PART 3: EPA rams through its reckless review scheme for new chemicals under TSCA, your health be damned

Richard Denison, Ph.D., is a Lead Senior Scientist.

Part 1               Part 2               Part 3

I’ve been blogging over the last week about how political appointees at EPA are starting to clear new chemicals to enter commerce based on a new – apparently unwritten and certainly not public – review process that ignores the law and will put the health of the public, workers and the environment at greater risk than even under the weak reviews conducted before Congress’ 2016 overhaul of the Toxic Substances Control Act (TSCA).

The first green light was given to a fragrance chemical intended for use in a wide array of everyday consumer products.  Here is its formal name, a simpler synonym used by its manufacturer, and an identifier called a CAS number:[pullquote]Late in the review process EPA switched to analogue chemicals to estimate toxicity that are 500-fold less toxic than the analogue identified by the company and EPA’s own models.[/pullquote]

  • Name: Oxirane, 2-methyl-, polymer with oxirane, bis[2-[(1-oxo-2-propen-1-yl)amino]propyl] ether
  • Synonym: Jeffamine diacrylamide
  • CAS: 1792208-65-1

In this post I want to look more at what is known – and not known – about the chemical’s hazards.  While the chemical is a polymer, it includes “low molecular weight (LMW) components” that are the primary hazard concern.  

First, very few measured health and environmental data were submitted by the manufacturer in its premanufacture notification (PMN), which EPA reviewed and approved without imposing any conditions or testing requirements.  This is typical of PMNs and, indeed, this PMN actually included more measured data than most do.  The submitted data, all commissioned by the manufacturer, consisted of three in vitro studies and one human patch test study to assess skin sensitization; and two in vitro studies to assess genotoxicity.  Results for those two health effects were reported to be negative.

All of the other health and environmental hazard information submitted by the company or developed and examined by EPA were estimated, not measured, data.  The estimates were derived either by:  1) using EPA models that predict a given characteristic (e.g., acute toxicity to fish) based on a chemical’s physical-chemical properties, or 2) identifying other chemicals that are considered structurally related to the LMW components of the PMN chemical – called “analogue chemicals” – and assuming the measured data available for those chemicals is the same for the LMW components of the PMN substance (this is sometimes called “read-across”).

The PMN submitter’s assessment

In this case, which is not common, the company that submitted the PMN actually, on its own, ran the EPA models and identified an analogue chemical, whereas usually EPA does that as part of its review.  Included with the PMN was a “summary assessment” that provided the company’s estimates for various health and environmental effects.  The company clearly noted that the best analogue for its chemical, based on structural similarity, was a chemical in the acrylamides category.  Indeed, it was the company that indicated that “Jeffamine diacrylamide” is a synonym for its chemical.

As I noted in the first post in this series, acrylamides are one of a set of “chemical categories” EPA uses to flag new chemicals that could pose serious health or environmental concerns.  I suggested that similarity to the acrylamides category is a red flag that likely is why EPA professional staff raised concerns about the PMN substance from the start and recommended restrictions.  Here is EPA’s description of the hazard concerns for this category (p. 5):

Based on analogy to acrylamide per se, members of the class are considered potential carcinogens, heritable mutagens, reproductive and developmental toxicants, and toxic to aquatic organisms.  Acrylamides are also potential neurotoxins based on data for a number of low molecular weight acrylamides.

In its own determination document clearing Jeffamine diacrylamide to enter commerce without restriction, EPA identifies these toxicities and more, noting “the potential for the following human health hazards: irritation, mutagenicity, developmental/reproductive toxicity, neurotoxicity, and carcinogenicity.”

The company’s summary assessment identified moderate concerns for persistence, aquatic toxicity hazard, cancer health hazard, and non-cancer health hazard.  The latter two rankings were based on the company’s selection and use of an analogue chemical from the acrylamides category.

EPA’s final assessment reverses course

All we have to go on, at least for now, to discern how EPA went from raising concerns to declaring this chemical safe for any use in any amount by any company is EPA’s determination document, which is, at best, a high-level summary of its risk assessment.  This short document omits critical information that would be needed for anything approaching an independent evaluation of its conclusions.  Whether EPA will make its full risk assessment and other analyses public or not remains to be seen.  Still, some things jump out as concerning even in reading the summary.

In its summary, EPA now says the PMN chemical’s LMW components  fall into two of EPA’s chemical categories:  acrylamides (as before) and acrylates.  EPA then says, with scant explanation and without providing any data or analysis, that it decided to use analogue chemicals from the acrylates category rather than from the acrylamides category, asserting they are “more predictive.”  This is convenient, given that EPA regards the former as posing little to no health concerns and the latter as posing among the greatest.  Here is what EPA’s chemical categories document says about acrylates (p. 8):

As a result of testing conducted under an agreement between the Agency and the Specialty Acrylates and Methacrylates (SAM) Panel of the Chemical Manufacturers Association (CMA), EPA no longer controls new chemical acrylates … as a category [of concern] based on health concerns.

The switch in analogues makes a monumental difference:  The toxicity data EPA cites in its determination document show that the acrylates analogues are 500-fold less toxic than is acrylamide. (EPA cites a No Observed Adverse Effect Level (NOAEL) for systemic toxicity of 250 milligrams/kilogram/day for the acrylate analogues.  The study EPA cites for acrylamide set a NOAEL of 0.5 mg/kg/day for the systemic effect of neurotoxicity.)

Clearly, EPA did not make the most health-protective choice.

Notably, the decision to switch analogues appears to have been EPA’s.  Just this past Tuesday, EPA added an asterisk and note to its “not likely” determination for this chemical (search for “P-16-0510” in the table here) that reads:

*In 2016, EPA’s original screening analysis for this PMN resulted in the interim finding listed to the Interim Status column. In response to EPA’s recommendation, the submitter requested EPA perform an assessment using read-across for a particular analogue (an acrylamide). In considering the submitter’s request, EPA identified additional analogues which it deemed more predictive (acrylates) than the submitter’s suggested analogue for conducting a reasoned evaluation for the PMN substance. EPA conducted risk evaluation with both analogues and based on the analogue EPA deemed most predictive concluded the PMN substance was not likely to present unreasonable risk.

Indeed, other documents submitted by the company based on running EPA’s own models identified acrylamides – and never acrylates – as the relevant category.  For example, a chemical profile the company generated using EPA’s PBT Profiler flagged the chemical as belonging to the acrylamides category as well as the ethylene glycol ethers category; and in applying EPA’s OncoLogic model, which predicts carcinogenic potential, the company also identified acrylamide functional groups as the relevant characteristic of its chemical.

So EPA’s selection of a different set of analogues radically departed from the key structural alert identified by both the company and EPA’s own assessment tools.  EPA then goes on to use hazard data for the far less toxic acrylates analogues to calculate risks and conclude the chemical is not likely to present an unreasonable risk.  EPA’s document claims it also calculated risks based on acrylamide as the analogue – but then only presents the results for the acrylates analogues, one more reason why EPA should make the full risk assessment public.

Another curious feature of the summary document is that, in presenting the final results of its human health risk assessment, EPA repeatedly notes that its calculated margins of exposure (MOEs) “exceed the benchmark MOEs” and hence are safe.  But EPA never states how large the exceedances are – critical to knowing whether EPA found the chemical quite safe or only marginally safe.  Why wouldn’t EPA provide the specific MOE values it calculated for the reader to be able to compare them to the benchmark MOEs it does specify?  This omission matters, given the substantial uncertainties associated with the calculation of MOEs in the first place.  I won’t speculate further on the reasons for the omission, but can also be remedied by EPA making its full risk assessment public.

Whatever one may think of EPA’s heavy use of analogue data as well as its decision to switch the analogues it used to do its risk assessment, the fact is that use of analogues inherently carries with it a significant degree of uncertainty.  Even subtle differences in chemical structure can mean significant differences in how a chemical behaves in the environment and in our bodies as well as its toxicity.

So what happens if EPA is wrong in its analogue choice?  What if the chemical it has just given carte blanche approval is actually more toxic than the analogues?

In the past, when EPA relied on analogue data for new chemicals and was concerned about their reliability, EPA would require testing of the actual new chemical to verify whether its choice of analogue was sound.  Such testing might be required up front, but more often as a condition imposed on a PMN submitter, for example, limiting its ability to expand manufacture of a chemical beyond a set limit without first submitting the required data.  That way, EPA could review the test results and determine whether any changes in conditions applicable to the chemical were needed.

Of course, all of this was achievable because the chemical was subject to an order.  Yet now, there is no order.  So not only are there no limits placed on the chemical’s production and use, there will be no testing to ascertain whether EPA’s risk assessment based on analogues was sufficiently protective.

The title to this blog series uses a term I did not choose lightly:  “reckless.”  I hope these posts help readers to understand just how reckless EPA’s new chemical review process has become.


This entry was posted in Health policy, TSCA reform and tagged , , , . Bookmark the permalink. Both comments and trackbacks are currently closed.

One Comment

  1. Jeff Cantin
    Posted August 9, 2018 at 6:49 pm | Permalink

    Thanks for exposing this and explaining it so clearly Richard. Your outrage is justified!