Chapter 8:  Chemical Risk Characterization and Uncertainty..
Fundamental issues and considerations
*risk characterization is the integration of toxicity and exposure assessment
**can only be as good as its parts
*should fully, openly, and clearly characterize risks and disclose analysis, uncertainties, and assumptions
**document sources for processes and protocols
Non-standard pollution groups
*care must be taken to ensure that assumptions about population parameters in the dose-response analysis are consistent with the population parameters used in the exposure analysis
Adjustment for chemical absorption
*may be necessary to ensure that the exposure estimate and the toxicity value being compared during the risk characterization are both expressed as absorbed or administered doses
**adjustments are made for vehicles of exposure, administered to absorbed dose for RfD and SFs
Aggregate effects of chemical mixtures
*when combining multi-chemical risk estimates for multiple chemical sources, it should be noted that
**if two sources do not affect the same individual or subpopulation, the sources’ individual risk should not be combined
**in mixture assessment (within the context for interactions):
***carcinogenic effects–segregated by weight-of-evidence category
***non-cancer effects–grouped by similar toxic endpoint/mode of action
Fundamental considerations in health assessment of carcinogens
*qualitative issues:
**weight-of-evidence–from case studies, epidemiological studies, long-term animal studies, long-term bioassays, short-term tests, and SARs
**mechanistic inference–multistage process that is characterized by a non-threshold assumption
***may not be true for each step, only for the process as a whole
**exposure route specificity–some chemicals may not pose a carcinogenic effect via ingestion, but may pose a carcinogenic risk via inhalation
**role of epidemiological data–can give more weight to a well-designed epi study rather than a well-designed animal study, remembering that the potential for association remains especially relevant for animal data
**sensitive and susceptible populations–independently address the associated health concerns
**structure activity relationships (SARs)–can be crucial in estimating toxicity factors for those chemicals that have little or no data available
**chemical interactions–we assume an additive effect, but we have to make a qualitative statement about the certainty of this with regard to mode of action
*quantitative issues
**dose scaling–from animal to human and when using epi studies
***individual sensitivities
**pharmakokinetics/pharmacodynamics–delivered dose rather than the exposure dose is important
***using exposure dose is conservative in most cases
**mechanistic considerations–extrapolations from high to low doses must consider which mathematical model to use
***most conservative model is best if there are equally plausible models
**individual vs. population risk–if possible experimental models, molecular biology, and epidemiology should be used to make the best estimate of risk
Carcinogenic risk effects
*calculated using one of two models, then take the sum for all carcinogens in the assessment
**linear low dose: CR = CDI * SF
***based on linearized multistage model assuming multiple stages for cancer
***good for low intake levels
**one-hit model: CR = 1 – exp(-CDI*SF)
***assumes single stage for cancer and that one molecular or radiation interaction induces malignant change
***good for high intake levels
**if risk is greater than 1 in 1 million, this is considered unacceptable
Non-carcinogenic risk effects
*hazard quotient
**HQ = E/RfD
**for each chemical
*hazard index
**HI = ?HQi
**summation of the HQ for each chemical in the assessment
*if the hazard index is greater than 1, this indicates unacceptable risk
Population excess cancer burden
*to assess the population cancer burden associated with a chemical exposure problem:
**one can estimate the number of cancer cases due to an exposure source within a given community (cancer burden)
Bgi = ? (Rgi * Pg)
***Bgi = excess cancer burden for a group
***Rgi = excess lifetime cancer risk for a group
***Pg = number of persons in exposed population group
Uncertainty and variablity
*variability–arises from true heterogeneity in characteristics such as
**dose-response differences
**differences within a population
**differences in exposure factors
**differences in exposure levels
*uncertainty–arises from lack of knowledge about factors
Types of variability
*spatial–across locations
**can be regional or local (i.e. fish intake)
*temporal–across time
*inter-receptor–amongst individual receptors
**human characteristics and behaviors
Types and nature of uncertainty
*parameter values
**incomplete or biased data
*parameter modeling
**model inadequacy or representation of situation
*degree of completeness
**representativeness of evaluation scenarios
Sources of uncertainty in human health assessment
*health effects
**toxicity data
*measuring or calculating exposure point concentrations
*calculating exposure dose
Commonly encountered limitations
*toxicity extrapolations
*adjustments in dose-response evaluation
*toxicity of mixtures
*background exposures
*representativeness of sampling plan
Uncertainty and variability analysis
*uncertainty/sensitivity analysis
**involves identifying parameters that will make the most contribution to uncertainty
**estimating the most appropriate value using distributions and non-point estimates
**provides full spectrum of information
Qualitative analysis of uncertainties
*determination of the general quality and reasonableness of the risk management data, parameters, and results
*most important for screening and preliminary assessments
*can contain errors, incomplete analysis, limitations, bad assumptions
Quantitative analysis of uncertainties
*Monte Carlo simulations
**assign a probability distribution to input variables
**estimate risk using an iterative process to calculate risk by changing the input variables in the risk equation
**produces a risk distribution instead of a point estimate
Sensitivity analysis
*measures relative response of output variables caused by variations of the input variables and parameters
*purpose is to identify the influential input variables and to develop bonds on the model output
*more resources can be directed to reduce the uncertainty in the most sensitive parameter
Leave a Reply
Your email address will not be published.