Richard Denison, Ph.D., is a Lead Senior Scientist.
I’ve been blogging about the deep problems surrounding the first draft risk evaluation the Trump Administration’s Environmental Protection Agency (EPA) has released under the recently amended Toxic Substances Control Act (TSCA). This risk evaluation, which is now out for public comment, is on a chemical commonly called Pigment Violet 29, or PV29. Among the many problems that immediately jumped out as we began our review of this draft evaluation are EPA’s reliance on clearly inadequate health and environmental hazard data to conclude the chemical is safe, as well as EPA’s illegal withholding from the public of the little hazard information it does have.[pullquote]I suppose if you start with almost no reliable data on a chemical, are dead set against using your enhanced authorities to get any more data, and are hell-bent on finding the chemical is safe, this is how you might choose to conduct a risk evaluation.[/pullquote]
This post will look at the other half of the risk equation, exposure. EPA has even less information on exposures to PV29 than it does on hazard. EPA has no actual data on the levels of PV29 released to or present in air, soil, sediment, surface water, people, other organisms, workplaces or products containing or made from the chemical. It lacks any data from, and hasn’t used its authorities to require, monitoring in workplaces or any environmental media.
So what does EPA have?
It has a personal communication from a conflicted industry source who claims “an approximate maximum workplace air concentration of 0.5 mg/m3 [milligrams per cubic meter] would be expected over a 12 hour shift” at a PV29 manufacturing facility (see page 22 of the draft risk evaluation). That number was provided to EPA by an employee of Sun Chemical, the only identified manufacturer of PV29 in the U.S., and – to state the obvious – an entity with a strong interest in having EPA find its chemical safe. All the public has to go on is a statement and reference in the risk evaluation that indicate a fellow at Sun Chemical named Robert C. Mott personally communicated this value to EPA on September 25, 2017. EPA has not even made the content of the personal communication public, nor has it provided or alluded to any actual data provided by Mott to support his statement.
EPA cites this value, but then acknowledges it does not know what this workplace air value actually represents, noting the personal communication was not even clear whether the claimed value was for Pigment Violet 29 itself or for total dust.
(Ironically, it appears that, in order to rely on this information, EPA had to exempt it from being scrutinized using its own TSCA Systematic Review approach. On page 18 of the draft risk evaluation, EPA states that its systematic review approach “is not well suited for the review” of such “correspondences with industry … used to inform the likelihood of exposure,” and “[a]s a result, formal data quality evaluation of these references according to the Application of Systematic Review in TSCA Risk Evaluations (U.S. EPA, 2018a) was not conducted.”)
But it’s what EPA does next with this unverified, conflicted single data point that is truly remarkable:
- First EPA uses the suspect value to calculate an “inhalation Potential Dose Rate (PDR) for workers.”
- Without providing any explanation or justification, EPA then combines that inhalation PDR with a no-effect level derived from a screening-level toxicity study in which animals were given doses of the chemical orally, not by inhalation. (EPA’s use of this oral toxicity study has its own problems; stay tuned for a future post.) EPA doesn’t even bother to mention that it is switching between routes of exposure or the uncertainty that injects into its analysis.
- Based on this questionable calculation, EPA concludes definitively that workers face no risk from PV29 inhalation.
- EPA then simply asserts without any data or analysis that all PV29 inhalation exposures to workers at downstream processing and use sites must be lower than those of manufacturing workers, and hence concludes there is no inhalation risk to any workers. (EPA constructs a similar, questionable line of argument for dermal exposure, which I won’t go into here.)
- EPA then asserts that all PV29 exposures to others must be lower than worker exposures, and hence that there are no risks to consumers or the general population.
- Finally, EPA argues that the hazard data it examined gave no evidence of increased susceptibility for any single group relative to the general population – despite the fact that nearly all of the studies it has and relied on did not even look for such evidence.
- Armed with this questionable finding, EPA concludes that “the exposure calculation for workers is … therefore protective of all other subpopulations, such as children and pregnant women in the general population, which are not expected be exposed to C.I. Pigment Violet 29 at similarly high levels.”
All of this, built layer-by-layer from that single, shaky workplace air value personally communicated to EPA from an employee of PV29’s sole U.S. manufacturer.
I suppose if you start with almost no reliable data on a chemical, are dead set against using your enhanced authorities to get any more data, and are hell-bent on finding the chemical is safe, this is how you might choose to conduct a risk evaluation.
But the house of cards EPA has built here certainly isn’t what Congress had in mind when it required that EPA use all reasonably available information (which EPA’s final risk evaluation rule defines to include information EPA “can reasonably generate”) and make decisions based on the best available science.