Richard Denison, Ph.D., is a Lead Senior Scientist.
In numerous assessment documents issued by the Environmental Protection Agency (EPA) developed in its implementation of the Toxic Substances Control Act (TSCA), EPA cites as a source the European Chemicals Agency (ECHA). ECHA is the agency that administers the European Union’s (EU) REACH Regulation, which (unlike TSCA) requires the registration of chemicals in commerce by companies that wish to continue to produce and use the chemicals in the EU.
When companies register chemicals under REACH, they are required to develop and submit a “dossier” of certain information on production and use as well as on physical-chemical properties, fate, hazard, exposures and risks. ECHA then makes information available on its website.
In its draft risk evaluations for the first 10 chemicals undergoing evaluations under TSCA, as well as in some of its support documents for high-priority substance designations under TSCA, EPA has heavily relied on these dossiers. But in doing so, EPA has grossly mischaracterized the source and nature of the data it references as coming from ECHA.
EDF has been raising concerns about EPA’s mischaracterizations for some time now (see section 1.E of our comments on EPA’s draft risk evaluation for 1,4-dioxane), but they persist. And as recently as yesterday, members of the peer review panel reviewing EPA’s draft documents have been led by EPA statements and citations to assume a degree of completeness and government review of these data that is simply false.
EPA needs to immediately cease and desist in its mischaracterizations.
EPA has hidden the industry authorship of information it cites as sourced from ECHA
EPA is asserting that ECHA dossiers are chemical assessments equivalent to EPA and ATSDR governmental assessments. For example, on p. 43 of its draft risk evaluation for 1-Bromopropane, EPA claims (emphasis added): “Examples of existing assessments are EPA’s chemical assessments (e.g. previous work plan risk assessments, problem formulation documents), ATSDR’s Toxicological Profiles, EPA’s IRIS assessments and ECHA’s dossiers.”
But in fact ECHA dossiers are not assessments and are not government documents. They are compilations of industry information submitted to ECHA that have not been evaluated for quality or reliability by ECHA or any other governmental entity. For EPA to equate them with EPA and ATSDR assessments is simply wrong. Among other things, such government assessments are typically subject to extensive internal, public and peer review.
ECHA’s posting of the dossiers themselves makes clear the lack of government review. At the bottom of each page of each dossier (see 1-Bromopropane here) is the following statement (emphasis added):
Information on Registered Substances comes from registration dossiers which have been assigned a registration number. The assignment of a registration number does however not guarantee that the information in the dossier is correct or that the dossier is compliant with Regulation (EC) No 1907/2006 (the REACH Regulation). This information has not been reviewed or verified by the Agency or any other authority.
EPA exacerbates the mischaracterization through its text references to the industry’s dossiers. EPA typically cites the dossiers posted on the ECHA website through a hyperlinked reference in the text of its documents that reads “ECHA, [date].” Clicking on that link takes the reader to EPA’s entry for that source in its Health & Environmental Research Online (HERO) data system. Those HERO entries then prominently list the “European Chemicals Agency” as the reference’s author. As an example, here is EPA’s HERO citation to the industry registrant’s 1-Bromopropane dossier.
Such text citations and HERO entries are false and highly misleading. All of these documents were prepared by the industry registrants, not ECHA. And the information has not been evaluated by ECHA or any other governmental entity.
While some chemicals do eventually undergo a “substance evaluation” by government authorities under REACH, none of the chemicals for which EPA has released draft risk evaluations has done so. Chemicals that an EU authority “has evaluated or will evaluate [] over the coming years” are listed here. Of the chemicals for which EPA has released draft risk evaluations, only Pigment Violet 29 (CAS # 81-33-4) is listed, and that is because it has been identified by the EU as a “suspected PBT [persistent, bioaccumulative and toxic] or vPvB” [very persistent, very bioaccumulative] chemical and as a result is scheduled to undergo a substance evaluation in 2021.
ECHA dossiers do not provide full studies, only summaries prepared by the industry registrants
EPA’s citations to and discussions of data it obtains through ECHA have also obscured the fact that the ECHA dossiers contain only summaries of studies, not the studies themselves, and that those summaries were prepared by the registrant, not ECHA. This was a major point of controversy and concern that has deeply tainted EPA’s first draft risk evaluation, for Pigment Violet 29.
It is now turning out that EPA itself frequently does not have access to the full studies and instead relies heavily or wholly on the industry-prepared summaries for its risk evaluation. In its charge questions to the peer review panel for 1-Bromopropane, EPA acknowledges this (emphases added):
Only a few environmental test data endpoints (including ECHA) are available in the public domain for 1-BP. Most are from the ECHA website. EPA attempted to obtain the full ECHA studies with no success. Since the studies were in French and Japanese (and no U.S.A. sponsor), EPA decided not to make further attempts to find the studies.
EPA’s decision not to pursue obtaining the full studies based on them not being in English stands in contrast to what EPA did to obtain the underlying industry studies for Pigment Violet 29, where it got copies of studies even though they were originally in German; see footnote 2 on page 1 here). Documents that EDF and other groups obtained through a FOIA request reveal that EPA went to great lengths to have the studies translated by the companies that owned the studies.
EPA needs access to underlying data to ascertain the accuracy of the industry’s information and its associated statements or conclusions, as well as to determine how much confidence or uncertainty should apply to a particular submission. Even the best study summaries are incomplete descriptions that do not allow for an independent examination of study quality and conclusions reached by authors.
EPA’s request to industry for the full studies on Pigment Violet 29 (see p. 1 here) says just this (emphasis added):
[S]ummary study results do not provide sufficient information upon which the hazard(s) and risk(s) from manufacture, distribution in commerce, processing, use, or disposal of this substance or any combination of such activities on health or the environment can reasonably be determined or predicted. *** EPA needs to review the full study reports to confirm the information in the summaries meets the scientific standards set forth in TSCA section 26.
One other reason why EPA (as well as public and peer reviewer) access to full studies is essential: As we blogged about previously, industry registrants can change – and have changed – their ECHA study summaries at will without any requirement that they even give notice of having done so.
EPA then gives the industry summaries kids-gloves treatment and special weight
First, the ECHA-sourced industry study summaries completely bypassed the data screening step of EPA’s literature search process (p. 44 of the draft risk evaluation), and were automatically designated by EPA as “key” sources of information:
These are key and supporting studies from existing assessments (e.g., EPA IRIS assessments, ATSDR assessments, ECHA dossiers) that were considered highly relevant for the TSCA risk evaluation. These studies bypassed the data screening step and moved directly to the data evaluation step.
Second, and especially ironically, EPA then used its lack of access to the full studies as an excuse not to subject the studies to systematic review (see p. 138 of the draft risk evaluation).
But, as described in our written comments provided to the peer review panel, this did not stop EPA from using the study summaries in its analysis. On the following pages of the draft risk evaluation, EPA described the results of these studies, even while indicating that, because EPA could not obtain the full studies, they were not reviewed for study quality. These studies comprise nearly all of the acute aquatic toxicity data EPA has.
EPA then relied almost entirely on those results to: a) conclude that 1-Bromopropane poses no acute aquatic risks; b) extrapolate from them to estimate chronic toxicity values, using a questionable methodology; c) conclude that 1-Bromopropane poses no chronic aquatic risks; and d) conclude that 1-Bromopropane poses no environmental risk at all.
Conclusion
To summarize:
- EPA hides the fact that the data it sources from ECHA are industry data never evaluated by ECHA. EPA instead falsely implies they are assessments equivalent to those of EPA or ATSDR.
- ECHA only provides industry-prepared summaries, not full studies. Despite EPA’s own acknowledgment that “EPA needs to review the full study reports to confirm the information in the summaries meets the scientific standards set forth in TSCA section 26,” EPA did not in fact obtain the full studies for 1-Bromopropane (nor, we suspect, for other chemicals it is evaluating).
- Nonetheless, EPA designates such industry summaries as “key” information, and allows them to skip out on both EPA’s data screening and its systematic review process.
- Despite not having evaluating the quality of the studies because it could not access them, EPA still proceeds to use them as nearly the entire basis for its sweeping conclusion that 1-Bromopropane poses no risk to the environment.
I ask you: Does this sound like an agency making decisions using the best available science?