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

Animal & Veterinary

  • Print
  • Share
  • E-mail

Challenges in Assessing and Regulating the Risk of Antimicrobial Use

Dr. Stephen Sundlof

DR. SUNDLOF: In the interest of time, I think I am going to go ahead and start my talk. Most of -- fortunately, I am going to try and move through this fairly quickly. One of the reasons is that most of the things that I was going to say, others have said. And I am talking about challenges in assessing and regulating the risk of antimicrobial use.


And I think just from the questions that have been raised this morning, people are pretty well in tune to some of the challenges that we face. First of all, risk assessment is something that I think the U.S. Government and the world government is beginning to embrace as a very useful process, a more precise process. It gives you better definition of the risk. It is a transparent process as you heard this morning. And it is being embraced I think on a worldwide basis.


But having said that, there really to our knowledge is not a -- there doesn't have a good history in terms of how these have been applied in terms of regulatory situations in the past. So we are really breaking new ground here by looking at a risk assessment and then trying to see how that might fit into a regulatory scheme. All of the regulators are talking about it. Nobody has really done it yet. So it is brand new territory.

We learned that there is not very many microbiological risk assessment models out there. Maybe half a dozen talked about Salmonella enteritidis, E. coli and Listeria as being some examples of recent risk assessments and the pros and cons of those and where their short-falls were. So it is a brand new area. It has great promise. But we are not really sure how we are going to incorporate these into the regulatory process yet.

The President's Food Safety Initiative certainly speaks considerably to the issue of risk assessment and that in order to help protect the food supply, that we need to be doing a lot more in government with risk assessments. Again, interesting, we are not really sure where we are going to go with those. And so that is going to be one of the great challenges in the upcoming years, is how do we actually use those risk assessments.

And I think there is agreement that the issue of antibiotic resistance as it relates to animal agriculture is a growing concern on a worldwide basis and not just in the United States. We are assuming that resistance develops from the use of antimicrobials, the transmission for the food-borne organisms that we are talking about, especially Salmonella and Campylobacter, are generally not from person to person and that the most likely source is from animals.

These are the assumptions that CVM is operating under at this time. And that certain antimicrobials are used empirically to treat patients that have developed food-borne bacterial infections. And so we have to consider all of those in the mix as to what is going to be the best public health policy.


The model does assume that resistance is due to antimicrobial use in animals. That was one question that was raised this morning during David Vose's talk. And there is evidence, there is epidemiologic evidence that seems to point in that direction.

Obviously, any additional information that we can get will help us in determining whether or not that is the right approach. But presently I think that the weight of the scientific evidence clearly points to the use of these drugs in animals as the cause of the resistance.

Incremental health risk to consumers from compromised therapy is the harm, one of the harms that we are talking. We say incremental risk. And what we are talking about there is that people are already ill at the time. And then failure of treatment results in prolongation of their -- of the disease that they already have. And that is what we are considering as the incremental increase in risk.

And how do you model that? What is the best way to model that? In the model, only the risk from the use of fluoroquinolones in chickens is assessed. There are all kinds of other antibiotic microbials. There are several different diseases of importance that may be implicated if there is resistance from the use of these drugs in animal agriculture.

And we have a pretty good example of Campylobacter. We actually had some access to some good data. What about some of these other ones? Can we apply the same kind of approach to other ones?

I can say that the risk assessment model really did help us to focus on what the critical issues were. And it helped us understand better the scientific limits than if we hadn't done the risk assessment. So the risk assessment, and I can say from CVM's point of view, was a very much a belying experience. We think we benefitted greatly from it. It has changed I think substantially the way we think about assessing the harm that may be due to use of animal drugs.

The mathematical part of the model, as David Vose indicated, is simple and it can be updated. And even though we have a lot of uncertainty within the model even at this point, that additional information can help reduce those uncertainties so the model can be a living model. It can learn as we obtain more information.

But it also has its limitations. And you heard about many of those limitations today, especially where we need additional information and some of the assumptions have to be studied a little bit more carefully.

Well, what is CVM facing then based on the risk assessment model and dealing with the whole entire issue of antimicrobial resistance in food-producing animals? Well, first of all, we have to develop a quantitative definition of acceptable level of risk if we are going to use a quantitative risk assessment approach.

I think one of the speakers earlier this morning said -- I think it was Doug Powell said that we had gotten away from talking about zero risk. And I think that has been a very important movement for the United States and a lot of the other countries, as well. Nothing is risk-free. I think we will all agree to that.

But once you have said that, then it is important to ask the next question, well, then what is acceptable; what is an acceptable level of risk. Many of the international treaties, especially things like the WTO's phyto-sanitary agreement, talk about the concept of acceptable level of risk within sovereign nations. How do we define that? What is an acceptable level of risk. These are issues that we are going to be struggling with.

Determination of the human health impact, we talked about the assumptions that were made in the risk assessment model about that -- and the assumptions were made that there would be prolonged public harm simply because people did not benefit from the drugs that were administered.

Is that the correct assessment? Are there better ways of assessing the human health impact? Are there other end points that we haven't thought of that might be more sensitive, might be better indicators?


The model can define -- you know, how do we define harm within the model? Is it just simply from resistant bacterial infections in people? Is it resistant infections in people that receive antibiotics?

Is it resistant infections in people that receive antibiotics and have an adverse effect that is measurable like prolonged illness or is it resistant infection and also those people receiving antibiotics that experience an adverse effect and for which there is no alternative drug treatment?

When we wrote the Framework Document, that last bullet there pretty much describes those drugs in Category 1 and drugs for which there are serious illnesses in people and for which there are no good alternative drugs. Those are the ones with the highest priority. So just defining what we mean by harm is going to be critically important in moving forward toward regulation.


Okay. And then David Vose talked about this, but -- showed you this slide about depending upon what the denominator is, the risk is different. So you can have one in 61,000 if you consider the entire U.S. chicken-eating population. Your chances as a normal citizen of being affected if you eat chicken is one in 60,000, versus the population that develops Campylobacter, versus the population that develops Campylobacter and seeks medical attention.

And so we have to make a decision as an agency, what is the proper denominator. Traditionally, I can tell you that FDA has spread the risk over the entire population. When we talk about the risk of cancer, we are talking about a one in a million risk of cancer for all citizens.

Recently, with the EPA, they have the new law, the Food Quality and Protection Act which looks at sub-populations, looks at women and children. Are we moving in this area? These are public decisions, public health decisions, policy decisions that at some point we are going to have to come to grips with.

And so part of the reason for having this meeting is to try to get some of these concepts out on the table, have people thinking about them. And David showed you these and how you would map that risk. And it shows the uncertainty or the confidence with which those point estimates were made. So we will go through that quickly.


And I will get to our conclusions then. So there is a clear difference. We are in a transitional stage in which we are shifting from risks that we traditionally have dealt with for chemical residues. And we are shifting to a different kind of risk which is antimicrobial resistance or just microbial contamination in general.

Very, very different. Very much more complicated. We are going to need all of the help that we can get in trying to get our hands around this issue.

The framework attempts to provide a mechanism to deal with this nontraditional risk. The risk assessment has helped us further along down that road. And we look forward to a lot of participation in helping us struggle with some of these very difficult issues. Thank you.