xvEPA
EP A-740-R-24-011
September 2024
United States Office of Chemical Safety and
Environmental Protection Agency Pollution Prevention
Tris(2-chloroethyl) Phosphate (TCEP); Regulation under the
Toxic Substances Control Act (TSCA)
EPA-HQ-OPPT-2023-0265
Comment Summary and Responses
September 2024
-------
This page intentionally left blank.
11
-------
Table of Contents
Acronyms and Abbreviations iv
Introduction vi
Table 1: Index of Comment Submissions Sorted by Submission Number vii
Section 1 - Overarching Comments 1
Section 1.1 - Scope of the draft risk evaluation 1
Section 1.2- Conditions of use 2
Section 1.4 - Peer review process 4
Section 2.2 - Environmental fate and transport (including comments on the approach to estimate
degradation in the absence of data) 7
Section 3.1 - Approach and methodology 10
Section 3.2 - Environmental releases 10
Section 3.3 - Concentrations of TCEP in the environment 11
Section 3.4 - Concentrations of TCEP in the indoor environment 12
Section 4 - Environmental Risk Assessment 12
Section 4.1 - Environmental exposures 12
Section 4.2 - Environmental hazards (including comments on the calculation of hazardous
concentration for 5% of species (HC05) and the calculation of concentration of concern (COC) for
aquatic organisms) 14
Section 4.3 - Environmental risk characterization 22
Section 5 - Human Health Risk Assessment 23
Section 5.1 - Human exposures (including comments on milk concentrations (Verner), dry and wet air
deposition, and disposal and groundwater pathway, aggregate exposure, exposure to certain PESS)... 23
Section 5.2- Human health hazard (including comments on confidence levels, male reproductive
effects from Chen et al. (2015), and benchmark response (BMR) of 5 percent for the benchmark dose
(BMD) modeling) 39
Section 5.3 - Human health risk characterization (including comments on types of PESS) 52
Section 6 - Unreasonable Risk Determination 53
Section 6.1- Unreasonable risk to human health 53
Section 6.2 - Unreasonable risk to the environment 54
Section 6.3 - Other comments 55
Section 7 - Systematic Review 57
Section 8 - Formatting and Editing 58
Section 9 - Other Comments on the Draft Risk Evaluation 59
in
-------
Acronyms and Abbreviations
1-BP
1- bromopropane
AERMOD
American Meteorological Society (AMS)/EPA Regulatory Model
ATSDR
Agency for Toxic Substances and Disease Registry
BCEP
Bis(2-chloroethyl) phosphate
BCHP
Bis(2-chloroethyl) hydrogen phosphate
BCGP
Glucuronide conjugate of bis(2-chloroethyl) 2-hydroxyethyl phosphate
BMD
Benchmark dose
BMDL
Benchmark dose (lower confidence) limit
BMR
Benchmark response
CDR
Chemical Data Reporting
COC
Concentration of concern
CPSC
Consumer Product Safety Commission
CYP
Cytochrome P450 monooxygenases
DRAS
Delisting Risk Assessment Software
ECso
Effect concentration at which 50 percent of test organisms exhibit an effect
ECEL
Existing chemical exposure limit
ECOSAR
Ecological Structure Activity Relationships (model)
EPA
U.S. Environmental Protection Agency
HBCD
Hexabromocyclododecane
HCos
Hazardous concentration for 5% of species
HLC
Henry's Law constant
HERO
Health & Environmental Research Online
IIOAC
Integrated Indoor-Outdoor Air Calculator
IL
Interleukin 6
KOC
Organic carbon-water partition coefficient
KOW
Octanol-water partition coefficient
LC50
Lethal concentration at which 50 percent of test organisms die
LCSA
Lautenberg Chemical Safety Act
MOA
Mode of action
MOE
Margin of exposure
NIOSH
National Institute for Occupational Safety and Health
NOAEL
No-observed-adverse-effect level
NTP
National Toxicology Program
OECD
Organisation for Economic Co-operation and Development
OES
Occupational exposure scenario
OPFR
Organophosphate flame retardants
OPPT
Office of Pollution Prevention and Toxics
OSHA
Occupational Safety and Health Administration
PBPK
Physiologically-based pharmacokinetic
PECO
Population, exposure, comparator, and outcome
PESS
Potentially exposed or susceptible subpopulations
POD
Point of departure
PPE
Personal protective equipment
PV-29
C.I. Pigment Violet 29
RACB
Reproductive assessment by continuous breeding
RQ
Risk quotient
IV
-------
SACC
Science Advisory Committee on Chemicals
SNUR
Significant New Use Rule
SSD
Species sensitivity distribution
TCEP
Tris(2-chloroethyl) phosphate
TCE
T richloroethylene
TERA
Toxicology Excellence for Risk Assessment
TRI
Toxics Release Inventory
TSCA
Toxic Substances Control Act
UF
Uncertainty factor
U.S.
United States
Web-ICE
Web-based Interspecies Correlation Estimation
V
-------
Introduction
On December 15, 2023, the U.S. Environmental Protection Agency (EPA) published the 2023 Draft Risk
Evaluation for Tris(2-chloroethyl) Phosphate (TCEP) and accepted public comment until February 13,
2024. Materials on the draft risk evaluation are available at www.regulations.gov in docket EPA-HQ-
OPPT-2023-0265. EPA then conducted a letter peer review to review the approach and methodologies
utilized in this document, which began on March 13, 2024, and ended on April 12, 2024. A preparatory
virtual public meeting was held on March 5, 2024, for reviewers and the public to comment on and ask
questions regarding the scope and clarity of the draft charge questions.
This document summarizes the public and external peer review comments from the letter peer review that
the EPA's Office of Pollution Prevention and Toxics (OPPT) received for the risk evaluation of TCEP. It
also provides EPA/OPPT's response to the comments received from the public and the peer review.
EPA/OPPT appreciates the valuable input provided by the public and peer review. The input resulted in
numerous revisions to the risk evaluation document. The peer review and public comments are
categorized by the nine themes listed below.
Additionally, within each theme comments that cover similar issues are presented together.
1. Overarching Comments
2. Chemistry, Fate, and Transport of TCEP
3. Releases and Concentrations of TCEP in the Environment
4. Environmental Risk Assessment
5. Human Health Risk Assessment
6. Unreasonable Risk Determination
7. Systematic Review
8. Formatting and Editing
9. Other Comments
VI
-------
Table 1: Index of Comment Submissions Sorted by Submission Number
Submission Number
Commenter Name
EPA-HO-OPPT-2023 -0265 -003 3
China WTO/TBT National Notification & Enquiry Center
EPA-HO-OPPT-2023 -0265 -003 5
Environmental Defense Fund (EDF)
EPA-HO-OPPT-2023 -0265 -003 6
Polyisocyanurate Insulation Manufacturers Association
(PIMA)"
EPA-HO-OPPT-2023 -0265 -003 7
Aceto US, LLC
EPA-HO-OPPT-2023 -0265 -003 8
American Chemistry Council (ACC) North American Flame
Retardant Alliance (NAFRA)
EPA-HO-OPPT-2023 -0265 -003 9
Environmental Protection Network (EPN)
EPA-HO-OPPT-2023 -0265 -0040
American Chemistry Council (ACC)
EPA-HO-OPPT-2023 -0265 -0041
National Tribal Toxics Council (NTTC)
EPA-HO-OPPT-2023 -0265 -0042
University of California, San Francisco Program on
Reproductive Health and the Environment
EPA-HO-OPPT-2023 -0265 -0043
Alliance for Automotive Innovation
EPA-HO-OPPT-2023 -0265 -0044
Earth) ustice et al
EPA-HO-OPPT-2023 -0265 -0045
Environmental Protection Network (EPN)
EPA-HO-OPPT-2023 -0265 -0046
Earth) ustice et al
EPA-HO-OPPT-2023 -0265 -0047
Earth) ustice et al
EPA-HO-OPPT-2023-0265-0048
Earth) ustice et al
EPA-HO-OPPT-2023-0265-0052
American Chemistry Council (ACC)
EPA-HO-OPPT-2023 -0265 -PR
Peer Review Comments
Vll
-------
Section 1 - Overarching Comments
General analysis comments
Summary: A public commenter (0044) praised EPA's risk evaluation for taking a whole chemical
approach, considering multiple exposure pathways, analyzing fenceline community exposures, and
recognizing Tribal communities as a potentially exposed or susceptible subpopulations (PESS).
However, the commenter also stated that the risk evaluation resulted in underestimations of exposure
and risk. Specific comments are summarized throughout in the following sections.
EPA Response: EPA thanks the commenter for the praise and looks forward to addressing the specific
comments.
Summary: A public commenter (0040) stated that EPA made overly conservative estimates in both the
hazard and exposure parts of the risk evaluation. The commenter characterized the TCEP risk
evaluation as a 'screening-level" evaluation and suggested that according to EPA's own guidance this
evaluation should not inform risk management without further refinement. The commenter
recommended that EPA incorporate feedback from the Science Advisory Committee on Chemicals
(SACC) and ad hoc reviewers.
EPA Response: EPA has reviewed and considered all submitted comments from public and peer
reviewers.
Summary: A peer reviewer (OC - Rl) stated that the draft risk evaluation is quite comprehensive in its
coverage and is an impressive work covering many aspects of the potential risks associated with TCEP.
EPA Response: The EPA thanks the reviewer for their comment.
Section 1.1 - Scope of the draft risk evaluation
Assessing flame retardants as a category
Summary: One public commenter (0044) recommended that instead of considering TCEP as a single
chemical, EPA should evaluate organohalogen flame retardants, a class of chemicals which, according
to the commenter, includes TCEP, polychlorinated biphenyls, polybrominated biphenyls, and
polybrominated biphenyl ethers. The commenter stated that considering TCEP alongside these
chemicals with similar properties and uses would prevent a cycle of replacing one hazardous chemical
with another. The commenter said that such an approach would not only be consistent with the TSCA
statute, but would align with scientific consensus as well as the regulatory approach taken in the
European Union. Similarly, a peer reviewer (OC - R5) expressed support for assessing multiple
organophosphorus flame retardants and using TCEP as a model to determine the best path forward.
EPA Response: EPA has the authority to prioritize categories of chemicals under TSCA section 6 and
would consider comments on whether to assess a chemical individually or as a category during the
prioritization stage of the TSCA Risk Evaluation process. Once a single chemical like TCEP is
designated high priority, the EPA must assess the risk of that particular chemical rather than a category
1
-------
to which it might belong. The 90 day comment period for TCEP designation as a high priority
substance was from August 22 to November 21 of 2019.
TCEP as an impurity
Summary: A public commenter (0037) suggested that EPA should address products that contain
TCEP in small quantities as an impurity and refer to it as an "impurity [condition of use]". The
commenter recommended that EPA consider the potential hazard of TCEP as an impurity in other
chemicals (such as 2,2-bis(Chloromethyl) trimethylene bis [bis(2- chloroethyl)phosphate] (CASRN
38051-10-4) (V6)) in a case-by-case basis for each chemical which may contain a TCEP impurity.
This, according to the commenter, would be consistent with the approach that it took for the
trichloroethylene (TCE) risk evaluation as a byproduct of dichloroethane. The commenter requested
that EPA make explicit note of such an exclusion in the risk evaluation and similarly exempt the import
of products with TCEP impurities from the proposed TCEP significant new use rule (SNUR).
EPA Response: EPA will not be making inapplicability an exemption for or referring to products in
the SNUR that may contain TCEP as an impurity as an "impurity [condition of use]" from SNUR
notice requirements because EPA has determined that a circumstance where TCEP is an impurity or a
byproduct is intended, known, or reasonably foreseen, and therefore falls within existing "conditions of
use" for TCEP in the risk evaluation. The conditions of use in this risk evaluation include
circumstances where TCEP is a byproduct and as an impurity; therefore, byproducts or impurities and
their contribution to the unreasonable risk of TCEP are considered in risk evaluation instead of being
considered as significant new uses in the SNUR. ,
Because TCEP is also known to co-occur in formulation with other flame retardants, such as V6, this
risk evaluation evaluates TCEP when it co-occurs with other flame retardants in commercial and
consumer products (e.g., when it co-occurs with V6). However, it does not evaluate uses of other flame
retardants where TCEP may occur as an impurity or as a byproduct.
As explained in the proposed SNURs for certain flame retardants, including TCEP, EPA did not
propose to make inapplicable the general exemptions from SNUR notice requirements that are
described in 40 CFR 721.45, including exemptions from notification requirements for persons
manufacturing or processing the chemical substance only as an impurity or certain byproduct. EPA
requested public comment on this proposal and will consider the public comments received before
finalizing the SNUR.
Section 1.2 - Conditions of use
General comments
Summary: Two public commenters (0042. 0044) stated that TSCA requires EPA to conduct risk
evaluations for all conditions of use. One commenter (0042) wrote that if a condition of use for TCEP
was voluntarily discontinued, there exists no reason that users could not resume their activity and
therefore EPA must consider this a reasonably foreseen use of TCEP. The commenters also stated that
EPA has been authorized to conduct tests and subpoena necessary information to address data gaps in
its risk evaluations. According to the commenters, EPA failed to take actions, such as lowering the
Chemical Data Reporting (CDR) threshold for TCEP or promptly adding TCEP to the Toxics Release
2
-------
Inventory (TRI) chemical list, which could have provided data to evaluate the unassessed conditions of
use. One commenter (0044) pointed out that EPA denied a 2017 petition to require more TCEP testing.
EPA Response: EPA disagrees that all voluntarily discontinued uses of TCEP should be considered
reasonably foreseen conditions of use, since the reasons for discontinuing the use might vary, including
technology developments or more cost-effective alternatives, and therefore it is not reasonably foreseen
to expect that all of them could resume. Also, certain discontinued uses are not reasonably foreseen
because it might not be possible for EPA to identify all conditions of use that were discontinued, since
they might not have been reported in previous CDR cycles or there is no reasonably available
information about the discontinued uses. However, if EPA fails to identify a future use of TCEP, such
possible future uses of TCEP are covered by the SNUR under TSCA. The SNUR under TSCA
proposed by EPA on April 8, 2024 includes a reporting requirement before starting a new use of TCEP
or resuming discontinued uses of TCEP, and can be found in the Federal Register under 89 FR 24398.
The CDR Rule, under TSCA, requires manufacturers (including importers) to provide EPA with
information on the production and use of chemicals in commerce. The information is collected every
four years for certain chemicals manufactured and/or imported generally when production volumes for
the chemical are 25,000 pounds or more for a specific reporting year. As explained in the draft risk
evaluation, the most recent updated 2020 CDR data showed that no company reported the manufacture
(including import) of TCEP in the United States from 2016 to 2019, and no company reported
processing and use information for the principal reporting year (2019). Under CDR requirements,
chemicals are subject to a lower reporting threshold of 2,500 pounds when they are the subject of
specific TSCA actions; however, being evaluated within the existing chemical prioritization program is
not one of those actions. However, manufacturers (including importers) will be required to report
TCEP production volumes of 2,500 lbs or more during the next CDR reporting cycle because of the
proposed SNUR for TCEP. In addition, because the CDR collection occurs so infrequently, even if the
reporting threshold was lowered for chemicals that enter the prioritization program, it is unlikely that
the lower volume information would be available in time for the identification of conditions of use.
However, EPA recognizes that the most recent reporting threshold for TCEP in CDR is 25,000 pounds
and that some manufacturing could be occurring below that threshold.
EPA issued TSCA Section 4 Test Orders for 2 of the 3 flame retardants (4.4'-( 1 -
methylethylidene)bis[2, 6-dibromophenol] (TBBPA) and triphenyl phosphate (TPP)) on the list of 20
chemicals currently undergoing risk evaluation. Test orders were not issued for TCEP because EPA
had not identified any domestic manufacturers or processors (which would be required in order to
determine who the recipients of an order would be). As the existing data were evaluated during the
systematic review process, it was noted that EPA did not have TCEP site/process specific inhalation
monitoring data, site/process specific dermal hand wipe sampling, and in vitro dermal absorption data.
EPA determined that modeled data from either our own Generic Scenario documents and other
standard EPA models for inhalation and dermal exposure could be used to develop the occupational
exposure data for the RE. There is also a 2009 EU Risk Assessment that provides a reasonable worst-
case modeling scenarios for inhalation, dermal, and oral exposures (HERO: 3809216).
EPA recently added TCEP to the Toxics Release Inventory (TRI) with the first year of reporting from
facilities due July 1, 2024. As of September 2024, there are no TRI entries for TCEP. A facility is
required to report to TRI if it meets all three threshold criteria, including that the facility is in a TRI-
covered industry, the facility has 10 or more employees, and the facility manufactures (defined to
include importing), processes or otherwise uses TCEP in quantities greater than established thresholds
during a calendar year. For TCEP these established threshold are 25,000 lbs for manufacturing, 25,000
lbs for processing, or 10,000 lbs for other use. Given the decline in production volume seen in the CDR
and Datamyne data in recent years, EPA anticipates receiving few-TRI submissions for TCEP due to
the manufacturing, processing, and use reporting thresholds.
3
-------
Comments on specific conditions of use
Summary: Two public commenters (0036. 0038) stated that the use of TCEP in polyisocyanurate
insulation has been phased out in both domestic production and importation. One commenter (0036)
further noted that this transition occurred in the mid-1990s to 2000s. According to the commenter,
enough time has passed that a majority of TCEP-containing polyisocyanurate insulation has been
replaced and any that remains has largely passed its "offgassing period" and therefore no longer
contributes to unreasonable human health risk.
EPA Response: EPA considered the off-gassing period in its risk characterization of TCEP containing
insulation in Figure 5-22 and accompanying discussion. Figure 5-22 in Section 5.3.5.2 provides a graph
of the mean gas phase concentration versus time in days. As seen in the figure, the gas phase
concentration has an initial spike followed by a precipitous decline after the first few months.
Section 1.4 - Peer review process
Comments on the letter peer review process
Summary: Multiple public commenters (0039. 0040. 0041. 0042. 0044. 0045. 0052) expressed
concern with EPA's decision to employ a letter peer review process. Several of the commenters (0039.
0040. 0045) made comparisons with EPA's approach to the review of the "Asbestos Part 2 White
Paper" which commenters claimed was poorly received by reviewers who wished for greater
transparency and opportunity for panel discussion. Generally, the commenters highlighted the
importance of the TCEP draft risk evaluation review due to the chemical's broad applications,
numerous exposure pathways, and high potential for human health impacts. The commenters
recommended EPA conduct a SACC panel peer review process in lieu of a letter peer review process to
improve transparency in the process and provide opportunity for collaborative interactions between the
reviewers and engagement with the public. Commenters also stated that a more robust peer review
process would save time in the long run.
EPA Response: Independent, expert peer reviews play an important role in EPA's development of
scientifically sound chemical risk evaluations conducted under TSCA. Selecting the right questions to
ask reviewers and the right type of peer review is an important part of effectively and efficiently
managing EPA's TSCA program. Letter peer review is one means for EPA to obtain independent
scientific advice to inform its evaluation of chemical risks.
EPA applies the criteria in its own Peer Review Handbook (https://www.epa.gov/osa/peer-review-
handbook-4th-edition-2015) to determine on a case-by-case basis which components of each risk
evaluation are peer reviewed. Per this handbook and OMB guidance, the EPA determined that a letter
peer review was an adequate mechanism for the TCEP Risk Evaluation, "giving due consideration to
the novelty and complexity of the science to be reviewed, the relevance of the information to decision
making, the extent of prior peer reviews, and the expected benefits and costs of additional review."
The Handbook states, "a letter review takes place when EPA seeks individual written peer review
comments from independent experts, typically in the form of correspondence to EPA from the peer
reviewer... Each reviewer evaluates the draft technical work product independently without
consultation with other reviewers. No collaborative or consensus peer review report is developed."
4
-------
Neither the Agency regulations nor OMB guidance require a specific type of peer review (e.g.,
committee consensus report or letter peer review) or a particular body to conduct the review.
Comments regarding the peer review charge questions
Summary: A public commenter (0044) questioned the scope of EPA's peer review charge questions.
The commenter said that the charge questions are narrow in scope, while ignoring major, overarching
concerns such as data gaps. The commenter recommended that EPA seek SACC's advice on
addressing data gaps in the risk evaluation.
Similarly, a peer reviewer (OC - R5) said that there are too many detailed charge questions to address
effectively without a full SACC dialogue.
EPA Response: The EPA limited the scope of the charge questions to subjects that had not previously
been peer reviewed within the TSCA framework. The charge questions focused on novel methods
used, and peer reviewers were chosen based on their expertise and ability to answer any one of the
subjects covered in the charge questions. However, peer reviewers have the freedom to comment on
other aspects of the risk evaluation if they so choose.
As stated previously, per EPA's Peer Review Handbook, the EPA determined that a letter peer review
was an adequate mechanism for the TCEP Risk Evaluation, "giving due consideration to the novelty
and complexity of the science to be reviewed, the relevance of the information to decision making, the
extent of prior peer reviews, and the expected benefits and costs of additional review."
Summary: A public commenter (0043) requested that EPA include the following questions in its peer
review charge questions:
Has EPA taken into consideration the lower exposure potential associated with a chemical that
is bound up in an article? For example, has EPA assessed the exposure to TCEP in cured paints
versus as a paint additive?
Has EPA appropriately assessed exposure to articles and replacement parts as directed by
TSCA Section 6(2)(D) and (E), specifically for automotive articles and replacement parts.
Suggest adding a COU for automotive equipment and replacement parts.
Is it appropriate for EPA to assume that no personal protective equipment (PPE) or other
exposure controls are in place in workplaces where TCEP is present? Does this approach
overestimate the risk? At a minimum, should EPA conduct exposure assessments that assess
risk both with and without the Occupational Safety and Health Administration (OSHA)-
required risk mitigation requirements?
EPA Response: By the time EPA received these comments, charge questions could not be added to
our letter peer review process as we were in the middle of the peer review.
Consumer articles were modeled using the various models within CEM 3.0. These models include:
E6 Emission from article placed in environment
A INH1 Inhalation from article placed in environment
A_ING1 Ingestion after inhalation
A_ING2 Ingestion of article mouthed
A ING3 Incidental ingestion of dust
A DERI Direct transfer from vapor phase to skin
5
-------
ADER2 Dermal dose from article where skin contact occurs
ADER3 Dermal dose from skin contact with dust
Each of these models are described in the EPA's CEM Version 3.0 User Guide and Appendices {U.S.
EPA, 2022, 11204170}. The TCEP risk evaluation briefly discusses uncertainties related to how TCEP
is bound to articles, in its discussion of curing of resins, whether TCEP is entrained in the polymer
matrix, and the discussion of blooming potential of TCEP to the surface of treated articles (Sections
4.3.6.2, 5.1.2.2, 5.1.2.4.1, 5.3.2.1.1, 5.3.2.3.2).
EPA evaluated automotive articles and replacement parts in the final risk evaluation of TCEP. The
changes made in the final risk evaluation are in the following conditions of use:
"Processing - incorporation into article - aerospace equipment and products" was edited to
"Processing - incorporation into article - aerospace equipment and products and automotive
articles and replacement parts containing TCEP;"
"Industrial use - other use - aerospace equipment and products" was edited to "Industrial use -
other use - aerospace equipment and products and automotive articles and replacement parts
containing TCEP;"
"Commercial use - other use - aerospace equipment and products" was edited to "Commercial
use - other use - aerospace equipment and products and automotive articles and replacement
parts containing TCEP;" and
"Consumer use - paints and coatings - paints and coatings" was edited to "Consumer use -
paints and coatings - paints and coatings, including those found on automotive articles and
replacement parts."
In addition to these changes, a new condition of use, "Industrial use - paints and coatings - automotive
paints and coatings" was included in the final risk evaluation of TCEP.
Most of these conditions of use were assessed qualitatively and similarly to how aerospace equipment
and products and commercial use of paints and coatings were evaluated. An overview of the risk
characterization approach for the aerospace equipment and products COUs can be found in Section 5 of
the draft risk evaluation of TCEP, Risk Characterization. EPA had previously assumed that workers
were provided and always used PPE in a manner that achieves the stated assigned protection factor
(APF) for respiratory protection or used impervious gloves for dermal protection. However, EPA
believes that the assumed use of PPE in a risk determination could lead to an underestimation of the
risk to workers. For example, workers may be highly exposed because they are not covered by OSHA
standards, such as self-employed individuals and public sector workers who are not covered by a State
Plan; their employers are out of compliance with OSHA standards; the PPE is not sufficient to address
the risk from the chemical; or their PPE does not fit or function properly.
Additionally, TSCA risk evaluations are subject to statutory science standards, an explicit requirement
to consider risks to potentially exposed or susceptible subpopulations, and a prohibition on considering
costs and other non-risk factors when determining whether a chemical presents an unreasonable risk
that warrants regulatory actionsall requirements that do not apply to development of OSHA
regulations.
Summary: A public commenter (0052) recommended that EPA include as a peer review charge
question whether the 2016 study on Japanese medaka constitutes sufficient evidence for a
determination of unreasonable risk to the environment. The commenter noted that the study found a
6
-------
significant effect on body length of medaka embryos but did not find impact on other markers of
embryonic development.
EPA Response: By the time EPA received these comments, charge questions could not be added to
our letter peer review process as we were in the middle of the peer review.
EPA has identified and reviewed additional studies with TCEP exposure to aquatic invertebrates and
invertebrates to improve the Environmental Hazards section of the TCEP Risk Evaluation and is
presented within Section 4.2.2 Aquatic Species Hazard. This work improves the representation of
species within the Web-ICE application and SSD for deriving an acute Concentration of Concern and
increased the number studies representing chronic exposures of TCEP for the identification of a
chronic Concentration of Concern for aquatic organisms.
Section 2.2 - Environmental fate and transport (including comments on the approach to estimate
degradation in the absence of data)
General comments
Summary: A peer reviewer (1.1.1 - R3) stated that, in the absence of relevant data, relying on
anaerobic degradation being 4 to 9 times slower than aerobic degradation is an appropriate assumption
and approach to estimate the degradation of TCEP.
EPA Response: The EPA thanks the reviewer for their comment. Additional anaerobic studies were
identified and are provided in Appendix F.2.3.2., and additional anaerobic degradation studies were
identified and are discussed in Appendix E.2.3.2
Summary: A peer reviewer (1.1.1 - R2) and a public commenter (0040) recommended that EPA
characterize TCEP as not biodegradable in aerobic or anaerobic conditions.
According to the public commenter (0040). the BIOWIN models in EPI Suite may not be
suited to analysis of TCEP given the limited data on which they are built. The commenter said
that there is not sufficient data to quantify the biodegradable half-life of TCEP and that such a
quantification is not material to the risk evaluation.
In addition, a peer reviewer (1.1.1 - R2) recommended that EPA conclude that TCEP is not
biodegradable under anaerobic conditions in sediments. The peer reviewer (1.1.1 - R2) stated
that there are large uncertainties regarding the biodegradation of TCEP in water and sediment,
and two studies on the biodegradation of TCEP in water found that TCEP is not readily
biodegradable, using Organisation for Economic Co-operation and Development (OECD)
30IB and OECD 30ID methods. The peer reviewer said that it was unclear why it was
necessary to default to a water half-life of 10,000 hours and then extrapolate aerobic to
anaerobic half-lives.
7
-------
EPA Response:
For the public commentor (0040) and peer reviewer (1.1.1 - R2):
The EPA thanks the reviewers for their comment. Additional anaerobic studies were identified
and are provided in Appendix F.2.3.2., and additional anaerobic degradation studies were
identified and are discussed in Appendix E.2.3.2
(see below).
"After systematic review concluded, four studies observing anaerobic biodegradation of TCEP
were identified. {Pang, 2018, 10228662@@author-year} studied the fate of TCEP in sewage
sludge with four different treatments, two of which were anaerobic digestion studies
(Treatments 3 and 4). The results of this study confirm the recalcitrance of TCEP under
anaerobic conditions and TCEP concentrations in both treatments were seen to increase over
time. The removal rates were -0.41 and -74.8 percent for Treatments 3 and 4, respectively.
The authors did not discuss the reasons for the increase in TCEP's final concentrations, but the
authors did note that the final concentrations of Tri(n-butyl) phosphate (TnBP) increased
during composting due to the ""inhomogencity" of the composts. Because of this uncertainty,
{Pang, 2018, 10228662@@author-year} is used qualitatively to support TCEP's persistence
in anaerobic environment and will not be used to derive a biodegradation half-life. {Yang,
2021, 8682618@@author-year} carried out biodegradation experiments using activated
sludges from aerobic and anerobic ponds of the Nanjing Chendong STP in China mixed with
kitchen garbage biomass and agricultural residues. The biodegradation of TCEP under
anaerobic conditions was observed to be very slow. The removal of TCEP did not reach 80
percent after 60 days. Because the media and inoculum used in {Yang, 2021,
8682618@@author-year} comprised a rich nutrient profile and high microbial biomass, the
direct use of the reported half-lives for estimating persistence in environmental sediments is
not appropriate; therefore, this study was excluded from use in this risk evaluation. {Yang,
2023, 11364896@@author-year} studied the degradation of TCEP using an anaerobic
enrichment culture from end-of-life vehicles dismantling sites in Guangzhou, China. The
study demonstrated that microbes, specifically Dehalococcoides sp., was able to anaerobically
transform TCEP. However, the microbial community was highly enriched and the system
highly controlled. This study will not appropriately represent TCEP's fate under anaerobic
conditions and will not be considered in this risk evaluation.
When considering the results of the {Pang, 2018, 10228662@@author-year} and {Life
Sciences Research Ltd, 1990, 6310865(ff>(ff>author-year} studies (Table 2-2), it is highly likely
that TCEP will not degrade under anaerobic conditions and be persistent in the sediment
compartment/'
Summary: A peer reviewer (1.1.1 - R5) wrote that anaerobic degradation tests of a compound of this
nature should be simple to complete and should be required for inclusion. The peer reviewer added
that, if these data are not provided, the upper limit of persistence should be incorporated into the risk
evaluation.
EPA Response: The EPA thanks the reviewer for their comment. The EPA was unable to use test
orders since there were no manufacturers to submit them to. The upper limit of persistence that the
EPA used is the 30,000-40,000 hours that was extrapolated using the results from an aerobic
biodegradation test.
8
-------
Comments on the references used
Summary: Two peer reviewers (1.1.1 - Rl. 1.1.1- R6) recommended EPA consider additional recent
publications. One peer reviewer (1.1.1 - Rl) stated that EPA identified one medium quality study in the
draft risk evaluation, but the study did not report any degradation of TCEP and was therefore less
useful. The peer reviewer stated that they found an article by Yang et al. (2021) that measured the
degradation of TCEP in activated sludge under anaerobic conditions, and the peer reviewer
recommended that EPA consider if this article could be used to provide a better estimate of the
anaerobic degradation of TCEP. Another peer reviewer (1.1.1- R6) recommended that EPA consider
the following recent publications describing anaerobic degradation:
Yang S, Wu J, Wang H, Yang Q, Zhang H, Yang L, Li D, Deng Y, Zhong Y, Peng P. 2023.
New dechlorination products and mechanisms of tris(2-chloroethyl) phosphate by an anaerobic
enrichment culture from a vehicle dismantling site. Environ Pollut 338:122704.
https://doi.Org/10.1016/j.envpol.2023.122704.
Yang X, Fan D, Gu W, Liu J, Shi L, Zhang Z, Zhou L, Ji G. 2021. Aerobic and Anaerobic
Biodegradability of Organophosphates in Activated Sludge Derived from Kitchen Garbage
Biomass and Agricultural Residues. Front Bioeng Biotechnol 9:649049.
tps://doi.org/10.3389/fbioe.2021.649049.
Pang L, Ge L, Yang P, He H, Zhang H. 2018. Degradation of organophosphate esters in
sewage sludge: Effects of aerobic/anaerobic treatments and bacterial community compositions.
Bioresour Technol 255:16-21. https://doi.Org/10.1016/i.biortech.2018.01.104.
Both peer reviewers (1.1.1 - Rl. 1.1.1- R6) expressed that, when limited studies are available, the
approach taken by EPA is acceptable.
EPA Response:
Yang, S. et al. (2023) demonstrated that microbes, specifically Dehalococcoides sp., was able
to anaerobically transform TCEP. However, because the microbial community was highly
enriched and the system highly controlled, this study does not appropriately represent the
anaerobic degradation of TCEP"s fate under standard environmental conditions and will not be
considered in this risk evaluation.
Yang, X. et al. (2021) carried out biodegradation experiments using activated sludges from
aerobic and anerobic ponds of the Nanjing Chendong STP in China mixed with kitchen
garbage biomass and agricultural residues. Because the media and inoculum comprised a rich
nutrient profile and high microbial biomass which does not represent standard environmental
conditions, the direct use of the reported half-lives for estimating persistence in environmental
sediments is not appropriate.
Pang et al. (2018) confirms the recalcitrance of TCEP under anaerobic conditions, and TCEP
concentrations in both treatments were seen to increase over time. The authors did not discuss
the reasons for the increase in final TCEP concentrations, but the authors did note that the
final concentrations of Tri(n-butyl) phosphate (TnBP) increased during composting due to the
""inhomogencity" of the composts. Because of this uncertainty, this study will be used
qualitatively to support TCEP's persistence in anaerobic environment and will not be used to
derive a biodegradation half-life.
Since the studies will not be used to derive an aerobic biodegradation half-life, the EPA will
use the extrapolated half-life derived from 64 FR 60197 and {U.S. EPA, 2000,
10286742@@author-year} as described in the draft risk evaluation.
9
-------
Summary: A peer reviewer (1.1.1- R4) stated that they were unable to track down the reference
supporting the assumption that anaerobic degradation is 4 to 9 times slower than aerobic degradation.
EPA Response: Guidance on estimating anaerobic degradation rates relative to aerobic degradation
rates can be found in Section III.D in 64 FR 60197. and in the documentation (Help files) for the Level
III fugacity model of EPI Suite{U.S. EPA, 2012, 2347246}. Note; anaerobic degradation rates were
not used as inputs for the AERMOD or DRAS models. EPA used the 30,000-40,000 hours that was
extrapolated using the results from an aerobic biodegradation test.
Summary: A peer reviewer (1.1.1 - R5) stated that the citations for chemical properties in Table 2-1
are deceptive. The peer reviewer expressed that EPA should acknowledge that the values used are
based on a single study, and the peer reviewer added that this type of dependence on a single value that
cascades into many references is a significant shortcoming of many data mining techniques. The peer
reviewer stated that safeguards to prevent such citation inflation should be established.
EPA Response: The EPA thanks the reviewer for their comment. The citations in Table 2-1 were
revised, as recommended.
Other specific comments
Summary: A peer reviewer (1.1.1 - R5) commented on line 1312 on page 42 and stated that TCEP
accumulation in soil may occur where rainfall is limited. The peer reviewer said that this could be
important for threatened or endangered species living in areas where airborne deposition is likely.
EPA Response: The EPA thanks the reviewer for the comment. Precipitation events, such as rain and
snow may enhance soil concentrations of TCEP, but accumulation in soil is expected to be unlikely due
to its water solubility (7,820 mg/L) and soil log Koc values (2.08-2.52). Dissolved TCEP was observed
to be mobile and migrated to groundwater by common soil transport processes such as advection and
diffusion.
Section 3.1 - Approach and methodology
Summary: A public commenter (0045) discussed EPA's approach to develop quantitative estimates of
exposure and hazards where there is a lack of empirical data. The commenter expressed support for the
use of vetted and validated models in evaluating exposure and hazards for risk assessments, stating that
they believed it would help fill in critical data gaps.
EPA Response: EPA thanks the reviewer for this comment.
Section 3.2 - Environmental releases
Summary: A public commenter (0044) stated that EPA underestimated the potential environmental
releases, occupational exposures, and consumer exposures to TCEP associated with e-waste recycling
in the assessment of industrial and commercial sources in the environment. The commenter stated that
EPA had not provided adequate information and justification for not considering e-waste recycling
10
-------
releases in the draft risk evaluation and suggested EPA consider the data from Gravel et al. (2019) and
Nguyen (2019).
EPA Response: EPA thanks the reviewer for their comment and reviewed the sources provided in the
comment. These sources did not provide new data that could be used to estimate releases that could
occur at electronic waste facilities. The rational for why EPA could not quantify e-waste recycling
releases is because data regarding releases was not available. Literature that contained data regarding
TCEP at e-waste facilities was limited to inhalation monitoring data, which represented very low levels
of exposure. Additionally, the total releases are expected to be low for several reasons: The volume of
TCEP in e-waste products is low; only a fraction of the products is recycled; and recycling will likely
be dispersed over many e-waste sites. Further explanation is located in sections 4.3.6.2 and 5.3.2.3.2.
Section 3.3 - Concentrations of TCEP in the environment
Disposal and groundwater pathways
Summary: In response to Charge Question 3.3, a public commenter (0040) stated that leachate from
poorly managed landfills represents a risk to human health regardless of its contents and thus, further
analysis into this exposure pathway is unwarranted in the draft risk evaluation.
EPA Response: Spills and natural disasters are generally not considered in the risk evaluations because
they are not known or reasonably foreseen. EPA is aware of older, unlined, and poorly managed
landfills across the United States which suggest that landfill leachate to groundwater may pose an
exposure to the populations living near landfills. Three sites, the Norman Landfill in OK, the Himco
dump in IN, and Ft. Devens, MA detected TCEP in groundwater plumes, suggesting landfill leachate as
a concern for nearby residents ({Barnes, 2004, 5469339; Buszka, 2009, 4912133; Hutchins, 1984,
1316091}). EPA used a range ofleachate concentrations from 1.0x10-4 to 1.0x103 mg/Lto account
for uncertainties. The top range of the leachate concentration is bounded by solubility.
Summary: Alternatively, other public commenters (0041. 0044) stated that TCEP in wastewater
effluent and leachate from landfills and dumps represent additional exposures to aquatic organisms and
should be included in risk management efforts. One public commenter (0044) suggested EPA consider
data published in studies from Los Angeles and the Washington Department of Ecology, stating that
EPA should revise the draft risk evaluation to incorporate accurate loading rates for TCEP-containing
waste disposed via landfilling, incineration, and wastewater to avoid substantially understating
disposal-related exposures and risks.
EPA Response: EPA appreciates notification of these additional references. The studies from Los
Angeles and the Washington Department of Ecology have been incorporated into our revised
assessment (section 3.3.3.7). The revised DRAS analysis considers waste volumes up to 817 m3
(2,500,000 lbs) to account for the uncertainty of TCEP waste volumes associated with past disposals of
TCEP. This value is two orders of magnitude higher than the 25,000 lbs CDR threshold.
Masoner et al. 2014 is more comprehensive in describing mass loading and leachate loading. The study
describes leachate loading rates for different regions of the U.S., for different ages of waste, for
different waste loads, leachate production and precipitation (Table 4).
11
-------
Section 3.4 - Concentrations of TCEP in the indoor environment
Summary: A public commenter (0044) said that EPA had understated the consumer, worker, and
general population exposures to TCEP from polyurethane foam products found in furniture and
automobile cushioning, urging EPA to use the higher concentrations found in the Ingerowski study
published in 2001.
EPA Response: EPA thanks the commenter for the comment. Although {Ingerowski, 2001,
32734@@author-year} recorded TCEP in polyurethane soft foam at 19,800 mg/kg (1.98 percent),
values from {Fang, 2013, 1676728@@author-year} were selected forthis furniture foam and auto
foam scenarios as they were thought to be more current to the amount of TCEP in polyurethane soft
foam in use across the U.S. .
Summary: A public commenter (0038) stated that the dust exposure pathway does not account for
product phase out or changes to the composition of products during the 19 years of data that were
evaluated and instead suggested EPA review a 2016 study that found decreased concentrations of flame
retardants in dust due to a transition in products.
EPA Response: Decreasing trends in the production volume of TCEP are described in Section
5.1.2.4.1. This section describes California TB 117-2013 which may be responsible forthe shift in
TCEP use in consumer articles. Section 3.4.2 describes the available indoor dust monitoring data
without regard to the changing trends in the monitoring data.
Summary: A peer reviewer (OC - R7) suggested that EPA revisit the calculations that derive the
indoor/outdoor ratio of 0.65, which was not presented in the draft risk evaluation. The reviewer said
that page 61 shows outdoor air measurements between 0.1 and lng/m3 with a maximum across studies
of approximately 50 ng/m3 in a study labelled as "near facility, highly exposed." However, on page 81,
EPA lists a range of indoor air concentrations between 10-100 ppm, with several approaching 1,000
ppm. Thus, the reviewer said it is difficult to see how a ratio of 0.65 could be predictive of indoor air.
The review said that a statistical summary of the available indoor air data and picking the 95th
percentile may be most appropriate.
EPA Response: The IIOAC model is a tool based on EPA's AERMOD that is used for assessing
releases to air and exposure potential to chemicals. An indoor-outdoor ratio of 0.65 is used forthe mean
ratio, whereas a value of 1 is used for the high-end ratio. Figure 19 in Section 10.3 in the IIOAC user
guide provides a summary of a range of indoor-outdoor ratios and justifies why 0.65 was used as the
default forthe screening model.
Indoor air concentrations due to use of consumer articles were modeled using the consumer exposure
model (CEM), in addition to the modeling of ambient air concentrations using IIOAC/AERMOD.
Section 4 - Environmental Risk Assessment
Comments associated with this issue are summarized in the subsections below.
Section 4.1 - Environmental exposures
Summary: A public commenter (0044) stated that EPA had not adequately assessed the risks to
wildlife from TCEP exposure, stating that EPA omitted conditions of use such as releases from
12
-------
consumer products and disregarded certain exposure pathways such as inhalation because they were
seen to present a lower risk than others like dietary exposures. The commenter also suggested that EPA
consider the impacts of microplastics and plastics debris ingestion when evaluating ingestion exposure.
In the final risk evaluation, the commenter requested that EPA consider aggregate and cumulative risks
to wildlife and the bioaccumulation exposure from birds ingesting fish and insects.
EPA Response: Considerations such as: air concentrations of modeled TCEP releases, monitored
TCEP within ambient air, and ecologically relevant hazard thresholds for TCEP indicated that
inhalation for wildlife is not a significant exposure route. TCEP in the gaseous phase is expected to
have a short half-life within the atmosphere (5.8 hours) with particle bound TCEP primarily removed
by wet and dry deposition. AERMOD modeling results indicated a maximum ambient air concentration
(95th percentile, MetHIGH) of 6.08/ 10 7 (ig/m3 at 1,000 m from a hypothetical facility forthe Use of
paints and coatings - spray application OES under the 2,500 lb/year production volume using the
Suburban forest land category scenario (see Section 3.3.1.2). Comparatively, the hazard threshold
derived for terrestrial mammals, represented as a toxicity reference value based on an oral route of
exposure, is 44 mg/kg-bw/day. TCEP sampled from ambient air at 13 urban sites in Chicago ranged
from 0.032 to 0.34 ng/m3 with an average 0.75 + 0.025 ng/m3 {Peverly 2015; 2939998}. In addition, the
human health portion of the TCEP risk assessment for general population risk estimates found non-
cancer MOEs were less than the acute and chronic benchmark MOEs for any COUs at 10 m from stack
or fugitive are releases, which are higher exposures than would be expected to wildlife would inhale
with farther residence to a TCEP fugitive or stack release.
The role of consumer products and environmental releases within section 3.2.2 detail the following,
"Although EPA acknowledges that there may be TCEP releases to the environment via the demolition
and disposal of consumer articles and the release of TCEP to wastewater via domestic laundry, the
Agency did not quantitatively assess these scenarios due to lack of reasonably available information.
EPA instead assessed them qualitatively. EPA uses anecdotal information in the peer-reviewed
literature to qualitatively describe potential releases from TCEP d to the environment from consumer
articles (Section 5.1.2.2.5)."
The impact of co-exposure with other chemicals is outside of the scope of the risk evaluation for TCEP.
The purpose of the risk evaluation under TSCA is to determine whether a chemical substance presents
an unreasonable risk of injury to health or the environment, under the conditions of use. Appendix
Section F.2.3.1 of the TCEP risk evaluation (RE) recognizes that microplastics can serve as both a
source and sink for TCEP and present an uncertainty for the source attribution of TCEP from an
original article.
Summary: A public commenter (0045) stated that EPA should include other possible exposure routes
from inhalation, deposition from air to surface water, and dermal contact for terrestrial species and
consider the impact of an aggregate exposure assessment to fully characterize environmental risks.
EPA Response: Inhalation exposure to wildlife has been addressed within a previous comment.
Dermal contact for terrestrial species is addressed within Section 4.3.5 with a section that addresses
both dermal contact and inhalation by wildlife. Based on the Guidance for Ecological Soil Screening
Levels (US EPA, 2003a, US EPA 2003b) document, for terrestrial wildlife, relative exposures
associated with inhalation and dermal exposure pathways are insignificant, even for volatile substances,
compared to direct ingestion and ingestion of food (by approximately 1,000- fold). In addition, TCEP is
not expected to bioaccumulate in tissues, and the screening level trophic transfer analysis indicated that
concentrations will not increase from prey to predator in either aquatic or terrestrial food webs.
13
-------
Deposition of TCEP from air to surface water was modeled and is detailed within Section 3.3.2.6 of the
TCEP risk evaluation. EPA used IIOAC and AERMOD to estimate air deposition from hypothetical
facility releases and to calculate pond water concentrations 1,000 m from the hypothetical facility. Pond
water concentrations from air deposition were estimated for the COUs with air releases. The annual
highest estimated 95th percentile pond water concentration from one year of deposition, across all
exposure scenarios, was 0.49 ppb for the commercial use of paints and coatings scenario at an annual
production volume of 2,500 lb per year.
Section 4.2 - Environmental hazards (including comments on the calculation of hazardous
concentration for 5% of species (HC05) and the calculation of concentration of concern (COC) for
aquatic organisms)
General comments
Summary: A public commenter (0040) recommended that EPA identify appropriate hazard endpoints
as part of the problem formulation of the risk evaluation in accordance with EPA guidance on
ecological risk assessment. The commenter discussed an example where they believed EPA had
misapplied the results of the Sun et al. (2016) study to establish chronic fish hazards. The commenter
also requested that EPA describe the weight of the scientific evidence for identified ecological hazards
and apply it to the determination of a chronic COC. Another commenter (0044) stated that EPA's
environmental hazard thresholds were not sufficiently protective of aquatic species and cited a study by
Zhao et al. (2021) that documented adverse effects to fish species at lower concentrations of TCEP.
EPA Response: The narrative summary of the Sun et al. 2016 study has been revised to emphasize that
the frequency of water changes, nominal concentrations (not analytically verified), and lack of other
significant adverse outcomes which are points of uncertainty that the commenter discussed within their
review of the study. The environmental hazard section of the TCEP risk evaluation (RE) has been
revised to reflect an increased number of acute and chronic exposure studies (Section 4.2.2 and 4.2.5)
that have improved the quality of the database, consistency, and biological relevance (Section 4.2.6.1).
The revised TCEP RE has additional studies with overall quality determinations of high for the apical
assessment endpoints of regulatory interest (i.e., impaired growth, survival, or reproduction).
Responses to Charge Question 2.1 related to data
Summary: In response to Charge Question 2.1, a public commenter (0040) noted that Dyer et al.
(2006) developed species sensitivity distributions using EPA's interspecies correlations estimation tool
but observed variability in the results based on the surrogate species applied. The commenter
recommended that EPA apply its method beyond fish species to include freshwater invertebrates and
evaluate the appropriateness of including algal taxa in such an approach. Another commenter (0045)
discussed the strengths and uncertainties of using HC05 in the TCEP draft risk evaluation and
suggested that EPA provide additional support to conduct testing early in the review process and
combining test results with modeling with other in silico, in vitro, and less whole-animal dependent
tools to help fill data gaps.
EPA Response: The application of Web-ICE and subsequent SSD in the draft TCEP RE has been
revised to include additional vertebrates and invertebrates. EPA calculated an SSD with the SSD
Toolbox using acute LC50 hazard data from systematic review and estimated data from the Web-ICE
14
-------
application (Appendix F.2.1.1) that included 25 fish, 1 amphibian, 12 aquatic invertebrates, and 15
benthic invertebrates. A second SSD was performed for the algae hazard data and was applied with the
5 empirical values and 3 predicted species values detailed in the previous section. The SSD is used to
calculate a hazardous concentration for 5 percent of species (HC05). The HC05 estimates the
concentration of TCEP that is expected to be protective for 95 percent of species. The SSD plot shows
the distribution of invertebrate and vertebrate species sensitivity to TCEP exposure using Bun-
distribution with the calculated HC05 of 31.6 mg/L with a 95 percent CI of 16.7 mg/L to 57.0 mg/L.
For algal species the HC05 for the logistic and normal distributions were 116.2 and 104.2, respectively,
with both distributions resulted in the same lower 95% CI of the HC05 at 66 mg/L.
Summary: Several peer reviewers (2.1.1 - Rl. 2.1.1 - R5. 2.1.1 - R6) expressed that having TCEP
data for more biological species to inform the assessment would be preferable. One peer reviewer
(2.1.1 - R6) specifically noted that the Weibull distribution had the best goodness-of-fit and the
method was appropriately used, but expressed concern about the species that were used to obtain the
lethal concentration at which 50 percent of test organisms die (LC50) toxicity data. The peer reviewer
stated that only rainbow trout and zebrafish studies were used to predict LC50 toxicity values for 18
additional aquatic organisms.
Another peer reviewer (2.1.1 - Rl) stated that the use of Web-based Interspecies Correlation
Estimation (Web-ICE) to estimate the acute toxicity of TCEP for species for which no data is available
seems reasonable, but the estimates are based on a small number of fish species which may not
accurately reflect the responses of other species. The peer reviewer said that they had found several
additional sources in a quick literature search and listed the following:
A recent article by Zhang et al. (2024) provides TCEP acute toxicity data for 8 marine species
ranging from diatoms to fish.
Acute toxicity values for 6 other freshwater species are presented in the Introduction to the
Zhang et al. (2024) article.
Qiao et al. (2022) presents tables with toxicity values for TCEP in a range of species and at
various durations of exposure.
The European Chemicals Agency database entry for TCEP shows an effect concentration at
which 50 percent of test organisms exhibit an effect (EC50) value for immobilization of
Daphnia magna of 170 mg/kg, and an algal growth rate inhibition value of 450 mg/L.
A recent article by Zhang et al. (2022) includes growth curves, which could be used to estimate
an EC50 value for the cyanobacteria Microcystic aeruginosa.
The peer reviewer (2.1.1 - Rl) recommended that EPA examine these sources and include this
information in its estimation of the species sensitivity distribution and HC05, to create a more robust
dataset. The peer reviewer expressed concern that they were able to find multiple relevant studies that
are missing from the document without difficulty, and the peer reviewer stated that this suggests
significant limitations with the current literature review approach.
EPA Response: EPA thanks the commenter for the comment and input to improve this risk evaluation.
EPA has been able to add additional studies from systematic review that increase our data landscape
for acute and chronic hazard. Revision of the Web-ICE model in the TCEP RE has taken place with the
additional acute hazard data. Revisions have been made for the additional empirical and predicted
values used in SSD and are documented within Appendix Section F.2.1.1 for Web-ICE and F.2.1.2 for
the SSD. Additional acute hazard values from Zhang et al. 2024 were placed through systematic review
15
-------
and were added to the Environmental Hazard section for vertebrates, invertebrates, green algae, and
diatoms.
Additional surrogate species allow for more predicted species from Web-ICE thus increasing our
representative species within the SSD. Details of this increases data landscape and incorporation into
the TCEP risk evaluation are detailed within Sections 4.2 (Environmental Hazard) and Appendix F.
Responses to Charge Question 2.1 related to the modeling approach
Summary: A peer reviewer (2.1.1 - R2) expressed that this is the first time Web-ICE has been used in
a TSCA risk evaluation, and the peer reviewer commented that this is commendable given the effort
invested in developing and maintaining this tool. The peer reviewer said that it is important for future
revisions to emphasize the scientific soundness of Web-ICE and its associated ICE models. The peer
reviewer added that the models have been vetted and validated by EPA scientists and the larger
scientific community. The peer reviewer specifically stated that the statement "'Model approaches such
as Web-ICE have more uncertainty than empirical data and are not substitutes for empirical data when
determining the hazard or risk" should be revised, as there is a substantial body of literature that
indicates that Web-ICE predictions can produce comparable hazard estimates as those derived from
empirical data.
EPA Response: The statement on the uncertainty of model approaches such as Web-ICE has been
removed from Section 4.2.6.1 Strengths, Limitations, Assumptions, and Key Sources of Uncertainty
for the Environmental Hazard Assessment.
Summary: A peer reviewer (2.1.1 - R2) stated that the risk evaluation summarized previously
recommended quality criteria for the inclusion of Web-ICE predictions in the assessment, with the
addition of narrow 95% confidence intervals. The peer reviewer expressed that this should be revised,
because a recent study (Raimondo et al., 2023) indicated that the range between the upper and lower
confidence interval is not correlated with the prediction accuracy. The peer reviewer added that an
interval range of 2-orders of magnitude has been recommended instead, because it minimizes the Type
I and Type II errors. The peer reviewer stated that making this modification would increase the
predicted acute toxicity values from 18 aquatic species to 30 aquatic species. In addition, the peer
reviewer said that the prediction for 4 fish species would be based on values from both surrogates
(rainbow trout) and zebrafish embryo. The peer reviewer included a table with the Web-ICE predicted
acute toxicity based on their suggested modifications. The peer reviewer commented that earlier
assessments have also suggested that it is preferable to develop species sensitivity distributions from
multiple, instead of single, surrogate species. The peer reviewer recommended EPA integrate
additional high-quality data into the assessments. Finally, the peer reviewer stated that EPA's approach
for calculating the HC05 is scientifically sound, and it follows probabilistic methods used elsewhere for
similar purposes.
EPA Response: EPA has reviewed the criteria from Raimondo et al. (2023) and has been in contact
with the study's authors. The revision of the Web-ICE and SSD with additional empirical data resulted
in 40 species represented in the SSD. The Web-ICE screening for predicted values has been revised to
include the reviewer's suggestion and peer-reviewed findings on the role of interval range from one to
two orders of magnitude. The number of empirical species represented in the overall SSD and the Web-
ICE model have increased from the inclusion of newly reviewed studies.
16
-------
Summary: A peer reviewer (2.1.1 - R3) stated that, while a formal systematic review criteria was used
to identify and categorize data and papers, EPA incorporated a further step to assess the quality of the
science. The peer reviewer wrote that papers were not automatically excluded due to lack of meeting
the systematic review automatic criteria, but rather assessed for possible quality contribution to the
assessment. Finally, the peer reviewer commented that the analysis was not done in isolation, and it
included other assessment reports from 2009.
EPA Response: Thank you for your comment and review.
Summary: A peer reviewer (2.1.1 - R4) stated that EPA clearly described the approach for selecting
the model that best fits the data distribution. The peer reviewer commented that the Weibull
distribution was determined to best fit the distribution of empirical data, but, at the lower end of the
distribution, the empirical data lie above the 1:1 line on the Q-Q plot, which indicates that the empirical
data points are higher than the predicted values. The peer reviewer stated that EPA could consider
modeling the lower end of the distribution separately from the entire distribution. The peer reviewer
also acknowledged that the 95% confidence interval includes the empirical data points.
EPA Response: Based on this comment the revised SSD has included the Q-Q plots from more fit
distributions to show these data. Revisions also included more narrative on selection criteria within
Appendix F.2.1.2 following the guidance detailed within {Etterson, 2020, 5085638}.
Rather than model the lower distribution from the whole, after screening, the predicted acute toxicity
values for 46 additional aquatic organisms (22 fish, 1 amphibian, 9 aquatic invertebrates, 14 benthic
invertebrate species) were added from the surrogate rainbow trout, zebrafish, and Daphnia magna data
(Table_Apx F 1). The toxicity data were then used to calculate the distribution of species sensitivity to
TCEP exposure through the SSD toolbox as shown in Figure Apx F 4 and Table 4 6 {Etterson, 2020,
5085638}. The distribution of acute hazard values from vertebrates and invertebrates was examined to
determine if these two groups had similar sensitivity to TCEP. The mean LC50 values for the 26
vertebrates and 27 invertebrates were 399 + 54 mg/L (+ SEM) and 197 + 51 mg/L and were not
significantly different (F = 1.87, df = 1,51, P = 0.17). Two predicted values, representing one vertebrate
(cape fear shiner) and one invertebrate (tubifex worm), were each four standard deviations from their
respective means. When statistical analysis was conducted with these two values omitted as outliers the
two groups were not significantly different (F = 6.24, df = 1,49, P = 0.015) with mean LC50 values of
253 + 30 mg/L and 152 + 27 mg/L for vertebrates and invertebrates, respectively. Acute hazard values
for invertebrates divided between benthic and water column species were not significantly different (F
= 1.01, df = 1,25, P = 0.32) with mean LC50 values of 243 + 90 mg/L and 138 + 27 mg/L, respectively.
Summary: A peer reviewer (2.1.1 - R5) stated that the EPA approach assumes that acute toxicity is
the basis upon which aquatic toxicity should be determined, and this is not realistic. The peer reviewer
wrote that chronic exposures are much more likely, and HC05s should be based on chronic toxicity
values. The peer reviewer stated that if acute toxicity data are used for estimates, adjustment factors are
needed. The peer reviewer noted that there are reasonable estimates of acute to chronic aquatic toxicity
values and referenced Kienzler et al. (2017). The peer reviewer stated that both acute to chronic
adjustments and interspecies extrapolation assessment factors are needed.
EPA Response: Thank you for your submission of Kienzler et al. (2017) and it is of particular interest
since the revised TCEP draft now contains acute and chronic hazard data for daphnia from the Japanese
Ministry of the Environment (Table 4-3). In contrast to the draft TCEP RE, the revised work also has
17
-------
five additional chronic exposure studies with fishes (Table 4-2) where the draft TCEP chronic COC
was derived from an empirical sub chronic hazard endpoint from a 14 exposure of TCEP to embryo
and larvae. The revised Environmental Hazard section of the TCEP risk evaluation now reviews these
chronic exposure studies and uses the apical endpoints (e.g., survival, reproduction, growth) in those
works to designate a COC based on empirical evidence.
Other comments in response to Charge Question 2.1
Summary: A peer reviewer (2.1.1 - R5) included several line-specific comments:
The peer reviewer asked if the data labeled 'Chronic" should be labeled 'sub-chronic" in Table
4-2 (page 106).
The peer reviewer stated that the description of chronic data in line 2563 on page 107 should
align with the data in Table 4-2 and with later statements.
The peer reviewer expressed that lines 2568-70 on page 107 should be written in past or
present tense, rather than future tense.
The peer reviewer stated that, in lines 2709-2717 on page 111, it would be helpful to state what
factors contributed to the final assessment factors in the LC50 estimates.
EPA Response: The term "sub-chronic" has been added to the table and the corresponding narrative to
better characterize the duration of the Sun et al. 2016 study. The description of chronic aquatic hazard
study has been revised within Section 4.2.2 beginning on line 2643 and 2730, for vertebrates and
invertebrates, respectively. The use of future tense in this section and others has been revised. The
section in the Draft Risk Evaluation in line 2709-2717 was demonstrating possible COCs from
ECOSAR values as a comparison to the limited empirical hazard data available. The acute and chronic
COCs has the considerations for applied assessment factors in the narrative but the algae COC
assessment factors were not described past the application of an AF of 100. Similar to the chronic COC
derived from ECOSAR data, the algae COC had an AF of 10 applied for aquatic plants and an
additional AF of 10 applied due to uncertainties associated with the use of this ECOSAR hazard value.
Ultimately, the presentation of COCs from ECOSAR data had been removed since the revised Risk
Evaluation has added empirical hazard data for both acute and chronic data and represents
invertebrates, vertebrates, and aquatic plants.
Responses to Charge Question 2.2 related to data
Summary: A peer reviewer (2.2.1 - Rl) stated that the use of the SSD toolbox combined with the
Web-ICE calculation to calculate the lower confidence limit on the HC05 seems like an acceptable
approach. The peer reviewer added that the use of the additional data mentioned in their response to
Charge Question 2.1.1 would improve the estimate and result in a narrower confidence interval.
EPA Response: EPA has been able to add additional studies from systematic review that increase our
data landscape for acute and chronic hazard. Revision of the Web-ICE model in the TCEP has taken
place with the additional acute hazard data. Revisions have been made for the additional empirical and
predicted values used in SSD and are documented within Appendix Section F.2.1.1 for Web-ICE and
F.2.1.2 for the SSD.
Summary: Several peer reviewers (2.2.1 - R2. 2.2.1 - R3. 2.2.1 - R5) expressed that the data used
was insufficient. One peer reviewer (2.2.1 - R2) wrote that, as currently presented, evaluations based
on the lower 95% confidence interval introduce unnecessary conservatism. The peer reviewer stated
that the incorporation of additional high-quality data could result in a refined HC05 value with a
18
-------
narrower 95% confidence interval, which would add certainty to the potential use of the HC05 (instead
of the lower 95% confidence interval) as the definitive acute COC. The peer reviewer included a graph
to depict this. Another peer reviewer (2.2.1 - R3) stated that EPA calculated the COC for aquatic
organisms with an assumption that the previously modeled Web-ICE data would be representative of
TCEP. The peer reviewer expressed that the absence of multiple species presents major uncertainties
and raises concerns about applying standard assumptions. Finally, a peer reviewer (2.2.1 - R5) stated
that the species distribution approach is appropriate, but there is insufficient data in all the datasets. The
peer reviewer added that the lack of data for amphibians, aquatic invertebrates, and plants is
unacceptable for a risk evaluation, risk determination, or risk assessment of compounds known to be
present in water. The peer reviewer expressed that if there is a desire to move away from animal
testing, in-vitro assessments would provide the best alternative. The peer reviewer additionally wrote
that all the data for TCEP toxicity to aquatic species was generated in a Japanese lab, which indicates
that the U.S. is providing no leadership in obtaining data that is essential for the processes required by
TSCA. The peer reviewer specifically commented that the lack of chronic data for fish presents a
problem.
EPA Response: Additional acute hazard data have been reviewed and integrated into the current RE.
Specifically, acute hazard data on vertebrates and invertebrates have been added to Table 4-2, and 4-3,
respectively, with accompanying narrative on the studies within Section 4.2.2. Chronic hazard data
have been added from studies representing several species of fishes with exposure durations ranging
from 21 to 120 days (Table 4-2). Chronic aquatic invertebrate hazard data have also been added to the
TCEP risk evaluation and is represented within Table 4-3. Aquatic plant hazard data have been added
to Table 4-4 also with accompanying narrative. These additional works have increased the data
landscape for both acute and chronic exposures and represent expanded taxa which increases
confidence in the overall COCs representing within the TCEP risk evaluation for ecologically relevant
receptors. It is important to recognize that although there are no empirical hazard endpoints for
amphibians, the use of WeblCE did result in the inclusion of a predicted acute hazard value for
bullfrog, which was included within the SSD.
The EPA has worked with the Japanese Ministry of the Environment to acquire, translate, and review
their studies on TCEP. The works were placed through the systematic review process and hazard
information has been integrated into the TCEP Risk Evaluation.
The inclusion of additional empirical acute hazard values in WeblCE and the resulting SSD did result
in decreasing the 95% confidence intervals associated with the HC05. The revised COC from these
additional studies did not change much but the representation of additional taxa and the details on the
distribution of taxa within the SSD are characterized within Appendix F 2.1.1. In addition to the
inclusion of more acute hazard data from Zhang et al. 2024, the SSD presented in that recent work
demonstrates strong agreement with the SSD performed within the TCEP Risk Evaluation.
Responses to Charge Question 2.2 related to the calculation of the COC
Summary: In response to Charge Question 2.2, a public commenter (0040) stated that the approach to
calculating the COC differs from other EPA offices such as the Pesticide Program and biases the risk
quotient (RQ) toward unreasonable risk. Another commenter (0045) discussed the use of a derived
COC, stating that the approach to model based on known data and reasonable assumptions seems to be
a "best possible solution" to the knowledge gaps.
19
-------
EPA Response: The TCEP Risk Evaluation derives acute COC from the SSD performed with
empirical and Web-ICE modeled hazard information and does not apply any assessment factor. The
chronic aquatic COC is represented by a ChV (geometric mean of a NOEC and LOEC for mortality)
from a 30-day exposure of TCEP with an assessment factor of 10. There are no RQs > 1 for the acute
aquatic COC in the draft or revised TCEP RE. The comment from (0040) is specific to the role of
Assessment Factors with the ECOSAR data with acute COCs. The revised TCEP RE does not use the
ECOSAR data for acute COCs nor did the draft TCEP which calculated an acute COC based on the
lower 95th percentile of the HC05 for an aquatic SSD.
Summary: In response to Charge Question 2.2, a peer reviewer (2.2.1 - R6) expressed that it is unclear
why a secondary acute COC was derived. The peer reviewer stated that the scenarios to use acute COC
or secondary acute COC are not provided in Section 4.2, and the peer reviewer expressed that the 10-
fold difference between the acute COC and secondary acute COC could be better explained.
EPA Response: This revision has been accepted as recommended. The secondary COCs were not
carried through to the Environmental Risk Characterization and were not used for the formation of Risk
Quotients. This review and comments have identified a critical gap in information on chronic duration
exposure studies with TCEP on aquatic organisms. The EPA has identified and evaluated several
chronic duration studies conducted on aquatic vertebrates and has updated the chronic hazard
assessment portions of the TCEP RE to reflect these newly identified studies.
Summary: A peer reviewer (2.2.1 - R2) expressed that the use of a data-driven acute COC for aquatic
organisms is an improvement over the application of traditional assessment factors. The peer reviewer
stated, however, that EPA did not provide a scientific justification for the lower 95% confidence
interval. The peer reviewer recommended that future updates to the risk evaluation clearly explain the
rationale behind the selection of the lower 95% confidence interval, even if the justification is based on
a policy decision. Additionally, the peer reviewer stated that revisions regarding the lower 95%
confidence interval should include appropriate citations. The peer reviewer noted that the complete
citation for U.S. EPA, 2022c is missing in the draft risk evaluation.
EPA Response: The justification for using the lower 95th percent confidence interval as the acute
Concentration of Concern over other options such as the HC05 or application of assessment factor has
been added to Section 4.2.4. A manuscript is currently under review within Integrated Environmental
Assessment and Management titled "Evaluation of Interspecies Correlation Estimation (ICE) models
for chemicals evaluated under the Toxic Substances Control Act". This submission evaluates the use of
the WeblCE model in TSCA chemical evaluations and includes TCEP along with other chemicals
within the TSCA program.
Summary: A peer reviewer (2.2.1 - R2) recommended EPA re-evaluate the chronic COC used in the
risk evaluation. The peer reviewer noted that the value was based on a single endpoint for a single
species (Sun et al., 2016), and the peer reviewer stated that there are two issues with this study that
were not flagged in the risk evaluation. First, the peer reviewer commented that the study did not
provide any details on the methodology used for quantifying the body length of the larval medaka.
Second, the peer reviewer stated that the study did not include chemical analyses of the exposure
media, so the concentrations used in the toxicity tests were not confirmed. The peer reviewer also said
that TCEP did not cause other adverse effects on any of the additional developmental endpoints in the
study. The peer reviewer stated that too much weight was given to a study that may not meet all the
20
-------
quality criteria needed for the risk evaluation, and the peer reviewer expressed that the conclusion of
unreasonable risk in Section 6.2.2 is not fully supported.
EPA Response: The review and critique of the Sun et al. (2016) is welcome and resulted in better
representation of chronic aquatic hazard for TCEP after this comment and others. As a result, more
chronic duration studies were identified and placed through the systematic review process. Ultimately,
six studies of fishes (including Sun et al. 2016) were available and represent TCEP exposure with
subchronic and chronic durations ranging from 14 to 120 days. Apical assessment endpoints of
regulatory interest (i.e., impaired growth, survival, or reproduction) in addition to organ level and
above endpoints are detailed within the narrative in 4.2.2 and summarized within table 4-2.
Summary: A peer reviewer (2.2.1 - R4) stated that EPA considered two approaches for calculating the
acute COC for aquatic organisms, and EPA chose the SSD approach because it relies on multiple data
points rather than the single daphnid study used in the Ecological Structure Activity Relationships
(model) (ECOSAR) approach. The peer reviewer said that EPA should provide further justification for
the selection, because the values are 20x different and the selected value is less protective.
EPA Response: Within the draft TCEP risk evaluation the ECOSAR COC would have been 3.4 mg/L
compared to the draft COC from the SSD of 85 mg/L. The potential use of an ECOSAR COC was
dervied from an estimated 48 hour LC50 (170 mg/L) from daphnia with an Assessment Factor (AF) of
50. This AF for the acute COC was represented with the application of an AF of 5 for acute
invertebrate hazard value and an additional AF of 10 for uncertainties associated with the use of an
ECOSAR hazard value for a water column invertebrate. Although more protective, the use of an
ECOSAR value for the acute COC would represent a more speculative approach as the AFs account for
wide gaps in information on a variety of unrepresented taxa when using a single ECOSAR value for
daphnia.
The presentation of ECOSAR modeling demonstrated potential TCEP hazard values for green algae
and daphnia that were not represented with empirical data in the DRAFT TCEP risk evaluation. Since
submission of the draft TCEP risk evaluation, EPA has acquired the acute and chronic daphnia studies
from the Japanese Ministry of the Environment (Table 4-3) in addition to algae hazard data (Table 4-4)
and additional hazard data both recommended from peer review and newly published works (Table 4-
2). The revised acute COC continues to use the SSD approach and now is represented with more
empirical hazard values from a greater variety of taxa. The revised acute COC of 16,700 ppb represents
empirical data as compared to the single ECOSAR value from a 48-hour daphnia of 170,000 ppb with
assessment factors (50) applied.
Other comments in response to Charge Question 2.2
Summary: A peer reviewer (2.2.1 - R5) recommended that EPA define the two hazard values in the
4th column of Tables 4.2 and 4.3.
EPA Response: These tables have been revised with a footnote in the fourth column (hazard value) to
indicate that the third column (endpoint) details the type of hazard endpoint value represented (e.g.,
LC50, EC50, NOEC, LOEC, etc).
Summary: A peer reviewer (2.2.1 - R5) specifically commented on lines 12395-12408. The peer
reviewer stated that the text describes a process in which measures of TCEP toxicity to fish were
21
-------
compared to the Web-ICE results, and the data were excluded from the dataset if the measured data did
not fit the Web-ICE model. The peer reviewer asked what dataset the data was excluded from. The peer
reviewer expressed that, if there are chemicals that do not fit the model, the solution is not to exclude
the chemicals. Additionally, the peer reviewer stated that TCEP falls into the category of very polar
general narcosis toxicants, and models like Web-ICE are likely to underestimate the toxicities of very
polar general narcosis toxicants.
EPA Response: WeblCE predicts toxicity values for species based on empirical hazard values and
there are model fit statistics associated with surrogate (empirical) and predicted pairs (R2, MSE, Slope).
This section (Page 460 within Appendix F.2.1.2) has been revised to be clearer about the model
selection criteria from Willming 2016 as it applies to the predicted hazard values from the WeblCE
model results. Predicted species and their corresponding hazard values that met these criteria are
presented within Table Apx F-l. The statistics and parameters of the WeblCE model information for
each WeblCE module are located here: https://www3.epa.gov/webice/iceDownloads.html .
The WeblCE v4.0 database has a total of 1517 species-chemical combinations, with narcosis
representing by 306 species-chemical combinations with an average inter-test range of acute toxicity
tests of 9.1 with 80th and 90th percentiles of 5.8 and 12.6, respectively. The model also includes other
MO As such as neurotoxicity (N = 172) and AChE inhibition (N = 255) with additional documentation
on the aquatic MOA models is available here: https://www3.epa.gov/webice/iceDownloads.html
The increased number of empirical hazard endpoints increases confidence that the resulting SSD does
not underestimate TCEP acute toxicity. The reviewer's comment on the role of very polar general
narcosis toxicants is noted and resonates with the inclusion of recent chronic duration studies on fishes
with hazard values presented in Table 4-2.
Section 4.3 - Environmental risk characterization
Summary: A public commenter (0040) suggested that EPA employ a tiered approach to risk
evaluation using a screening-level deterministic RQ approach first and following up with more refined
approaches if the RQ is greater than 1.0, similar to the approach used by the EPA Office of Pesticide
Programs. The commenter stated that the aquatic risk characterization was based on "flawed and
implausible assumptions", including the volume of estimated releases from a single facility to surface
water.
EPA Response: TCEP is not listed on the National Emissions Inventory (NEI) and was only recently
added to TRI, with the first year of reporting from facilities due July 1, 2024. As of September 2024,
there are currently no TRI entries for TCEP. The 2016 CDR data {U.S. EPA, 2019, 6277143} included
a single reporting site, Aceto Corporation in Port Washington, New York, importing TCEP, with no
downstream industry sectors identified. TCEP was not reported in the 2020 CDR {U.S. EPA, 2020,
6277143}. EPA modeled environmental releases and occupational exposures for these hypothetical
scenarios. For each OES, where monitoring data were not available, daily releases were estimated per
media of release based on EPA Standard Models, Generic Scenarios (GSs), and/or Emission Scenario
Documents (ESDs) to generate annual releases and for the estimation of associated release days.
Chronic RQ values for COUs with releases to water are presented with several scenarios for flow and
production volume with revisions presented within Section 4.3.6.1 and additional Tables within
Appendix G-l. For example, 7Q10 flow values have been calculated to represent the 50th and 90th
percentile of the distribution of SIC codes of the industry sectors. Results from both central tendency
and high-end estimate scenarios at each flow regime are now represented to demonstrate the
importance of receiving waters flow in TCEP concentrations and days of exceedance reported within
22
-------
the PSC model. For the Environmental Risk Characterization for TCEP EPA did employ a screening
level trophic transfer analysis as a tiered approach for determining potential risk with terrestrial and
aquatic pathways to relevant terrestrial biota.
Revisions to the chronic COC include a value dervied from 30-day TCEP exposure, resulting in a
longer duration above the COC concentrations to demonstrate days of exceedance. Please note that this
COC is not limited to embryo/larval development as the previous COC was derived from a 14-day
exposure in embryo/larvae.
Summary: A public commenter (0044) stated that EPA had not evaluated sufficient diversity of bird
species and life stages and did not consider risks to wildlife species that are listed as threatened or
endangered under the Endangered Species Act such as salmon, marine mammals, and loggerhead sea
turtles, contrary to EPA's own guidance on risk assessments.
EPA Response: The environmental risk characterization was performed with the landscape of
reasonably available hazard data from TCEP toxicity studies to evaluate risk to environmental
receptors. The use of the WeblCE model for acute TCEP toxicity to aquatic organisms has allowed for
the inclusion of several salmonid species and numerous freshwater mussels, both taxa representing
species of conservation concern (Appendix F.2.1.1).
Summary: Another commenter (0045) suggested that EPA has understated the hazard and risk to
aquatic invertebrates from chronic exposure to TCEP because no actual empirical exposure duration
data for fish were available and suggested that EPA use a cold freshwater fish species and a testing
procedure that incorporates a longer exposure duration, which the commenter described as more
appropriate to address chronic exposure for aquatic vertebrates.
EPA Response: Additional chronic hazard studies on TCEP have been identified and reviewed within
our Systematic Review Process. These studies were integrated into the Environmental Hazard section
of the RE and are summarized in the narrative of Section 4.2.2 and Table 4-2. Rainbow trout are
represented with empirical acute hazard data within our SSD for the formation of an acute COC for the
TCEP risk evaluation. Although the species represented are not characterized as cold-water fishes, sub
chronic exposures from one study and chronic exposure effects from 5 studies within aquatic
vertebrates represent a variety of exposure durations, life history stages, and fish species (Table 4-2).
Sub chronic and chronic duration exposures of TCEP to aquatic vertebrates produced effects including:
mortality {Hu, 2022, 11365033; Wang, 2022, 11365040}; growth/development {Sun, 2016, 4292102;
Hu, 2022, 11365033; Hu, 2023, 11365083; Wang, 2022, 11365040}; organ level effects within liver
{Sutha, 2020, 6772951; Hu, 2022, 11365033; Hu, 2023, 11365083}, gills {Sutha, 2020, 6772951},
kidney {Sutha, 2020, 6772951}, and heart rate {Wang, 2022, 11365040}.
Section 5 - Human Health Risk Assessment
Comments associated with this issue are summarized in the subsections below.
Section 5.1 - Human exposures (including comments on milk concentrations (Verner), dry and wet
air deposition, and disposal and groundwater pathway, aggregate exposure, exposure to certain
PESS)
Exposure scenarios for Tribal populations
23
-------
Summary: Several public commenters (0035. 0041. 0044) expressed support for EPA's inclusion of
Tribal populations in PESS and the Agency's efforts to quantify their greater exposures via higher fish
consumption, but said EPA failed to comprehensively assess the factors that increase Tribal
population's exposures to and susceptibility to harm from TCEP. One of the public commenters (0035)
stated that though EPA claimed that quantifying unique exposure scenarios for Tribal populations was
not possible, Harper et al. (2012) published extensive guidance on conducting risk assessments in
Tribes and developed specific exposure factors for Tribal lifeways. The commenter suggested that EPA
draw on these data. Another commenter (0041) stated that EPA needs to aggregate these exposures
across pathways and conditions of use by including background exposures from non-TSCA uses,
accounting for increased Tribal susceptibility, and considering cumulative impacts on Tribal people
from TCEP exposure. The commenter, along with another public commenter (0044). also said that the
exposures of Tribal people via fish ingestion are underestimated in the risk evaluation by:
using a mean current fish consumption rate vs. a 95th percentile rate for Tribal populations,
when both 50th and 95th percentile exposure values are used in determining occupational and
general population exposures, and when 95th percentile values are available for many different
Tribes and they are much higher than 216 g/day;
using a fish consumption rate from only one Tribe, when much higher rates and for many
different Tribes are available in peer-reviewed literature;
disregarding the higher mean current fish consumption rate of 770 g/day for Alaskan
communities published in Chapter 10 of the EPA Exposure Factors Handbook with no
explanation;
using the lower fish bioaccumulation factor in determining risk; and
using the 33-year value for residential mobility which does not apply to Tribal members who
typically spent most, if not all, of their lives living on their Tribal lands.
EPA Response: The mean current ingestion rate was used to quantify a central tendency scenario that
may represent current use patterns for tribal populations. However, EPA also integrated the heritage
fish consumption rate, which is much higher than any of the available 95th percentile current ingestion
rates in the suggested literature. The 770 g/day for the Alaskan communities is a maximum rather than
a mean ingestion rate and still lower than the heritage consumption rate of 1646 g/day.
For the bioaccumulation factors, risk determination was based on the lower value because of the
considerable uncertainties with the higher value, as explained in Section 5.1.3.7.1 and 5.3.5. The study
that reported the higher BAF value for walleye collected surface water and fish tissue at different years,
thus it is difficult to hypothesize if TCEP surface water conditions at the time of collection influenced
BAF values. Both BAF values resulted in fish tissue concentrations that were several magnitudes
higher than both the monitored data and the calculations based on the bioconcentration factor that the
environmental assessment conducted (Table 5-31). Lastly, Tables 5-61, 5-63, and 5-65 present acute
noncancer, chronic noncancer, and cancer risk estimates, respectively, for the different BAFs and fish
ingestion rates. For Tribal populations, the risk estimates vary by 1-2 magnitudes based on the BAF
value. These different estimates do not result in significantly different risk conclusions.
EPA thanks the commenter for highlighting the 33-year value for residential mobility. EPA has
changed the 33-exposure duration to 78 years (the full lifetime) for lifetime drinking water and lifetime
ambient air inhalation risk estimates. For the fish ingestion pathway, the fish ingestion rates used in this
RE were specific to adult tribal members aged 16-78 years. EPA did not factor in residential mobility
and used 62 years, rather not 33, as the exposure duration.
24
-------
Summary: Conversely, a public commenter, in two submissions to the docket (0040. 0052). said that it
is unclear why EPA selected the highest estimated fish ingestion rate for the Kootenai Tribe when it
was acknowledged that this may overestimate fish consumption.
EPA Response: EPA acknowledged this uncertainty. However, the fish ingestion rate for the Kootenai
Tribe was selected as the most appropriate fish ingestion rate from the reasonably available data to
characterize a sentinel exposure scenario. Other rates were not considered primarily because it is
unclear if they represented fish ingestion only or fish used for trade and other purposes.
Summary: A public commenter (0044) said that EPA must account for the impact of solid waste
disposal practices on Tribal communities, which is increased due to substandard landfill infrastructure,
open dumps, and open burning of solid waste. Additionally, the commenter said that EPA overlooked
TCEP concentrations in the Arctic environment due to contaminated air and other environmental
media, as well as organophosphate flame retardants, brominated flame retardants, and other persistent
organic pollutants. The commenter suggested EPA look at a recent study, Moran et al. (2023), which
reported vapor phase TCEP concentrations in ambient air in the Yupik community of Sivuqaq that are
significantly higher than the levels reported in literature cited in the draft risk evaluation.
EPA Response: EPA appreciates the commenter sharing the Moran et al. 2023 reference. This study
has now been incorporated into the ambient air and deposition sections (3.3.1 and 3.3.2.6), as well as
the discussion regarding disposals for TCEP (5.1.2.2.5).
The air concentrations and deposition values reported in Moran et al. 2023 are lower than the
maximum concentrations and deposition values evaluated in the TCEP risk evaluation. EPA has
identified Tribal populations as a PESS in the TCEP risk evaluation and has done an analysis on Tribal
fish ingestion exposure (Section 5.13.4.4). While EPA has not done a specific Tribal analysis related to
substandard landfill infrastructure, open dumps or the open burning of solid waste, the TCEP risk
evaluation does evaluate the potential for TCEP to leach from TCEP containing landfills (Section
3.3.3.7), and does calculate inhalation risks from breathing ambient air (Section 5.1.3.2) and dermal
and ingestion risk for children playing in soils of nearby facilities emitting TCEP (Sections 5.1.3.3.2,
and 5.1.3.4.5).
Without a full characterization of open burning of solid wastes containing TCEP, and lack of
monitoring data describing open burning of TCEP containing solid wastes, the data needed to produce
quantitative risk estimates from open burning is not reasonably available.
Aggregate exposure to TCEP
Summary: A public commenter (0035) expressed support for EPA's decision to aggregate exposures
for each consumer condition of use across dermal, ingestion, and inhalation routes of exposure.
However, the commenter said that EPA considered each condition of use and scenario in isolation,
which does not provide an accurate picture of humans" real-world exposure to TCEP. The commenter
recommended that EPA aggregate exposures across occupational and consumer conditions of use and
stated that EPA did not justify its decision to forego an aggregate exposure assessment. Additionally,
the commenter, along with another commenter (0044). suggested that EPA should aggregate exposure
across routes and pathways in the general population, even though EPA claimed that the individual
exposure estimates were based on release estimates assuming a product volume of 2,500 pounds and an
aggregation would double count the production volume. Aggregating soil ingestion and drinking water
25
-------
exposures would total an estimated 5,000 lbs, which is still well below the CDR threshold and within
the range of estimated exposures included by EPA in the draft risk evaluation, according to the
commenter. The commenters (0035. 0044) said that failing to aggregate exposures is in violation of
TSCA.
One of the commenters (0044) further stated that under TSCA, EPA must evaluate those reasonably
foreseen combinations of exposures and EPA cannot claim to lack reasonably availably data indicating
co-exposures. By failing to address aggregate exposures to lactating individuals, the commenter
claimed that EPA underestimated infants" and children's exposure to TCEP. The commenter also wrote
that EPA underestimated this exposure by assuming
that infants breastfeed for a maximum of one year despite contrary evidence; and
that infants mouth toys and products containing TCEP for seven to ten minutes per hour,
significantly less than the duration that EPA's Exposure Factors Handbook recommends for
use in risk assessment.
Finally, the commenter questioned EPA's claim that when a single exposure pathway or condition of
use is sufficient to establish unreasonable risk, the consideration of additional, aggregate exposures is
unnecessary, which misunderstands the purpose of a risk evaluation. If EPA has not considered
aggregate exposures to determine the full extent of a chemical's unreasonable risks, then it will not
have sufficient information to properly manage the chemical and ensure that such risks are eliminated.
Another commenter (0045) similarly expressed disappointment that the Agency has not incorporated
aggregate exposure into its risk characterization, and stated that even if there appears to be a
predominate source of exposure, all routes should still be assessed and combined when determining if a
condition of use poses an unreasonable risk or not. The commenter recommended that EPA develop
aggregate exposure assessments and risk estimates for each condition of use. A peer reviewer (OC -
R6) said that EPA stated several times that they did not aggregate exposures, thereby missing
opportunities to specifically quantify exposure estimates among sensitive or highly exposed sub-groups
including infants and workers. The commenter said that EPA's comment on Line 5125 that background
levels of TCEP in indoor air and indoor dust are not considered or aggregated in this assessment does
not reflect real world exposure scenarios and EPA should update their methodology.
EPA Response: A full discussion of EPA's consideration of aggregate exposure is provided in Section
5.1.4 of the risk evaluation and details the conditions of aggregating exposure across routes within the
consumer assessment. Because the health outcomes are systemic and EPA did not identify evidence of
differences in toxicokinetics across exposure routes, considers it is possible to aggregate risks across
exposure routes for all exposure durations and endpoints for the selected PODs identified in Sections
5.2.6.1 and 5.2.6.2 but only if exposure scenarios indicate that aggregation is reasonable. Forthe milk
pathway (from page 583 of RE): EPA combined exposures forthe milk pathway across all routes for
each COUs/OESs within workers and consumers. However, forthe general population, EPA only
assessed the oral route when assessing the milk pathway because exposure estimates showed that oral
doses were several magnitudes higher than dermal or inhalation doses.
Because data were not reasonably available to indicate that co-exposures of multiple TCEP-containing
activities or products in the occupational and indoor environment, EPA did not assess aggregate
exposure across consumer, commercial, or industrial COUs, see section 5.1.4.
TSCA Section 6(b)(4)(F)(ii) directs EPA to "describe whether aggregate or sentinel exposures to a
chemical substance under the conditions of use were considered, and the basis for that consideration"
in risk evaluations. Therefore, EPA may, but is not required to, consider aggregate exposure in the risk
26
-------
evaluations. 40 CFR 702.33 defines aggregate exposure as "the combined exposures to an individual
from a single chemical substance across multiple routes (i.e., dermal, inhalation, or oral) and across
multiple pathways (i.e.. exposure from different sources)". Therefore, aggregate exposure includes
combined exposures from a single chemical across multiple routes and multiple pathways. 40 CFR
702.33 defines sentinel exposure as "the exposure from a single chemical substance that represents the
plausible upper bound of exposure relative to all other exposures within a broad category of similar or
related exposures/1
Susceptible subpopulations
Summary: A public commenter (0044) said that EPA failed to evaluate as PESS people who
experience "greater susceptibility" to harm from TCEP exposures because of their cumulative
exposures to multiple chemicals and non-chemical stressors, which the commenter described as a
violation of TSCA. The National Academies have repeatedly called for the consideration of cumulative
exposures in chemical risk evaluations, according to the commenter, and asked agencies to "move
beyond source-by-source and pollutant-by-pollutant... risk assessment". Yet, EPA did not evaluate co-
exposures to TCEP and microplastics, benzo-a-pyrene, or other flame retardants, according to the
commenter.
The commenter also stated that EPA failed to calculate TCEP's increased risks to truck drivers,
students residing in dormitories, and firefighters. First, TCEP is used in automotive foam, posing
increased risks to long-haul truck drivers who spend long periods of time in their vehicles, according to
the commenter, but this was never considered in the risk evaluation. Second, the commenter stated that
though EPA claims that it conducted a qualitative assessment for firefighters, it never determined
whether firefighters are exposed to unreasonable risk or the extent of those risks. Finally, the
commenter said that high levels of TCEP have been detected in dust in student dorms and common
areas, and EPA needs to consider this in the risk evaluation.
EPA Response: EPA thanks the commenter for highlighting the concern about co-exposures, non-
chemical stressors and cumulative exposures. The impact of co-exposure with other chemicals is
outside of the scope of the risk evaluation for TCEP. At present, the purpose of the risk evaluation
under TSCA is to determine whether a chemical substance presents an unreasonable risk of injury to
health or the environment, under the conditions of use. EPA accounted for the following PESS groups:
infants exposed through human milk from exposed individuals, children and male adolescents who use
consumer articles or are among the exposed general population, subsistence fishers, Tribal populations,
pregnant women, workers and consumers who experience aggregated or sentinel exposures, fence line
communities who live near facilities that emit TCEP, and firefighters.
While EPA did not provide quantitative estimates for all of these groups, where data and assessments
approaches were readily available, EPA quantified exposures for the various COUs.
A summary of the information found on firefighters is presented in Section 5.3.3. EPA has
incorporated additional information of TCEP found in automobiles with the revision of the paints and
coatings COU, and the incorporation of additional literature discussing TCEP in automobiles (Hoehn et
al. 2024 in Section 5.1.2.2). The consumer assessment included an analysis of exposure from auto
foam containing TCEP (Section 5.1.2).
27
-------
Dermal exposure assessment
Summary: A public commenter (0040) stated that the Abdallah et al. (2016) study selected by EPA
may over-predict dermal absorption for both workers and the general populations because it used
acetone and 20% Tween-BO (aka Polysorbate 80) as vehicles, which may function as penetration
enhancers. Additionally, the commenter said that EPA relied upon individual deterministic modeling
point estimates for its dermal exposure assessments, which led to a substantial degree of uncertainty in
the Agency's risk conclusions, and did not apply a tiered approach to its inhalation, dermal, or
ingestion exposure assessments for any consumer exposure assessments.
The commenter also stated that in several instances, EPA used the OPPT Dermal Exposures to Volatile
Liquid Model with default central tendency and upper bound loading rates for the dermal exposure
estimation. The commenter said that these loading rates are not specific to any scenario assessed in the
risk evaluation, nor are they specific to TCEP and chemicals similar to TCEP, but rather are default
values used in the absence of any other scenario-specific information. Per EPA's Draft Protocol, using
a generic default value should result in the weighing of "slight" in the weight of evidence assessment,
not "moderate" and the commenter stated that the reason for this discrepancy is unclear.
EPA Response: Thank you for the comment. EPA used values from a technical report, A Laboratory
Method to Determine the Retention of Liquids on the Surface of the Hands (U.S. EPA, 1992b) to
estimate dermal loading values. These are based on the activity/scenario that is most likely involved for
each process (i.e., routine/incidental contact and immersion). Dermal loading is covered in Appendix
D.2.2. in the Engineering Supplemental File.
Recycling of e-waste (processing)
Summary: A public commenter (0040) stated that for the OES regarding recycling of e-waste, EPA
directly used monitoring data for TCEP but employed a data summary approach that appears to have
overestimated inhalation exposures by 1,000-fold or greater. To correct this, the commenter said that
AIHA recommends using the 95th percentile of a dataset relevant to a similar exposure groups as a
decision metric, and using the 95% upper confidence limit to determine a certainty level.
EPA Response: The available monitoring data only provided summary statistics (minimum,
maximum, median) rather than discrete samples. Therefore, EPA could not create a full distribution of
monitoring results across the sources to use in estimating central tendency and high-end exposures
(From Supplemental file Section 3.7.4.3)
Verner physiologically based pharmacokinetic (PBPK) model
Summary: A public commenter (0040) questioned the use of the Verner et al. model due to the
differences in tissue distribution and half-lives of elimination. The commenter suggested that studies by
Toxicology Excellence for Risk Assessment (TERA) and Minegishi et al. further inform this issue. For
example, a report by TERA (2016) prepared for the Consumer Product Safety Commission (CPSC)
described the elimination of TCEP as biphasic with a half-life of elimination in adipose tissue of 87
hours. The TERA report also referenced a study by Minegishi et al. (1987) which provides a detailed
toxicokinetic study of radiolabeled flame retardants. The commenter expressed concern that EPA did
not identify this highly informative study even though it's located in the health & environmental
research online (HERO) database.
28
-------
EPA Response: EPA added a summary of {Minegishi, 1988, 656593@@author-year} to Section 5.2.2
and Appendix J. 1, which is consistent with other studies" findings that most of TCEP (>75%) is
eliminated within the first 24 hours. No changes were made in how EPA modeled TCEP concentrations
in human milk. A challenge with using radiolabeled compounds for ADME studies, such as the one by
{Minegishi, 1988, 65 65 93 author-year}. is that half-lives of the parent compound can be
overestimated. After administration, whether the 14C is still tagged to the parent compound or a
metabolite cannot be determined. It is also possible that the functional group containing the
radiolabeled carbon could have reattached to another substance that is no longer TCEP. Despite
limitations in applying the Verner model to TCEP, it is the best approach based on reasonably available
information for estimating infant exposure through human milk ingestion.
Summary: Several peer reviewers (3.1.1 - Rl. 3.1.1 - R2. 3.1.1 - R3. 3.1.1 - R4. 3.1.1 - R6)
expressed support for the use of the Verner model and said that it is clearly and transparently described.
However, one of the reviewers (3.1.1 - Rl) said that one may raise caution with regards to assumptions
of milk intake rates used and thus the actual dose that a child may receive across different ages of
development, composition of breast milk, and the social constraints that limit use of breastfeeding.
Therefore, it is anticipated that the estimated exposure will be relatively higher than what will occur.
Another reviewer (3.1.1 - R4) stated that in addition to the sensitivity analysis for the input parameters
half-life, KOW, age at pregnancy, and lipid fraction, EPA should describe the uncertainties associated
with the maternal oral doses used. Further, the reviewer suggested that EPA justify the exposure
duration of one year, as some infants breastfeed for much longer. Finally, one of the reviewers (3.1.1 -
R6) mentioned two other studies that described TCEP concentrations in breast milk in the US: Ma et al.
(2019) and Yao et al. (2023).
EPA Response: This risk evaluation used the milk intake rates from the Exposure Factors Handbook
(EFH), which stratified the rates for different age groups from birth up to the first year of life. EPA did
not consider exposure beyond one year as a result (Section 5.1.3.4.7). Social and physiological or
biological components that may affect the estimated infant dose were only considered to the extent that
the EFH factored those elements into the milk intake rates. Regarding the uncertainties associated with
the maternal oral doses, Section 5.1.3.7.2 describes that route-to-route extrapolation to an oral-
equivalent dose can potentially underestimate exposure because of the first-pass effect. However,
TCEP's relatively slow clearance rate of approximately 18 hours gives it time to partition out of the
liver and mitigate the effect of TCEP concentrations in milk.
EPA thanks the reviewer for identifying those two studies. Ma et al. (2019) has been added to the
summary of biomonitoring data in Section 5.1.3.4.7 and Appendix H.5.3. Yao et al. (2023) only
measured TCEP concentrations in dairy milk and was not included for further consideration.
29
-------
Summary: A peer reviewer (3.1.1 - R5) provided line by line suggestions regarding the Verner model:
Line 13216 - The reviewer stated that that the term "milk concentration" is incorrect as these
are TCEP concentrations in milk. The reviewer cautioned EPA against using this type of
wording and said it should be carefully attended throughout the draft risk evaluation and all
agency documents.
Line 13220 - The reviewer said that the listed pKa of -9.1 seems unlikely as it does not make
sense that TCEP is an acid, but perhaps a base. Additionally, NR, N/R, and N/A are not
defined in the table.
Lines 13229-13232 - The reviewer said that the Verner model only addresses persistent
organic pollutants exposures for Inuit Peoples, and that there are no data for compounds as
polar as TCEP. The commenter said that modeling of compounds with different functionality
can be problematic. The commenter recommended that validation with an organophosphorus
compound is needed for this model to be useful.
Lines 13330-13335 - the reviewer questions the detection or quantification limits for the
different studies, and asked if the means from data sets with non-detects include those data in
calculation of means or not.
Lines 13337-13338 - the reviewer said that the magnitude of confidence bands clearly
indicates that more measured data are needed to ground truth and better parameterize the
model for use in estimating TCEP partitioning.
EPA Response: The following revisions have been made: milk concentrations corrected to "TCEP
concentrations in milk" throughout the document, acronyms defined at the bottom of Table H-l 1,
removed the pKa with an explanation in Table H-l7 (i.e., no pKa value was identified), and addition of
how non-detects were handled in Section H.5.3. While the range of modeled TCEP concentrations in
milk are several magnitudes lower and higher than the minimum and maximum of measured
concentrations, respectively, the infant's dose is consistently below the mother's. Although the model
was originally built for persistent organic pollutants that are less polar than TCEP, membrane
permeability of TCEP is still possible because its topological polar surface area (PSA) is below 140 A.
Monitoring data that measured TCEP in human milk also supports the conclusion that TCEP can
accumulate in human milk despite its polarity.
Furthermore, a new study identified by another peer reviewer was added to compare modeled and
measured TCEP concentrations in human milk {Ma, 2019, 7268788}. That study collected 100
samples from mothers across the U.S., and TCEP was detected in 37% of the samples. The maximum
measured concentration is lower than the upper bound of the modeled TCEP concentrations in human
milk. Although the sample size is small, it adds confidence that protecting the mother from exposure to
TCEP will also protect the nursing infant. Lastly, the large spread in modeled concentrations can be
attributed to the wide range of maternal doses and not necessarily the model itself.
Summary: A peer reviewer (3.1.1 - R7) also discussed the lack of calibration or validation data for
TCEP within this modeling framework and said it is a substantial barrier to placing confidence in the
modeling results. In order to build confidence in the Verner model, the reviewer recommended that
EPA evaluate whether a clearance rate can be established in rats and mice based upon the cited studies
and then scale that metabolic rate to humans. Additionally, the reviewer said that the nursing exposure
predictions appear to be highly uncertain, as they are several orders of magnitude lower than the
maternal doses.
EPA Response: EPA is unable to find clearance rates in rodent studies except for a value from a
completed assessment. {TERA, 2013, 5155526@/@/author-year} cited a 1994 study that measured a
30
-------
clearance rate after intravenous administration; however, the primary study could not be located for
further evaluation on data quality and applicability.
The nursing exposure predictions are 1-2 orders of magnitude lower than the maternal doses. EPA
revised the tables in Section H 5.4 to make the units consistent and improve clarity.
Extrapolating from the inhalation to oral routes of exposure
Summary: A peer reviewer (3.1.2 - Rl) stated that the approach seems reasonable.
EPA Response: EPA thanks for the reviewer for this comment.
Summary: Several peer reviewers (3.1.2 - R2. 3.1.2 - R3. 3.1.2 - R4. 3.1.2 - R5. 3.1.2 - R6. 3.1.2 -
R8) expressed concern with extrapolating from the inhalation to oral route of exposure.. One reviewer
(3.1.2 - R2) stated that while estimating based on a young adult physiology provides information for
extrapolating, transferring that estimate to the young, elderly, or health compromised may not be
accurate, and these age and health related differences were not incorporated into the estimate.
Similarly, another reviewer (3.1.2 - R3) stated that EPA does not provide any evidence that route-to-
route extrapolation is appropriate, and several limitations were not addressed such as differences in
metabolism via inhalation exposure vs oral exposure, and differences in absorption from lungs
compared to gastrointestinal tract. Two reviewers (3.1.2 - R4. 3.1.2 - R8) stated that assuming
inhalation exposure would result in the same absorbed dose as would oral exposure is incorrect, while
another peer reviewer (3.1.2 - R6) stated that significant errors can occur due to this assumption. The
reviewer also stated that when exposure is via inhalation, the blood may move directly from the lung to
major organs so the toxicant may encounter sensitive nervous system tissue or the liver before moving
to digestive organs and kidneys. Thus, the reviewer said that simple dosimetry is inappropriate without
understanding the toxicokinetics and toxicodynamics of the compound in specific organisms. Finally,
the reviewer said that it is important to understand the membrane partitioning/transition of TCEP
relative to other toxicants, especially polar ones. Another reviewer (3.1.2 - R7) said that the exposure
frequency and working years fixed in Equation 5-23 may not be relevant if the target population is
women of reproductive age.
EPA Response: EPA acknowledges that route-to-route extrapolation introduces uncertainties to the
modeled TCEP concentrations in human milk. However, the exposure estimates for the nursing infant
are consistently lower than the mother's by at least one order of magnitude. EPA does not expect the
conclusion that the mother is more sensitive than the infant to change despite the assumption that
inhalation exposure would result in the same absorbed oral dose. Furthermore, all the toxicity values to
estimate risks from TCEP exposure were derived from oral studies, and EPA assumed absorption via
the oral route to be 100 percent.
The exposure frequency is considered the maximum number of days a worker is exposed for the dose
metric of interest and assumes 5 days/week, 50 weeks/year. Those values are relevant to all workers.
The 40 working years may be too high for women of reproductive age, but 40 years is also averaging
time for exposure years of 40. This cancel the inputs such that the working years becomes irrelevant.
31
-------
Intrinsic hepatic clearance
Summary: A peer reviewer (3.1.3 - Rl) recommended against assuming that intrinsic hepatic
clearance would remain constant throughout all life stages. The reviewer stated that while few studies
have been conducted on the metabolism of TCEP and its possible bioactivation, there is evidence that
metabolism by the cytochrome P450 monooxygenases (CYPs) play an important role, as reported in
Burka et al. (1990). It is also known that aggregate cytochrome P450 levels are much lower at birth and
increase fairly quickly with age (Hart and Timbrell, 1979). The reviewer suggested that EPA assume
that the CYP metabolism of infants at birth are approximately 30% of their adult level (Ginsberg et al.,
2004) and suggested that the percent metabolism be modeled to increase in a linear or stepwise fashion
until an adult level was reached at approximately 12 months of age (Alcorn and McNamara, 2003;
Ginsberg et al., 2004), after which it would remain constant through the remainder of life (Klotz,
2009). Another reviewer (3.1.3 - R4) also recommended that EPA look into the literature that
addresses enzyme ontogeny and development in humans, such as Cytochrome P450 isozymes.
Another peer reviewer (3.1.3 - R2) stated that the documented age-related changes in hepatic clearance
clearly support the need to scale any estimate of compound half-life and clearance based on relevant
physiological factors as identified in the literature. Such factors include organ weight, blood flow, age
(immature or elderly). Similarly, another peer reviewer (3.1.3 - R4) said that scaling according to body
weight is a good alternative. On the other hand, another reviewer (3.1.3 - R5) suggested scaling by age
since body weight can have a large variation in the same age.
EPA Response: EPA added a discussion of CYP levels and metabolic rate to Appendix H.5.6. In brief,
it described how EPA increased the half-life of TCEP as a proxy for an infant's reduced CYP levels
and thus longer clearance rate. The higher half-life value was used throughout the first year of life, and
assuming an infant's metabolism remains at 30% of an adult for the entire simulation period is hence
protective. Only the half-life as a proxy for metabolism rather than other physiological parameters was
explored because it did not require changing any of the model inputs determined by the model's
original authors. Varying the half-life still did not change EPA's conclusion that the nursing infant is
exposed to lower levels of TCEP through human milk ingestion than the mother.
Summary: A peer reviewer (3.1.3 - R3) said it was reasonable to assume constant clearance. Another
peer reviewer (3.1.3 - R6) stated that while scaling metabolic clearance across species based upon
body size is a useful tool, it is well known that this does not work well in humans. Thus, the reviewer
said that it would be speculative and highly uncertain to add to the modeling complexity by simulating
anything but a uniform intrinsic clearance.
EPA Response: EPA thanks the reviewers for their comments.
Gaseous phase vs. particle phase
Summary: A peer reviewer (3.2.1 - Rl) stated that Wolschke et al. (2016) is a well-designed study
and the use of 82%/18% proportion is appropriate for this risk evaluation.
EPA Response: EPA thanks the reviewer for their comments.
32
-------
Summary: A peer reviewer (3.2.1 - R2) provided line-by-line comments regarding TCEP partitioning
between the gaseous phase and particulate phase:
Lines 1665-1670 - the reviewer said that there is a need for more data to be generated to
inform these types of evaluations, and that sampling artifacts could contribute significantly to
uncertainties in vapor/particle ratios.
Lines 1942-1956 - the reviewer said the explanation that snowmelt controls the upper end of
TCEP concentrations in soil may hold an approach for evaluating the airborne contributions
and possibly forms of TCEP.
Lines 1681-1682 - the reviewer stated that the process of selecting the 82%/18% ratio is not
described in Appendix H.3.3 and there is no mention of Wolschke et al. (2016) in any text
within the Appendices. The reviewer stated that a brief description is needed along with why
data from this study was chosen to the exclusion of others.
EPA Response: EPA agrees with the commenters that the uncertainty in the vapor/particle ratio and
sampling artifacts could significantly contribute to uncertainties in the assessment. EPA summarized
the available literature on particle size fractions, and gas versus vapor partitioning as described in
section 3.3.1.2.1. EPA has further clarified in section 3.3.1.2.1 that the suggestion from {Okeme,
2018, 5165658@@author-year} was considered when adopting the gaseous phase and particulate
phase proportion from {Wolschke, 2016, 3374439@@author-year}.
EPA thanks the reviewer for the suggestion that snowmelt example may hold an approach for
evaluating the airborne concentrations and forms of TCEP. {Mihajlovic, 2012, 2662833@@author-
year} indicates that the meltwater generated at the snow surface percolated downwards due to gravity
picking up chemicals present at the snow grain edge, and that for this reason amplified concentrations
of TCEP were detected in soil samples.
Summary: A peer reviewer (3.2.1 - R3) suggested that EPA look to relevant US papers if they are
available, like Liu et al. (2023).
EPA Response: EPA has reviewed Liu et al 2023, which observed the gas-particle phase portioning of
TCEP in indoor and outdoor air environments in an urban environment in China. Liu et al 2023 utilizes
quartz fiber filter (QFF) rather than the glass fiber filters (GFF)suggested by Okeme et al., 2018 and the
vales reported may be vulnerable to sampling artifacts.
Leaf cuticular resistance
Summary: A peer reviewer (3.2.2 - Rl) stated that there is no indication of what measure of vapor
pressure is being modeled in Figure Apx_H.5. According to the reviewer, high variances in the
function for the leaf cuticular resistance uptake to lipids (rcl) values are problematic for the use of these
data for modeling. The reviewer said it would be helpful to determine if compounds with these higher
rcl have any structural similarities or experimental anomalies. If there is no way to improve this
estimation method, the reviewer suggested the mean and worst case 95% CL should be used in
modeling TCEP partitioning. Additionally, the reviewer suggested that it would be helpful to depict the
range of TCEP vapor pressures on graph Apx. H-5 by shading or defining the range on the graph. The
reviewer said there are no unit designations for either vapor pressure or resistance on the axes of this
figure.
33
-------
EPA Response: Figure Apx_H.5 has been labeled to include the unit of measure for vapor pressure
(Pascals). The chemical dataset utilized to calculate the relationship for rcl did not include values for
TCEP. EPA utilized this approach as it was the best approach reasonably available.
EPA appreciates the suggestion to use a 95% CL when modeling TCEP portioning. EPA has not
included a central tendency and high-end estimate for the rcl parameter as it is only one parameter,
among a number of parameters that have been used to estimate gaseous deposition (Table_Apx H-8).
However, EPA has included 10th, 50th and 95th percentiles estimates for air concentrations and
deposition at various distances to account for uncertainties in the ambient air modeling.
Particle size deposition parameters
Summary: A peer reviewer (3.2.3 - Rl) stated that the mass fraction 2.5 microns or smaller has units
of length, which seems incorrect, and said that the value should be reported as a fraction. The fraction
of 0.4 is consistent with what is presented in Lee and Patterson (1969) according to the reviewer.
Another reviewer (3.2.3 - R2) commented that there is no mention of mass fractions 2.5 microns or
smaller in Appendix H3.3 and there is no reasonable description of particle sizes in Section 3.3.1.
EPA Response: EPA thanks the commenter for the comment. Table Apx H-9 has been updated to
remove the pm from the mass fraction value. MMAD refers to the median mass aerodynamic diameters
which is typically reported in micrometers (pm). a measure of length. Many studies report mass
fractions of TCEP that is found among various dust sample or particulate matter sizes ({He, 2018,
4728480} and {Schreder, 2016, 3222316}).
A brief discussion on the mass fractions and particles sizes of TCEP is available in Section 3.3.1.2.1.
EPA used the defaults for phosphates suggested in AERMOD for mass fraction 2.5 pm or smaller (0.4)
which cites {Delumyea, 1979, 774505@@author-year} and {Lee, 1969, 33980@@author-year}, and
the default MMAD for phosphates which was listed as 2.2 [am.
Exposure controls and PPE use
Summary: A public commenter (0040) stated that EPA did not assess other types of exposure controls
throughout all exposure assessments for TCEP, including the use of fume hoods, general dilution
ventilation and chemical safety cabinets. The commenter said that the assumption of a lack of these
exposure controls is not realistic.
Similarly, a public commenter (0043) stated that by adopting an assumption during the risk assessment
phase that no PPE is used, there is a greater potential to overestimate risk. The commenter said that this
approach does not appear to fix a perceived problem, but rather replaces it with a potentially greater
problem creating a false and misleading perception of worker risk. The commenter recommended that
EPA ensure that OSHA workplace standards and requirements of the OSHA general duty clause be
taken into consideration when assessing the potential exposures associated with any industrial use of
TCEP.
EPA Response: EPA had previously assumed that workers were provided and always used PPE in a
manner that achieves the stated assigned protection factor (APF) for respiratory protection or used
impervious gloves for dermal protection. However, EPA believes that the assumed use of PPE in a risk
determination could lead to an underestimation of the risk to workers. For example, workers may be
highly exposed because they are not covered by OSHA standards, such as self-employed individuals
34
-------
and public sector workers who are not covered by a State Plan; their employers are out of compliance
with OSHA standards; the PPE is not sufficient to address the risk from the chemical; or their PPE does
not fit or function properly.
Additionally, TSCA risk evaluations are subject to statutory science standards, an explicit requirement
to consider risks to potentially exposed or susceptible subpopulations, and a prohibition on considering
costs and other non-risk factors when determining whether a chemical presents an unreasonable risk
that warrants regulatory actionsall requirements that do not apply to development of OSHA
regulations.
Production volume estimate
Summary: Two public commenters (0035. 0044) stated that EPA's risk evaluation of TCEP relied on
assumptions that they claim may underestimate exposure and risk for both occupational and general
populations. Both commenters said that EPA's TCEP production volume estimate lacks adequate
conservatism and noted the significance of this figure which informs subsequent risk calculations.
Citing historical production volumes much higher than the EPA's estimate of 2,500 lb as well as the
25,000 lb CDR threshold, the commenters stated that EPA does not adequately justify its estimate and
suggested that the actual production volume may significantly exceed it. One of the commenters (0044)
listed several possible sources of error in the Datamyne shipping data used by EPA, namely the
mislabeling of TCEP and the exclusion of TCEP containing articles. Both commenters recommended
that, in its risk evaluation, EPA use the upper bound estimate of 25,000 lb in place of the 2,500 lb
central estimate. They suggested that this would provide necessary conservatism for PESSs and better
account for high historic production levels which may have resulted in lingering effects.
EPA Response: The results from both the 2,500 lb and 25,000 lb PVs are included in the
Supplemental Files.
Duration of residence
Summary: A public commenter (0044) stated that EPA should use a longer duration of residence than
33 years to assess the exposure resulting from living near air and water contamination and suggested
EPA expand this duration to exposures over a lifetime.
EPA Response: EPA thanks the commenter for the comment. EPA has revised its exposure durations
to 78 years for inhalation estimates from ambient air, and drinking water estimates.
Disposal
Summary: A public commenter (0044) stated that in the draft risk evaluation EPA has not adequately
assessed the risks to health and environment posed by TCEP disposal. The commenter stated that
according to EPA's findings, the disposal of TCEP-containing products would continue to be a
significant source of TCEP releases over a long time horizon. They stated that EPA's approach of
considering waste disposal within each condition of use is flawed because it will not fully account for
the exposures for people working in or living near disposal facilities.
EPA Response: Uncertainties for disposal are described in Section 4.3.6.2 that include:
Without a full characterization of non-hazardous landfill (e.g., Norman Landfill) conditions
and historical wastes (e.g., Himco Dump and Fort Devens) around the country, the data needed
35
-------
to produce quantitative risk estimates for disposal is not reasonably available. EPA does not
have data representing municipal and managed landfills and is uncertain how often
contaminant migration occurs given modern practices of non-hazardous landfill and historical
site management. Source attribution of the consumer uses to the leaching concentration
exhibited within Sections 3.3.3.7 and 3.3.3.8 are not reasonably available; therefore, it is
unknown if these concentrations are the result of consumer and/or commercial disposal.
For the commercial uses that have been phased out, any currently used products that contain
TCEP are expected to be disposed of in landfills but will represent just a fraction of previous
amounts from when TCEP was used more widely. Further data are lacking with which to
estimate exposure and risk from disposal or waste treatment activities for these COUs and EPA
has not quantified such risks. EPA's confidence in these exposures is indeterminate and cannot
quantify risk for the disposal or waste treatment activities for these COUs. EPA acknowledges
that while some releases and exposures could occur during the disposal of the wide variety of
items that TCEP has found its way into, based on a review of the limited information on TCEP
within groundwater at landfills and wastewater runoff presented in the section above, these are
expected to be minimal and dispersed, and exposures are expected to be negligible.
TSCA Section 6(b)(4)(F)(ii) directs EPA to "describe whether aggregate or sentinel exposures to a
chemical substance under the conditions of use were considered, and the basis for that consideration"
in risk evaluations. Therefore, EPA may, but is not required to, consider aggregate exposure in the risk
evaluations. 40 CFR 702.33 defines aggregate exposure as "the combined exposures to an individual
from a single chemical substance across multiple routes (i.e., dermal, inhalation, or oral) and across
multiple pathways (i.e.. exposure from different sources)". Therefore, aggregate exposure includes
combined exposures from a single chemical across multiple routes and multiple pathways.
A full discussion of EPA's consideration of aggregate exposure is provided in Section 5.1.4 of the risk
evaluation and details the conditions of aggregating exposure across routes. Included are discussions of
the limitations for aggregating across COUs, and across exposures scenarios. For example, Section
5.1.4 describes the different ways in which exposures can be aggregated (e.g., across COUs, across
Exposure Scenarios) Without any information indicating co-exposures to multiple COUs, or exposure
across multiple exposure scenarios, EPA is unable to conduct an aggregate exposure assessment.
Furthermore, as described in section 5.1.4, it would be misleading for EPA to aggregate exposures
across exposure scenarios when the individual scenarios were based on the total production volume of
2,500 lbs per PES, and aggregating would double count the production volume.
DRAS model
Summary: A peer reviewer (3.3.1 - Rl) discussed EPA's use of the Hazardous Waste DRAS to assess
groundwater concentrations of TCEP after TCEP-containing consumer products have been disposed of
into a non-hazardous landfill. The reviewer remarked that there are less than two pages of text to
describe the new approach, which is inadequate, and Appendix H.5 offers a paragraph to introduce the
DRAS model before sending readers to old technical support for the DRAS software. The reviewer
concluded that there is insufficient information in the draft risk evaluation for anyone other than an
experienced DRAS user. Additionally, regarding Line 2112, the reviewer commented that regulatory
decisions that influence public and environmental health should not be made using a "back of the
envelope computation".
Another peer reviewer (3.3.1 - R2) discussed the landfill modeling and said that a more scientific
approach would be to consider typical TCEP levels in cushions, fabrics, construction, and demolition
36
-------
waste streams and then estimate the total TCEP loaded into a municipal landfill based upon a
population disposal model. The reviewer also said that Table 3-7 is difficult to interpret and is not
related to actual landfill leachate data, so it doesn't appear that EPA tried to calibrate the DRAS model
against actual data. Finally, the commenter suggested 0.25 miles as the default distance to the nearest
private well.
EPA Response: EPA thanks the reviewer for the comment. The language in sections 3.3.3.8 has been
revised to help clarify the selection of leachate concentrations and loading rates for the DRAS analysis.
EPA has added additional language in Section 3.3.3.8 that calibrate the DRAS model the reported
groundwater monitoring concentrations in the monitoring literature and water quality portal. In
addition, EPA has increased the range of the DRAS analysis, bounding the leachate concentration by
TCEP's solubility, and increasing the loading rate to account for past disposal practices. To account for
the uncertainties in estimating current loading rates, EPA varied loading rates over four orders of
magnitude. While current production volumes are anticipated to be lower (2,500 lb), usage of TCEP
was higher in the past and the leaching occurring from landfills may be a result of past disposal
practices.
While EPA appreciates the comment regarding estimating the total TCEP loaded to a municipal landfill
based on a population disposal model, there is limited information available on municipal disposal of
TCEP making it difficult to estimate TCEP loading. Rather EPA, conducts a bounding exercise using
an EPA peer reviewed model (DRAS) to estimate TCEP groundwater concentrations from landfill
leachate.
Note, no default distance was 'selected" for the DRAS analysis. The DRAS analysis is a simulation of
more than 700 different landfills, and the estimates are "within" 1 mile of a poorly managed landfill.
The 3-4 RCRA Delisting Technical Support Document Chapter 3: Exposure Scenario Selection
Revised October 2008 survey data are entered in the EPACMTP model as an empirical distribution:
minimum = 0 m, median = 427 m, and maximum = 1,610 m (approximately 1 mile) (U.S. EPA 1997e).
Summary: A peer reviewer (3.3.2 - Rl) stated that given the paucity of information in the draft risk
evaluation, it is difficult to ascertain a domain of applicability of the DRAS model. The reviewer
suggested that if sufficient data are available for monitored sites, the domain of applicability could be
chemicals of similar structure and function as those measured near waste disposal sites that are present
in aquifers with properties similar to those in the monitored group. Estimates for chemicals in such
aquifers could be sorption, degradation, and movement, according to the reviewer. The reviewer said
that the bounds of these domain need to be clearly stated in the draft risk evaluation.
EPA Response: EPA has revised the DRAS analysis to reflect the paucity of information surrounding
TCEP loading, and leachate concentrations. EPA has increased the range of the DRAS analysis,
bounding the leachate concentration by TCEP's solubility, and increasing the loading rate to account
for past disposal practices. To account for the uncertainties in estimating current loading rates, EPA
varied loading rates over four orders of magnitude. While current production volumes are anticipated to
be lower (2,500 lbs), usage of TCEP was higher in the past and the leaching occurring from landfills
may be a result of past disposal practices.
37
-------
Exposure values for cancer assessment
Summary: A peer reviewer (OC - R1) stated that the use of TCEP has dropped considerably in recent
years and most exposures will likely follow this trend. The peer reviewer expressed that the cancer
dose calculations (e.g., Lifetime Average Daily Dose) need to account for these changes in exposure.
EPA Response: EPA thanks the reviewer for the comment. Lifetime Average Daily Dose equations
are calculated by substituting the value of an individual's lifetime for the AT (averaging time) variable
in the dose equation. It is true that the phasing out of a chemical would mean that an individual would
have greater exposures earlier in life and lower exposures later in life. In such cases the 'ED" (exposure
duration) variable would be less than the AT term.
In the TCEP risk evaluation, EPA estimated exposures to TCEP from current and ongoing conditions
of use that are expected to occur currently. It is difficult to predict whether these exposures would
continue to decrease as they have done in the past decade or whether they would increase in the future.
Therefore, since information is unavailable EPA assumed 78 years for ED and 78 years for AT for the
environmental analysis (e.g., drinking water exposures, ambient air inhalation exposures) and
consumer analysis.
The Supplemental TCEP Consumer Modeling Results includes a table of Lifetime Cancer Risk
estimates by a range of exposure durations. For articles that are used over a shorter exposure duration,
the lifetime cancer risk estimate can be adjusted by ED/AT. For example, if a carpet is used in a
household for 20 years, the lifetime cancer risk estimate can be adjusted by 20/78. CEM 3.2 assumes
that the consumer article containing TCEP would be replaced with a similar article containing TCEP.
Organophosphate esters such as TCEP have begun phasing out from consumer articles. Thus the 78/78
assumption for ED/AT is conservative. Hence, this table provides the risk calculations with a range of
exposure durations (1, 5, 20, 33, 57 years).
General comments on the exposure assessment
Summary: A public commenter (0040) suggested that EPA conduct a sensitivity analysis for key
parameters used in its exposure models and mitigate the potential for unrealistic results that may occur
when multiple upper-bound default values are used simultaneously. To start, the commenter suggested
that EPA should apply a similar Monte Carlo analysis for the key inputs to its dermal model, including
the TCEP concentration, skin loading, fraction of skin exposed, exposure duration and exposure
frequency inputs. Additionally, the commenter recommended that EPA use probabilistic methods for
determining inputs for parameters that also account for the use of engineering controls, administrative
controls, and PPE.
EPA Response: EPA follows a hierarchy of using relevant data, if available, or modelling when data
are not available. Additionally, EPA is always interested in improving our models and appreciates the
commentor's suggested improvements; as improved models and/or methods become available, they
will be incorporated into future Risk Evaluations.
Summary: A public commenter (0040) suggested that EPA should mitigate uncertainty in exposure
estimates for OESs by performing targeted analyses to develop similar exposure groups within an OES
when the OESs consist of multiple exposure groups.
EPA Response: For occupational exposures, EPA analyzes workers and occupational non-users
(ONUs) as separate exposure groups. Where reasonably available data allows, EPA provides separate
38
-------
estimates of exposures that each group could be subjected to during the activities that are reasonably
expected to occur during a given PES.
Summary: A peer reviewer (OC - Rl) stated that it seems unreasonable to assume that sentinel
exposures (e.g., subsistence fishers) may not be exposed to TCEP via other exposures scenarios. The
result, according to the commenter, may be an underestimation of risk for some PESS.
EPA Response: In the risk evaluation procedural rule, EPA defines sentinel exposure as "the exposure
to a single chemical substance that represents the plausible upper bound of exposure relative to all
other exposures within a broad category of similar or related exposures" (40 CFR 702.33).
As described in Section 5.1.5, the sentinel exposure for these general population exposure scenarios
was fish ingestion for subsistence fisherman and fishers who are members of Tribes. The sentinel
exposure for the consumer assessments by route were inhalation from building and construction
materials (roofing insulation) for consumers, oral ingestion of TCEP from children's mouthing of foam
seating and bedding products (foam toy blocks), and children's dermal absorption of TCEP from wood
resin products (wood flooring).
Section 5.1.4 describes the different ways in which exposures can be aggregated (e.g., across COUs,
across Exposure Scenarios) Without any information indicating co-exposures to multiple COUs, or
exposure across multiple exposure scenarios, EPA is unable to conduct an aggregate exposure
assessment. Furthermore, as described in section 5.1.4, it would be misleading for EPA to aggregate
exposures across exposure scenarios when the individual scenarios were based on the total production
volume of 2,500 lbs per OES, and aggregating would double count the production volume.
While it is plausible that co-exposures across scenarios could occur, EPA did not quantitively assess
aggregate exposure across exposure scenarios because no data was available indicating the co-exposure
of TCEP from multiple exposure scenarios.
Section 5.2 - Human health hazard (including comments on confidence levels, male reproductive
effects from Chen et al. (2015), and benchmark response (BMR) of 5 percent for the benchmark
dose (BMD) modeling)
Confidence levels
Summary: A public commenter (0040) said they applied the proposed Judgement Levels for Evidence
Integration of the SACC and obtained results no stronger than "evidence suggests" for all cancer and
non-cancer endpoints for TCEP which differs from EPA's "likely" conclusions.
EPA Response: EPA uses the framework in the 2021 Draft Systematic Review Protocol {U.S. EPA,
2021, 10415760}. As an example, EPA determined that the reproductive endpoint has an indeterminate
judgement for human data, moderate judgement for animal data, and slight mechanistic evidence.
Using the 2021 Draft Protocol, these evidence-stream judgements lead to the overall judgment that
"TCEP likely causes reproductive toxicity in humans under relevant exposure circumstances." The
ORD StaffHandbook for Developing IRIS Assessments {U.S. EPA, 2022, 10367891} also states that
this combination of evidence-stream judgements can also indicate an overall judgement that TCEP
likely causes reproductive toxicity in humans given sufficient exposure circumstances.
39
-------
Summary: A peer reviewer (2.3.1 - R4) stated that, based on the confidence summary for each field,
the overall hazard confidence values seemed reasonable.
EPA Response: EPA thanks the commenter for the comment.
Summary: A peer reviewer (2.3.1 - R7) stated that "consistency/concordance" should be considered in
assigning confidence ratings for a given endpoint. For example, Bradford Hill evidence criteria were
mentioned as evidence criteria, but ultimately did not make it onto the list of considerations. The
reviewer stated that the overall rating of "moderate" undervalues the level of evidence that TCEP
causes these various effects. The reviewer suggested adding a new column to provide a level of
confidence in the outcome and Table 5-53 would reflect the level of confidence in the endpoint itself.
Finally, the reviewer said that given that there are always uncertainties in modeling dose response
information or evaluating PESS variabilities, "it's hard to imagine anything higher than moderate".
EPA Response: EPA did consider the Bradford Hill considerations when integrating the evidence for
human health hazard, as stated in Section 5.2.1: "The Agency integrated data from these evidence
streams to arrive at an overall evidence integration conclusion for each health outcome category (e.g.,
reproductive toxicity). When weighing and integrating evidence to estimate the potential that TCEP
may cause a given human health hazard outcome, EPA uses several factors adapted from Sir Bradford
Hill {Hill, 1965, 71664}. These elements include consistency, dose-response relationship, strength of
the association, temporal relationship, biological plausibility, and coherence, among other
considerations." Table 7-13 in the 2021 Draft Systematic Review Protocol {U.S. EPA, 2021,
10415760} describes the use of Bradford Hill criteria more completely when determining evidence
integration j udgments.
Also, the specific Bradford Hill considerations are listed, as relevant to an evidence stream and hazard
endpoint for TCEP, in the evidence integration tables in Appendix K of the risk evaluation as factors
that increase or decrease the strength of the evidence. Table 5-53 in the RE includes a column called
"Evidence Integration Conclusion," which is the sum of all the Bradford Hill considerations. In the
introduction to Section 5.2.7, EPA added reference to Section 5.2.1 where the use of Bradford Hill
criteria is described.
Uncertainty factors
Summary: A public commenter (0044) stated that, when calculating TCEP's hazards, EPA
disregarded the best available science, as well as EPA's own risk assessment guidance, concerning the
use of uncertainty factors (UFs). The commenter said that EPA failed to apply two critical UFs and
misapplied another. First, the commenter (0044), along with other commenters (0042. 0045). stated
that EPA omitted critical UFs when characterizing non-cancer risk in the draft risk evaluation. The
commenter said that the risk evaluation makes no mention of the subchronic-to-chronic study duration
UF that is usually applied to account for the lower dose that may produce the same effect if a chronic
study were conducted. One of the commenters (0042) said that the subchronic UF would increase the
benchmark margin of exposure (MOE) and in turn lower EPA's threshold for identifying risks of
concern by a factor of 3 to 10, which results in a significant underestimation of risk, according to the
commenter. One of the commenters (0045) said a UF of 100 should serve as the benchmark MOE for
chronic exposure scenario risk characterizations.
40
-------
Second, the commenter (0044) said that EPA failed to apply the recommended UFs for database
deficiencies, despite acknowledging significant gaps in EPA's understanding of TCEP's exposures and
hazards. And finally, the commenter (0044) said that EPA misapplies the intraspecies UF which is
intended to account for "variations in susceptibility" within the general population. The commenter
said that this UF is designed to cover the myriad sources of variations within the general population,
and not the increased risks faced by discrete PESS, which is how EPA applied it.
EPA Response: {NTP, 1991, 5469669'a. c/au t hor-ycar} evaluated histopathology of male reproductive
organs after two years. The continuous breeding study {NTP, 1991, 10603716 @@author-year} that
used a chronic exposure duration also evaluated sperm effects at the highest dose. Although it is
possible that an additional chronic study would result in more sensitive effects on seminiferous tubules,
EPA considers that the existing chronic studies adequately evaluated male reproductive effects and a
subchronic to chronic duration uncertainty factor is not needed.
Database UF is applied for deficiencies in the toxicological database that might lead to a lower POD.
Under TSCA, there is no universal list of hazard data that is required to consider a database sufficient
to conduct risk evaluation, nor is there a minimum set of data required to conduct a risk evaluation. The
decision to incorporate UF database in TSCA risk evaluations is determined on a case-by-case basis,
and for TCEP, EPA determined that the database of studies did not identify deficiencies. Multiple
studies evaluated short-term, subchronic, and chronic effects and covered major organ systems. In
addition, studies evaluated various aspects of neurotoxicity or neurotoxicity-related endpoints in both
rats and mice (clinical signs, histopathology, neurobehavioral changes, cholinesterase, and other
biochemical changes). In particular, {Yang, 2018, 5469245@@author-year} conducted the Morris
water maze in rats after 60 days of exposure; the Morris water maze measures spatial learning and
memory, and was an important assessment given that lesions were found in the hippocampus. In
addition, newly identified epidemiological studies examined neurobehavior and IQ in children and
depression and stress in mothers. Multiple reproductive outcomes are well covered. The reproductive
assessment and continuous breeding study {NTP, 1991, 10603716} includes endpoints recommended
for evaluation in OECD Test Guideline 416 (2-Generation Reproduction Toxicity Study). {NTP, 1991,
10603716@@author-year} examined F0 and F1 adults (changes in reproductive organs, sperm
parameters, and estrous cyclicity) as well as reproductive performance of both generations (fertility,
viability of F1 and F2 offspring, sex of offspring, birthweight). Reproductive endpoints were also
evaluated in subchronic and chronic studies. {Chen, 2015, 4199395@@author-year} examined
reproductive endpoints in adolescent male mice and males appear to be more sensitive than females.
Regarding the intraspecies UF, EPA has used all chemical-specific information that is available. For
example, male reproductive effects are the basis of one of the PODs, and they represent a susceptible
subpopulation. When chemical-specific information is not available, EPA relies on using the
intraspecies UF value of 10. EPA recognizes that the intra-species UF is not specific, and Section 5.3.3,
Table 5-79: Summary of PESS Considerations Incorporated into the Risk Evaluation acknowledges
that it is not known whether the UF of 10 would cover various subpopulations (such as individuals with
Klinefelter" s, which is one of the examples identified by the commenter). EPA has added the following
statement to Appendix E for more clarity: "In addition, given that EPA is using a default UFh in the
absence of data regarding whether adverse effects from TCEP exposure differ for certain
subpopulations (such as those with genetic polymorphisms or underlying diseases), it is also not known
whether the chosen default UFh would fully cover pre-existing diseases or disorders {U.S. EPA, 2002,
88824(gj(gjauthor-year}
41
-------
Male reproductive effects (Chen etal.)
Summary: A public commenter, in two submissions to the docket (0040. 0052). said they applied
SACC's recommendations for evidence integration judgment which resulted in an Evidence Integration
Judgment of "suggestive" for male reproductive effects, therefore, the mail reproductive effects
identified by EPA are likely insufficient to support causality for risk evaluation.
EPA Response: EPA uses the framework in the 2021 Draft Systematic Review Protocol {U.S. EPA,
2021, 10415760} to arrive at evidence integration judgments. As an example, EPA determined that the
reproductive endpoint has an indeterminate judgement for human data, moderate judgement for animal
data, and that there is slight mechanistic evidence. Using the 2021 Protocol, these evidence-stream
judgements lead to the overall judgment that "TCEP likely causes reproductive toxicity in humans
under relevant exposure circumstances/' The ORD Staff Handbook for Developing IRIS Assessments
{U.S. EPA, 2022, 10367891} also states that this combination of evidence-stream judgements can also
indicate an overall judgement that TCEP likely causes reproductive toxicity in humans given sufficient
exposure circumstances.
Summary: A public commenter (0042) stated that EPA continues to rely on scientifically deficient
methods for non-cancer dose-response analysis and risk characterization. For example, EPA used the
MOE approach which is a scientifically inappropriate approach for characterizing risk and is
inconsistent with amended TSCA's requirements to use the "best available science"', according to the
commenter. The commenter went on to state that the MOE does not estimate the proportion of the
exposed population projected to experience a specified health endpoint or the number of individuals
affected. The commenter stated that the National Academies and the World Health Organization have
outlined more robust methods for risk estimation that more accurately account for variability and
vulnerability across the human population. The commenter applied the World Health Organization
methodology to estimate risks of male reproductive effects and found that the risk level is 25,000 times
higher than the target range that EPA typically applies for protection of carcinogenic risks. The
commenter recommended that EPA include this approach to non-cancer dose-response and risk
characterization in the final TCEP risk evaluation.
EPA Response: EPA thanks the commenter for the comment. EPA does not want to a priori preclude
the use of any methods or data types, in order to allow its risk evaluations to change as science
advances. EPA will utilize current policies, models, and screening methods, but is committed to being
consistent with the best available science and weight of the scientific evidence approaches to guide the
Agency in using this information. EPA recognizes the advancing science to inform risk evaluation and
will not discourage the use of new methods as long as they are consistent with the standards in section
26 of TSCA. EPA also recognizes that different approaches require different types and amounts of data
and will select and employ methods that are fit for purpose within the context of a particular risk
evaluation. In some cases, it may be necessary to utilize default parameters in modeling and risk
calculations, and to utilize conservative assumptions, whereas in other cases assumptions may be
replaced with specific or specialized data. It should also be noted, in addition, their use will be peer
reviewed, and the public will have the opportunity to comment on them during the public comment
periods. EPA has utilized the MOE approach in previous risk assessments, citing its utility. However,
EPA does agree with comments that there are numerous ways to characterize risk, of which MOE is
just one. There will be risk scenarios where one approach may be better than another and, as
commenters correctly pointed out, the science of risk characterization is still evolving, particularly for
non-cancer hazards. Hence, OPPT will use risk characterization approach(es) suitable for the purpose
42
-------
of the risk evaluation and that the best available science and data support. EPA does not agree with the
commenter to use another approach for the TCEP risk evaluation.
Summary: A public commenter (0045) stated that EPA used as a point of departure the BMDL5
instead of the frequently defaulted BMDL10 to accommodate for the nature and severity of the effects
observed in the Chen et al. study. The commenter stated that this requires an additional 3X UF to
account for subchronic extrapolation.
EPA Response:
Use of 5% BMR to set a BMDL5: In consultation with a reproductive/developmental toxicologist and
another EPA office that routinely conducts dose-response modeling, EPA concluded that a BMR of 5
percent to be appropriate for the severity of effects that can result in significant reproductive effects,
including decreased fertility and viability observed in {NTP, 1991, 1060371@@author-year}. EPA
acknowledges that the modeling is not ideal because the BMD is low on the dose-response curve
compared with the doses tested in the study. However, BMD modelers identified the importance of
choosing a BMR a priori. EPA has already cited publications supporting the use of this BMR
{Blessinger, 2020, 6966510; Lanning, 2002, 2850042}.
In the risk evaluation, EPA expanded the justification for the severity of effects that warrants a BMR of
5 percent. Even if the effects are not life-threatening to the parents, the possibility of decreased
viability in offspring is of concern. Furthermore, EPA expects that humans may be more sensitive to
male reproductive toxicants than rodents.
Most peer reviewers supported EPA's use of the 5 percent BMR because of the severity of the effect.
Subchronic to chronic UF: {NTP, 1991, 5469669 <7, <7, author-year} evaluated histopathology of male
reproductive organs after two years. The continuous breeding study {NTP, 1991, 10603716
V/ V/,author-year} that used a chronic exposure duration also evaluated sperm effects at the highest
dose. Although it is possible that an additional chronic study would result in more sensitive effects on
seminiferous tubules, EPA considers that the existing chronic studies adequately evaluated male
reproductive effects and a subchronic to chronic duration uncertainty factor is not needed.
Summary: Two peer reviewers (2.3.2 - Rl. 2.3.2 - R3) stated that they agree with the choice to use
Chen et al. for both short-term and chronic exposures scenarios. One reviewer (2.3.2 - Rl) said that it
looks to be a reliable study which was conducted during a potentially sensitive life stage which results
in a health-protective POD for the risk evaluation. Therefore, this reviewer asserted that this POD will
likely be protective of the other adverse effects identified in the TCEP toxicity studies of other non-
cancer endpoints. The second reviewer (2.3.2 - R3) commented that one limitation of this study is that
the feed consumption was not reported, therefore, the dose is uncertain. This reviewer stated that EPA
may want to consider using an additional uncertainty factor to address the dose uncertainty.
EPA Response: Because the feed consumption was not reported, it is not clear whether the intake of
TCEP could be higher or lower than the reported doses. Therefore, EPA is not planning to add an
uncertainty factor to this study.
Summary: A peer reviewer (2.3.2 - R2) stated that the data obtained from the National Toxicology
Program (NTP) studies and Chen et al. clearly demonstrated the potential for species and strain
43
-------
differences (and possibly age differences) in the manifestation of effects on the male reproductive
system. The reviewer said that there were points of comparative concern with the absence of exposure
related differences in testes weights or pathology of male reproductive organs at dose levels in the
reproductive assessment by continuous breeding (RACB) FO adult CD-I mice that were seen in the
Chen et al. study. However, the reviewer stated that the use of male reproductive effects reported in
Chen et al. were in line with the available findings and appropriate for use in this assessment.
EPA Response: EPA thanks the commenter for the comment. EPA has described these uncertainties
related to differences in results among studies within the risk evaluation.
Summary: A peer reviewer (2.3.2 - R4) stated that they are hesitant to endorse the use of the Chen et
al. study and would not use it as a part of the quantitative risk assessment. The reviewer said that the
limitations described by EPA in Table 5-48 are serious and the use of nominal doses in the absence of
food consumption and bodyweight data are questionable.
EPA Response: EPA agrees that there are important limitations to {Chen, 2015, 4199395@@author-
year}, as identified by the peer reviewer. However, EPA has chosen to use this study because: 1) TCEP
has a limited data set and this study has an adequate overall quality determination; 2) other studies also
have limitations; 3) human males are sensitive to reproductive effects and this study identified
significant degeneration of seminiferous tubules; and 4) it is possible that if additional toxicological
studies were to be conducted on male reproductive effects of TCEP, similar effects might be observed.
Summary: A peer reviewer (2.3.2 - R5) stated that the Chen et al. study is considered high-quality and
that the reproductive toxicity observed in male rats is considered to be a sensitive endpoint relevant to
humans. The reviewer stated that the Chen et al. study was also used for the chronic exposure scenario
and the calculated Human Equivalent Dose (HED) is within an order of magnitude of the HEDs
calculated for short-term exposure for neurotoxicity and developmental toxicity. The reviewer stated
that the use of the same study for subchronic and chronic exposures seems appropriate. However, the
reviewer asserted that EPA failed to apply an uncertainty factor for the chronic exposure scenario,
which typically relies on a lower value than the subchronic exposure scenario. As such, EPA should
justify the lack of an uncertainty factor adjustment.
EPA Response: {NTP, 1991, 5469669@@author-year} evaluated histopathology of male reproductive
organs after two years. The continuous breeding study {NTP, 1991, 10603716 @@author-year} that
used a chronic exposure duration also evaluated sperm effects at the highest dose. Although it is
possible that an additional chronic study would result in more sensitive effects on seminiferous tubules,
EPA considers that the existing chronic studies adequately evaluated male reproductive effects and a
subchronic to chronic duration uncertainty factor is not needed.
Summary: A peer reviewer (2.3.2 - R6) stated that the use of Chen et al. is generally supported by the
strengths of the study which also has the lowest HED among the other studies considered by EPA. The
reviewer stated that the limited dose groups used in the Chen et al. study, only at 100 and 300 mg/kg-
day, were a weakness to derive BMDL5 as the reduction in the numbers of seminiferous tubules was
22% and 41%, respectively, in these dose groups. The reviewer said that, for the chronic exposure
scenario, the Chen et al. study had a much shorter exposure duration of 35 days, giving it a
disadvantage in its use in providing the HED. However, the reviewer stated that the advantage of the
Chen et al. study is that it identified histopathological changes in the seminiferous tubules which was
44
-------
not seen in the NTP (1991a) study even though the NTP study had a much longer exposure duration.
The reviewer suggested that EPA should provide additional details on the potential explanation of
differences between these study findings in order to better understand this comparison.
EPA Response: EPA does not have additional details to explain the differences. In the absence of
these details, EPA is using a sensitive study (more sensitive than chronic studies) to evaluate risks from
TCEP. This is recommended by the EPA guidelines for reproductive toxicity risk assessment {U.S.
EPA, 1996, 30019}, which suggests using the most sensitive species.
Summary: A peer reviewer (2.3.2 - R7) stated that the use of the Chen et al. dose response to define
both short-term and chronic PODS is reasonable. The reviewer commented that TCEP is clearly a male
reproductive toxicant and supported by the RACB crossover studies by NTP. The reviewer said that the
NTP 1991 RACB protocol should have detected the anti-male effect at 175 mg/kg/day given that it
covered critical windows of development, but only found male reproductive effects at higher doses.
The reviewer reasoned that there is less likelihood or concern that the Chen et al. study missed a critical
window of vulnerability and that the calculated BMDL HED from Chen et al. is in line with other
reproductive endpoints. The reviewer agreed with EPA's decision to use this endpoint (histopathologic
evidence of decrease seminiferous tubules) as the most sensitive indicator of TCEP non-cancer effect
for the purposes of POD development across endpoints and exposure timeframe,
EPA Response: EPA thanks the commenter for the comment.
Cancer risks
Summary: A public commenter (0044) stated that EPA understated TCEP's cancer risks. When
calculating TCEP's cancer risks, EPA relied exclusively on kidney tumor data even though TCEP
exposures have been associated with other cancer sites as well. The commenter stated that EPA must
consider all known cancer sites in its dose-response analysis and its determination of whether TCEP
presents unreasonable risk and recommended that EPA add non-kidney cancers associated with TCEP
exposure to its analysis of cancer risk.
EPA Response: EPA considered information supporting evidence of other tumor types (mononuclear
cell leukemia, thyroid, Harderian gland, and liver) in rodents {NTP, 1991, 5469669} and two studies
of thyroid cancer in humans {Hoffman, 2017, 4161719;Liu, 2022, 11364830} to be slight.
Mononuclear cell leukemia is a common spontaneous cancer in F344 rats and most incidence (other
than females at the highest dose) were within historical controls. Statistically significant increased
thyroid cancer was seen in female rats but not male rats or mice; humans showed inconsistent results
in two studies. Increased Harderian tumor incidence was only seen in female B6C3F1 mice but not in
male mice or in either sex of rats. Dose-related trend in liver tumor incidence was seen only in one sex
of one species (male B6C3F1 mice), with only borderline statistical significance when comparing the
high dose to controls. Due to the slight evidence associated with these tumors, EPA did not add them
to the incidence of kidney cancer.
Summary: A public commenter (0042) also stated that though EPA has appropriately modeled cancer
dose-response as a linear relationship with no threshold, EPA incorrectly states that because TCEP
does not act through a known mutagenic mode of action (MOA), there is a threshold below which there
is no cancer risk. The commenter stated that EPA provides no evidence to support its speculation that
there is a threshold for TCEP's cancer risk, and the absence of a mutagenic MOA is not sufficient
45
-------
evidence. The commenter recommended that EPA remove the statements regarding a cancer threshold
for TCEP from the draft risk evaluation before it is finalized.
EPA Response: EPA has revised certain language in sections 5.2.5.3; 5.2.6.2; and 5.2.7.1.3 to indicate
that the dose-response is not clear based on lack of adequate data to determine the mechanisms.
Summary: A peer reviewer (2.3.1 - Rl) agreed with EPA's conclusions concerning each identified
tumor type, that the available data indicate that TCEP has little genotoxic potential, and that it is
unlikely that TCEP acts through a mutagenic MOA. The reviewer stated that EPA should note that the
dataset used is incomplete and lacks an in vivo genotoxicity study of acceptable quality. The reviewer
recommended that the statement indicating that TCEP has little genotoxic potential should be qualified.
EPA Response: In the evidence integration section (5.2.5.4), EPA added language to note that in vivo
studies are limited: "The mechanistic evidence for carcinogenesis is slight. Available data indicates
that TCEP has little if any genotoxic potential, but data are limited to assess in vivo genotoxicity/1
Benchmark dose
Summary: A public commenter (0040) said that though EPA attempted to justify a 5% BMR, the
modeled response is not predictive, as indicated by empirical data, and the BMR and/or selected model
must be re-examined. Additionally, the commenter said that for the cancer endpoint, EPA's BMD
modeling does not comport with EPA's BMD Guidance and is possibly biased toward a more
precautionary approach due to EPA's a priori exclusion of animals expiring prior to the study
conclusion.
Similarly, a peer reviewer (2.3.3 - R4) stated that the doses and response data used for the modeling
presented in Table 1-1 are concerning and that they are uncomfortable basing the BMDL on a BMR of
5%. The reviewer reasoned that the biological data from the Chen et al. study showed the BMR
associated with decreases of 22.2 and 40.7 percent at 100 and 300 mg/kg-day, which are lower than the
actual study results. The reviewer stated that they do not understand the scientific basis for choosing
5% given these data and refers the reader to their previous comment for Charge Question 2.3.2.
EPA Response: Five percent BMR: In consultation with a reproductive/developmental toxicologist
and another EPA office that routinely conducts dose-response modeling, EPA determined a BMR of 5
percent to be appropriate for the severity of effects that can result in significant reproductive effects,
including decreased fertility and viability observed in {NTP, 1991, 10603716 V/ V/ author-year}. EPA
acknowledges that the modeling is not ideal because BMD is low on the dose-response curve compared
with the doses tested in the study. However, BMD modelers identified the importance of choosing a
BMR a priori. EPA has already cited publications supporting the use of this BMR {Blessinger, 2020,
6966510; Lanning, 2002, 2850042}.
In the final risk evaluation, EPA expanded the justification for the severity of effects that warrants a
BMR of 5 percent. Even if the effects are not life-threatening to the parents, the possibility of decreased
viability in offspring is of concern. Furthermore, EPA expects that humans are likely to be more
sensitive to male reproductive toxicants than rodents.
Most peer reviewers supported our use of the 5 percent BMR because of the severity of the effect.
46
-------
Cancer - exclusion of animals expiring: The com mcntcr's concern appears to be that EPA adjusted
the cancer data for mortality a priori without justification. This is not the case. Survival was reduced in
high-dose male and female rats in the NTP study. It was noted that female rats dying early frequently
showed brain lesions (associated with TCEP exposure), suggesting a cause of early death unrelated to
the observed kidney tumors. The adjustment was made to try to account for the observed difference in
mortality across treatment groups in the study. This practice conforms with EPA BMD and cancer
guidelines, as well as the FDA guidance quoted by the commenter.
Summary: Two peer reviewers (2.3.3 - Rl. 2.3.3 - R5) stated that the use of a BMDL05 seems
appropriately justified. One of the reviewers (2.3.3 - Rl) said this is due to the severity of the testicular
effects, the potential for high sensitivity of humans, and the indication from the description of an
unavailable study (Shepel'skaia and Dyshginevich (1981)) that TCEP may exhibit reproductive toxicity
at low doses. However, one of the peer reviewers (2.3.3 - R3) expressed concern that the BMD and
BMDL fall quite close to the control group and are lower than the biologically and statistically adverse
response, increasing uncertainty about the values. The reviewer stated that, while not ideal, the other
studies had limitations as well. Another peer reviewer (2.3.3 - R2) stated that the use of Chen et al. for
male reproductive effects (decreased numbers of seminiferous tubules) was an appropriate endpoint for
BMD modeling. The reviewer expressed concern as to whether this reflects the functional outcomes
observed in the NTP studies given that similar pathology was not reported.
EPA Response: EPA thanks the commenter for the comment. EPA has consulted with reproductive
and dose-response modeling experts to choose the BMR of 5 percent based on severity. EPA also
discusses the uncertainties, including that the BMD and BMDL are below the doses used in the study.
Given the limited database and concern that changes in sperm measures are expected to have greater
impacts in humans than rodents, EPA considers the BMR of 5 percent to be appropriate for the risk
evaluation of TCEP even though it does fall below the responses within the study at 100 and 300
mg/kg-day; no NOAEL was identified in the study so if the study had included lower doses, one of
them may have shown a response similar to 5 percent.
EPA acknowledges the differences in outcomes among studies. EPA chose the more sensitive endpoint
in the absence of information to definitively identify the more appropriate study.
Summary: A peer reviewer (2.3.3 - R3) stated that the BMR of 5 % seems appropriate based on the
severity of the effect accompanied by disintegration of tubules, decreased testosterone levels, and testes
weights at the highest dose. The reviewer expressed concern that the BMD and BMDL fall quite
closely to the control group and are lower than the biologically and statistically adverse response,
increasing uncertainty about the values. The reviewer stated that, while not ideal, the other studies had
limitations as well. Another peer reviewer (2.3.3 - R6) stated that a BMR of 5% is protective of public
health, given the severity of the impact on fertility (reduction in the number of seminiferous tubules),
that infertility is increasingly of concern in US, and the lack of strong human study data. A peer
reviewer (2.3.3 - R7) stated that a BMR of 5% is supportable in general for reproductive and
developmental endpoints given their severity, as long as the data and model fit are supportive of
predicting a reliable 5% response. The reviewer noted that the data were "well behaved" within the
experimental and extrapolated range, and this was shown in the Benchmark Dose Modeling
Supplemental File, in which "acceptable data fits were obtained for several models with predictions of
47
-------
the selected model fit for BMDL5 within a reasonable range (some estimates lower, some higher)
across model fits/'
EPA Response: EPA thanks the commenter for the comment.
Neurotoxicity risks
Summary: A public commenter (0044) said that EPA failed to use the most sensitive study, Tilson et
al. (1990), when calculating TCEP's neurotoxic effects. The commenter said that none of EPA's
arguments against using the study justifies the decision to understate TCEP's neurotoxicity risks, and
stated that the study has been favorably cited in an analysis conducted for the CPSC, by Environment
Canada and Health Canada, and previously by EPA itself.
EPA Response: {Moser, 2015, 3008543@@author-year} remains the best (but not ideal) study given
that the effects were seen beyond the technical definition of acute (< 24 hr). As stated on line 7750 of
the RE, EPA identified a NOAEL from Moser et al. 2015 with effects seen only at the highest dose.
The other acute studies identified only a NOAEL or LOAEL with effects observed only at the highest
dose or the only dose in the study. {Moser, 2015, 3008543@@author-year} resulted in the most
sensitive endpoint with an identified NOAEL, and EPA considers this to be appropriate for the acute
endpoint.
Summary: A peer reviewer (2.3.1 - Rl) agreed with EPA's summary statement on the confidence
level for neurotoxicity and stated that neurotoxic effects were seen in multiple quality studies in both
mice and rats. Another peer reviewer (2.3.1 - R2) stated that neurotoxicity was clearly demonstrated
across the various studies and the apical endpoints were suggestive of excitotoxicity. A third reviewer
(2.3.1 - R3) stated that the evidence for neurotoxicity in animals is robust with several studies showing
consistent neurotoxic clinical and behavioral findings. However, this reviewer (2.3.1 - R3) added that
the inference across evidence streams for an overall judgement of "likely causes
neurological/behavioral effects in humans under relevant exposure circumstances" is unclear because
of the limited mechanistic and human data. The reviewer also said that it is unlikely humans would be
exposed to a similar exposure scenario as a bolus gavage dose in rodents. In the discussion of the
strengths, limitations, etc. for acute non-cancer endpoints, the reviewer said there was no discussion
beyond the Moser and Hazleton Laboratories study which are reproductive/development neurotoxicity
studies. According to the reviewer, the Yang (2018) study was not discussed, though it was considered
for endpoint selection. This reviewer remarked that it was unclear why short-term and chronic exposure
scenarios were assessed together.
EPA Response: EPA thanks the commenter for the comment. EPA uses the framework in the draft
systematic review protocol published in December 2021 for inferencing across evidence streams. Table
7-14. Classification for Weight of the Scientific Evidence for Causal Determinations for Characterizing
Potential Human Health Hazards provides guidance and example scenario for overall evidence
integration judgement in narrative. Yang et al was discussed in section 5.2.3.1.1 in both the
"Laboratory Animals" and in the "Mechanistic information" sections.
Summary: A peer reviewer (OC - R2) expressed that there are several statements related to
neurotoxicity in the draft risk evaluation that should be edited for accuracy:
48
-------
The peer reviewer stated that the statement regarding tremors in line 6448 of the RE is
misleading.
The peer reviewer wrote that the document states incorrectly in line 6385 of the RE that 13 rats
showed tremors. The peer reviewer said that only two rats showed tremors and were
terminated, and the rest were decreased to 90 mg/kg.
The peer reviewer commented on line 6392 and requested that EPA check how many animals
showed ataxia, salivation, gasping, convulsions, and occasional hyperactivity. The peer
reviewer stated that it was not the full cohort.
EPA Response: Although {Moser, 2015, 3008543(ff>(ff>author-year} state that they euthanized two
dams due to tremors and decreased the dose from 125 to 90 mg/kg-day, the study does not specifically
state how many dams showed tremors. The authors provided an excel spreadsheet to EPA that showed
13 dams with tremors ranging from very slight to greater than moderate. The two dams with greater
than moderate tremors were sacrificed on day 3.
EPA summarized the information on female rats that showed ataxia, salivation, gasping as {NTP,
1991, 5469669(ff>(ff>author-year} presented the information (line 6392); the number of female rats with
these signs was not identified. Therefore, there is no revision necessary.
Reproductive toxicity
Summary: A peer reviewer (2.3.1 - Rl) agreed with EPA's summary statement on the confidence
level for reproductive toxicity and stated that almost all of the evidence comes from one species
(mouse).
EPA Response: EPA thanks the commenter for the comment.
Developmental toxicity
Summary: A peer reviewer (2.3.1 - Rl) agreed with EPA's summary statement on the confidence
level for developmental toxicity and stated that several studies in one NTP report provide evidence for
developmental toxicity and that the evidence primarily comes from one species (mouse).
EPA Response: Thank you for your comments. However, based on comments from other reviewers,
EPA has decided that the evidence is suggestive but not sufficient to indicate that TCEP causes
developmental toxicity. Section 5.2.3.1.3 of the risk evaluation has been updated.
Summary: A peer reviewer (2.3.1 - R2) stated that it was not clear which studies were used as the
basis to make a determination of "moderate." For example, the reviewer said that the Moser et al
(2015) study did not interpret the data to represent developmental toxicity, the Hazleton Laboratories
1983 study did not demonstrate altered viability or growth of mouse offspring, and the effects observed
from the NTP RACB study did not suggest developmental toxicity. If EPA assigned a ranking of
"moderate" to developmental toxicity, the reviewer recommended additional supporting data. A second
peer reviewer (2.3.1 - R3) stated that the evidence suggested in Appendix K is not consistent with a
moderate developmental toxicity health hazard. This reviewer noted that most of the evidence of
toxicity from the Moser and Hazleton Laboratories indicate reproductive toxicity.
EPA Response: Although EPA believes that differences in study protocols between the RACB and
prenatal studies may explain differences in outcomes, EPA agrees with the commenter given lack of
49
-------
evidence in several prenatal/postnatal studies and in an additional study recently translated to English
{Kawashima, 1983, 4992702}. Therefore, EPA has changed the evidence for animal studies to slight.
EPA has also recently identified human epidemiology studies that resulted in slight evidence. Overall,
the evidence integration conclusion in Section 5.2.3.1.3 has been changed to "suggestive but not
sufficient to indicate that TCEP causes developmental toxicity." Given this change in the evidence
integration conclusion, the information for developmental toxicity was not considered for dose
response modeling (Section 5.2.6) or summarized in the weight of scientific evidence section (5.2.7).
Nephrotoxicity
Summary: Two peer reviewers (2.3.1 - Rl. 2.3.1 - R3) agreed with EPA's summary statement on the
confidence level for kidney effects and stated that neurotoxic effects were seen in quality studies in
both mice and rats. One reviewer (2.3.1 - R3) stated that it was unclear if kidney weight was a relative
or absolute value, and that the distinction should be made.
EPA Response: There was a statistically significant 12% and 10% increase in absolute kidney weight
and an 8% and 10% increase in relative to body kidney weights at 175 and 350 mg/kg/day, respectively
in male rats during the 16-day study. Absolute kidney weights were significantly increased 21.9% in
males at 350 mg/kg/day. Relative-to-body kidney weights were increased 24.9% and absolute kidney
weights were significantly increased 7%, 7%, 16.9%, and 46.5% and relative-to-body kidney weights
were significantly increased 9.2%, 11.1%, 13.3%, and 22.2% in females given 44, 88, 175,and350
mg/kg/day, respectively. During the 16 weeks study, male mice receiving 700 mg/kg/day had
significantly reduced absolute kidney weights, decreased 19.4% compared to the controls. Relative-to-
body kidney weights were decreased at 175, 350, and 700 mg/kg/day by 13.3%, 16.0%, and 14.1%
compared to controls during the 16 weeks study.
Summary: A peer reviewer (2.3.1 - R2) agreed that the data support a determination of "moderate" for
kidney toxicity and kidney adenomas and carcinomas but stated that it is not clear whether that
represents an independent call of "moderate" for kidney toxicity and one for cancer. The reviewer
suggested that ranking cancer as "moderate" requires additional supporting information.
EPA Response: EPA thanks the commenter for the comment. The confidence levels of moderate were
chosen separately for non-cancer kidney toxicity and for kidney cancers (see section 5.2.3.1.4. for
kidney toxicity and section 5.2.5.4 for cancer, as well as separate evidence integration tables in
Appendix K in the risk evaluation), although some consideration was given across endpoints when
assigning the individual evidence integration conclusions and confidence levels.
Liver toxicity
Summary: Two peer reviewers (2.3.1 - Rl. 2.3.1 - R3) agreed with EPA's summary statement on the
confidence level for liver effects. One reviewer (2.3.1 - R3) noted that it was unclear if liver weight
was a relative or absolute value, and that the distinction should be made.
EPA Response: EPA thanks the commenter for the comment. In section 5.2.3.2, EPA revised the
evidence integration section to identify whether liver weight changes were absolute or relative to body
weight.
50
-------
General comments on hazard identification
Summary: A public commenter (0045) stated that no adequate human data were identified for use in
the assessment of potential noncancer or cancer risk of TCEP exposure to humans, and none of the
available evidence demonstrated that TCEP causes the effects of concern in humans. However, the
commenter agreed that the available data do identify neurotoxicity, reproductive toxicity,
developmental toxicity, and kidney toxicity including cancer as the most likely and sensitive potential
adverse human health hazard outcomes associated with TCEP exposure.
EPA Response: EPA thanks the commenter for the comment. EPA has since identified additional
human data and has added this information to the TCEP RE. In Section 5.2.3.1, EPA added studies on
neurological, kidney, immune/hematological, thyroid, lung, and developmental effects, and in Section
5.2.5.1, EPA added additional cancer studies.
Summary: A public commenter (0045) also stated that they agree with the approaches taken to derive
the oral and dermal human equivalent doses, inhalation Human Equivalent Concentrations and Cancer
Slope Factors Inhalation Unit Risks. Finally, the commenter said that they disagree with the weight-of-
evidence assessment and stated that there was not clear evidence of renal tubule cells carcinomas in
rats.
EPA Response: EPA agrees with the commenter that the evidence is based on adenomas and has
revised section 5.2.5.4. Evidence Integration Summary from "clear evidence of renal adenomas and
carcinomas in rats" to the following: "clear evidence of carcinogenic activity in rats based on renal
tubule adenomas," which matches the conclusion from {NTP, 1991, 5469669@@author-year}. The
commenter still agrees that the tumors can be used for dose response, and EPA also agrees with this
conclusion. Furthermore, most peer reviewers agree with the weight of the scientific evidence for
cancer and EPA will keep the overall weight of scientific evidence the same.
Summary: A peer reviewer (2.3.1 - R7) stated that they agree with EPA's separation of likely TCEP
outcomes (kidney cancer, neurotoxicity, reproductive toxicity, developmental toxicity, kidney toxicity)
from endpoints with only suggestive or indeterminate evidence (e.g., liver, thyroid). This reviewer
stated that it is unclear how EPA judges "Endpoint/POD" sensitivity and such sensitivity
determinations should be clarified.
EPA Response: EPA thanks the commenter for the comment. If the endpoint is relevant as an apical
endpoint for use in the RE document, considerations regarding sensitivity are focused primarily on the
dose at which the effect occurs, with the lower the dose associated with greater sensitivity, without
explicit consideration of the magnitude of uncertainty factors/margins of exposure.
Summary: A peer reviewer (OC - R6) suggested that EPA conduct a non-cancer risk quantification in
the draft risk evaluation.
EPA Response: EPA did quantify risks from non-cancer endpoints as presented in Section 5.3. EPA
did not provide a probabilistic assessment of non-cancer effects.
51
-------
Section 5.3 - Human health risk characterization (including comments on types of PESS)
PESS
Summary: A public commenter (0042) stated that to date, EPA has not employed a consistent or
structured approach to identifying PESS in TSCA risk evaluations, including scope documents. One
inconsistency was a difference in whether health conditions related to a chemical's hazards were
considered and whether fenceline communities were included. For example, fenceline communities
were identified as PESS in hexabromocyclododecane (HBCD), but not for 1,4-dioxane, 1-
bromopropane (1-BP), or C.I. Pigment Violet 29 (PV-29). The commenter stated that Rayasam et al.
(2022) recommended that EPA prepare a comprehensive methodology to identify PESS and quantify
their risks consistently within risk evaluations. To do this, the commenter recommended that first, EPA
focus on identifying susceptible subpopulations based on either chemical-specific evidence or the
broader literature on intrinsic and extrinsic susceptibility factors. Once appropriate subgroups are
identified, EPA can then consider the availability of chemical-specific data. The commenter also
suggested EPA focus on lifestage, pre-existing disease, individual activities, occupational exposures,
geographic factors, socio-demographic factors, nutrition, genetic, aggregate exposure, and other
chemical and non-chemical stressors to properly identify PESS.
A peer reviewer (OC - R6) similarly stated that EPA did not consider individuals with pre-existing
disease, genetic factors, lifestyle factors, or exposures to other chemical and non-chemical stressors
that may increase susceptibility to harm from TCEP exposure.
EPA Response: EPA conducted a comprehensive literature search to identify any chemical-specific
human health hazard and exposure data and evaluated these data for data quality and described the
available TCEP-specific PESS information in Appendix D, with a summary in Section 5.3.3. For
human health hazards, EPA also identified mechanistic and toxicokinetic data and summarized any
PESS information in the TCEP risk evaluation. EPA similarly identified and summarized chemical-
specific PESS information resulting from the comprehensive exposure literature search.
For non-cancer effects, EPA applied an intraspecies uncertainty factor of 10 to account for more
sensitive populations (e.g., pre-existing disease, etc). EPA also considered exposure to other chemical
and non-chemical stressors in Appendix D and section 5.3.3.
Risk characterization
Summary: A public commenter (0045) stated that upon examination of Table 5-57 in the draft risk
evaluation, which presents the risk estimates for each condition of use, the Agency failed to implement
aggregate exposure approaches but provided estimates separately by route of exposure. The commenter
stated that this practice assures an underestimation of the real world risk in which actual exposure is
occurring.
The commenter also discussed Tables 5-58 "Acute and Chronic Non-cancer Consumer Risk Summary"
and Table 5-59 "Lifetime Cancer Consumer Risk Summary" and said that they present risk
characterization on a route-by-route basis rather than in the aggregate. The commenter said that these
tables are incomplete and contrary to the Agency's stated goal of increased transparency. Further, the
commenter discussed multiple tables in Section 5 of the draft risk evaluation and again expressed
disappointment that these tables do not present aggregate MOEs or cancer risk estimates for scenarios
52
-------
reflecting multi route exposures, and all the tables also include findings only for conditions of use that
exceed an acceptable threshold.
EPA Response: EPA thanks the commenter for the comment. A full discussion of EPA's
consideration of aggregate exposure is provided in Section 5.1.4 of the risk evaluation and details the
conditions of aggregating exposure across routes within the consumer assessment.
Because data were not reasonably available data to indicate co-exposures of multiple TCEP containing
activities or products in the occupational and indoor environment, EPA did not assess aggregate
exposure across consumer, commercial, or industrial CPUs.
Section 6 - Unreasonable Risk Determination
Comments associated with this issue are summarized in the subsections below.
Section 6.1 - Unreasonable risk to human health
Summary: A public commenter (0044) stated that EPA's unreasonable risk thresholds are unsupported
and under-protective. First, the commenter stated that EPA has historically supported the use of a 1-in-
1,000,000 cancer risk threshold under TSCA but in this draft risk evaluation, EPA replaced the
benchmark with a range of extra cancer risk from l-in-10,000 to l-in-1,000,000. EPA set this threshold
for unreasonable occupational cancer risks at the less protective end of that range, claiming it is
consistent with The National Institute for Occupational Safety and Health (NIOSH) guidance.
However, the commenter stated that EPA misinterpreted NIOSH guidelines, which described l-in-
10,000 not as a target cancer risk level but rather a "starting point for continually reducing exposures in
order to reduce the remaining risk". For consumers, the general public, and fenceline communities,
EPA also claims that it typically considers the benchmark for cancer risk to be within the range of 1 -in-
10,000 to l-in-1,000,000, which, according to the commenter, is a departure from the approach set
forth in EPA's Draft TSCA Screening Level Approach for Assessing Ambient Air and Water
Exposures to Fenceline Communities. The commenter claimed that this "range" of risk levels is up to
100 times less protective than EPA's prior approach and far less transparent.
However, citing the draft risk evaluation's findings for cancer and non-cancer risk from both
occupational and general public exposures, the commenter expressed support for EPA's determination
that TCEP presents unreasonable risk to human health.
EPA Response: EPA appreciates the comment and clarified the unreasonable risk determination to
make clear what cancer benchmarks EPA considers in the unreasonable risk determination. As
explained in Section 5.3.1 of the risk evaluation of TCEP, "EPA considers a range of extra cancer risk
from 1 x 10 4 to 1 /10 " to be relevant benchmarks for risk assessment (U.S. EPA, 2017a). Consistent
with NIOSH guidance (Whittaker et al., 2016), under TSCA, EPA typically applies a 1 x 10 4
benchmark for occupational scenarios in industrial and commercial work environments subject to
OSHA requirements. EPA typically considers the general population and consumer benchmark for
cancer risk to be within the range of 1 x 10 4 and 1 x 10 Again, it is important to note that these
benchmarks are not bright lines and EPA has discretion to find unreasonable risks based on other risk-
related considerations based on analysis. Exposure-related considerations (e.g., duration, magnitude,
population exposed) can affect EPA's estimates of the excess lifetime cancer risk."
The following paragraph in Section 6.1.2 of the draft risk evaluation of TCEP has been edited:
53
-------
"The health risk estimates for workers, ONUs, consumers, the general population, and infants
through the milk pathway are presented in Section 5.3.2. For consumer and general population
exposures, risk estimates are provided in Sections 5.3.2.2 and 5.3.2.3 of this draft risk
evaluation only when margins of exposure (MOEs) were smaller than benchmark MOEs for
non-cancer effects or when cancer risks exceeded benchmark risk levels of 1 in 1,000,000
(1 x 10 "). A complete list of health risk estimates for consumers and the general population is
in the following supplemental files of the draft risk evaluation (see also Appendix C): Draft
Risk Evaluation for Tris(2-chloroethyl) Phosphate (TCEP) - Supplemental Information File:
E-FASTModeling Results, Draft Risk Evaluation for Tris(2-chloroethyl) Phosphate (TCEP) -
Supplemental Information File: Exposure Air Concentration Risk Calculations, and Draft Risk
Evaluation for Tris(2-chloroethyl) Phosphate (TCEP) - Supplemental Information File: TCEP
Consumer Modeling Results, Risk Calculations and Sensitivity Analysis "
The revised paragraph reads:
"The health risk estimates for workers, ONUs, consumers, the general population, and infants
through the milk pathway are presented in Section 5.3.2. For occupational, consumer, and
general population exposures, risk estimates are provided in Sections 5.3.2.1, 5.3.2.2, and
5.3.2.3 of this draft risk evaluation only when margins of exposure (MOEs) were smaller than
benchmark MOEs for non-cancer effects or when cancer risks exceeded benchmark risk levels.
Standard cancer benchmarks used by EPA and other regulatory agencies are an increased
cancer risk above benchmarks ranging from 1 in 1,000,000 to 1 in 10,000 (i.e., 1/10 6 to
1 x 10 4) depending on the subpopulation exposed. In this draft risk evaluation, EPA used
1 x 10~4 as the benchmark for the cancer risk to individuals in industrial and commercial
workplaces, and for consumer and general population exposures EPA used a benchmark of 1 in
1,000,000 (1 x 10 ") for cancer risk. A complete list of health risk estimates for workers,
consumers, and the general population is in the following supplemental files of the draft risk
evaluation (see also Appendix C): Draft Risk Evaluation for Tris(2-chloroethyl) Phosphate
(TCEP) - Supplemental Information File: E-FAST Modeling Results, Draft Risk Evaluation for
Tris(2-chloroethyl) Phosphate (TCEP) - Supplemental Information File: Exposure Air
Concentration Risk Calculations, and Draft Risk Evaluation for Tris(2-chloroethyl) Phosphate
(TCEP) - Supplemental Information File: TCEP Consumer Modeling Results, Risk
Calculations and Sensitivity Analysis
The fenceline analysis mentioned by the commenter, Draft TSCA Screening Level Approach for
Assessing Ambient Air and Water Exposures to Fenceline Communities, was designed to serve as a
screening level analysis to identify the highest possible risk, and therefore a 1 in 1,000,000 range was
used. However, the draft analysis was not designed to determine unreasonable risk. In this risk
evaluation of TCEP, EPA is considering the range of 1 x 10~4 and lxl() " for consumers and the general
population and brings other risk-related considerations that can affect the excess lifetime cancer risk for
consumers and general population. Based on such risk characterizations, EPA is determining how the
COUs contribute to the unreasonable risk of TCEP.
Section 6.2 - Unreasonable risk to the environment
Summary: A public commenter (0040) opined that EPA did not use the best available science in
determining whether TCEP poses unreasonable risk to the environment. They stated that EPA's use of
a risk quotient is more appropriate for a screening-level risk assessment but is overly conservative for a
complete risk determination.
54
-------
EPA Response: For each occupational exposure scenario, where monitoring data were not available,
daily releases were estimated per media of release based on EPA Standard Models, Generic Scenarios
(GSs), and/or Emission Scenario Documents (ESDs) to generate annual releases and for the estimation
of associated release days. The representation of these releases were coupled with flow conditions
comprising TCEP SIC codes with 50th percentile and 90th percentile 7Q10 flows detailed within
Appendix H2.
Concentrations of TCEP in surface water and pore water were modeled based on environmental
releases to determine TCEP concentrations for these relevant aquatic media. Like the first 10 chemicals
to undergo risk evaluation under the amended Toxic Substances Control Act, RQs were presented in
addition to the metric of the number of days of exceedance denoting the duration of the chemical
exceeding the COC within flowing receiving water. The role of days of exceedance accompanies each
aquatic based RQ value. The days of exceedance must also be greater than the experimental exposure
duration associated within the hazard, in this case 30 days.
Understanding the concern from the public commenter, within the draft TCEP RE, the Environmental
Risk Characterization section has included tables that represent RQs and days of exceedance for
surface water and pore water at both production volume (2,500 lbs/year and 25,000 lbs/year) under
high-end and central tendency scenarios. Additionally, the Environmental Risk Characterization within
the final TCEP RE provides combinations of production inputs coupled with 7Q10 flow values to
represent both the 50th percentile and 90th percentile of TCEP SIC codes. The representation of
different production values, releases, and flows provides perspective on modeled concentrations
compared to aquatic COCs and the duration of those concentrations.
EPA characterized the environmental risk of TCEP using RQs. As described in EPA's Guidelines for
Ecological Assessment, risk quotients are commonly used for chemical stressors, where reference or
benchmark toxicity values are widely available. Also, as explained in Section 6.2.3 Basis for EPA's
Determination of Unreasonable Risk of Injury to the Environment, the RQ is not treated as a bright-
line, and other risk-based factors may be considered, including confidence in the hazard and exposure
characterization, duration, magnitude and uncertainty, for purposes of making an unreasonable risk
determination.
EPA has moderate confidence in the chronic aquatic hazards and aquatic exposures contributing to
unreasonable risk. Additionally, the Agency has moderate to robust confidence in the terrestrial
exposures and hazards, which do not contribute to unreasonable risk. Because exposure via soil and the
terrestrial food web was determined to be the driver of exposure, EPA does not expect exposure to
TCEP via air or surface water to contribute to unreasonable risk to terrestrial organisms. Similarly,
EPA does not expect exposure to TCEP via biosolids to contribute to unreasonable risk to the
environment. The Agency's overall environmental risk characterization confidence levels were varied
and are summarized in Table 4-23.
Section 6.3 - Other comments
Summary: A public commenter (0043) stated that the Lautenberg Chemical Safety Act (LCSA)
specifically prescribes assessment requirements for "articles and replacement parts", which, according
to the commenter, requires that EPA consider these conditions of use separately from the consideration
of the chemical as a whole. The commenter stated that the LCSA includes this provision for articles
and replacement parts to account for the lower exposure risk presented by these items compared to
direct chemical exposure. The commenter stated that members of the automotive industry are waiting
55
-------
on EPA to clarify how it will approach these assessment requirements. Once the determination is made,
the industry can begin an effort to address conditions of use of concern.
Referencing the discussion in the draft risk evaluation regarding aircraft and aerospace articles, the
commenter stated that automotive articles share similar limited exposure profile. The commenter
recommended that EPA include manufacture of automobiles and automobile replacement parts in the
no unreasonable risk determination.
EPA Response: EPA evaluated automotive articles and replacement parts in the final risk evaluation
of TCEP. The changes in the final risk evaluation are in the following conditions of use:
"Processing - incorporation into article - aerospace equipment and products" was edited to
"Processing - incorporation into article - aerospace equipment and products and automotive
articles and replacement parts containing TCEP;"
"Industrial use - other use - aerospace equipment and products" was edited to "Industrial use -
other use - aerospace equipment and products and automotive articles and replacement parts
containing TCEP;"
"Commercial use - other use - aerospace equipment and products" was edited to "Commercial
use - other use - aerospace equipment and products and automotive articles and replacement
parts containing TCEP;" and
"Consumer use - paints and coatings - paints and coatings" was edited to "Consumer use -
paints and coatings - paints and coatings, including those found on automotive articles and
replacement parts."
In addition to these changes, a new condition of use, "Industrial use - paints and coatings - automotive
paints and coatings" was assessed in the final risk evaluation of TCEP. These conditions of use were
assessed qualitatively and similarly to how aerospace equipment and products and commercial use of
paints and coatings were evaluated. An overview of the risk characterization approach for the
aerospace equipment and products COUs can be found in Section 5 of the draft risk evaluation, Risk
Characterization.
The risk determination of the final risk evaluation indicates how these COUs contribute (or not) to the
unreasonable risk of TCEP.
Summary: Citing the draft risk evaluation's findings for cancer and non-cancer risk from both
occupational and general public exposures, a public commenter (0044) expressed support for EPA's
determination that TCEP presents unreasonable risk to human health and the environment. The
commenter also stated that the draft risk evaluation contains several important improvements from
EPA's prior TSCA risk evaluations:
EPA made a risk determination for TCEP as a whole chemical;
EPA calculated the aggregate risks to consumers who are exposed to TCEP from multiple
exposure routes;
EPA evaluated TCEP's risk to fenceline communities;
And EPA correctly identified Tribal populations as PESS.
EPA Response: EPA appreciates the supporting comment.
56
-------
Section 7 - Systematic Review
Summary: In response to Charge Question 2.3.1, public commenters (0040. 0045) requested that EPA
address how the SACC's recommendations on evidence integration will be adopted, as it impacts the
derived toxicity values for risk evaluations, and suggested these recommendations be incorporated via
finalizing the Draft Systematic Review Protocol.
EPA Response: EPA relied on the Draft 2021 Systematic Review Protocol {U.S. EPA, 2021,
10415760} for guidance and will continue to review and update the TSCA systematic review processes
for future chemicals.
Summary: Two public commenters (0042. 0045) suggested that EPA fully incorporate
recommendations provided by organizations such as the National Academies of Sciences, Engineering,
and Medicine and the SACC, including disregarding numeric quality scores and changes to the study
quality evaluation metrics, on the 2021 Draft Systematic Review Protocol to reflect TSCA's "best
available science" requirement.
EPA Response: During development of the systematic review process overall, EPA consulted many
existing systematic review frameworks and experts and has relied on best available science throughout
the process. One primary goal was to develop data quality criteria that were consistent across
disciplines as much as possible. For example, among the hazard disciplines (animal toxicity,
ecotoxicity, epidemiology) EPA updated criteria in the 2021 Draft Systematic Review Protocol {U.S.
EPA, 2021, 10415760} to increase consistency, which is of primary importance because some studies
are used for both disciplines. Thus, EPA necessarily drew from frameworks for both disciplines rather
than relying only on frameworks (e.g., IRIS) for the human health hazard discipline.
EPA implemented several recommendations from NASEM when preparing the 2021 Draft Systematic
Review Protocol {U.S. EPA, 2021, 10415760}.
For future chemicals, EPA is implementing a domain-based set of criteria for human health animal
toxicity studies for TSCA risk evaluations that further responds to peer-review comments. Thus, EPA
under TSCA continues to use best available science in updates to the systematic review process. Until
there is an update to the 2021 Draft Systematic Review Protocol {U.S. EPA, 2021, 10415760} under
TSCA, EPA will present any updates to the systematic review processes in the individual chemical
protocol documents that accompany the risk evaluations.
Summary: A public commenter (0042) stated that EPA references inconsistent population, exposure,
comparator, and outcome (PECO) statements to identify relevant health effects studies that may
inappropriately exclude no-apical effects such as cellular-level outcomes and recommended revising
the PECO statement used in the TCEP risk evaluation and include that PECO statement in the
chemical-specific systematic review protocol. The commenter further requested that EPA release a new
TSCA systematic review methodology document that states how consistent practices will be applied to
all TSCA rulemakings and for EPA to prepare a chemical-specific review protocol for each TSCA risk
evaluation it conducts that does not reference the 2021 Draft Protocol for critical elements such as
PECO statements.
EPA Response: EPA has considered all relevant health effects from studies identified as supplemental
(including cellular level effects) when conducting evidence integration for TCEP.
57
-------
For future chemicals, EPA's PECO statements will identify all in vivo animal toxicity studies as
meeting PECO so that studies with less than organ level responses can be immediately identified and
evaluated if they are relevant for further consideration in the risk evaluation.
Section 8 - Formatting and Editing
Summary: A public commenter (0037) stated that EPA misattributed TCEP activity to Aceto US,
L.L.C. According to the commenter, Aceto Corporation produced TCEP but declared bankruptcy and
ceased operations in 2019. Aceto US, L.L.C is a separate legal entity, but EPA incorrectly attributed
activity by Aceto Corporation to Aceto US, L.L.C. in three instances on page 24 and once on page 87
of the draft risk evaluation. The commenter requested that EPA correct these misattributions.
EPA Response: EPA has edited the draft risk evaluation of TCEP to correctly attribute activity to
Aceto Corporation rather than Aceto US LLC. The following edits have been made to Section 1.1.1,
Life Cycle and Production Volume, of the final risk evaluation:
"The production volumes for TCEP reported to CDR for years 2012 to 2015 were all from one
company, Aceto US LLC, a chemical manufacturer and supplier importing TCEP in chemical
form. Aceto US LLC indicated to EPA that TCEP was imported and used as a flame retardant
for unsaturated polyester resins and for aircraft furniture {U.S. EPA, 2020, 10617335}" was
edited to "The production volumes for TCEP reported to CDR for years 2012 to 2015 were all
from one company, Aceto Corporation, a chemical manufacturer and supplier importing TCEP
in chemical form. Aceto Corporation indicated to EPA that TCEP was imported and used as a
flame retardant for unsaturated polyester resins and for aircraft furniture {U.S. EPA, 2020,
10617335};"
"Note that for 2014, the Aceto US LLC data is included in the total production volume for
CDR and Datamyne" was edited to "Note that for 2014, the Aceto Corporation data is included
in the total production volume for CDR and Datamyne;" and
"For 2014, Aceto US LLC's production volume is included in both the CDR data and the
Datamyne data" was edited to "For 2014, Aceto Corporation's production volume is included
in both the CDR data and the Datamyne data."
EPA found no other instances of misattributing activity to Aceto US LLC throughout the draft risk
evaluation for TCEP.
Summary: A peer reviewer (OC - R1) wrote that, based on its structure in the Agency for Toxic
Substances and Disease Registry (ATSDR) toxicological profile and in Burka et al. (1991), the
abbreviation BCGP should refer to the glucuronide conjugate of bis(2-chloroethyl) 2-hydroxyethyl
phosphate rather than bis(2-chloroethyl) 2-hydroxyethyl phosphate itself, as indicated on pages 249,
402, and 525. The peer reviewer stated that the text should be corrected.
EPA Response: EPA revised the names in these three sections of the RE.
Summary: A peer reviewer (OC - Rl) stated that the statement "were identified as hydrogen
phosphate (BCHP)" in line 13697 on page 524 should read "were identified as bis(2-chloroethyl)
hydrogen phosphate (BCHP)." In addition, the peer reviewer recommended adding "and alcohol
dehydrogenase pathways" after "glucuronidation" in line 13694 on page 524.
EPA Response: EPA has made this revision.
58
-------
Summary: A peer reviewer (OC - R1) wrote that the HERO link for "Chen, G; Jin, Y; Wu, Y; Liu, L;
Fu, Z. (2015a). Exposure of male mice to two kinds of organophosphate flame retardants (OPFRs)
induced oxidative stress and endocrine disruption. Environ Toxicol. Pharmacol. 40: 310-318" in the
document leads to a different citation.
EPA Response: EPA thanks the commenter for the comment. The hyper link has been corrected.
Summary: A peer reviewer (OC - R1) commented that the metabolite bis(2-chloroethyl) phosphate
(BCEP) should be included in the list of abbreviations. The peer reviewer added that it should be
prominently noted in both the metabolism section and elsewhere that BCEP and BCHP are names for
the same metabolite.
EPA Response: EPA added BCEP to the list of abbreviations in Appendix A. Section 5.2.2 and
Appendix J. 1.3. now state that BCHP and BCEP are synonyms of each other.
Summary: A peer reviewer (OC - R1) stated that Table 560 should have a footnote explaining what
Drinking Water (Diluted) means.
EPA Response: EPA thanks the commenter for the comment. A footnote describing the diluted
estimates has been added to Table 5-40 and Table 5-41 of the final TCEP Risk Evaluation: "A method
was derived to incorporate a dilution factor to estimate TCEP concentrations at drinking water
locations downstream from surface water release points. Since no location information was available
for facilities releasing TCEP, a dilution factor and distances to drinking water intake was estimated for
each relevant SIC code. Table 5-26 provides the 50th quantile distances and 50th quantile dilution
factors for the relevant SIC codes."
Summary: A peer reviewer (OC - Rl) expressed that it would be helpful to the reader if the MOEs
that were less than the target MOE were shown in bold so that they can be easily identified. In addition,
the peer reviewer stated that it would be helpful if cancer risks that exceed the targeted risk value were
bolded or marked in another easily identifiable manner.
EPA Response: EPA does not bold these low MOEs or high cancer risk values because they may not
be considered unreasonable risks. The reader should focus on risks included in Chapter 6, where the
unreasonable risk determinations are presented.
Section 9 - Other Comments on the Draft Risk Evaluation
Existing chemical exposure limit (ECEL)
Summary: Two public commenters (0038. 0040) expressed concern regarding the ECEL proposed for
TCEP and stated that an inhalation exposure limit is unnecessary for TCEP as dermal exposures are the
predominant exposure risk.
Another public commenter (0044) stated that establishing an ECEL during the risk evaluation phase is
premature and could be viewed as "prejudging" risk management regulations. The commenter
suggested that the proposed ECEL is not protective enough, failed to consider multiple exposure routes
59
-------
for workers, and is inconsistent with TSCA's mandate to regulate to the extent necessary to eliminate
unreasonable risk.
EPA Response: EPA routinely establishes occupational exposure values to accompany the TSCA risk
evaluations as recommended in comments on previous TSCA risk evaluations. These values identify
levels EPA considers to be without appreciable risk and are useful when EPA considers options for
managing risks from TCEP. Although the commenter is correct that dermal exposures are predominant,
EPA must consider risk via all relevant exposure routes. The establishment of inhalation values is
standard practice and has been implemented by many United States agencies and organizations for use
in managing and/or communicating inhalation risks.
EPA realizes that the draft appendix was confusing and has changed the title of Appendix N to
Occupational Exposure Value Derivation and Analytical Methods Used to Detect TCEP. The
value in this appendix is calculated based on the POD from the risk evaluation and exposure
assumptions for workers and is not meant to be a regulatory limit. As described in the updated
appendix: "Any existing chemical exposure limit (ECEL) used for occupational safety risk
management purposes could differ from the occupational exposure value presented in this appendix
based on additional consideration of exposures and non-risk factors consistent with TSCA section
6(c)."
The calculated occupational exposure value for TCEP represents the inhalation exposure concentration
below which workers and occupational non-users are not expected to exhibit any appreciable risk of
adverse toxicological outcomes, accounting for PESS. Although dermal exposures contribute to the
majority of risk to workers, including occupational non users, EPA determined worker inhalation
exposure from one COU (commercial use of paints and coatings) contributes to the unreasonable risk to
TCEP, and so EPA derived an occupational exposure value for TCEP. It is derived based on the most
sensitive human health effect (i.e., cancer) relative to benchmarks and standard occupational scenario
assumptions of 8 hours per day, 5 days per week exposures for a total of 250 days exposure per year,
and a 40-year working life. EPA expects that at the occupational exposure value of 0.008 ppm (0.09
mg/m3), a worker or occupational non-user also would be protected against neurotoxicity from acute
occupational exposure as well as male reproductive effects from short-term and chronic occupational
exposures if ambient exposures are kept below this draft occupational exposure value. EPA has not
separately calculated a draft short-term (i.e.. 15-minute) occupational exposure value because EPA did
not identify hazards for TCEP associated with this very short duration.
If EPA chooses to set an ECEL to manage risks from TCEP, the feasibility of compliance will be
considered when setting the final ECEL during risk management. Any ECEL used for occupational
safety risk management purposes could differ from the occupational exposure value presented in the
risk evaluation based on additional consideration of exposures and non-risk factors, consistent with
TSCA section 6(c).
De minimis
Summary: A public commenter (0043) recommended that EPA identify and establish a de minimis
level for TCEP and other TSCA chemicals below which EPA has no reasonable basis to conclude that
the chemicals pose a risk. The commenter also requested that EPA issue a TSCA Section 6(i)(l) order
for conditions of use that have been determined to pose no unreasonable risk once the final risk
evaluation is released.
60
-------
EPA Response: As explained in the Procedures for Chemical Risk Evaluation Under the TSCA final
rule (89 FR 37028. May 3. 2024). EPA will make a single determination as to whether the chemical
substance presents an unreasonable risk of injury to health or the environment under the conditions of
use. In the final risk evaluation for TCEP, EPA is concluding that TCEP presents unreasonable risk as a
chemical substance. Pursuant to 15 U.S.C. 2605(a) EPA will propose risk management actions to
address the unreasonable risk of TCEP. Since EPA is not concluding that TCEP presents no
unreasonable risk, EPA will not be issuing a section 6(i)(l) order. However, in the final determination
of unreasonable risk, EPA has identified conditions of use that do not significantly contribute to the
unreasonable risk of TCEP.
EPA has included additional narrative based on weight fractions in Sections 5.3.5.1 and 5.3.5.2.
Public access
Summary: A public commenter (0044) requested that EPA prepare a non-technical summary
document of the draft risk evaluation in multiple languages that follows Interstate Technology and
Regulatory Council risk communication principles to provide for meaningful outreach, public access,
and process transparency.
EPA Response: EPA publishes a non-technical summary with the final TCEP Risk Evaluation with
information about what TCEP is, how it is used, how you might be exposed to it, etc. in plain language.
EPA will consider translating this document as the need arises.
Registration evaluation
Summary: A public commenter (0033) suggested that EPA list the other countries that have completed
a registration evaluation under the principle of equivalence under the Technical Barriers to Trade
Agreement to avoid repeated evaluations.
EPA Response: EPA used reasonably available information, defined in 40 CFR 702.33, in a fit-for-
purpose approach, to develop a risk evaluation that relies on the best available science and is based on
the weight of the scientific evidence in accordance with TSCA sections 6 and 26. EPA reviewed
reasonably available information and evaluated the quality of the methods and reporting of results of
the individual studies using the evaluation strategies described in the 2021 Draft Systematic Review
Protocol (U.S. EPA, 2021) and the TCEP Systematic Review Protocol (U.S. EPA, 2023n).
EPA also identified key assessments conducted by other EPA programs and other U.S. and
international organizations. Depending on the source, these assessments may include information on
COUs (or the equivalent), hazards, exposures, and potentially exposed or susceptible subpopulations.
Some of the most pertinent assessments that were consulted for TCEP include the following:
U.S. EPA's 2009 Provisional Peer-Reviewed Toxicity Values (PPRTV) for Tris(2-
chloroethyl)phosphate (TCEP) (CASRN115-96-8);
2009 European Union Risk Assessment Report: CAS: 115-96-8: Tris (2-chloroethyl)
phosphate, TCEP;
Environment Canada and Health Canada's 2009 Screening Assessment for the Challenge
Ethanol, 2-chloro-, phosphate (3:1) (Tris(2-chlrorethyl) phosphate [TCEP]);
Australia's 2016 Ethanol 2-chloro-, phosphate (3:1): Human health tier II assessment;
Australia's 2017 Ethanol, 2-chloro-, phosphate (3:1): Human health tier III assessment;
ATSDR's 2012 Toxicological Profile for Phosphate Ester Flame Retardants;
61
-------
NTP's 1991 Technical Report on Toxicology and Carcinogenesis Studies ofTris(2-
chloroethyl) Phosphate (CASRN115-96-8) in F344/N Rats and B6C3F1 Mice (Gavage
Studies); and
I ARC's 1999 Monographs on the Evaluation of Carcinogenic Risks to Humans Volume 71.
Literature review
Summary: A peer reviewer (OC - Rl) noted that the literature review on TCEP related to human
health effects took place in 2019, and the peer reviewer was unsure when the literature review occurred
for other topics. The peer reviewer stated that updates should be performed to ensure that important
new studies have not been missed. A list of post-2019 literature was provided by the reviewer,
organized into three categories: neurotoxicity, developmental toxicity, and kidney toxicity.
EPA Response: EPA updated the literature search from 2019 to target animal toxicity studies
conducted via the inhalation route as well as epidemiological studies of any human health effects
(inclusive of inhalation, dermal, and oral routes). No animal toxicity studies were found. EPA did
identify several epidemiological studies that were screened and then evaluated for data quality along
with additional studies identified by peer reviewers. EPA then incorporated the epidemiological studies
into the TCEP RE. In Section 5.2.3.1, EPA added epidemiology studies on neurological, kidney,
immune/hematological, thyroid, lung, and developmental effects, and in Section 5.2.5.1, EPA added
epidemiological studies on cancer.
62
------- |