• 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

Strengths and Limitations of Threshold Approach

SLIDE 1: Thresholds for Antimicrobial Resistance: Issues in controlling anticipated harm

January 24, 2001

Robert V. Tauxe, M.D., M.P.H.
Foodborne and Diarrheal Diseases Branch,
Division of Bacterial and Mycotic Diseases,
National Center for Infectious Diseases
CDC, Atlanta, GA


SLIDE 2: Public health need for thresholds: controlling anticipated harm

  • An important and welcome discussion
  • Antimicrobials agents critical to human and veterinary medicine
  • They have a particular property – using them tends to select for resistance to them
  • Challenge: We anticipate there may be harm, which is not demonstrable when the agent is first approved. We need a strategy to recognize and mitigate this harm
  • Defined thresholds to control the anticipated harm

SLIDE 3: The threshold approach

  • Monitoring the appearance of resistance that will affect the public health
    • in bacteria in food animals
    • in bacteria in the food itself
    • in bacteria isolated from humans
  • Resistance to a group of agents of concern
  • Human infectious disease specialty needed
  • Set points that identify harm to the public health
  • Mitigation or control actions linked to those set points


SLIDE 4: Strengths of the threshold approach

  • Straightforward set points for mitigation efforts
  • Using publicly available surveillance data
  • Visible targets, available to all
  • Have two sources of surveillance data now for human health:
    • National Antimicrobial Resistance Monitoring Program - Enteric Bacteria (NARMS-EB)
    • FoodNet
  • Have potential source of surveillance data for isolates from meat, clinical animal isolates (NARM-EB)


SLIDE 5: Concerns about the threshold approach

  • Presumes that resistance is reversible
    • Bacterial antimicrobial resistance may not disappear as rapidly as it emerges
    • Argues for conservative (low) thresholds
  • Challenge in setting “acceptable” levels of harm
    • Argues for basing thresholds on resistance levels, rather than on numbers of persons seriously harmed
    • Indicators of harm based on isolates from foods or food animals provides “earlier warning” than isolates from people
  • Lack of systematic data on pathogen resistance in foods or animals, and of comparison with isolates from people.


SLIDE 6: The flow of resistant pathogens from animals to humans

  • Many important bacterial foodborne pathogens have food animal reservoirs
  • Salmonella, Campylobacter, E. coli O157:H7, Yersinia enterocolitica, Vibrios
  • Through the food chain, consumers are exposed to and infected with those pathogens
  • Use of antimicrobial agents in animals can lead to resistance in those pathogens
  • Resistant as well as susceptible strains can cause illness in the consumer


SLIDE 7: Types of harm when resistant pathogens arise in food animal reservoirs

  • Direct clinical harm. Infection with a foodborne pathogen in human is more difficult, expensive and less successful, because of resistance to the agents of choice for that pathogen.
  • Indirect clinical harm due to disease promotion: Illness with a resistant foodborne infection complicates the treatment of another infection (similar to antibiotic-associated colitis caused by Clostridium difficile). Result: More sporadic cases and larger outbreaks.


SLIDE 8: Direct clinical harm

  • Scenario: A 78 year old woman hospitalized with high fever, and diarrheal illness. Empiric treatment with fluoroquinolone has little impact. After 24 hours, with falling blood pressure, started on ceftriaxone and gentamicin. Slowly improves, but has complicated two month hospitalization. Admission stool culture yields Salmonella, shown to be FQ resistant 48 hours after admission.
  • Mechanism: Antimicrobial treatment of invasive salmonellosis or campylobacteriosis can be lifesaving. Initial choice of agents is critical. “Treatment of choice” = fluoroquinolones, new cephalosporins. If strain is already resistant to those agents,
    • Initial treatment is less likely to be effective
    • Will need to change to other agents, less effective, more expensive
    • Longer illness, more complications, higher mortality
  • Agents of concern: Treatments of choice for those infections.


SLIDE 9: Indirect clinical harm through disease promotion

  • Scenario: A child begins treatment for confirmed strep throat with amoxicillin. 24 hours later, the throat is improving, but the child has a higher fever, severe diarrhea, and abdominal cramps. A stool culture grows multiresistant Salmonella.
  • Mechanism: Low dose exposure to Salmonella often results in silent infections, as the normal gut flora restrain the impact of Salmonella. During that time, treatment for another infection can interfere with the normal flora. If the Salmonella strain is resistant, it has a selective advantage in this, and can increase and cause illness
  • Agents of concern: Those in common human medical use that interfere with anaerobic gut flora.

SLIDE 10: A low dose exposure to Salmonella in a healthy host usually results in inapparent infection

A graph


SLIDE 11: That inapparent infection can be converted to clinical illness by intercurrent treatment for another infection, if the Salmonella is resistant

A graph


SLIDE 12: Index for monitoring such harms: Direct clinical harm:

  • Person treated for infection with an appropriate agent to which the infecting organism is resistant. Could construct an index based on
    • Frequency of human treatment with “agents of choice”
    • Frequency of resistance to “agents of choice” in foodborne pathogens isolated from people OR
    • From foods or food animals
  • Likelihood could be expressed as likely number of such mistreatments among all cases of salmonellosis.
  • High risk populations, risk of severe illness, suggests need for safety factor. This harm would need to be extremely rare


SLIDE 13: Index for monitoring such harms: Indirect clinical harm through disease promotion

  • Person becomes ill with resistant Salmonella infections in association with treatment of other infections. Could construct an index based on
    • Frequency of human treatment with specific agents
    • Frequency of resistance in Salmonella to those agents
      • Isolated from humans OR
      • Isolated from foods or food animals
  • Entire population at risk
    • Likelihood could be expressed as likelihood of such an illness occurring among the entire population


SLIDE 14: Other harms, less easy to quantify

  • Because resistant strains will spread more easily in animals taking antimicrobials, contamination of food with Salmonella may increase. Result: an increase in the number of foodborne infections.
  • The selection of a resistance gene in a commensal food animal with subsequent transfer to a pathogen in the human host. Result: A newly resistant pathogen in the human.
  • Mobilization of Shiga toxin in phage by exposure of Shiga toxigenic E. coli organisms to low dose antimicrobials. Result: A new potential pathogen.


SLIDE 15: Surveillance for antimicrobial resistance and its public health impact: NARMS

  • National Antimicrobial Resistance Monitoring System for Enteric Bacteria (NARMS)
  • National sample of clinical human strains of
    • Salmonella
    • Campylobacter
    • E. coli O157:H7
    • Shigella
    • Enterococcus
  • Also collection of animal isolates of Salmonella
    • Systematic samples obtained at slaughter
    • Clinical and research isolates submitted to NVSL
  • Could add collection of samples from retail foods

SLIDE 16: Surveillance for antimicrobial resistance and its public health impact: FoodNet

  • FoodNet: A nine site active surveillance system for diagnosed cases of specific foodborne infections, including Salmonella and Campylobacter. Collaboration between CDC and State health departments, USDA, FDA
  • Special studies: Case studies to measure frequency of treatment with specific agents, frequency of hospitalizations, complications in a series of cases
  • Special studies: Case-control studies to determine risk-factors for resistant infections, including measurement of illness promotion by treatment of other infections


SLIDE 17: Arenas where direct clinical harm could occur

  • Campylobacter resistant to fluoroquinolones (FQ)
    • 1999, NARMS data: FQ resistance in C. jejuni = 17.6%
    • 1999, FoodNet – special study: 53% of diagnosed Campylobacter cases treated with FQ
    • 1999, FoodNet – special study: FQ resistant infections associated with eating poultry, foreign travel
  • Salmonella resistant to FQ or to ceftriaxone
    • 1999: NARMS data
    • 1 (0.1%) of 1499 isolates have ciprofloxacin MIC > 4 ug/ml
    • 6 (0.4%) of 1499 isolates have ceftriaxone MIC > 64 ug/ml


SLIDE 18: Arenas where indirect clinical harm through disease promotion could occur

  • Salmonella Typhimurium DT104 – ACSSuT pattern
    • 6.8% of all salmonellosis in the United States*
    • Resistant to Amp, Chloro, Strep, Sulfa, Tet
    • About 45% of cases appear as complications of treatment of other infections
  • Salmonella Typhimurium – AKSSuT pattern
    • 2.6%% of all salmonellosis in the United States*
  • Salmonella Newport – ACSSuT pattern
    • 1.1% of all salmonellosis in the United States*
    • Similar resistance to DT104
  • All three strains have known bovine reservoirs


SLIDE 19: Attributing risk to different foods and food animal species

  • Several tools available to do this
    • Distribution of sources of outbreaks
    • Case-control studies of sporadic cases
    • Comparison of subtypes of strains from humans and various animal reservoirs
  • Denmark has successfully used subtypes to attribute risk of various sorts of Salmonella across food animal species.
  • Critical need: Ability to compare systematically collected strains from foods with those found in humans.

SLIDE 20: Strengthening existing surveillance

  • In FoodNet sites, begin collecting routine data on treatment of Salmonella and Campylobacter infections, and on exposure to antimicrobial agents before was infection apparent
  • Begin molecular comparisons of similar strains from human and animal sources
  • Increase coverage of NARMS-EB for human infections, and contamination of retail foods
  • Begin collecting data on amounts of antimicrobials used in animals

SLIDE 21: Conclusions

  • Thresholds are a concrete advance to the challenges of managing the anticipated harm associated with the use of antibiotics in agriculture
  • Several types of harm, may mean different thresholds, and different denominators for each. Direct clinical harm should have the most conservative threshold
  • Indices based on isolates from food or animals may predict harm before it is documented in humans
  • Existing surveillance systems can be used to monitor the potential for harm, with some strengthening

SLIDE 22: Issue of multiple linked resistances

  • Particular challenge of the multiply resistant organism, like penta resistant Salmonella Typhimurium DT104 (ACSSuT)
  • Indirect clinical harm, means using a threshold for those agents commonly used in human medicine
  • Use of any one of several different agents in animals could provide selective advantage
  • Mitigation would need to focus on the agents commonly used in agriculture:
    • Ampicillin and other pen derivatives
    • Florfenicol (causes cross resistance to chloramphenicol)
    • Streptomycin
    • Sulfa derivatives
    • Tetracyclines


SLIDE 23: Challenges to measuring and monitoring effects

  • Need comparison of strains from humans food and animals
  • Need measures of resistance in food itself, not mixed with the clinical isolates from ill animals
  • Need data on the actual amounts of each agent used in agriculture
  • Determining the proportion of risk attributable to various food sources.