• 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

Role of NARMS in Assessing Risk of Antimicrobial Agents in Food Animals

by Dr. Jeffrey M. Gilbert, Supervisory Microbiologist and Team Leader on Microbial Food Safety Team, Office of New Animal Drug Evaluation; Dr. David G. White, Director, Division of Animal and Food Microbiology, Office of Research; and Dr. Patrick F. McDermott, Team Leader, CVM’s NARMS program, Office of Research
FDA Veterinarian Newsletter 2007 Volume XXII, No VI

An undesired consequence of antimicrobial use in animals is the development of antimicrobial-resistant foodborne bacte-ria that are human pathogens and the subsequent transmission of those bacteria to humans via food. To help address this concern, the Food and Drug Administration’s Center for Veterinary Medicine in 1996 launched the National Antimicrobial Resistance Monitoring System (NARMS) for enteric bacteria (bacteria found in the digestive tract). NARMS is part of a strategy to assess the effect of antimicrobial use in animal agriculture on the evolution of antimicrobial resistance in hu-man clinical bacterial isolates.

NARMS was designed to generate data on antimicrobial resistance trends in enteric bacteria from food animals, foods of animal origin, and humans. The data are used to inform physicians, veterinarians, and public health authorities on anti-bacterial drug resistance levels, including new or atypical patterns of resistance. The information can be used to design epidemiology and bacteriology research studies to understand the emergence and dissemination of resistance. In addi-tion, the information is important for the development of public health recommendations on the use of antimicrobial drugs in food animals and humans. NARMS provides a national repository of isolates for use in research such as diagnostic test development, discovery of new genes and molecular mechanisms associated with resistance, the study of mobile gene elements, and the study of bacterial virulence and colonization.

NARMS researchers test isolates of the foodborne bacteria Escherichia Coli, Salmonella, Enterococcus, and Campylo-bacter from animals, humans, and retail meats to look for changes in the bacteria’s resistance to antimicrobial drugs that are important in human and veterinary medicine.

NARMS is a national monitoring program that combines the activities of CVM, the Centers for Disease Control and Prevention (CDC), and the U.S. Department of Agriculture (USDA). These three components collect data on bacterial iso-lates from retail meats, human clinical cases, and food animals, respectively.

Retail meat samples are collected from grocery stores in states (California, Colorado, Connecticut, Georgia, Maryland, Minnesota, New Mexico, New York, Tennessee, and Oregon) that are participating in CDC’s Foodborne Diseases Active Surveillance Network. The laboratories in those participating states forward the isolates recovered from the retail meat samples to CVM’s Office of Research in Laurel, MD, for further analysis. Participating state and local health departments from all 50 states send isolates from human clinical cases to the CDC National Center for Infectious Diseases in Atlanta, GA, for testing.

USDA employees gather bacteria samples from healthy farm animals, from carcasses of food animals at slaughter, and from clinical specimens collected from animals undergoing medical treatment for infections. In addition, USDA collects samples from ground meat products at federally inspected slaughter and processing facilities. Researchers at the Agricul-tural Research Service/Antimicrobial Resistance Research Unit of USDA in Athens, GA, test the isolates from animals and from slaughter facilities for resistance.

Molecular epidemiology of resistance genes

Information about antimicrobial resistance phenotypes derived from the NARMS program does not indicate which of several genetic elements may underlie resistance. Therefore, researchers must characterize transmissible resistance genes at the nucleotide sequence level to determine the nature and extent to which specific gene transfer occurs among different bacteria, the consequence of selection pressure in the drug use environment, and the spread of resistance through the food-production continuum.

Genetic studies provide data needed to develop risk-assessment models and to aid in the regulatory decision-making process with regard to antimicrobial use in food animals. Transmission to humans via food is implied when the same genes are found in animal, food, and human isolates. Researchers from all three NARMS components (CVM, CDC, and USDA) perform genetic studies focusing on resistance mechanisms relevant to approved animal drugs, those conferring resistance to important classes used in human medicine, and unusual resistance phenotypes among isolates to charac-terize the resistance genes at the nucleotide level.

Researchers use a process called pulsed-field gel electrophoresis (PFGE) to characterize Salmonella and Campylo-bacter isolates received from the NARMS program. The PFGE process allows researchers to identify bacteria subtypes by creating a genetic “fingerprint” of the individual strain. Researchers use this information along with antimicrobial resis-tance data to determine how different strains are related, whether they came from the same sources, and how they were disseminated.

The information on the genetic fingerprint is also used to populate the “PulseNet” database. CDC created the PulseNet program in 1995 in cooperation with the Association of Public Health Laboratories and state public health laboratories. Public health authorities can use the information in PulseNet in their epidemiological investigations of foodborne disease outbreaks to determine if different cases of foodborne illness are related.

Also, researchers have exploited NARMS isolates to investigate novel molecular typing tools to help determine the animal origin of foodborne bacterial pathogens. To date, more than 2,000 isolates of Salmonella and Campylobacter have been characterized using a combination of two or more biochemical and genetic typing methods. Results from these tests indicate that certain bacteria serotypes are associated with only certain food animal groups. Antibiotics susceptibility pro-files have shown certain resistance phenotypes to be occurring with particular animal hosts. And PFGE profiles coupled with antibiotic susceptibility profiles and additional genetic tests have shown specific strain types associated with a par-ticular animal host. These data can help attribute different resistance profiles within a specific animal host, making it pos-sible to identify sources of resistance.

Public health authorities have addressed several important epidemiological issues using NARMS data and the NARMS isolate collections, including the issues of “burden-of-illness estimates,” case-control studies, emergence of new pheno-types, and antimicrobial resistance trends. Data from NARMS has helped to explain trends in antimicrobial susceptibility among Salmonella, Campylobacter, and Shigella; estimate the public health burden due to antimicrobial resistance in Salmonella and Campylobacter; and identify risk factors for Campylobacter infection. The data have also helped re-searchers understand the epidemiology of resistance in rare Salmonella serotypes and the emergence of resistance to the antimicrobial ceftriaxone.

To bolster NARMS data used to support FDA risk assessment models and to better understand causes of resistance, researchers have conducted studies to evaluate the effects of antimicrobial use on the evolution of resistance in food-borne bacteria residing within the target food animal species. These studies include longitudinal on-farm studies and ex-perimental animal-group studies that include a control group.

An example of how NARMS data have been used in a risk assessment model was the use of the data to indicate a rise in resistance of Campylobacter to the antimicrobial fluoroquinolone. That finding prompted research designed to directly measure the impact of fluoroquinolone use in broilers, which are a major reservoir of Campylobacter. This type of targeted research is a byproduct of the NARMS program and is needed to fully evaluate NARMS phenotypic data.

NARMS data and antimicrobial drug applications

In 2003, CVM updated its regulatory policy to include a microbial food safety assessment for all new antimicrobial prod-ucts proposed for use in food-producing animals. FDA published Guidance for Industry #152, “Evaluating the Safety of Antimicrobial New Animal Drugs with Regard to Their Microbiological Effects on Bacteria of Human Health Concern”. This guidance, which provides a framework for approaching microbial food safety assessments in an organized manner, does not cause FDA to limit its consideration with regard to microbial food safety. In accordance with the Federal Food, Drug, and Cosmetic Act, the Agency’s decision regarding whether to approve a new animal drug application is driven by factors that include: (1) whether the application includes adequate tests to determine whether the drug is safe, (2) whether the results of these tests show the drug is unsafe or fails to dem-onstrate the drug is safe, or (3) whether, based on information either in the application or otherwise available to the Agency, there is sufficient information to determine that the drug is safe.

The guidance contains non-binding recommendations to sponsors concerning the approval of antimicrobials for food-producing animals. It provides sponsors of antimicrobial new animal drugs with an example of what would address FDA’s concerns about emergence and selection of antimicrobial resistant bacteria in or on food-producing animals as a result of the use of a sponsor’s drug product in those target animals. FDA is concerned that public health may be adversely af-fected, as humans are exposed to foodborne bacteria that are pathogens to human and that become resistant to antim-icrobials used to treat illness in humans.

The Guidance for Industry is comprised of subparts which, when considered as a whole, constitute a microbial food safety assessment.

The first part of the assessment is a hazard characterization. Presenting the Agency with a hazard characterization is a good choice for sponsors to consider if they have a product that has been previously approved and they are proposing only minor changes to the conditions of use (or original conditions of approval), such as adding a new claim, defining a bacterial organism in the indication, adding a new target animal class, or making minor changes to excipients.

The hazard characterization normally contains basic information on:

  • Specifics of the drug (chemical class, structure, mechanism of action, spectrum of activity, etc.);
  • Antimicrobial resistance information (species and strains of bacteria of public health concern, and phenotypic/genotypic resistance characteristics associated with the identified bacteria of public health concern); and
  • Data gaps that might be of interest to the overall picture and the extent to which they are relevant.

Full microbial food safety assessment

Release Assessment: If the hazard characterization presented is not sufficient or substantial gaps in the data exist be-tween what was presented by the sponsor and what FDA may have concerns about, a full microbial food safety assess-ment (i.e., qualitative risk assessment) may be required. The assessment is composed of three parts. The first part of the full microbial food safety assessment is a sub-assessment referred to as the release assessment.

Here, sponsors will provide information from a variety of sources to answer more in-depth questions, compared to the hazard characterization, about their product, including:

  • Active/inactive ingredients
  • Conditions of use
  • Drug description
  • Mechanism of action
  • Mode of antibacterial action
  • Spectrum of activity
  • Specific susceptibility data
  • Pharmacokinetics and pharmacodynamics of the drug
  • Additional effects of the drug (first-exposure effects, post-antibiotic effects, sub-minimum inhibitory concentrations ef-fects, concentration and/or time-dependent effects, etc.).

Information on resistance mechanisms and genetics, such as known mechanism(s) of resistance in animal and human pathogens (e.g., antimicrobial inactivation, alteration of the drug target, reduced uptake, efflux of the antimicrobial drug, etc.), and the transmissibility of resistance determinants (e.g., plasmid-mediated or chromosomal; present on transposon, integron, or phage) should be supplied.

The occurrence and rate of transfer of resistance determinants should also be described if known.

Sponsors should characterize the relative magnitude of selection pressure for resistance that may exist for their particu-lar drug, including information on other antimicrobials that may co-select for resistance, and information regarding cross-resistance to other antimicrobial drugs approved in veterinary and human medicine.

Finally, information on baseline prevalence of resistance should be provided.

Sponsors should describe available epidemiological data on existing resistance to their drug and/or related drugs in target pathogens and commensal intestinal flora. These data may be newly generated or come from existing sources, such as current literature or other reliable surveillance sources, such as NARMS.

NARMS data can be used especially in the release assessment, since NARMS monitors changes in antimicrobial drug susceptibilities among enteric organisms of public health concern. Because NARMS provides a national source of enteric bacterial isolates that could be used for research on the characterization of molecular mechanisms of resistance, and for studying mobile gene elements, sponsors of new animal antimicrobial drugs may be able to partner with NARMS re-searchers to identify resistance traits relevant to their specific drug product. Through this partnership, they also could bet-ter understand how their drug product will affect resistance in such a pool of readily available organisms, giving them an idea of what resistance impact may occur should their product be approved and used in food-producing animals.

Exposure Assessment: Following the release assessment is the exposure assessment, the second part. It describes the likelihood of human exposure through animal-derived food products to foodborne bacteria of human health concern. Evaluating new animal antimicrobial drug microbial food safety relative to the most significant exposure pathway (i.e., foodborne pathway) is the best way to qualitatively assess the risk of antimicrobial drug use in food-producing animals.

Contemporary survey data about a contamination (low, medium, or high) of a food commodity associated with the tar-get animal species/class and the level of consumption (low, medium, or high) may support a qualitative ranking of the probability of human exposure to the given bacteria via a particular food commodity. Retail meat monitoring done through NARMS provides data at a point of exposure close to consumers.

When combined with data from slaughter plants and on-farm studies, these contemporary survey data provide insight into the prevalence of antimicrobial resistance in foodborne pathogens originating from food-producing animals.

The NARMS retail meat program can be especially helpful in determining exposure assessments. NARMS retail meat data might be used to address concerns about exposure in areas where high contamination and high consumption of food products from target animals present a particular level of risk. In addition, it is difficult to track the spread of foodborne pathogens in or on treated animals that are still on the farm, as they are processed into food, and eventually as the food products arrive at the retail point of sale. NARMS retail meat data may offer compelling insight on the extent to which pathogens are presented to humans in the retail setting.

Consequence Assessment: The consequence assessment, the third part of the microbial food safety assessment, rec-ognizes that some antimicrobials are more important for treating human infections. Thus, the use of a proposed new ani-mal antimicrobial drug is compared against a ranking of the same or similar antimicrobial drugs used in human medicine. Human antimicrobials are ranked by FDA with respect to importance to human medicine and determined to be “impor-tant,” “highly important,” or “critically important.” These rankings were established based on a set of criteria developed with input from the animal health industry, farmers, consumers, human medical community, and others. Sponsors can refer to Appendix A of the Guidance for Industry to see where their drug product(s) will fall on the importance continuum.

Integration, risk ranking, and risk mitigation

Outcomes from the release, exposure, and consequence assessments are integrated to derive an overall risk ranking of low, medium, or high. This risk ranking corresponds to default (but flexible) risk mitigation strategies under categories that describe the safest conditions of use possible for a drug (or drug class) in target animals with respect to antimicrobial resistance emergence and selection.

One of the tenets of the categorization is a need for active post-approval monitoring for resistance. The use of NARMS, which tracks antimicrobial agents representing several antimicrobial classes, is a readily available mitigation step. Fur-thermore, NARMS can easily be updated to include new antimicrobial drugs or to give priority to screening for antimicro-bial drugs of particular interest as related to a particular approval.


Sponsors of antimicrobial new animal drugs can follow the outline set forth in the Guidance for Industry to gauge how their product might be regarded as a contributor to antimicrobial resistance among pathogens of interest (and, therefore, its contribution to the hazard), as well as the risk (as a probability of occurrence of the hazard) it might present to public health. This determination is based on product characteristics, its particular conditions of use in target food animals, and the importance of the subject drug (or drugs in the same class) in human medicine.

Further, a sponsor can use the guidance to anticipate possible risk management mitigations that could be applied to their product as safeguards against resistance selection.

The qualitative risk assessment in FDA’s Guidance for Industry outlines one approach for addressing concerns about antimicrobial resistance as applicable to original or supplemental new animal antimicrobial drug applications in food-producing animals. Data gathered in NARMS can be used to help determine the public health burden posed by resistant pathogens, measure the impact of interventions, and to make informed medical and regulatory decisions.

In addition, the development of risk assessments and mathematical models for foodborne disease epidemiology is im-portant for prioritizing the use of limited public health resources. NARMS surveillance and research activities are designed to supply the data needed to inform and prioritize science-based approaches to ensuring food safety, and to minimize public health concerns with regards to antimicrobial use in food animals.