Microbial Risk Assessment Scenarios, Causality, and Uncertainty

1
Microbial Risk Assessment Scenarios, Causality,
and Uncertainty

• M. E. Coleman, B. K. Hope, H. G. Claycamp, and J.
  T. Cohen

2
It is difficult to establish causality in
biological systems

• These difficulties affect efforts to assess risks
  in microbiology
• Likelihoods of infectious disease
• Development of antibiotic resistance
• von Pettenkofers challenge
• Drank Vibrio cholerae bacteria
• Failed to become ill
• Concluded that V. cholerae didnt
• cause disease
• ?Microbial risk cannot be assessed
• on the basis of one factor

von Pettenkofer
3
Risk Assessment

• Multiple factors determine the likelihood and
  severity of adverse effects
• Ingested or inhaled dose, etc
• For both chemical and microbial hazards there is
  a widely accepted framework for risk assessment
• Four main elements (1) hazard identification,
  (2) exposure assessment, (3) dose-response
  assessment, and (4) risk characterization
• Dose-response predictions are uncertain!

4
Assessing Risk of Infectious Disease

• Why didnt von Pettenkofer develop cholera
  after ingesting V. cholerae?
• Host defenses
• Virulence and physiological state of organisms
• Environment of the flask and GI tract
• Combination of these factors

V. cholerae
5
Assessing Risk of Infectious Disease

• Formal dose-response assessment would present
  evidence for the factors that cause or control
  virulence and pathogenesis

6
Risk Assessments and Antimicrobial Resistance

• Risk Assessment frameworks provide a formal means
  for attributing causality and thus strengthen the
  scientific basis of inferences
• FDA did not have an established framework for
  determining causality of antimicrobial resistance
  over the past decade
• Issue Does antimicrobial resistance develop in
  bacteria that colonize food animals that are
  exposed to antimicrobial drugs?

7
Risk Assessments and Antimicrobial Resistance

• The Center for Veterinary Medicine (CVM)
  described a possible causal pathway for human
  health concerns from food animal uses of
  antimicrobials
• Use Monte Carlo simulations to estimate risk and
  uncertainty

Monte Carlo simulation - a method for
iteratively evaluating a deterministic model
using sets of random numbers as inputs. This
method is often used when the model is complex,
nonlinear, or involves more than just a couple
uncertain parameters. A simulation can typically
involve over 10,000 evaluations of the model, a
task which in the past was only practical using
super computers.
8
Problems with Risk Assessments in the context of
microbial resistance

• Simulation of a sequence of possible events does
  not provide sufficient demonstration of
  cause-and-effect relationships, particularly when
  this effort depends on subjective judgments to
  assign parameter values
• Although particular events may appear to be
  associated or correlated, they may not be
  causally related
• Risk analysis is an analytical-deliberative
  process one that describes probable behavior
  based on systematic testing of the possible
  behavior
• Risk assessments can provide objective
  estimates of risk and uncertainty when they
  characterize appropriate alternative scenarios

9
Microbial Risk Assessment Practices

• Efforts address a standard set of questions
• What can go wrong?
• How likely is it to go wrong?
• What would the consequences be?
• Analytic process involves
• Compiling and validating evidence and models
• Developing assumptions and extrapolations
• Making predictions for complex systems
• Assembling interdisciplinary teams whose members
  exercise a good deal of judgment

10
Difficulties inherent to conducting microbial
risk assessments

• Little guidance for providing enough transparency
  to distinguish scientific data from assumptions
  and judgments
• Hypothetical nature of many microbial risks
• Scientific data are insufficient for predicting
  many potential adverse effects from infectious
  agents of current interest
• Likelihood and rate of transmission of bovine
  spongiform encephalopathy (BSE), avian flu, and
  diseases from intentional or accidental release
  of various biothreat agents

11
So.. What should we do?

• Stan Kaplan of Bayesian Systems, Inc.
• Should elicit information from experts, NOT their
  opinions about parameter values or models
• Expert evidence could then be compiled into a
  common knowledge base for use in
• risk assessment and formal inferencing

Dr. Kaplan
12
Challenges of Assessing Risk of Infectious Disease

• Documenting accurate dose-response relationships
  is a major challenge
• Variability is understated because data on host
  defenses is not routinely available
• Making comprehensive models simple, reliable, and
  easy to update is very difficult
• Consider the risk assessment for bovine BSE that
  officials of the USDA commissioned in 1998

13
Bovine Spongiform Encephalopathy (BSE)

• BSE is a transmissible, neurodegenerative, and
  fatal brain disease of cattle
• The nature of the causative agent remains a
  controversy
• Dr. Stanley Prusiner (UCSF) believes causative
  agent is a prion (misfolded proteins, no nucleic
  acid)

Cow with BSE
Prion affected tissue
14
Bovine Spongiform Encephalopathy (BSE)

• Uncertainties regarding the BSE infectious agent
  in the USDA risk assessment because of
• Spontaneous mutations in cattle that develop
  symptoms
• Transmission to cattle from domestic sheep with
  scrapie
• Importation of infected cattle, meat products,
  and feed
• Transmission to cattle from deer, elk, mink, or
  pigs infected with chronic wasting diseases
• Transmission to cattle from domestic feed
• Consumption of materials from prion-contaminated
  bovine carcasses
• Model stops short of claiming a causal
  relationship between these possible sources of
  risk and BSE
• The formal risk assessment findings are that BSE
  and other encephalopathies are not understood
  sufficiently to predict the likelihood of
  possible future cases

15
Principle of Iterative Risk Assessment

• Risk assessments are by nature iterative if
  direct evidence existed, risk could be calculated
  directly
• However, knowledge is almost always incomplete,
  indirect, and ambiguous.

16
von Pettenkofer challenge, version 2.0

• In a more recent challenge, 38 healthy
  individuals administered 100 million bacterial
  cells of a virulent strian of EI Tor Vibrio
  cholerae
• 3 volunteers did not develop illness
• ? Any risk model for cholera therefore should
  include variables for host resistance to account
  for this outcome

17
Advances in methodologies for host- pathogen
interactions..

• In Vivo Induced Antigen Technology (IVIAT)
• a technique that identifies pathogen antigens
  that are immunogenic and expressed in vivo during
  human infection. Genes and gene pathways
  identified by IVIAT may play a role in virulence
  or pathogenesis during human infection, and may
  be appropriate for inclusion in therapeutic,
  vaccine or diagnostic applications

18
(No Transcript)