VetNet, we send all of our isolates to our pulsed field lab where they undergo Pulsed Field Gel Electrophoresis. And then it goes on to preliminary data and then year-end analysis.


At the request of FSIS, who has a need now to send information back to the plant on a regular basis, we have looked at how we can compress our time frame in getting this information back to FSIS that they might be able to use. So, particularly for our slaughter isolates now, we are looking at if everything works right in life and we have perfect gels and perfectly susceptibility tests, about a two-week turnaround time from receipt of isolates to a susceptibility pattern and a PFGE gel picture that we could send back to FSIS which they, in turn, we understand, will send back to the plants.

Now, in an testament to the compliance testing success, we see that the number of isolates from 2000 to 2004 has gone down from about 3,500 to 2,500 or so that we receive for a year from FSIS. The number of isolates in toto that we receive for diagnostic are on-farm is also about another 2,000 to 3,000 isolates for any particular year.


For generic E. coli this did not start for us until 2001. We do have some diagnostic sources. We have on-farm sources, which we have not included as part of NARMS. But we have those isolates available that are for our own independent tests, and again, these are from the spent FSIS rinsates and only represent poultry. Particularly, chicken.

There are thought to be reservoirs of resistance genes, which is why we are still testing them as part of the program.


And the isolation procedure is almost identical to Salmonella, except that we use com-agar* for our isolation. I will put a caveat in here; that we have done extensive and exhaustive testing of the different medias that you use for isolation and we can say that you will have different results on occasion, dependent upon the isolation methods that you are using.

So, for instance, if you use Campy line agar for a Campylobacter as opposed to cephex, you will get different populations than you will get of Campylobacter jejuni or coli. The ratios will be different, and within those different populations from those different plates you will have different antimicrobial susceptibility patterns, and I think that is important to remember.

From 1998, which is our Campy started, through 2000 we had Campy obtained as an actual isolate, because FSIS was taking care of the methods. The isolation. They were using nalidixic acid in their screening and at that time we believe that it probably under represented the percent resistance to quinolones and fluoroquinolones.

We changed the method in 2001 when our baselines actually stopped. Again, these isolates originate from spent FSIS rinsates and we believe that now at least the susceptibility testing more accurately reflects the resistance of the total population.


It is very similar, although there is a difference in the amount of manipulation that has to go on with Campylobacter, and again, it depends on if we get them on-farm and we actually have to do the isolation too if we get the isolates from diagnostic sources. And now we are doing all of the enriching for slaughter. So, the time from receipt of a sample is several months before we are ready to have some preliminary data.


For Enterococci this started in 2001 also. Again, we are at the FSIS rinsate origin. Our methods were tested and developed in 2003, and dependent upon which incubation temperature and media that you use, you can obtain different populations of Enterococci. We will look at some of this. We do have a paper that has been published on it. Like E. coli, they are thought to be reservoirs of resistance genes.


I will say that USDA started USDA VetNet in 2003 and this was established in collaboration with CDC and FDA and is essentially the animal arm of PulseNet. It was originally known as PulseVet, VetNet-Animal and then the under secretary decided they liked USDA VetNet and that is what we went with, because they were doing the blessing.

All of our personnel are trained and certified at CDC, but the program actually resides in Athens, Georgia.


We aim to capture PFGE patterns of Salmonella isolates submitted to NARMS, and we also have a branch for our CAHFSE program. We compare VetNet and PulseNet PFGE patterns and we would like to say that eventually what we would like to do is to be able to use this data to assist in surveillance and then the study of the ecology of organisms along the food chain and for us to be able to assist in the investigation of animal illness outbreaks on-farm, as well as foodborne outbreaks when CDC has them.


This is modeled after PulseNet. We have two independent databases because we have two independent programs, the CAHFSE and the NARMS Program. These databases by the end of this year will be moved to RRC and they will be able to be accessed much like PulseNet does now through an online server. Currently the gels are being sent -- currently everything is being run and analyzed by USDA-ARS, but that will change with the submission of gels from the University of Penn Lab.


So, as of May 5th we had 82 serotypes. We have 3,596 isolates. We had 1,452 unique patterns, all done with just one enzyme, XbaI. That has since gone up to well over 4,000 isolates now.


One of the things, when we look at our PFGE distribution, is these are the top serotypes that we have, and you will see that in the rank, 1, 2, 3 through 10, the "R" stands for isolates that are also seen from the retail arm of NARMS.

Each is indicative of isolates that are serotypes that are also seen from the human arm of NARMS. These are the numbers that we have so far for these top 10 in our database, and you can see the numbers with unique patterns. One of the things that we have been able to do in comparison of the VetNet and the PulseNet databases already is to see that there is much more heterogeneity amongst the animal isolates than there are amongst the human isolates. So, we have many, many, many diverse patterns in the animal arm.


When we look at the top 10 patterns that we have to date, we can see that, in fact, there is an overlap with particularly four of these, and they are coming from Enteritidis, Heidelberg and Typhimurium. And you can see the PulseNet on the right and the VetNet on the left.


And we can generate patterns much like this, and we intend to be putting this on for our reports in the coming years too so that we will be able to actually have available what the pattern looks like and what the serotype is; how many isolates we are generating from that and what the XbaI pattern is.


Now, we go through this debate and I know that Dave just put up the PFGE pattern, and you can see that this is for Salmonella Newport, and I am not sophisticated enough like Dave to draw up those nice little boxes that come up. But you can see that we have significant clustering of one particular clonal type and then we have some independent.

I always have this debate with Pat and Dave. If you look particularly at 44 and 42, they both came from cattle. They are both in the same cluster, but they have different antibiograms*. And so, yes, they have similar PFGE patterns, but are they, in fact, the same isolate because they have -- are expressing different resistance patterns?

And depending upon how many microbiologists you have in a room and how many molecular biologists you have in the room, that is dependent upon what the opinion is of the day. So, whether you can actually use it for an analysis to say this came from a human versus this came from an actual animal source, I think that we can generate some interesting discussion about that.

Certainly, when you have matches that are all the way through and you can identify the same genes, the same resistance patterns, the same PFGE pattern is the same, then that would be an indistinguishable and a confirmation.


When we look at VetNet, we see NARMS, we see CAHFSE. We have -- NVSL is interested in submitting, and they are going to begin later on this year submitting diagnostic TET files to VetNet. The American Meat arm of USDA, from their feeding programs for the food school lunch program and the military, will start submitting their files.

Veterinary diagnostic labs, as in Shelley Rankin, will begin submitting her files, and then we hope to have -- EPA has expressed some interest; other government related agencies. Perhaps some university labs and some Ag State labs at some point in time, making this a very comprehensive system.


So, when we look at what NARMS has evolved to over time, these colored regions are actually representative of the FSIS regions where we have isolates coming from the compliance testing. When we see -- and that would be in region one, two, three, four and five.

When we look at Dn and the diagnostic laboratories, the red Dns now are labs that have not submitted isolates in 2005, but we have isolates from the past. The blue is where we are in discussion for getting isolates. Vn is the VetNet lab. This is the VetNet hub. This is where we will begin getting isolates for the VetNet program.

And "C" overlays our CAHFSE and our future CAHFSE site. I think that what this shows is that -- the point that I wanted to make is that with NARMS we have a very comprehensive look at what is going on in animal production.


One of our most exciting things, and this was the killer last night that caused the major snafu over the dial up, is the advent and development of microarray technology. We now have a chip that has about 150 resistance genes on it, and you can use any bug. Campy, Salmonella, Enterococci and E. coli. It will pick up all of the resistance genes.

We also have about 900 virulence genes that are also placed on this. So, we hope to be able to use this, particularly when we have questions about isolates that may be part of an outbreak so we can tell, with a higher degree of certainty, what their degree of relatedness is over time.


Other studies that we undertake as part of NARMS. We have optimized isolation protocols between labs, and we have, just about every 18 months, a methods meeting. Either CDC comes up and FDA comes down or we go someplace where we talk about how we are doing everything and what the optimized method is and how it might be different between labs and effect our results.

Campy isolation protocols for Enterococci in particular and E. coli. We also conduct additional molecular studies, like the retail arm does in CDC on plasmids. Mechanisms of resistance. Dr. Charlene Jackson has identified several new mechanisms of resistance in Enterococci, we look for integrons and we have developed some multiplex PCR assays.


When we go to reporting, I think a soft critique is that this the part of NARMS that I am least happy with. Our reports are posted to the website now. We don’t do a hard report. A big part of this is because our -- one, we are short staffed, as far as any people spending time to get this done.

But our reports now, if we put out just what is on the website right now, you would be printing about 329 pages of must Salmonella information and then another 100 pages of all of the rest. Since 2003 we have worked diligently to try to update our reporting so that it is more timely.

One of the things is that as soon as our reports now are completed or as soon as our data is completed and cross checked, then it starts going up. Within the next two weeks you will start seeing tables and graphs going back up for 2004. So, I think that we don’t have text. That is one of the big lacks. We don’t do an interpretation, but the data is essentially all there.

The biggest problem with the data for all of the arms that I see right now is the lack of flexibility in the way that you look at this data.

John Stelling and Terrell Miller have been working diligently to have a unified reporting database and this I am really excited about. Tom Chiller and I spent a lot of time kidding each other about this, but we have had many meetings with -- the CPAR group has come down now, CDC has some over; FDA has come down.

We have gotten together so that we have -- the goal is to have similar reports so that if you click on a table one for CDC and look at Salmonella, you should be able to click on the same table for all of the other arms and get a similar report up. I think that is spectacular, and that is where we need to go.

However, given enough money and time, what I would really like to have is a dynamic database, much like they have for TSN and some of the others where we have drop down menus so you can just go in, if you want to look at Salmonella or Typhimurium, a particular region, a particular drug, a particular year, you just go in and drop down all of that, click go and it gives it to you.

I think that that, in fact, will be when we have reached our pinnacle of success. So that is what we are working toward, and I see that particular drop down menu option as probably being two years away.


If you go to our website, this is the first thing that comes up. ARS, in its infinite wisdom, said that they were going to consolidate and make things easier. No one can find anything now.

But if you go to our website, what we did get them to do was allow us to put a separate box up and you will see that there are NARMS reports available in that box.


If you click on that, then you get the report screen that comes up, and you can click on each individual report that you want for every year.


One of the things that the report doesn’t do is it doesn’t give the type of presentations or it doesn’t the give the summary of information that we can give when we give presentations. One of the other things that we are working toward is before the end of the year we will have a presentation where everything will be converted to PDX and every presentation that anyone in the lab gives associated with NARMS will be up there so that people can look at the way we have manipulated the data in a different way, which I think will provide an extra layer of information.

But you can see that this is the distribution. All sources, all clinical status for these particular serotypes. Again, the "R" is retail or human sources that they show up on their particular reports. You can see the frequency. Salmonella Kentucky is, by far, our most serotype; 9.2 percent of our total number of isolates, and a majority of them tend to be pan susceptible. Then there is about five percent that is resistant to five or more antimicrobial.

This is the type of information that people request. One of the things that I have always tried to encourage, and people have made use of this, is if they want information presented a different way, they just call and we get that information to them. So, it is, by no means, a static system.


We can look at our isolates and this is part of our report where we actually go by source. These are from slaughter: Cattle, chick and swine and turkey. You can see the top 10 serotypes within each, cumulative for 1997 through 2003, and the blue highlights will be indicative of those serotypes that also show up in the human arm.


We can compare now with the human serotypes that are coming up. This is an example of dairy cattle serotypes. We think it is absolutely critical. Here is diagnostic, here is on-farm, here is the human, because you can see that there will be some differences, particularly whether we are getting them on-farm versus diagnostic. This gives us a much better handle on what might be happening over time.


And you can see then that it changes for each source, and that is why it is absolutely critical. You have to look at cattle isolates separate from swine isolates, separate from chicken isolates and by each of the different serotypes.


We had two post-docs and they have, cumulative, six publications that will be coming out this year. We had the first ciprofloxacin resistant isolate identified in 2003. I’m sorry. In 2000. And you can see that Salmonella Niakar* was in 2000. We had Newport and Typhimurium in 2002 in Kentucky, but we have very little fluoroquinolone resistance amongst our Salmonella isolates.


We can look at our Salmonella Newport isolates, and I think that is what this is -- particularly important to look at is these are the years when we didn’t have any farm isolates. So it is not necessarily that there weren’t any isolates. It is just that we didn’t have any studies conducted by APHIS during those years, and those are critical gaps that we are missing.

But you can see compared non-diagnostic, which would be slaughter sources to diagnostic. The majority of our Newport comes from diagnostic sources.


Dr. Jenetta Tankson then has characterized a number of these Salmonella Newport and we can look at the presence of entigrons, plasmids in the CMY2 gene, and we find that most of the multi drug resistant Newport harbor the plasmids and CMY2 genes.


Our reports also contain information on DT104. Quite interestingly, our isolates primarily originate from swine for DT104 and not cattle. Cattle is in this green and this purple. Forty-three percent are from swine, versus about 38 percent from cattle total.


We can look at the various sources or sites then, whether they are diagnostic, slaughter or on-farm, and you can see that particularly in dairy. They are primarily diagnostic isolates.


We then have the ability now to look at phage types. This is not included as part of our report in the past. These are being done for all of the various species, and they will be presented this coming year as an ancillary box that you can click on in our report page so you can see how, in fact, the various serotypes emerge, stay and then seem to disappear over times in the various years amongst sources.

We think that this will help also with the comparison between human outbreaks.


For Campylobacter this is the type of information that we can generate now, and you can see for Campy and coli we have more resistance there for more drugs than we see for C. jejuni. In our other talks that we have given this is where our methods change. Up to this point in time was when we were receiving isolates where FSIS was using nalidixic acid in the screening. After it changed then you saw an increase in our resistance to the fluoroquinolones.


We can look at E. coli now, and we have this broken down by year. You can see that in 2004 we did several thousand isolates, and we have seen very little change over time, except in 2004. We see our first ciprofloxacin resistant E. coli.


We also have E. coli in cats, dogs, eggs, dairy cattle and chicken in 2003, and we can see the differences in resistance patterns amongst the various animal sources.


When we look at Enterococci, by far faecalis is our primary serotype, although we also have a significant number of faecium, durans and hirae and gallinarium, casseliflavus and then it decreases from there.


When we look at resistance for all of them, we get a snapshot that is indicative of bacitracin, tetracycline being prominent resistant patterns.


But when you look by serotype, you get different results and this is why -- for species this is why it is important to have all of this broken down over time.


So, what are our needs? We have tested, as part of NARMS, almost in the 11 years that we have been around I think, we are close to 80,000 isolates now, and where do we see our greatest needs? Well, our greatest needs I think are in permanent staff.

With the receipt of the interagency agreement the USDA does not recognize this as permanent dollars, and what we really need is to have some type of language or have some type of agreement between the agencies themselves that says that this money is going to be stable, at least at a particular level. So that would enable us to hire the staff permanently.

Unfortunately, at USDA staff have to be rotated after four years, and if we lose that continuity, particularly from the people that we have now, we lose a lot of ability to maintain the numbers of isolates that we are able to analyze right now.

Permanent staff is also required for the computer end of things. We know that CDC has gone through several programmers. We have John Stelling, and John is only paid on a part-time basis because he is off and about on his many other journeys across the globe. I think it would be important for the system to have someone looks at the computer and the reporting and is able to meet all of our needs comprehensively over time.

Funding again and the coordination between the three agencies. Continual reviews. I think that we have our own annual reviews that we do internally, but it would be important for us to have annual reviews.

Right now we have 18 month where we have a NARMS get-together and we have presentations, but I think that we really need to have these reviews where we all, the people who are doing all of the work, get in a room, say this is how we are doing it and then have somebody just ask us the hard questions; so that we are constantly then asked how we can make this better, how we can change to meet everybody’s needs.

We have, for the past four years, for some outside oversight, and we have been told that it is difficult because of regulatory agencies being involved, but I really think that it would be important to have many stakeholders from many different areas who can give us input on a regular basis. A core panel of people who can look at the program, who will be able to accurately and adequately analyze and discuss the needs, both within and without of the industry and internationally to tell us how we are doing.

We have lost Dr. Marcia Hedrick, who was the coordinator, and I think it is absolutely critical that we have a coordinator. This coordinator has to have the trust of all three agencies.

In my opinion, I believe that we, FDA, CDC and USDA, should be the people who select this coordinator, because this would be somebody who would be our spokesperson and who would be moving the system forward and meeting all of our needs, because we also have individual needs within each of the agencies too. And they should be able to have freedom to make decisions regarding both monies and studies that need to be done.


There are many, many people who are involved with this, and I just want to acknowledge the staff. This is just a mere snapshot of the staff that is back at USDA.


Again, this is an interagency. The animal arm of NARMS is only as good as the human and the retail arms are too. This is the best example of a three-legged dog that makes it in veterinary medicine. So, with that, I’ll entertain questions.

DR. SAHM: I want to get back to something that helps us formulate the questions around these points of discussions, and it is something that Sue brought up earlier regarding the goal of the NARMS program.

When this was started we weren’t quite sure to what extent antimicrobial use in animals presented a human health threat. Is that true?

DR. FEDORKA-CRAY: That is right.

DR. SAHM: And we haven’t done the analysis yet to completely show what the extent of that threat is. Correct?

DR. FEDORKA-CRAY: That is correct.

DR. SAHM: And then, with NARMS -- what I am getting to is if that threat or link was established, then there would be a better case for maintaining ongoing surveillance or what have you. Right?


DR. SAHM: So the question is are these studies eventually going to be -- when are we going to get to finding out what this data tells us in terms of the threat to the human population, because that will help make the case as to whether or not it is worth it. Do you continually survey these animals and at what stage, which types or what have you?

So, that is one question as to what is the staging of the NARMS goals.

DR. FEDORKA-CRAY: Well, I think that is a lofty goal, and that is one that we have all wanted to work towards. With the limited resources that we have had with the NARMS program and with the monies that we have been able to obtain from the interagency agreement, you have to go with what you have.

There isn’t in the U.S. a comprehensive program

on-farm that allows us to look at on-farm use of antimicrobials, the isolates that are obtained there over a period of time, follow those through, those same animals to the slaughter and follow that then with what may end up in retail.

Because of this and -- you know, this is the NARMS review, but USDA -- I will just mention that USDA underwent some soul searching and came up with the CAHFSE program that I discussed, which sets out to meet those missing needs. Until the CAHFSE program though goes from its infancy to a fully mature system, we have to continue with the isolates that we have available to us.

Right now the only isolates from the veterinary side, in fact, are the diagnostic isolates that we are getting now, the slaughter isolates from the compliance testing and any on-farm studies that may be conducted either through NARMS or other epidemiologic studies. I think it requires more of a discussion than we have time for here.

DR. SAHM: Right. I agree. The reason why I pose that question that way, Paula, is because from Dave’s previous presentation I am wondering what you have learned with all of this extensive work already can be used to start weeding out and narrowing the focus of the program.

For example, one could make the argument that the poultry that was brought up earlier is the key. The other thing I noticed about your data is on the human side 70 percent of enterococcis faecium in the U.S. population among humans are resistant to vancomycin and about three percent faecalis.

Of all the enterococci you are isolating from these animals you are not seeing vancomycin resistance?


DR. SAHM: So, is there a plan to take that kind of information and say, okay, this is probably not the level of threat and let’s turn more of our attention to Campylobacter in poultry or something like that and start to narrow the focus and redesign the focus, given the data that you have colleted and analyzed already?

DR. FEDORKA-CRAY: I think that discussions that Dave White, Tom Chiller and myself have had would go along the lines of being able to sit down and do a self-critique of a lot of what we are testing now. We haven’t done it up to this point in time.

A lot of us were waiting for this review to see what would come out of this. I think that that is a very appropriate task. Do we need that level of testing for, particularly, the commensal organisms? That could be something that could be addressed more along our research lines and less of an emphasis on our NARMS testing where, in fact, then we would look at either getting more on farm isolates and following that through or changing the dynamics of the system.

DR. SAHM: Right. I agree. I think you all have learned a very -- we have learned a lot from the NARMS efforts, and it is building on that rather than continuing it in all aspects as it was originally designed.

DR. FEDORKA-CRAY: Right. Exactly. It needs to be more robust and it needs -- I think that the biggest problem is -- like I said, it is all independent. So, you make an inference here, you make an inference there.

If you see an isolate, N = 1 or 2 or 3 for a particular year, you could say it is emerging, but you don’t see it again for another year. So yes. I think we can do it better. I agree.

DR. MILLER: Jumping along what Dan was saying, one way you might be able to focus down is by looking at the commonalities across the three surveillance systems. And if your ultimate goal is to link to human health, that is another way to limit down.

I just ask you, Paula, -- I mean, from the pet species, the cats and dogs, has that work led to any sort of hypotheses about human disease at all that have been followed up? Or is that something necessary to continue on?

DR. FEDORKA-CRAY: When we go to NVSL, you can see that our numbers of pet and dog isolates are very small. We take those mostly because they don’t get very many. We have some years where there is 25 isolates or 50 isolates.

So, we take those as more of an interest, and they are a very, very small percentage. I can’t tell you the percentage exactly off of the top of my head, but they are small numbers.

When we look at saying -- I think there have been a number of studies that have come out now, particularly recently. For instance, the pocket pets have been a problem. There have been some reports with dogs harboring some of the Salmonellas.

There are some, but that is such a small part of what we have here. It is there as a piece of information for people, but the focus of NARMS is not on that number of isolates whatsoever, and if we didn’t do that number of isolates, that is something that we would pick up as just a research type of look-see. But that is not the major focus of the program

DR. ALTEKRUSE: In addition to this overlay of the public health impact where we talked poultry isolates perhaps having more importance, there is another -- it is not a scientific consideration as much as it a policy consideration. I think it is important to point out that the FSIS slaughter isolates are going to be used in a very specific way in the future.

Rather than just counting the number of positives in a Salmonella performance set, we intend to provide that information back to the establishments so they can make some adjustments in terms of their HACCP plans about antibiotic resistance and serotypes.

And so, the timeliness of getting that information is important. You mentioned that, that we are going to try to improve timeliness. But also, those isolates have the potential to have to be used for enforcement purposes, in which case chain of command -- chain of custody issues are essential for those isolates.

They really do rise to a high level of priority in terms of what needs to be done and how they are handled.

DR. FEDORKA-CRAY: The FSIS isolates are always our priority in NARMS. I mean those are isolates that historically are always done first. Diagnostic isolates are filled in on the side. NVS isolates are only obtained twice a year. It is only from NVSL that we get the small number of dog and cat isolates that we do.

We have discussed with Dr. Pat McKaskey all of your needs, and we will be implementing and undergoing chain of custody training so that we can provide the appropriate paperwork to you. We also have all of these isolates that are available and that have been stored for the past 11 years. We have worked with you and with the lab to take this information and compress it from months down to several weeks.

We think that if we are providing that information probably back to you, that that would be -- we will have to look at ways that we can also probably put that on the website on a more timely manner too so that people are seeing all of the information as it is coming out more real time. And that is the whole key, is the real time.

DR. ALTEKRUSE: I just really want you to know that we really appreciate the value of the collaboration too.

DR. FEDORKA-CRAY: It will cost you a lot of pizza over time.

DR. YOUNGMAN: If we can hold the rest of the questions, we need to press ahead, because lunch is only during a certain time. So, we want to hold Tom Chiller’s talk until after lunch, and if we can, continue with questions afterward.

Otherwise, you might not get lunch. So, we can break for lunch now. Thank you, Paula.

DR. FEDORKA-CRAY: Thank you.

DR. YOUNGMAN: And Aleta will give you some information about lunch.

(Whereupon, at 12:12 p.m.,a lunch recess was taken.)


(1:14 p.m.)

DR. YOUNGMAN: In the interest of time, are there any questions that people had from before we went to lunch, Dr. Cray question would be happy to answer them.

DR. VOGEL: I guess I raised my hand before, so I will start off here.

I guess first a compliment. Even though you criticized yourself for the lack of timeliness of the reporting, I note that the USDA report is one year more recent than the retail meat or the human arm. So, you have got 2003 on your website, whereas the other two only have 2002. So, I compliment you for that.

And I also thank you for your offer of allowing us to give you a call and ask you for a presentation of data in a different format, because I have tried to put data together that is the same for all three arms and I find it difficult. So, I will be calling you to ask for some of that data.

thirdly, just a minor kind of question I guess in the tone of uniformity of reporting. I notice on the USDA reporting you differentiate the Salmonella Typhimurium variety of Copenhagen from Salmonella Typhimurium, which is different than the other two arms.

Is there a reason why you do that and do we need to continue that? Or can we consolidate it and report it thesame?

DR. FEDORKA-CRAY: We have the -- well, we receive our serotype information from NVSL, and NVSL differentiates between the two. Historically, if you look at a comparison of data between Typhimurium versus Typhimurium Copenhagen, you will notice a tendency toward more multi drug resistance and a higher resistance levels amongst the Copenhagen.

Additionally, there are different PFGE patterns associated with the Copenhagens as opposed to the Typhimuriums. I would be less in favor of consolidating them than I would in maintaining their separate, and I would encourage the other two arms to look at differentiating theirs.

Additionally, on a number of our Copenhagens, more than so the Typhimuriums, end up being the DT104. My understanding from NVSL is that this requires one additional o-antigen differentiation and it could be done. So, I would be less inclined to combine them than I would be to keep them together or to keep them separate.

DR. WHITE: I think it is something we have to talk about because I am more inclined to do the opposite, because our pulse field shows that they are indeed the same when you look at Copenhagens and Typhimuriums. But I think it is something that we can talk about. We speciate down to Copenhagen as well, but we combine them together. But I think that is something the three arms can together and agree on.

DR. BARRETT: This is Tim Barrett, from CDC. We have had the same experience. There are two issues for us. One is that we see them the same by pulse field. The other issue is that we get data from the state health departments and they don’t all test with the 05 antigen. We don’t always know whether they have tested or not. It is not clear.

We don’t feel that we can necessarily trust that information to separate them. But Paula is another situation obviously.

DR. FEDORKA-CRAY: The other thing that I should make clear is that the slides that I present are less in number than you received as your handout. This was due -- I gave you additional information and additional examples of the way we break some data out by species and sources.

So, the addition that you are going to see is a data one. Additionally, FDA gave us some questions to consider at one point in time and you will also see those on there also, and we have tried to address those over time.

The last thing that I would like to do is that I brought with me a -- what I call our unit presentation package, which is information that deals with our activities from our unit. Our accomplishments, the publications, whether they have been submitted or whether they are actually in press with page numbers and complete journal citations and then some of our other program information and three papers on NARMS, VetNet and the CAHFSE program. I will provide that to each of the panelists.

DR. MILLER: Paula, I just want to thank you for all of the work that you have done. It is an impressive nine plus years of data collection and toughing it out and making the connections with the diagnostic labs and through NAHMS and so forth.

Perhaps some of what I am asking is in the blue folder that you are about to hand out, but maybe you can summarize since a mature surveillance system really should have some impact on public policy and public health practice. And maybe, could you reflect for us what you have done and even the other parts of NARMS and any kind of public health outcome decisions, secondary questions that have been asked that have been important with public health follow up? That sort of thing.

DR. FEDORKA-CRAY: I think that some of our endeavors have certainly been with the commodity groups in the development and publication and acceptance and putting into practice actually judicious guidelines. We have done a lot of work with the industries, with the particular commodity groups, providing information back to them to develop these guidelines that they would go back and use.

We have had probably not as much of an impact on policy and that -- I am a little puzzled by that from the fact when -- for instance, when the NLOH was being developed, we provided all of the information to FDA for use from the animal arm, and the animal arm was actually excluded as part of the package. The CDC human arm was used in a lot of the decision making process at the time.

I think where we have significant inroads is in our relationship with other countries, both in an international level and with other programs within the United States. We have had the CIPARS group down, and we actually had them down four or five years ago when they were going to launch their CIPARS program.

And Rebecca Irwin said that rather than reinvent the wheel, she would come down and improve the wheel, and we agreed. I thought that was very good on her part, because we were able to share all of our mistakes and some of where we would go, and I think that you see some of that reflected in the completeness of their reports.

Rebecca has had tremendous success in integrating all of the programs, primarily because she is the head of all of that up there. So they have the ability to look more at their data. They know some of our frustrations that we have undergone and they were able to go back and talk to some of their commodity groups about that.

They recognize the importance of having on-farm antimicrobial use data from on-farm as part of the big picture. They came down and spent -- on several occasions, Anne Muckle and Anne Decker came down and then now -- I forget her name. Lucy. Yes. Comes down with the data analysis.

They looked at how we were doing our database, and we talked about what would be changed in it, and they have gone ahead and done that. By the same token then, we have learned from them. I mean, we now have reports like -- we can actually punch out some of the MIC data the way they report it, and we are excited to be able to put that up, hopefully, in the 2000 report so it can reflect; you can look then at the CIPARS data and our data.

By the same token then, we were instrumental in getting the Resistvet Project started in Mexico. It was our lab that did all of the training, like for CIPARS. We trained people and they needed training in micro broth dilution and other methodologies. We shared all of that.

We did the same thing for the Mexico project and brought that lab up, trained them; went down on several occasions. So, I believe that we have had that type of impact. We have had interaction and we provided our protocols to numerous labs, veterinary diagnostic labs in the U.S., who now follow our protocols for both isolation and testing.

All of this information then is used to make decisions in some way. So, it is more of a subtle impact than a -- I think one of the things that we haven’t been good at is touting our own horn. I think we have had tremendous impact, except when it has actually come down to using data for policy.

DR. SAHM: Paula, could you briefly describe the interconnections between PulseNet for the retail food and the human isolates in VetNet? Is it 100 percent comparisons across the two systems? Or is it somewhat selective?

DR. FEDORKA-CRAY: No. Our system is 100 percent comparable to PulseNet, in that the training is the same. The protocol is the same. One of our agreements with Bali Swami Nathan, who is the head of PulseNet, was that we would be certified by CDC to run gels. So we cannot submit gels even to our own database until we are certified and receive that certificate from CDC.

The University of Penn, who will be our first sentinel site, and NVSL, who will be a second sentinel site, will need to be certified by CDC to be able to submit TET files for inclusion in our database. Our databases are set up exactly the same way, with the exception that we have also learned from their beginning history.

So we have now started naming all of our isolates a certain way and could go back then and match them with PulseNet’s isolates. But our isolate I.D. numbers will differ from PulseNet’s. We have a mechanism by which we will be ableto have cross talk between those so we can identify human and animal.

I think some of the differences is that -- well, I think the retail arm does all of their isolates for PFGE, and we will do all of the slaughter isolates initially because of monies, but right now limit it to all of the slaughter isolates. And pending monies and personnel increases, then we will do diagnostic isolates, although we have done those already on special requests. Particularly from FSIS as an interest in SE or Kentucky or some other serotype.

So, there is complete comparability between the two systems, and we would like to be able to have the same type of program where you could scan -- you know, there will be different levels of access. So, there will be administrative level of access, then there will be a collaborator level of access and then there will be a public level of access.

So all of the NARMS participants will have a collaborator level of access, internal then to USDA we will have an administrator level of access and then the rest of the public will have access to it too. They should be able

-- the hope is that they should be able to look for a pattern on ours; look for similar patterns on PulseNet’s.

What we hope to be able to do though and what we are already working towards is linking all of our isolates with all of the resistance data. So you should be able to pull up all of the resistance data, you should be able to pull up all of the demographic data, species, source, region; all of the descriptors like that from a public level.

DR. KOTARSKI: Paula, I am impressed with the breadth of this program. I wanted to ask you. If I understand correctly, the diagnostic sampling program is passive surveillance. What I would like to ask from you is in terms of that sampling strategy how does it meet the objectives for your program and what would you do to change the current system to meet the objectives and does it meet the objectives of the program?

DR. FEDORKA-CRAY: Thank you for the question. I think that the diagnostic isolates provide one level of information to us in meeting our objectives in how resistance might be arising, because those isolates from the diagnostic lab tend to be more resistant than we see from slaughter. So, if you are looking at resistance emerging, you would want to look at diagnostic isolates first I think, because that is where we see most of the multi resistance especially.

The other thing is if we talk about the concern about transfer of genetic elements between organisms, I think that is also going to be one of the first places that we probably catch these multi drug resistant plasmids that might be moving, and simply because when you use -- when you treat animals, then you know that you have the ability to maintain a resistance population more so than in healthy animals.

So, I think that on some level it does meet. What we did when we started that is we sent out invitation letters to all of the veterinary diagnostic labs, and in my discussions we specifically targeted some labs. We have had -- everybody has had the opportunity to participate.

We pay $10 an isolate actually to receive those isolates, and we give them all the data back. Some states are reluctant to do that because they fear some type of regulatory or punitive action being taken against them from different places.

Some diagnostic labs don’t have the personnel to be able to participate, and some diagnostic labs have chosen to participate with other people because of friendships that may have been developed or -- and so, they don’t send isolates to us. They will send isolates elsewhere. And some have decided to just keep everything in-house and do their own work.

We have talked with Chin-Chin Woo at Purdue. Again, it all comes down to money, but having enough money to do everything, we would move toward integrating all of the diagnostic labs so that we actually wouldn’t do any testing. What we would do is compile data. We would ask them to either switch their testing methodologies to Sensititre and try to facilitate some way getting that equipment into them so that they could then move to the same type of testing so that we have continuity of testing between all of the protocols.

Or we would ask them to send us their data or send it to a central repository where we can then access it and then we can do an analysis. If they are still doing

Kirby-Bower, we know how that compares to broth microdilution or to agar dilution or to E-Test and then do an analysis that way.

Under the idea situation APHIS would facilitate that, and we actually have a list with Chin-Chin and Dave Dargantz at APHIS. That has all been compiled. We have methodology complied from the majority of these. We have a lack of funds to implement something like that at this particular time, but that is what I would like to implement.

DR. KOTARSKI: A follow up question is between the sentinel sites from which you get diagnostics and the NVSL isolates that you receive, is there a possibility of duplication? Do those laboratories also contribute to NVSL?

DR. FEDORKA-CRAY: Those laboratories do contribute to NVSL, and what we do is we do not take isolates from those states. For instance, within New York, New York has all of the New York isolates. They will have some isolates that come from Pennsylvania, if they don’t go to Penn. From the northern region of Pennsylvania. They will have some Vermont isolates that come down there too, particularly from the state public health lab. We have some codes that we can look so that we try not to get those at all. From New York in particular, because that is the one state where we get 500 isolates from. They get about 500 isolates a year from their diagnostic lab.

Otherwise, we do not -- like we would never go to NVSL and take an isolate -- we can actually see what state they are from. We would never take an isolate from New York.

Now, what that might do is preclude us from taking an isolate from a smaller diagnostic lab in New York that might be submitting isolates. We are trying to exclude all of the Cornell isolates. This way we are also excluding some other isolates from New York too, but we feel that is safer than get duplication.

So we believe, to the best of our ability, we have voided duplication of isolates from the diagnostic labs.

DR. KOTARSKI: One more question. For the diagnostic isolates then are you getting samplings that is more representative of animal populations? Are you getting sampling that reflects outbreaks? Are you getting sampling strategy that ultimate reflects human population dense areas? Or do we know?

DR. FEDORKA-CRAY: I don’t think that I’ve done -- I mean, I know I haven’t done that kind of an analysis and I don’t think that we can say. What we do try to do is that -- we know that NVSL gets in "x" thousands of isolates. We actually get these in printouts. We don’t get it electronic. We get stacks of paper like this.

We go through and cross out all of the states that we don’t -- we aren’t going to take isolates from at all. Then we look at how many isolates are left and then from there we try to do a random, every fifth. And then we also look down in there, because you will noticed that some diagnostic labs will send NVSL.

For instance, 50 isolates on one day that may all come from the same herd, and you can actually see the accession number on there and you can see the accession date. What we will do in that instance is that instead of going every fifth or every 10th isolate or however our selection is for that particular commodity, we will actually look down through those 50 then.

We will see if any of the serotypes deviate. If none of the serotypes deviate, we will take no more than two from that one submission, and we do that simply to make sure that we don’t see differences between the isolates because we know that just because the serotypes are the same it doesn’t necessarily mean that the isolates are the same.

Then we will go to the end of that submission and then start again our every fifth or 10th or 20th isolate that we are taking. So, I think we got to some fairly extensive and extraordinary manual means to insure that we don’t over represent one submission or over represent a particular state or an outbreak.

DR. MILLER: Along those same lines of the diagnostic samples, you mentioned the rationale for continuing to collect and analyze them is they can be an early warning of emergence of resistance that will later show up in the food chain. Have you actually had examples of that where you have seen it or you were suspecting something was emerging and then you saw it later in the food?

DR. FEDORKA-CRAY: I will say that we have not used that as much as we could have in the past, but we are starting to use that and now paying more attention to what the serotypes are that are coming out. In addition, we are on a list serve for the American Veterinary Microbiologists, and they will send around from the diagnostic labs that, gee, they are starting to see, for instance, a significant increase in Uganda and are we seeing this.

And we will go back and take a look at it and see if we are starting to see this too. We really need to pay more attention to that over time, because when we go back and do retrospective analyses now, we could say, yeah, it was a good question. But yes. It is something that we need to improve.

Sue, did you have another question?

DR. KOTARSKI: No. That is fine.

DR. YOUNGMAN: Thank you very much.

DR. FEDORKA-CRAY: Thank you. And I will pass out these blue folders for each of you. I have a few extra if someone wants one.

DR. YOUNGMAN: I would now like to turn the podium over to Dr. Tom Chiller, from the Centers for Disease Control, who is going to be talking about the human arm of NARMS.

Back to Top