Скачать 0.86 Mb.
The Animal Drug Approval Process for Antimicrobial Agents|
DR. MILLER: Good morning. I am Dr. Margaret Miller. I go by "Peggy." I am Deputy Director for Human Food Safety and Consultative Services in the Office of New Animal Drug Evaluation at CVM.
What I want to do today is talk a little bit about the studies that we require in the approval of a new animal drug, new antimicrobial drug; how we evaluate these studies; and how we use these studies to make a prediction of whether or not the product is safe; and then talk a little bit about how we could apply these techniques or similar techniques to making a determination about the safety in the microbiological area.
Before any new animal drug is approved for use in the United States, the drug sponsor must have an approved new animal drug application. In the new animal drug application the drug sponsor provides data to show that the drug is efficacious, that it is safe for the target animal, that it is safe for the environment, and that it can be manufactured to uniform standards of purity, strength and identity. If the drug is going to be used in a food producing animal, the drug sponsor must also provide data to show that the drug is safe for humans.
In the area of environmental safety the agency uses an exposure threshold approach to determine when environmental fate and effect testing are needed. Environmental studies are not needed for compounds that have limited environmental introductions. When an environmental assessment is needed the drug sponsor conducts laboratory toxicity studies and in vertebrates, plants and microbes representative of the environmental compartment of concern. The no observed effect level, or MIC in the case of the microbes, is divided by a safety factor to arrive at a predicted environmental no effect level.
The predicted environmental concentration of the drug is then calculated, and we compare the predicted environmental concentration, which is referred to as PEC, with the predicted environmental no effect level to come up with a PEC/PNEC ratio. If this ratio is less than 1 the agency concludes that the compound is safe for the environment or that there will be no significant environmental effects from the use of the drug.
To determine the human food safety of residues of an antimicrobial product the drug sponsor conducts a standard battery of toxicology tests. The standard battery of toxicology tests looks at systemic toxicity, genotoxicity, mutagenicity, reproductive toxicity and developmental toxicity. Information on these endpoints is required for all drugs which require an acceptable daily intake or a food safety assessment.
Additional food safety studies may be required if we have additional human health concerns. For example, if a product tends to bioaccumulate the agency might ask for chronic feeding study in order to establish a no effect level for that compound.
The toxicology studies are designed to show a dose that causes a toxic effect and a dose that causes no effect. The no observed effect level is not always a classical tox endpoint. CVM considers the development of diarrhea following treatment with an antibiotic as an adverse effect although clinically this is generally considered a side effect of the drug. The Center views the results of toxicity tests conservatively because we believe that consumers should experience no effects from drug residues in their food.
Once we have established the no effect level for all endpoints, the most sensitive effect in the most predictive species and by that we mean predictive of man is established. This no effect level is divided by a safety factor, and the safety factor takes into account uncertainty, that is, the extrapolation between the animal model and the human as well as variability, which is the difference among individuals. After dividing by the safety factor we calculate an acceptable daily intake, and the acceptable daily intake is defined as the level of drug residue that can be safely consumed daily for a lifetime.
There are special food safety concerns for residues of antimicrobial drugs. It is well known that therapeutic doses of antimicrobials can cause adverse effects on the human intestinal microflora. The agency has identified the selection of resistance, perturbation of the barrier effect, changes in enzyme activity and alteration in bacterial counts as potential impacts of antimicrobial drug residues on the human intestinal microflora that are a public health concern.
The perturbation of barrier effect is of concern because normally the gut flora prevent the overgrowth and invasion of pathogenic bacteria. When the normal flora is disturbed by an antibiotic, for example, overgrowth of pathogens can occur and infections.
While the adverse effects of therapeutic doses of antimicrobials on the human intestinal microflora have been well documented, in most cases the lowest dose at which these effects occur have not been established. Based on the literature available at the time and the advice of experts in the field, CVM established an exposure threshold for concern of 25 mcg/person/day. For antimicrobial products meeting an acceptable daily intake of greater than 25 mcg/person today the food safety evaluation must include an examination of the effect of the drug on the human intestinal microflora in addition to the standard battery of toxicology tests.
Recognizing that model systems used to evaluate the effects of antimicrobials on the human intestinal microflora were only research methods, CVM funded research to validate an in vitro human fecal culture system and a human flora associated mouse model. Many of the techniques developed for validating these model systems, especially those to look at the development of resistance and the disruption of the barrier effect, can be applied to assess the development of resistance and changes in pathogen load in the target animals following antimicrobial treatment.
Now, as was mentioned by Dr. Sundlof, we have asked for microbial safety studies in the past for antibiotics that were administered in feed for more than 14 days. These studies, which are often referred to as 558.15 studies, were performed to look at the level of drug resistant bacteria and the level of pathogenic bacteria.
There were two studies generally performed in this battery. The first study looked at the effect of the drug on excretion of Salmonella in the feces of animals artificially infected with a laboratory strain of Salmonella. This study is referred to as the Salmonella shedding study. The other study was a coliform resistance study. This monitored the effect of the drug on the resistance pattern of E. coli present in the endogenous flora.
In the Salmonella shedding study between 7 12 animals were infected with a laboratory strain of Salmonella typhimurium which was known to accept plasmids. The animals were treated with drug for eight weeks and fecal samples were collected weekly. The laboratory strain of Salmonella was isolated from the fecal samples and examined for resistance patterns, as well as shedding quantity, duration and prevalence.
The design of the coliform study was similar to that of the Salmonella shedding study, except that the animals were not inoculated with bacteria. Rather, the effect of the drug on the endogenous E. coli was evaluated.
Now, because it is difficult to measure a change in resistance against a high background, it was necessary to use animals with less than 20 percent resistance in their endogenous E. coli. A change in coliform susceptibility between the drug treated and control groups indicated a drug effect.
I want to say that we do not have standardized protocols developed for the microbial safety studies mentioned in the framework document. However, the techniques that have been used to measure the effect of antimicrobial drugs and residues on the human intestinal microflora, together with a modification of the traditional 558.15 studies, could serve as a basis for developing protocols for these studies, and we are seeking scientific input on both the design and interpretation of these studies and feel that the protocols will be improved if we have significant public input into the process.
As discussed in the framework document, we intend to look at pathogen load issues on an exposure based threshold. Then we will determine, based on the amount of the exposure, when a drug sponsor will need to determine if their product alters the level of pathogenic bacteria.
Now, the design of the colonization resistance studies that we did in the human gut flora was similar to the design of the Salmonella shedding study, and it could serve as a prototype for how these studies would be designed to look at pathogen load in the target animal.
Basically, what we are doing in the gut flora studies is that animals are inoculated with a bacterial strain that is resistant to the antibiotic being tested. Also, inoculated bacteria has a propensity to proliferate when the barrier is perturbed. The animals are then treated with increasing doses of antibiotics and the number of indicator bacteria are measured.
One could propose that if there is a margin of safety between the dose intended for use in animals and the dose that causes a proliferation of the indicator bacteria that the product may be considered safe. Alternatively, if the indicator organism or the pathogen proliferates at the intended dose the study could be continued for a recovery period to determine the amount of time required for the endogenous flora to recover from the antibiotic perturbation.
The framework document discusses that we intend to use human health concern to determine when studies will be needed to determine resistance. The objective of these studies is to characterize the development of resistance so that we can make some prediction about the product's safety. To accomplish this, we will need to make several modifications to the traditional 558.15 studies. For example, the traditional 558.15 studies were designed like a bioequivalence study. They were designed to show no difference between the treated and control groups. In order to characterize the development of resistance it will be necessary to design the studies such that the null hypothesis states that there is no difference, and the alternative hypothesis states that there is a drug effect. This type of design will facilitate statistical analysis and improve our ability to make a prediction from the study.
The traditional 558.15 studies were done in the target species, and we suggest that the new pre approval studies should continue to use the target species. However, we believe that there need to be more numbers in order to improve the power of the test and to actually show the development of resistance, how that is going to occur.
In the past we extrapolated data from chickens to pigs to cattle. I think this approach is still acceptable provided that the first study provides a more protective standard than the subsequent species.
In the traditional 558.15 studies all the studies lasted eight weeks. It seems that in the future the treatment period may need to be extended. Basically, the study duration should be sufficient to establish a baseline level of resistance, allow for resistance development and to look at the persistence of the resistant bacteria.
In the traditional 558.15 studies animals were housed individually in separate treatment facilities. This requirement severely limits the number of animals that can be used in the study. The new study will need to look at different approaches for separating treatment and control animals, and for dealing with the problem of cross contamination.
As far as dosing, in the traditional 558.15 studies animals were dosed continuously throughout the eight week treatment period, and this is because it was assumed that for feed administration the animal would be continuously exposed to the antibiotic. For products that are intended for food producing animals by therapeutic routes the continuous administration is not appropriate. Perhaps some type of short term repeat dosing regime, using the dose and route of administration intended in the target animal, would be more appropriate. One could assume that we would do repeat dosing to cover the maximum amount of doses that an animal is likely to encounter under field conditions.
Also, in the traditional 558.15 studies fecal samples were collected weekly. In the new pre approval studies it seems that the sampling times would need to be tailored based upon the target animal species, the dosing regime and the study duration.
Finally, we come to indicator organisms. In the traditional 558.15 studies we looked at the development of resistance in Salmonella, E. coli and, in some cases, enterococci. It seems to me that having one set of indicator organisms for all antibiotics may not be appropriate. We may need to change what indicator organism we are looking at depending on the antibiotic. We might have to have drug sponsors provide a justification for what indicator organism they are choosing. Alternatively, we could look at a panel of indicator organisms as we are in the gut flora studies. In those studies the indicator organisms cover both anaerobes and aerobic bacteria.
Bacterial load issues in order to look at a susceptibility change in an indicator organism you need to have sufficient quantities of the bacteria there to make an accurate measurement. In the 558.15 studies animals were inoculated with a laboratory strain of Salmonella to ensure that they had sufficient quantities of the pathogen present to measure the drug effect.
Ideally, the study should be conducted with a more normal bacterial load. However, to ensure that there are sufficient numbers of indicator organisms present we may need to do something like use a CDER animal, or provide some other means for establishing sufficient number of bacterial in the animal.
As I mentioned before, the 558.15 studies relied on no difference between the treated and control groups to predict that the use of the antimicrobial would not affect antimicrobial resistance or pathogen load. The new studies really should be designed to characterize the differences between the treated and control groups using standard statistical procedures. In this way, we will have information that we can use to make some prediction about the likelihood of resistance development and transfer to humans.
I want to reemphasize that there will be numerous opportunities for comment on how these studies should be designed and interpreted but, conceivably, we could develop a safety assessment, a risk assessment process similar to that used to do safety assessments in the area of environmental and residue. For example, we could look at the level of resistance development seen in the pre approval study and compare that to a threshold level in order to make a prediction of safety. The threshold level then would represent the level of resistance that causes an adverse public health outcome.
So, to summarize then, we have seen that the use of antimicrobial drugs in food producing animals represents a public health concern, both in terms of the development of resistant bacteria and in pathogen load.
The framework document lays out an approach for when we would look at the studies to address these different areas and, as I have just talked about, one way of trying to do predictions in this area would be to apply the safety assessment procedures used in other areas, to make a modification of that to look at the public health and help ensure product safety.
DR. STERNER: Do any of the panel members have questions of Dr. Miller? Dr. O'Brien?
DR. O'BRIEN: I would just make one comment. One difficulty with this general type of study is that if one looks back at the antimicrobial agents that did cause selective overgrowth of resistant bacteria that came over the years to cause this problem, for almost none of them would it have been detected at the time when the drugs were new.
The problem is that the antibiotic resistance genes development is a considered effort of the world's total bacterial populations apparently, and it sometimes takes years or decades for the resistance gene to emerge. Then, after that does happen the selection process by the agent is quite different than it was before.
So, the general problem and I don't know how one could approach it in testing a new agent is that in any experimental model when the agent is new the resistance gene is unlikely to exist and, therefore, the new agent will have no selection for resistance strains. There is nothing to select. Again, I think this has to be at least recognized as a general problem for new agents. And, the general issue that runs throughout this is that it is hard for us to know what the bacteria are going to do.
DR. MILLER: Yes, I don't think that pre approval studies can supplant the need for continuing monitoring, and Dr. Tollefson will talk about monitoring in a minute. But I do think that they can provide us some information about what we should be monitoring; what indicator organisms we should be looking at. And, I do think that if resistance develops in a very short or relatively short time frame, I would have some real concerns about recommending approval of that product. So, without this type of information I can't make any predictions that can help even in following this along.
DR. STERNER: Other questions? Steve?
DR. BARKER: I would like to agree with Dr. O'Brien's comments that, indeed, it is the entire population of bacteria globally that has to be considered as well, and I am sure at some point we will address imports.
The environmental aspects of the approval for antibiotics, the environmental safety studies that are done for microbes currently address the MIC picture. Given that the soil and environmental bacteria that become a component of normal gut flora are exposed to a range of antibiotics through urine and feces dilution in the environment, what contribution to the development of drug resistance might environmental bacteria be adding to the picture, and is anyone examining that?
DR. MILLER: I think the way we are looking at that, and I just briefly alluded to it on the slide, is cross contamination issues. If we bring clean animals into a dirty facility for subsequent dosing, you know, are they then picking up resistant organisms from the environment? I mean, I am open to suggestions as to how to address all of these issues, but we thought that might be the most convenient way.
The traditional environmental fate and effect studies look at the actual drug entity. So, we haven't gotten into environmental effects of the organisms. That would be handled under these pre approval studies in the microbiological area. I am looking at it as an environmental cross contamination issue.
DR. BARKER: Just to follow up, that certainly is a component of controlling your studies but I think my question goes a little bit further than that about what contribution this might have just to the general production of resistant bacteria in the environment.
DR. MILLER: So, you are suggesting that as part of the environmental safety studies that we not just look at MIC values but we look to see whether we are selecting for resistant organisms, resistant soil microorganisms?
DR. BARKER: It is just another question of what the use of antibiotics and their effect in the environment generation of resistance, not only in the animals that are actually treated with the drugs but the bacteria that are in the environment that eventually become part of the normal gut microflora of these animals, what effects these drugs may be having there, and how that might be assessed as part of the overall picture.
DR. STERNER: Thank you. We have to draw this to a close. Dr. Linda Tollefson, from the Center of Veterinary Medicine, is going to discuss national monitoring surveillance issues.
|Israel journal of veterinary medicine vol. 63 No. 1 2008||Ranch hand advisory committee|
|Advisory committee on immunization practices||Medical Devices Advisory Committee|
|External Advisory Committee on Cities and Communities||Wildlife Diversity Policy Advisory Committee|
|National Vaccine Advisory Committee (nvac)||Peer reviewed by the Arizona Department of Commerce Economic Research Advisory Committee|
|College of veterinary medicine north carolina state university m. Christine McGahan, Professor and Department Head||College of veterinary medicine north carolina state university m. Christine McGahan, Professor and Department Head|