How to evaluate a non-cancer risk assessment
From
Contents |
Overview
The key to non-cancer risk assessment is the development of a reference dose (RfD) or a reference concentration (RfC), which essentially represent regulatory safety points, not actual safety points. Exposure to agents below the RfD/RfC will generally be determined to be safe and exposures above the RfD/RfC will generally be determined to be not safe.
The process for developing an RfD/RfC begins with determining one of the following:
- No-observed-adverse-effect-level (NOAEL);
- Lowest-observed-adverse-effect-level (LOAEL); or
- Benchmark dose (BMD).
Once the NOAEL/LOAEL/BMD is determined, it is divided by an uncertainty factor (UF) to produce the RfD/RfC.
The process of developing the RfD/RfC is not entirely scientific in nature as it involves several subjective (non-scientific) decisions, including:
- Determining the health endpoint of concern, which leads to the establishment of the NOAEL/LOAEL/BMD; and
- Applying the uncertainty factor to the NOAEL/LOAEL/BMD.
While risk assessors tend to characterize these subjective decisions as "judgment-based," they may in fact be entirely arbitrary and designed to facilitate the outcome desired by regulators.
The last step of a non-cancer risk assessment is to assess the RfD/RfC in the context of expected exposures to produce the risk characterization.
Step 1: What is the health endpoint of concern?
The determination of the health endpoint of concern (or "critical effect" in EPA-speak) is crucial to the outcome of non-cancer risk assessment. Regulators may select the health endpoint that will result in the lowest RfD/RfC. There are several aspects of this subjective determination that should be considered.
Step 1.1: Is it a genuine adverse health effect or a mere biological response?
Just because an agent can be associated with some physiological change in a laboratory experiment, that does not mean that the physiological change necessarily constitutes an adverse health effect; it may very well be a harmless biological response that shouldn’t be of concern to regulators.
The way to think about this point is as follows. Imagine an experiment where researchers are gradually administering doses of some agent to the study subjects, either animals or humans. The researchers are carefully monitoring the subjects, looking for changes in blood chemistry, heart rate and other physical characteristics.
At some point as the doses gradually rise, the study subjects’ bodies will begin to respond to the agent. Does this first response amount to an adverse health effect, perhaps indicating a NOAEL or a LOAEL. Or is this first response more aptly described as a mere biological response that, while indicative of cause-and-effect does not rise to a health effect of concern? A classic example of this situation involves certain pesticides (carbamates and organophospates) and cholinesterase inhibition.
Step 1.2: Is interspecies extrapolation appropriate?
The determination of a NOAEL/LOAEL/BMD is typically accomplished through laboratory experiments. These experiments are typically conducted on laboratory animals because of the expense, risk and regulatory hurdles associated with human testing. So the determination of a NOAEL/LOAEL/BMD often involves interspecies extrapolation – i.e., the assumption that results from animal experiments are relevant to humans. This assumption may not be appropriate, especially when it is based on the most sensitive species.
Step 2: is the uncertainty factor justified?
Uncertainty factors come in multiplicative powers of 10, so if a NOAEL/LOAEL/BMD is subjected to four UFs, the RfD/RfC will be a factor of 10,000 (10 x 10 x 10 x 10) lower than the experimentally determined NOAEL/LOAEL/BMD.
Regulators claim to apply uncertainty factors no a NOAEL/LOAEL/BMD in order to:
- Account for the uncertainty of extrapolating experimental results to broader human populations and between species; and
- Provide a margin of safety to the RfD/RfC.
While these sound like plausible reasons – and may very well be some of the time – they also provide regulators with an arbitrary mechanism for reducing the RfD/RfC and making it much more stringent than warranted. A too low RfD/RfC could result, for example, in a de facto ban of a chemical.
When evaluating the application, keep the following in mind:
- The valuation of each UF at 10 is an arbitrary decision and is not supported by scientific evidence.
- The practice of multiplying or compounding UFs has not been validated by scientifically.
While a great deal of uncertainty exists when extrapolating from animal experiments to human reality, the regulatory assumptions are that
- the agent will have similar effects on humans; and that
- humans are more sensitive to the agent than animals.
In reality, however, the uncertainty is whether the experimental animal results are at all relevant to real-life human experiences.
Step 3: Is the exposure assessment realistic?
Exposure assessments often involve a great deal of uncertainty leading to regulatory decisions to rely on upper-bound exposures rather than realistic or expected exposures.
Do you want to evaluate an exposure assessment?
Step 4: Conclusion
Non-cancer risk assessment can be dramatically affected by non-scientific determinations that are biased toward exaggerating risk.
Though regulators often try to pass off these assumptions as "scientific judgment" that bridges gaps and uncertainties in scientific data and knowledge, in practice they are an arbitrary means of shaping science to fit pre-determined policy.
Return to the Step-by-step guide for debunking.
Go to the DEBUNKosaurus™ Main Page.
Use the search function to find your topic.
