• 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

Alternatives to Establishing Thresholds

Dr. Richard Carnevale, Animal Health Institute

January 24, 2001

I want to thank CVM for inviting me to take part in the program today. Since I represent the Animal Health Industry, I have been tasked with talking about alternatives to the CVM threshold concept we have been discussing in some detail the last two days.

Before I approach the topic of alternatives, I would like to make some general comments on the proposed CVM threshold approach. I appreciate the effort CVM has undertaken in an attempt to develop an objective means for regulating antibiotic resistance, however, because of the complexity of the issue we believe CVM will find it difficult to apply such a precise approach.

I am not a risk assessor and, truthfully, I have had a very difficult time understanding the concepts and mathematical relationships expressed in the document. But I do know that the threshold concept is grounded in the risk assessment model CVM developed for Campylobacter in chicken and, therefore, suffers from some of the problems that AHI has identified in that model. From the comments at this meeting I sense that the agency is having difficulty understanding the ramifications of the document as well.

We have several concerns with the model being proposed. CVM suggests that the value of this approach is that it is simple. But the process that they are attempting to model is far from simple. Therefore, we question whether it will actually reflect real conditions in both exposure and in medical impact.

  • It treats all chicken the same in that it doesn’t take into account that not all chicken is delivered to the consumer in the raw state. A large proportion of chicken is pre-cooked which virtually eliminates any risk of pathogen bacteria.
  • It treats any level of bacteria as equally capable of causing disease. We know that the infectious dose may differ for different types of bacteria. Because only the percentage of samples with resistant bacteria are considered, prevalence estimates will dependent on what units of measurement are used and not on the actual level of contamination.
  • It bases a threshold determination on a database that has not been statistically designed to represent a true national prevalence of pathogens in meat and poultry. While AHI strongly supports the NARMS program, it was never intended to be used to establish hard regulatory thresholds.
  • It assumes an absolute relationship between resistance in animal isolates and resistance in human isolates. As we have heard at this workshop, this relationship is not clear – certainly not clear enough to set an absolute “bright line.”

Because of these shortcomings it is our opinion that the tight proportional relationship CVM has proposed to arrive at a resistance percentage threshold may not at all result in regulatory limits that are truly indicative of human health risk. We are dealing with biological systems with a great deal of uncertainty and many assumptions. The analysis frankly seems more appropriate for a project in engineering where we can precisely measure the relationships between the various factors.

It was stated by CVM that exceeding the threshold would cause immediate withdrawal of products or species claims. We don’t know how CVM would accomplish this as we don’t believe FDA has the authority under the FD& C Act to deny a company due process to challenge the action, except in the case of a declaration by the Department that an imminent hazard exists. We are not aware that CVM has suggested this to be a hazard of this magnitude. If a company is voluntarily asked to agree to such thresholds prior to approval then it will clearly be a major disincentive for them to invest the millions of dollars and years of development and research into a product that can be summarily withdrawn on likely arbitrary grounds.

For these reasons, AHI is against the establishment of thresholds that will be used to decide on whether a product remains on the market or is withdrawn. This is not to say that AHI wold necessary object to the agency establishing some level of resistance as a guideline that triggers further investigations.

It appears from comments at this meeting that both public interest groups and the industry are both against thresholds but clearly for different reasons. On the one hand some believe that no thresholds can be set for certain antibiotics of importance to human medicine, and therefore these drugs shouldn’t be approved at all in food animals. As far as the industry is concerned they don’t believe that such hard numbers can be set which are other than arbitrary and which could jeopardize millions of dollars worth of product development in a very short time.

So what can be done? I don’t come to you with any magical solutions because there aren’t any. However, I believe that we must all recognize that antibiotic resistance may not fit the usual paradigms that FDA operates under to regulate chemical residues. The intent of the agency to establish resistance “tolerances” is clearly problematic.

AHI believes that a more reasoned approach to controlling antibiotic resistance relies on a careful evaluation of an adequately designed and conducted surveillance system accompanied by certain regulatory controls and judicious use initiatives:

  1. We should strengthen the post- marketing surveillance program, NARMS, to include additional antibacterials, pathogens such as Campylobacter and, if possible, redesign the program to represent a national prevalence. If possible, NARMS could also be structured so that regional data could be gathered and analyzed to examine geographical differences. We cannot the support the use of retail sampling as has been suggested as the basis for judging prevalence of resistance from animal drug use. There are too many other opportunities for contamination post-slaughter plant that would bias the results.
  2. Establish official resistance breakpoints that correspond to likely clinical outcomes in humans for all important foodborne bacteria
  3. Appoint an independent expert third party group composed of microbiologists, veterinarians, epidemiologists, and other public health experts to evaluate the NARMS data on a periodic basis and advise FDA of trends in resistance that may be of concern. This group could recommend to FDA that follow-up investigations to determine the cause of the resistance trends might be indicated. We don’t understand why the immediate reaction of the agency should be to propose banning a product. If residue violations occur, the agency investigates the cause of the violation and takes appropriate action on the violator. For resistance, the NARMS database is small. It may be that an increase in resistance prevalence may have been due to isolates from one region of the country. This is where regionally derived data could be useful.
  4. Given a properly developed risk assessment model, the resistance data could be plugged in to the model to determine what level of risk the observed resistance is likely to represent. Based on a determination of whether the risk appears to be low, moderate, or high, then the agency will be in a better position to determine if an action is needed along the lines of altering the product label, restricting certain uses, educational efforts, or in the worst case, temporary or permanent product removal.
  5. Develop a comprehensive program in conjunction with the AVMA, industry and producer groups to embed prudent use guidelines into the usual practice of veterinary medicine. Educational programs in veterinary schools and colleges of animal science should be strongly considered.
  6. Therapeutic antimicrobials are restricted to use by or on the order of licensed veterinarian or through a veterinary feed directive. The industry understands that drugs that have importance to human medicine may also face other restrictions such as prohibiting extra-label uses.

These are some brief thoughts on the subject. I recognize that this will not satisfy those that want some more absolute action but I don’t believe there is a quick fix. We need to emphasize that the threshold concept is being directed to therapeutic veterinary prescription drugs. I have heard a lot of agreement at this meeting that therapeutic antibiotics are necessary for veterinarians to treat animal disease. It seems to me that we should be able to evaluate the continued microbial safety of these compounds without having to establish unyielding standards that place both the agency and the industry in inflexible positions.