&EHV
United States Office of Chemical Safety and Pollution Prevention
Environmental Protection Agency December 2020
Summary of External Peer Review and Public Comments and
Disposition for n-Methylpyrrolidone (NMP)
Response to Support Risk Evaluation of
n-Methylpyrrolidone (NMP)
December 2020
-------
Table of Contents
1. List of Comments 4
2. Environmental Fate and Exposure 6
3. Environmental Hazard and Risk Characterization 20
4. Occupational and Consumer Exposure 31
5. Human Health Effects 92
6. Dose-Response Assessments 110
7. Risk Characterization 137
8. Content and Organization 175
-------
This document summarizes the public and external peer review comments that EPA's Office of
Pollution Prevention and Toxics (OPPT) received for the risk evaluation of n-methyl pyrrol i done
(NMP). It also provides EPA's response to the comments received from the public and the peer
review panel.
EPA appreciates the valuable input provided by the public and peer review panel. The input
resulted in numerous revisions to the hazard summary.
Peer review charge questions1 are used to categorize the peer review and public comments into
specific issues related to the main themes.
1. Environmental Exposure Assessment, Including Environmental Fate and Transport and
Environmental Release Assessment
2. Ecological Exposure, Hazard Assessment, and Risk Characterization
3. Occupational and Consumer Exposure Assessment
4. Human Health Hazard
5. Human Health Dose-response Assessment
6. Risk Characterization
7. Content, Organization and Clarity of the Document
All peer review comments for the seven charge questions are presented first, organized by charge
question in the following section. These are followed by the public comments. For each theme,
general comments pertaining to all chemicals are presented first, and then additional comments
pertaining to only one or several chemicals follows.
1 These are the questions that EPA submitted to the panel to guide the peer review process.
Page 3 of205
-------
1. List of Comments
#
Docket I'ilc
Sn hm iller
31
EP A-HO-OPPT-2019-023 6-0031
David Isaacs, Semiconductor Industry Association (SIA)
32
EP A-HG-GPPT-2019-02:
Suzanne Hartigan, Senior Director, Regulatory and Technical Affairs, American
Chemistry Council (ACC)
33
EP A-HO-OPPT-2019-023 6-003 3
Kathleen M. Roberts, NMP Producers Group, Inc.
34
EP A-HO-OPPT-2019-023 6-0034
Jonathan Kalmuss-Katz, Staff Attorney, Earthjustice (12-04-2019)
37
EP A-HO-OPPT-2019-02:
Sharon Shindel, Corporate Industrial Hygienist, Intel Corporation
38
EP A-HO-OPPT-2019-023 6-003 8
Jennifer Sass, Senior Scientist, Natural Resources Defense Council (NRDC)
39
EP A-HO-OPPT-2019-02:
Weihsueh A. Chiu, Professor, Veterinary Integrative Biosciences, Texas A&M
University
40
EP A-HO-OPPT-2019-023 6-0040
Veena Singla, Associate Director, Program on Reproductive Health and the
Environment, School of Medicine, University of California, San Francisco
42
EP A-HO-OPPT-2019-023 6-0042
Eric Berg, Deputy Chief, California Division of Occupational Safety and Health
(Cal/OSHA)
44
EP A-HO-OPPT-20
19-0236-0044
Anonymous
45
19-02
Sheryl Beauvais, Senior Compliance Analyst, Hach Company
46
EP A-HO-OPPT-20
19-0236-0046
Attorneys General of New York, Illinois, Maine, Maryland, Massachusetts,
Minnesota, New Jersey, Oregon, Vermont, and Washington
47
EP A-HO-OPPT-2019-023 6-0047
Mark Kohorst, Director, Environment Health & Safety, National Electrical
Manufacturers Association (NEMA)
48
EP A-HO-OPPT-2019-023 6-0048
Swati Rayasam, Science Associate, Program on Reproductive Health and the
Environment, University of California, San Francisco
49
EP A-HO-OPPT-2019-023 6-0049
Aaron Rice, Environmental Health & Safety Director, EaglePicher Technologies,
LLC
50
EP A-HO-OPPT-2019-023 6-0050
Janet M. Carlock, EHS Regulatory Manager, FUJIFILM Holdings America
Corporation
51
EP A-HO-OPPT-2019-023 6-0051
Safer Chemicals Healthy Families, Environmental Health Strategy Center,
Natural Resources Defense Council, and Earthjustice
52
EP A-HO-OPPT-2019-023 6-0052
David Isaacs, SIA
Page 4 of205
-------
53
EP A-HO-OPPT-2019-023 6-0053
Riaz Zaman, Counsel, Government Affairs, American Coatings Association
(ACA)
54
EP A-HO-OPPT-2019-023 6-0054
Suzanne Hartigan, Senior Director, Regulatory and Technical Affairs, ACC
55
EP A-HO-OPPT-2019-02:
Dianne C. Barton, Chair, National Tribal Toxics Council (NTTC)
56
EP A-HO-OPPT-2019-023 6-0056
Martha Marrapese, Wiley Rein LLP on behalf of the Lithium Cell Manufacturers'
Coalition
57
EPA-HO-OPPT-2019-023 6-005 7
Kathleen M. Roberts, Manager, NMP Producers Group
58
EPA-HO-OPPT-2019-023 6-005 8
Kathleen M. Roberts, Manager, NMP Producers Group (Attachments)
59
EP A-HO-OPPT-2019-02:
Richard A. Denison, Lead Senior Scientist, Environmental Defense Fund (EDF)
60
EP A-HO-OPPT-2019-023 6-0060
Hesham M. Soliman, Senior Product Steward, Global Chemical Control,
Lyondell Chemical Company
61
EP A-HO-OPPT-2019-023 6-0061
Earthjustice and the Occupational Safety & Health Law Project on behalf of the
American Federation of Labor and Congress of Industrial Organizations (AFL-
CIO); International Union, United Automobile, Aerospace, and Agricultural
Implement Workers of America (UAW); North America's Building Trades
Unions (NABTU); and United Steel, Paper and Forestry, Rubber, Manufacturing,
Energy, Allied Industrial and Service Workers International Union (United
Steelworkers)
62
EP A-HO-OPPT-2019-023 6-0062
Earthjustice and the Occupational Safety & Health Law Project on behalf of the
American Federation of Labor and Congress of Industrial Organizations (AFL-
CIO); International Union, United Automobile, Aerospace, and Agricultural
Implement Workers of America (UAW); North America's Building Trades
Unions (NABTU); and United Steel, Paper and Forestry, Rubber, Manufacturing,
Energy, Allied Industrial and Service Workers International Union (United
Steelworkers) (Exhibits)
63
EP A-HO-OPPT-2019-023 6-0063
Lawrence E. Culleen, Arnold & Porter on behalf of Chemical Users Coalition
(CUC)
64
EP A-HO-OPPT-2019-023 6-0064
John Currier, Corporate EHS TSCA Program Manager, Intel Corporation
Page 5 of205
-------
2. Environmental Fate and Exposure
Environmental Eate and Exposure
Charge Question I.I: Please comment mi lhe data. approaches and or methods used to characterize exposure to aquatic receptors
#
Summary of Comments for Specific Issues Related
to Charge Question 1
EPA Response
Scope ol
'fate and exposure assessment - Other regulatory programs
SACC
SACC COMMENTS:
Recommendation: Provide a summary of the focus,
status, and results of NMP assessments completed
or progressing under other EPA regulatory
programs.
The evaluation must clearly state which
environmental releases are covered by other
regulations and are not associated with TSCA.
EPA provides a summary of assessments that have been
previously completed for NMP in Table 1-5 and a summary of the
regulatory history of NMP in Appendix A. EPA has added
Section 1.4.2, which describes exposure pathways and risks that
fall under the jurisdiction of other EPA administered statutes or
regulatory programs. As described in Section 1.4.2, EPA believes
it is both reasonable and prudent to tailor TSCA risk evaluations
when other EPA offices have expertise and experience to address
specific environmental media, rather than attempt to evaluate and
regulate potential exposures and risks from those media under
TSCA. EPA believes that coordinated action on exposure
pathways and risks addressed by other EPA-administered statutes
and regulatory programs is consistent with statutory text and
legislative history, particularly as they pertain to TSCA's function
as a "gap-filling" statute, and also furthers EPA aims to
efficiently use Agency resources, avoid duplicating efforts taken
pursuant to other Agency programs, and meet the statutory
deadline for completing risk evaluations. EPA has therefore
tailored the scope of the risk evaluation for NMP using authorities
in TSCA Sections 6(b) and 9(b)(1). Pathways that are within the
scope of the risk evaluation are described in Section 1.4.3.
Scope of fate and exposure assessment- Exposures/receptors not assessed
SACC
SACC COMMENTS:
Recommendation: Determine potential NMP
exposures to threatened and endangered species
and to honeybees.
The TSCA risk evaluation focuses on exposures for environmental
receptors associated with conditions of use for the NMP. The
assessment focuses on environmental receptors that may be
exposed to NMP as a result of the conditions of use and associated
Page 6 of205
-------
NMP has demonstrated toxicity to honeybees (Fine
and Mullin, 2017).
No assessment of exposures to threatened or
endangered species is provided in the draft risk
evaluation. EFAST has a feature that allows
"searching for endangered species in the vicinity of
specific facilities."
hazards to those affected species.
SACC
SACC COMMENTS:
Recommendation: Discuss the evidence for uptake
of NMP by terrestrial plants and the potential for
trophic transfer to herbivores.
There is evidence that terrestrial plants take up
NMP (Doucette et al., 2018; Dettenmaier et al.,
2009), and hence uptake and exposure to terrestrial
plants should be assessed, particularly as part of a
bioconcentration process (NMP Risk Evaluation, p.
48). Such uptake would provide a route for trophic
transfer of NMP to herbivores. Additional data are
needed to better understand these risks and to
determine if plant uptake poses unreasonable risks
to herbivores.
During problem formulation, EPA identified several pathways
(including sediment, ambient water, land-applied biosolids, and
ambient air) that did not require further analysis because
environmental fate properties and first-tier analysis of
environmental release data indicated that exposures were well
below levels of concern. Terrestrial environmental receptors were
therefore not further evaluated. EPA's conclusions about risks
from exposure through ambient air, ambient water, sediment, and
land-applied biosolids are summarized in the Problem
Formulation and in Section 4.6.2.3 of the risk evaluation.
Include a mass balance analysis
SACC
SACC COMMENTS:
Recommendation: Provide an analysis matching
annual imports and manufactured NMP amounts to
NMP releases and amounts used in
products/processes (a mass balance analysis),
incorporate analysis findings into the life cycle
diagram, and discuss how these findings impact the
estimated water releases used in this draft risk
evaluation..
The life cycle discussion should be expanded to
paint a more complete picture on sources of NMP
EPA developed an approach and conducted a mass balance
analysis for NMP. The analysis, which accounted for 83% of the
NMP production volume, is summarized in Section 1.4.1 and
details are provided in Appendix C.
Page 7 of205
-------
emissions, distribution, and sinks useful for
providing a mass balance analysis for all NMP
produced and imported. Additional information
about NMP could be added to Fig. 1-1 to increase
utility and transparency (e.g., some uses are
missing, as noted above; the disposal box is
uninformative).
Clarity/rationale needed
SACC
SACC COMMENTS:
The statement that discharges to air, water,
sediment, land, and biosolids were all evaluated is
misleading (NMP Risk Evaluation, p. 56). Water is
the only ambient media considered in the
environmental assessments in the NMP risk
evaluation.
Consideration of environmental releases to water
alone provides an inadequate picture of risk from
NMP, as 99% of all environmental releases for
NMP remain unassessed. Ten million pounds of
chemical are unaccounted for, an amount
representing over 70 times the releases modeled in
the environmental exposures section. At a
minimum, the evaluation should discuss, and
preferably assess, all releases to water, soils, and
the ambient atmosphere.
The decision to not further analyze aquatic
exposures, based on a preliminary analysis that
water releases were not expected to exceed
concentrations of concern, limits the completeness
of the risk evaluation.
Information on why individual media were
dismissed from further evaluation is not provided.
The rationale and the process used in deciding to
As described above, EPA has tailored the scope of the risk
evaluation for NMP using authorities in TSCA Sections 6(b) and
9(b)(1). Section 1.4.2 of the final risk evaluation describes
exposure pathways and risks that fall under the jurisdiction of
other EPA-administered statutes or regulatory programs. The
rationale for the selection of environmental compartments for
assessment is also presented in the Problem Formulation
document, as referenced in Section 1.4 of the risk evaluation.
During problem formulation, EPA identified several pathways
(including sediment, ambient water, land-applied biosolids, and
ambient air) that did not require further analysis because
environmental fate properties and first-tier analysis of
environmental release data indicate that exposures are well below
levels of concern. The statement cited in the first bullet point of
the comment specifies that during problem formulation air, water,
sediment, and biosolids were analyzed. As described in the
problem formulation document, the air pathway was not further
analyzed because inhalation exposure and bioaccumulation
potential are expected to be low, and the solids pathways were not
further analyzed because NMP is not expected to adsorb to
suspended solids or sediment due to its high water solubility and
estimated soil organic carbon/water partition coefficient (log Koc
= 0.9). EPA's conclusions about ambient air, ambient water,
sediment, and land-applied biosolids are summarized in the
Page 8 of205
-------
exclude non-aqueous media from the
environmental assessment should be further
discussed in the risk evaluation.
Problem Formulation and in Section 4.6.2.3 of the risk evaluation.
In the risk evaluation, EPA updated the screening level analysis
for the ambient water pathway using surface water concentrations
modeled based on 2018 TRI data. EPA evaluated potential risks
for aquatic species by comparing surface water concentrations to
concentrations of concern for aquatic species. EPA performed a
screening level evaluation of potential human health risks by
comparing exposures expected from incidental ingestion of
surface water and dermal contact from swimming to human
health points of departure.
SACC
SACC COMMENTS:
A clearer rationale is needed for the selection of
modeled versus empirically measured data.
All data selections were made according to the Application of
Systematic Review in TSCA Risk Evaluations. Under the
systematic review guidelines, data sources are assigned overall
quality scores based on strict and clearly defined criteria. While
these criteria are different for measured and modeled data
sources, the scoring system is designed to permit direct
comparison between different types of data sources. In some
cases this may lead to greater confidence being given to a model
result (for instance, an output from one of the modules contained
in the EPI Suite package) than to an empirical study whose
design is lacking in one or more important aspects. Please refer to
the Data Evaluation for detailed scores (with justifications) for
individual data sources.
SACC
SACC COMMENTS:
Recommendation: Clarify the issues related to the
assumed percent reduction of NMP discharges to
POTWs in Table 2-1 and Appendix D.
The text regarding NMP transformation in publicly
owned treatment works (POTWs) (NMP Risk
Evaluation, p. 370) needs clarification. Table 2-1
indicates 45% reduction of NMP in POTWs, but in
Appendix D, 92% removal is indicated. Data in
Table D-2 suggest that the 45% value is used to
EPA revised Appendix E (previously Appendix D) describing
POTW releases to surface water so as to clarify that the model
input of 92%) was used to estimate POTW removal of NMP for
indirect dischargers.
Page 9 of205
-------
derive the information in the "PDM; input
Loadings" column. A modification is reported in
the stream concentration column that is difficult to
follow, but could change the assumed use of the
45% value.
SACC
SACC COMMENTS:
Recommendation: Clarify text regarding NMP
transformation in POTWs (p. 370) to explain
differences in NMP degradation that are listed in
Table 2-1, Appendix D, and EPI Suite
calculations.
One Committee member commented on the
difference in the assumed fraction of NMP
removed in wastewater treatment plants (WWTPs)
as compared to assumed NMP removed in POTWs.
It is difficult to follow the discussion in the text on
Table 2-1, Appendix D of the draft risk evaluation
and in the related Estimation Programs Interface
Suite (EPI Suite) calculations. It appears that
the differences are the result of using different
residency times for each type of facility in the EPI
Suite runs.
Data in Table Apx D-2 suggest a median NMP
degradation rate in activated sewage sludge of 17%
in 24 hours and 61% in 120 hours. It seems
reasonable to use the more robust 24-hour
transformation rate estimate of 17% to describe
NMP dissipation from POTWs instead of the 45%
that appears to have been used in calculations. If
this were done, the original hazard quotient (HQ)
would be raised from 0.85 to 1.56 (0.85/0.45*0.87)
and 1.85. Using this value and combining the
effluent from the Oregon facilities, an HQ of 2.1 is
The POTW removal presented in Section 2.1.1 (>90%) was
predicted by EPI Suite and is consistent with the value used for
POTW removal as described in Appendix E (92%). Table 2-1
does not contain a value for POTW removal. It does contain the
results of biodegradation studies, which are conducted under very
different conditions from those present in a POTW and should not
be viewed as equivalent.
EPA updated Appendix E (previously Appendix D) to include
2018 reported data and made revisions to clarify that the NMP
removal efficiency for wastewater treatment plants and for
POTWs is the same at 92%.
EPA was unable to replicate or identify the source of the
degradation rate values cited in the third bullet. They do not
appear in Table Apx D-2.
Page 10 of 205
-------
calculated. Using EPI Suite degradation
estimates, the HQ would be larger.
Water releases and exposures
SACC
SACC COMMENTS:
Recommendation: Provide better estimates of the
amount of NMP released to waters, or alternatively,
consider increasing the estimate of releases to
water to include all, or a substantial part, of the 1.4
million pounds of NMP unaccounted in the
emission estimates.
The Committee had difficulty in tracking emissions
and disposal amounts by media and type of
disposal. Release data presented in Tables 1-3 and
1-4 do not line up. For 2017, Table 1-3 identified
10.4 million pounds of NMP releases, while Table
1-4 listed 1.53 million pounds to air, 7.55 million
pounds to land, and 0.02 million pounds to water,
leaving 1.4 million pounds unaccounted for. The
unaccounted-for fraction is almost 2 orders of
magnitude greater than that reported as discharge to
water. In the absence of monitoring data to the
contrary, with so little of total NMP discharges
reported as releases to water and other media, and
given that all other media are not being considered
in this analysis, several Committee members
concluded that it is reasonable and conservative to
assume the unallocated fraction are water releases.
EPA's water release analysis uses TRI data to estimate the highest
local per site water releases of NMP and is not intended to
estimate overall releases. EPA does not expect any higher per site
local water releases beyond those reported to TRI.
EPA developed an approach and conducted a mass balance
analysis for NMP. The analysis, which accounted for 83% of the
NMP production volume and includes all releases reported to
TRI, is summarized in Section 1.4.1 and details are provided in
Appendix C.
SACC
SACC COMMENTS:
Recommendation: Verify uncertainties in surface water
concentrations (NMP Risk Evaluation, pp. 58, 59, and
370) and assumptions made regarding the fate of NMP
in these surface waters by obtaining surface water
monitoring data from public or private organizations
EPA obtained and considered reasonably available information,
defined as information that EPA possesses, or can reasonably
obtain and synthesize for use in risk evaluations, considering the
deadlines for completing the evaluation. No surface water
monitoring data were available. EPA gathered the amount of NMP
released to surface waters as reported in TRI for 2015 for the draft
Page 11 of 205
-------
through TSCA-provided authority.
risk assessment and then updated to include more recent, 2018
reported TRI data. Appendix E lists the facilities that discharge
NMP directly to surface waters and the top 10-12 facilities that
discharge indirectly, that is, discharge to a wastewater treatment
facility that discharges to surface waters after treatment. EPA
estimated surface water concentrations using E-FAST 2014
model. This model is considered conservative as the estimated
surface water concentrations consider receiving water dilution but
do not include fate processes such as volatilization or
biodegradation.
SACC
SACC COMMENTS:
Recommendation: Provide better justification for
why so little environmental data covering only a
small fraction of environmental releases of NMP
were used in this draft risk evaluation.
The draft risk evaluation was limited to direct
releases that enter surface water, which represent
<0.2% of total releases (NMP Risk Evaluation, p.
369). This leaves >99% of NMP releases
unassessed. The rationale for only assessing 12
discharge sites (NMP Risk Evaluation, p. 59)
should be explicitly provided.
The risk evaluation should identify and discuss
available monitoring data (at least at the point of
discharge) from commercial users and producers.
EPA's water release analysis uses TRI data to estimate the highest
site-specific water releases of NMP and is not intended to
estimate overall releases.
EPA included both direct dischargers and indirect dischargers
(facilities that report transferring wastewater to another treatment
facility such as a POTW) in both the draft and final risk
evaluations. In addition, EPA updated the facility data as reported
to TRI in 2018 to include all of the direct dischargers (8 facilities)
and the top 12 indirect dischargers (representing 87% of total
annual NMP discharges). If more than one facility discharged to a
POTW the influent NMP mass was combined to estimate total
NMP surface water concentrations.
EPA developed an approach and conducted a mass balance
analysis for NMP. The analysis, which accounted for 83% of the
NMP production volume and includes all releases reported to
TRI, is summarized in Section 1.4.1 and details are provided in
Appendix C.
Beyond releases reported to TRI, EPA did not identify any
additional reasonably available environmental release or
monitoring data for NMP.
SACC
SACC COMMENTS:
Recommendation: Consider discharges relative to
stream flow volumes because the largest releases
EPA used the E-FAST 2014 model to estimate NMP surface
water concentration resulting from facility discharges of NMP.
Using the release data from TRI, including facility location, EPA
Page 12 of 205
-------
may not present the largest ratios of discharge to
stream flow.
Data that the Agency references for the NMP
release amounts (NMP Risk Evaluation, Table D-l:
Appendix D) are exclusively from the Toxics
Release Inventory (TRI) database. TRI reports do
not estimate releases as a proportion of stream
flow. Stream flow data are potentially available
from other sources, however. Provide releases per
stream flow for (as many as practical of) the 124
facilities reporting releases. Estimating releases as
a function of stream flow could provide additional
insight into the impact of releases. It is unclear
whether the largest absolute releases are not also
the largest releases relative to stream flow, and
hence also represent the highest exposure profiles
for aquatic organisms. Reporting the discharge
target waters would also be helpful.
could determine for some facilities the receiving waters,
particularly if the facility had an associated National Pollutant
Discharge Elimination System (NPDES) permit identification.
The E-FAST model consists of a database of NPDES facilities
and the corresponding receiving water stream flow data. Thus, the
concentration of NMP in surface water is a function of the
amount released over a given time period (12 days per year or 250
or 300 days per year) and the receiving water 7Q10 stream flow
{i.e., 7 consecutive days of lowest flow over a 10-year period).
Thus, receiving water stream flow as well as the NMP discharge
amount are both important factors EPA considered in estimating
NMP surface water concentrations.
SACC;
64
SACC COMMENTS:
Recommendation: Add NMP releases to the single
POTW from the two facilities co-located in
Hillsboro and Aloha, OR to better estimate the
highest potential water release estimates. Summing
releases from facilities in Hillsboro (1,496 (J,g/L)
and Aloha (499 (J,g/L) gives a predicted in-stream
concentration of 1,995 |ig/L, an increase of 33% in
the value used for HQ determination. This increase
would raise HQ for amphibians from 0.85 to 1.13
(0.85*1.33). This omission must be corrected.
One facility in the assessment discharged NMP-
containing water to a treatment system that reuses
partially treated water as process water (NMP Risk
Evaluation, p. 370). The agency should state where
EPA has revised the risk evaluation to present updated discharges
in 2018 and has revised the summary table to estimate the
combination of the two facilities (Intel - Aloha Campus and Intel
- Ronler Acres Campus) discharging to the same POTW (Rock
Creek STP) in Hillsboro, Oregon.
EPA updated the evaluation of risks to aquatic receptors using
surface water concentrations estimated for the combine releases.
While the RQ for the combined releases in 2015 was just over
one, there were less than 20 days of exceedance. In addition, the
combined releases based on the more recent 2018 TRI data did
not result in an RQ greater than one. After updating the analysis
in response to this comment, EPA did not identify unreasonable
risks to aquatic receptors.
Page 13 of 205
-------
this reused process water is discharged. If it returns
to a POTW, then the modeled residual NMP would
be estimated differently, specifically as 17%
[0.15*(l+0.85*0.08)], rather than 15% as reported
in the draft risk evaluation.
PUBLIC COMMENT:
Table Apx D-2 indicates that Intel's Aloha and
Hillsboro facilities discharge water to different
POTWs. However, discharges from these facilities
flow to the same POTW and should be combined in
the table.
Intel has built a new onsite comprehensive
wastewater treatment and recycling system at the
Hillsboro facility that will greatly reduce future
NMP discharges relative to the figures reported to
TRI in prior years that are used in the draft risk
evaluation. The new treatment facility became
operational in 2019. Intel expects that by June
2020, all wastewater that might contain NMP at the
Hillsboro facility will be treated onsite. Onsite
NMP removal by the new wastewater treatment
and recycling system is expected to equal or exceed
the 92% NMP removal efficiency assumed by EPA
for POTW treatment. Water treated at the onsite
wastewater treatment and recycling system will
then flow to the POTW, where it will be treated
again and reduced by another 92%. Intel is
collecting data to demonstrate the removal
efficiency for NMP at the Hillsboro onsite
treatment system and offered to share the data with
EPA.
Applying the expected benefit of onsite wastewater
treatment to 2015 TRI data, the estimated
Page 14 of 205
-------
concentration of NMP from the combined
Hillsboro and Aloha facilities to the receiving body
of water would be reduced from 1,995 to 619 |ig/L,
This demonstrates that on a going-forward basis,
NMP releases from Intel's Oregon facilities will
result in NMP surface water concentrations that are
well below the threshold values assumed for
purposes of the Agency's draft risk evaluation, and
will not present an unreasonable risk to the
environment.
SACC
SACC COMMENTS:
The EPI Suite modeling used in this part of the
assessment should not use the default of equal
emissions to air, soil, and water but should be set
appropriately for NMP. Applying the assumed
release amounts to air and water into a standard
level 3 fugacity model shows that partitioning from
air into the water accounts for 1/3 to 1/2 of the
NMP estimated to be discharged to water. This
suggest that releases to water may be
underestimated by 30-50%. This partitioning from
air to water points out the significant inadequacy of
evaluations that fail to include atmospheric
emissions, especially when releases to the
atmosphere dwarf those to water, as is the case
with NMP.
Recommendation: Use the releases reported in the
draft risk evaluation in EPI Suite modeling
(Fugacity Level 3) rather than defaults of equal
emissions to air, soil, and water.
No fugacity modeling was used in EPA's risk evaluation for
NMP. As described in Section 2.3.2 and Appendix E, the surface
water assessment relied on a combination of Toxics Release
Inventory data and the Agency's Exposure and Fate Assessment
Screening Tool (E-FAST). EPA believes this approach is more
robust than relying on fugacity modeling to predict environmental
concentrations.
SACC
SACC COMMENTS:
Recommendation: Discuss the potential for surface
EPA considered non-point source releases for the pathway of
NMP remaining in land applied biosolids. EPA does not have
Page 15 of 205
-------
runoff and other non-point source releases of NMP
to water.
The focus on TRI reporting (NMP Risk Evaluation,
p. 58) may have inadvertently excluded
consideration of surface runoff and other non-point
source releases of NMP to water.
specific monitoring data but based on NMP fate properties, EPA
does not expect biosolids application and either migration through
soil to groundwater or NMP runoff from subsequent precipitation
to significantly contribute to NMP surface water concentrations.
EPA was able to use facility-specific release date to quantitatively
estimate NMP surface water concentrations. During problem
formulation, EPA considered exposures from land-applied
biosolids, one source of non-point source releases to water. Based
on fate properties (described in more detail in response to the
comment below) EPA concluded that no further analysis of this
pathway was needed.
Fate
SACC
SACC COMMENTS:
Recommendation: Add a discussion of land
application of NMP and discuss the potential for
movement to groundwater as well as the potential
for degradation in subsurface soils.
NMP and its equitoxic major transformation
products are known to be mobile and water soluble.
Land application could provide a mechanism for
movement to groundwater, much of which is not
regulated/monitored under the Safe Drinking Water
Act (SDWA). The Committee was provided no
information on drinking water surveillance to
confirm that this is not an issue. Two million
pounds a year are disposed of in non-hazardous
waste landfills and five million pounds in
underground injection wells (NMP Risk
Evaluation, Table 4-1). Yet, no data were provided
on occurrence of NMP in groundwater, including
near disposal facilities. Whether NMP degrades in
the subsurface is an unanswered question.
As described in Sections 2.1.1 and 4.6.2.3 of the risk evaluation,
EPA considered exposures from land-applied biosolids during
problem formulation and concluded that no further analysis of
this pathway was needed. In the NMP Problem Formulation, EPA
explains that "NMP exhibits high water solubility (1000 g/L) and
limited potential for adsorption to organic matter (estimated log
Koc = 0.9); therefore, land releases will ultimately partition to the
aqueous phase (i.e., biosolids associated waste water and soil pore
water) upon release into the environment. Because NMP readily
biodegrades in environments with active microbial populations,
NMP residues that remain following wastewater treatment are not
expected to persist. NMP concentrations in biosolids-associated
water are expected to decrease, primarily via aerobic degradation,
during transport, processing (including dewatering), handling, and
land application of biosolids (which may include spraying)."
EPA believes it is both reasonable and prudent to tailor TSCA
risk evaluations when other EPA offices have expertise and
experience to address specific environmental media, rather than
attempt to evaluate and regulate potential exposures and risks
from those media under TSCA. EPA has therefore tailored the
Page 16 of 205
-------
scope of the risk evaluation for NMP using authorities in TSCA
Sections 6(b) and 9(b)(1).
As described in Section 1.4.2 of the risk evaluation, EPA did not
include exposures via the drinking water pathway or disposal to
underground injection, RCRA Subtitle C hazardous waste
landfills, or RCRA Subtitle D municipal solid waste (MSW)
landfills in this risk evaluation, as these exposure pathways fall
under the jurisdiction of other EPA-administered statutes and
associated regulatory programs.
SACC
SACC COMMENTS:
The risk evaluation should be consistent in how
major fate processes are discussed. For example,
the draft risk evaluation states that NMP does not
persist in the environment and does not volatilize
into the air, but later (NMP Risk Evaluation, p. 56)
explicitly discusses releases to the atmosphere.
The aqueous persistence evaluation is missing an
underlying citation in the primary literature. The
citation in the draft risk evaluation is for the
Agency Work Plan (EPA, 2015), but that document
provided no underlying rationale for making this
assertion.
Direct atmospheric releases and volatilization from surfaces can
be thought of as distinct processes. The fate language is intended
to describe general tendencies and should not be interpreted as
categorically excluding these processes, only minimizing their
relative importance.
A citation regarding hydrolysis has been added to the narrative in
Section 2.1.1.
Consider potential exposure to metabolites of NIMP
SACC
SACC COMMENTS:
Recommendation: Include major NMP degradation
products (e.g., N-methyl succinimide [NMS]) when
estimating potential aquatic exposures.
The fate of major metabolites (a carbonyl
compound and NMS, Chemical Abstracts Service
Registry Number [CASRN] 1121-07-9) should be
considered as part of the environmental fate
discussion of the chemical. This could involve
summing the estimated concentrations of the parent
A discussion of the fate of major NMP metabolites and
degradants has been added to Section 2.1.1. Note that NMS (also
called MSI) is primarily of interest as a metabolite of NMP and is
not expected to persist as a degradant in the environment. As
described in Section 2.1.1, while NMS is a potential product of
atmospheric oxidation of NMP, it is likely transitory in the
atmosphere, being subject to oxidation by hydroxyl radicals with
an estimated half-life on the order of hours. Based on these
properties, EPA does not expect NMP releases to the environment
to result in substantial concentrations of NMS in surface water.
Page 17 of 205
-------
and metabolites, and weighting them by relative
toxicity. Not considering NMP metabolite
(particularly NMS) concentrations in
environmental media results in aquatic exposure
estimates that are biased low. Measuring and
summing the amount of these two analytes in
discharge waters would improve understanding of
aquatic environmental exposures to the
combination of equitoxic NMP and NMS.
Physical-chemical properties
SACC
SACC COMMENTS:
Recommendation: Add to the Quality Review a
discussion of the quality of estimates used for
physical-chemical properties. Discuss methods to
assess the quality of these data and why the
estimates used were chosen over others available.
Recommendation: Add the Koa and dimensionless
Henry's Law constant to the list of physical-
chemical properties regularly reported in TSCA
chemical evaluations.
Please see Supplemental File IB for a complete summary of the
data quality evaluation for physical and chemical properties.
For the sake of conciseness and consistency with other Agency
assessments, EPA prefers not to include Koa and the
dimensionless Henry's law constant, both of which can be
calculated from the values provided.
SACC
SACC COMMENTS:
Table 2-1 (NMP Risk Evaluation, p. 57) lists the
EPI Suite estimate of the indirect photolysis
(photodegradation) half-life as 5.8 hours. This
value is indicated as "estimated for atmospheric
degradation" but may more appropriately be
termed "photo-oxidation." It is also unclear
whether this value is a half-life or an atmospheric
lifetime.
The language in Table 2-1 has been revised for clarity.
SACC
SACC COMMENTS:
The term "sorption" should be used instead of
"adsorption."
The document has been revised accordingly.
Page 18 of 205
-------
59
PUBLIC COMMENTS:
Full access to the three studies relevant to the fate
of NMP in the environment is needed; the
European Chemical Agency (ECHA) dossiers are
cited (pp. 338-39).
EPA has replaced the ECHA study summaries by their respective
Drimarv sources: Shaver (1984) for the first Gerike and Fischer
(1979) and Krizek et al. (1 ' ฆ. ) for the second, and U.S. EPA
( ) (i.e., EPI Suite) for the third.
Climate change considerations
34, 51
PUBLIC COMMENTS:
Elevated temperatures due to climate change are
expected to influence vapor pressure, water
solubility, and Henry's Law constants, and these
scenarios should be considered in exposures where
inhalation is considered.
Elevated temperatures due to climate change are
likely to affect stream flow rates (15-30-year-old
stream flow data were used to calculate surface
water concentrations for NMP) and contaminant
fate and transport.
To the extent that specific impacts of climate
change are difficult to predict, EPA may account
for that uncertainty through sensitivity analyses, a
broader range of temperature-related assumptions,
or additional UFs. Foreseen changes in
temperatures and their impacts on the risk
evaluation process must be considered by EPA in
this risk evaluation.
Preliminary calculations indicate this temperature increase would
increase vapor pressure by only 0.1%. Water solubility would
increase to some extent, but NMP is already fully miscible at
25ฐC. NMP's enthalpy of solvation, needed to correct its air-water
partition coefficient for an increase in temperature, is not readily
available. However, without performing a sensitivity analysis,
such a correction seems qualitatively unlikely to alter the
conclusions of the fate assessment. For these reasons, an analysis
of the influence of increasing temperatures on NMP exposures
was not included in the risk evaluation.
Page 19 of 205
-------
3. Environmental Hazard and Risk Characterization
Kiivironiiicnliil lla/ard and Risk C'h;iriiclcrixnlion
Charge Question 2.1: EPA determined that there are no environmental risks based on a screening level assessment of risk using
environmental hazard data, TRI exposure data, fate information, and physical/chemical properties. Please comment on whether the
information presented supports the analysis in the draft environmental hazard section and the findings outlined in the draft risk
characterization section.
#
Summary orCommenls lor Specific Issues Related to
Charge Question 2
KPA Response
The committee was unable to review some data due to technical issues
SACC
SACC COMMENTS:
Some toxicity data for aquatic organisms were not
available to the Committee, and as a result, the quality
and accuracy of the aquatic toxicity benchmarks could
not be evaluated.
Weisbrod and Seyring (1980) cited in the NMP paint
strippers work plan (EPA, 2015) was not in the reference
list.
EPA remains committed to a transparent and reproducible
systematic review process to ensure that the information the
Agency relies on in its risk evaluations meets the scientific
requirements in TSCA Section 26. EPA did provide access
to all studies upon which the environmental risk evaluation
was based. Access to the BASF and GAF studies were
provided to reviewers through the HERO website. The peer
review panel did not identify any issues with accessing this
information either prior to or during the formal peer review
panel meeting. In the paint stripper work plan document,
Weisbrod and Sevring (1980) is cited in an appendix table
based on information reported in OECD (2007), but is not
included as a primary reference in the reference list. This
citation was also not identified as part of the literature
search and systematic review process for this risk
evaluation.
Assessment Factors (AFs)
SACC
SACC COMMENTS:
Recommendation: Increase the AF used in the
calculation of the COC from daphnia toxicity data to 100
to account for limited acute and chronic testing and use
of nominal levels in testing.
EPA acknowledges that there is uncertainty regarding the
use of a single assessment factor to estimate hazards from
chronic exposure to NMP. Additional context has been
added to Section 4.3.4 of the final risk evaluation to describe
the protectiveness of assessment factors.
EPA acknowledges that several of the aquatic toxicity
Page 20 of 205
-------
The Committee was troubled that 4 of 5 studies used to
establish environmental hazard were conducted at
nominal concentrations, which are normally higher than
actual test concentrations due to chemical losses to
vaporization, sorption to test chamber walls, and
transformation.
Daphnia appears to be the only species that was tested
with chronic exposures. It is suggested that the numbers
of aquatic species used for the acute and chronic
assessments be specified on lines 133-136, 3787, and
3792 in the risk evaluation.
When chronic data are available for only one species, as
for NMP, then the AF should be 100 to protect 93% of
aquatic species (see Keinzler et al., 2017; Figure 8).
Keinzler et al. (2017) also provides ratios for
extrapolating daphnia toxicity to fish acute and chronic
toxicities (also see Ahlers et al., 2009).
Use of a chronic AF of 10 rather than 100 may
underestimate the risk that NMP poses to fish. Using an
AF of 100, the chemical of concern (COC) would be 177
|ig/L and the risk quotient (RQ) would be 8.5. This
change, combined with underestimates of chemical
concentrations, lead to an RQ estimate that approaches
20, even without accounting for the large
underestimation of discharges to water noted by SACC
in response to CQ1.
studies used nominal concentrations. This factor is
considered in the systematic review process and data
evaluation for the environmental hazard data. These studies
were identified to be of sufficient quality for use in the RE.
While an AF of 10 may not be protective for all chemicals
and trophic levels, the use of 10 to calculate a concentration
of concern for acute and chronic exposures to environmental
receptors is consistent with existing EPA methodology for
the screening-level assessment of new chemical substances.
EPA is in the process of evaluating the body of reasonably
available literature on the subject in order to determine
whether to revise standards for application of AF and ACRs
for the next 20 high-priority substances undergoing risk
evaluation. EPA will consider the Keinzler et al., 2017 study
in its assessment.
Until the body of scientific evidence for assessment factors
is evaluated, EPA will continue to use OPPT methodology
as cited in the risk evaluation and apply an AF of 5 for acute
and 10 for chronic aquatic invertebrate data.
SACC,
34, 51
SACC COMMENTS:
Recommendation: Clarify the selection of AFs and
include UFs in the estimate of risks for aquatic
organisms.
The assessment appears erroneous for aquatic receptors
based on chronic toxicity data, due to the incorrect
Additional context has been added to uncertainty section
4.3.4 of the final risk evaluation to acknowledge the
uncertainty associated with the use of AFs and ACRs. As
described above, EPA is evaluating the body of reasonably
available literature on the subject in order to determine
whether to revise standards for application of AFs and ACRs
for the next 20 high-priority substances undergoing risk
Page 21 of 205
-------
application of AFs and failure to apply uncertainty
factors (UFs).
PUBLIC COMMENTS:
In absence of chronic data for fish, EPA divided the
acute median lethal dose (LC50) by 10 to develop a
chronic hazard value. EPA should apply a higher ACR or
additional AFs, as recommended by SACC for 1-BP. A
recent study of approximately 200 industrial chemicals
reported a median fish ACR of 12.8, with a 90th
percentile ACR of 102.4 and a maximum ACR of
1,370.6.
evaluation. EPA will consider the Keinzler et al., 2017 study
in its assessment.
Until the body of scientific evidence for assessment factors
is evaluated, EPA will continue to use OPPT methodology
as cited in the risk evaluation and apply an AF of 5 for acute
and 10 for chronic aquatic invertebrate data.
SACC
SACC COMMENTS:
Several Committee members suggested that future TSCA
chemical evaluations consider the use of Species
Sensitivity Distributions in setting AFs.
o It was noted that this requires toxicity evaluations on
>5 organisms and is therefore not applicable to the
current evaluation.
Adequate data to compute Species Sensitivity
Distributions should be required for the next 20
chemicals scheduled for TSCA evaluation.
A member suggested that information about threatened
and endangered species could help incorporate other UFs
into the risk evaluation.
Thank you for the recommendation. EPA does use robust
statistical methodologies including species sensitivity
distributions when enough toxicity data are available for
each taxonomic group. While insufficient data were
available to do so for NMP, EPA will consider the use of
species sensitivity distributions for the next 20 chemical
undergoing TSCA evaluation, where possible.
The TSCA risk evaluation focuses on exposures to
particular species and environmental receptors, and
appropriately considered impacts to affected species.
Conclusions on environmental risks - Scope of assessment
SACC
SACC COMMENTS:
Recommendation: Provide a scientific or regulatory
justification for why exposures and risks to terrestrial
receptors should not be assessed.
The Committee concluded that statements regarding "no
environmental risks" are misleading and must be
modified. Only risks posed through surface waters were
During problem formulation, EPA identified several
pathways (including sediment, ambient water, land-applied
biosolids, and ambient air) that did not require further
analysis because environmental fate properties and first-tier
analysis of environmental release data indicate that
exposures are well below levels of concern. As described in
the problem formulation document, the solids pathways
Page 22 of 205
-------
considered for environmental receptors in this draft risk
evaluation. The NMP risk evaluation needs to be more
specific in describing what risks were assessed and
identifying what risks were expected but not assessed.
Terrestrial receptors should have been assessed (NMP
risk evaluation, pp. 164-166) given the large amount of
waste disposed in this manner. Terrestrial organisms are
mentioned in the environmental hazards section (NMP
risk evaluation, lines 3671-3673), but no further.
Decisions made during problem formulation did not
consider equitoxic transformation products of NMP
when assessing potential risks to soil- and sediment-
dwelling organisms.
were not further analyzed because NMP is not expected to
adsorb to suspended solids or sediment due to its high water
solubility and estimated soil organic carbon/water partition
coefficient (log Koc = 0.9). EPA's conclusions about
ambient air, ambient water, sediment, and land-applied
biosolids are summarized in the Problem Formulation and
in Section 4.6.2.3 of the risk evaluation.
EPA considered transformation products of NMP, including
biodegradation products and metabolites. As described in
Section 2.1, based on qualitative analysis of reasonably
available information, EPA concludes that these products
are unlikely to pose risk to the aquatic environment.
In the risk evaluation, EPA updated the screening level
analysis for the ambient water pathway using surface water
concentrations modeled based on 2018 TRI data. EPA
evaluated potential risks for aquatic species by comparing
surface water concentrations to concentrations of concern for
aquatic species. Based on these analyses, EPA did not
identify an unreasonable risk to environmental receptors
from these pathways.
SACC
The Committee determined that information presented
was insufficient to support the conclusion that NMP does
not present an unreasonable risk to environmental
receptors through surface water exposure pathways.
Issues that support this concern are: (1) the determination
was based on limited chronic toxicity data; (2) only
aquatic receptors were evaluated; (3) too small of an
Assessment Factor (AF) was used (10 used instead of
100; see separate comment below); and (4) discharges to
the same POTW were analyzed as separate events
instead of a single larger discharge. These issues tend to
EPA has acknowledged the committee's concerns by adding
additional context to Section 4.3.4 of the final risk
evaluation to describe the uncertainty associated with the
lack of chronic toxicity data, and the use of assessment
factors (AFs) and acute to chronic ratios (ACRs).
During problem formulation, EPA considered fate properties
of NMP and performed a first tier analysis of environmental
risks from NMP exposure through sediment, land-applied
biosolids, and ambient air. EPA did not identify
environmental risks from these pathways.
Page 23 of 205
-------
support a conclusion of there being a reasonable
probability of hazard to aquatic receptors from NMP
exposures and that subsequent risk estimates are
underestimated.
In response to the fourth concern, EPA updated the
evaluation of risks to aquatic receptors using surface water
concentrations estimated for the combined releases of the
two facilities (Intel - Aloha Campus and Intel - Ronler
Acres Campus) discharging to the same POTW (Rock Creek
STP) in Hillsboro, Oregon. While the RQ for the combined
releases in 2015 was just over one, there were less than 20
days of exceedance. In addition, the combined releases based
on the more recent 2018 TRI data did not result in an RQ
greater than one.
Data for determination of ecological hazard
34, 51
PUBLIC COMMENTS:
The draft risk evaluation excludes the studies that
demonstrate the greatest environmental risk, obscures the
results of the studies that it does consider, and disregards
risk quotients >100 times greater than EPA's
unreasonable risk threshold. For aquatic invertebrates,
EPA did not consider the most sensitive data, resulting in
an underestimate of NMP's ecological risks. In the draft
risk evaluation, EPA reports an ECso/LCso of 1,107-
4,897 mg/L for aquatic invertebrates, based on a 1979
study of Daphnia magna. However, a 2004 study cited in
the NMP Problem Formulation document reported an
LCso of 1.23 ml/L, approximately 1,000 times lower, for
that same species.
EPA lacks adequate data to evaluate ecological risk. EPA
does not have any studies of NMP's effects on terrestrial-
or sediment-dwelling species, and no chronic aquatic
toxicity data for NMP in fish. EPA should use its TSCA
authority to collect or generate missing data on NMP's
toxicity.
EPA obtained and considered reasonably available
information, defined as information that EPA possesses, or
can reasonably obtain and synthesize for use in risk
evaluations, considering the deadlines for completing the
evaluation. Given the timeframe for conducting risk
evaluations on the first 10 chemicals, use of TSCA data
gathering authorities has been limited in scope. In general,
EPA intends to utilize TSCA data gathering authorities more
routinely for the next 20 risk evaluations.
During problem formulation, EPA considered fate properties
of NMP and performed a first tier analysis of environmental
risks from NMP exposure through sediment, land-applied
biosolids, and ambient air. EPA did not identify
environmental risks from these pathways.
EPA remains committed to a transparent and reproducible
systematic review process to ensure that the information the
Agency relies on in its risk evaluations meets the scientific
requirements in TSCA Section 26. EPA did provide public
access to all studies upon which the draft risk evaluation was
Page 24 of 205
-------
Where data do exist, EPA has not provided public access
to the studies that it relied on for its environmental risk
evaluations (e.g., study reports obtained from the NMP
Producer's Group [BASF and GAF]), leaving the public
unable to verify or critically evaluate EPA's conclusions.
based. Access to the BASF and GAF studies were provided
through the HERO website. There were no NMP Producers
Group studies used for evaluation of environmental hazard.
Consider toxicity of degradation products
SACC
SACC COMMENTS:
Recommendation: Expand the discussion on the potential
risk from NMS exposure to environmental receptors and
include the lack of NMS concentration data in surface
waters as a source of uncertainty in the evaluation of
environmental risks.
The Committee expressed concern that the potential risk
to aquatic organisms from exposure to NMS, a
degradation product of NMP, is not discussed in the draft
risk evaluation. Data suggest that NMS could be as toxic
as NMP to daphnia.
ซ-methyl-succinimide (NMS) is a metabolite of NMP, but it
is not expected to persist as a degradation product in the
environment. As described in Section 2.1, while NMS is a
potential product of atmospheric oxidation of NMP, it is
likely transitory in the atmosphere, being subject to
oxidation by hydroxyl radicals with an estimated half-life on
the order of hours. Based on these properties, EPA does not
expect NMP releases to the environment to result in
substantial concentrations of NMS in surface water.
Cite Environment Canada reference
SACC
SACC COMMENTS:
Recommendation: Include references to the Canadian
determinations on bioaccumulation of NMP.
The draft risk evaluation states that NMP exhibits low
potential for bioaccumulation (NMP risk evaluation, p.
58, line 977). One Committee member noted that Canada
made a similar determination (Environment Canada,
1994), but that this seems to be in contradiction with
toxicity data on microbiota (Campbell et al., 1999).
The NMP physical and chemical properties and
environmental fate characteristics used in the RE are based
on EPI Suite estimations and reasonably available fate
data to characterize the environmental fate and transport of
NMP. During problem formulation, EPA also analyzed the
air, water, sediment, land and biosolids pathways. These
results are described in the NMP Problem Formulation
document (U.S. EPA. 2018). EPA identified and evaluated
environmental fate data quality studies according to the
TSCA systematic review process. Environmental fate data
from acceptable studies were extracted and integrated during
risk evaluation. Based on the results obtained from the data
quality evaluation process EPA has high confidence in the
studies used to characterize the environmental fate of NMP.
This data may or may not be included in the Canadian
Page 25 of 205
-------
assessment.
Modeling issues
SACC
SACC COMMENTS:
Recommendation: Include confidence bounds for E-
FAST predicted concentrations and reduce the number of
significant digits reported in tables and text from 4 to 2,
based on precision of input values.
The E-FAST model predicted concentration (1,496 (J,g/L)
is close to the chronic value of 1,768 (J,g/L. The
Committee recommended including estimates of upper
and lower 95% confidence limits on model-estimated
values to help assess if these two values are statistically
the same.
EPA has revised the surface water estimates to reduce the
number of significant figures.
EPA used the E-FAST model to predict site-specific stream
concentrations of NMP given TRI releases. However, EPA
also used the Probabilistic Dilution Model (PDM) portion
of E-FAST 2014 for free-flowing water bodies. The PDM
predicts the number of days/yr a chemical's COC in an
ambient water body will be exceeded. COCs are threshold
concentrations below which adverse effects on aquatic life
are expected to be minimal.
PDM calculates the COC exceedance probability using a
stochastic procedure developed bv Di Toro (1984). This
approach requires the means and coefficients of variation of
stream flow, effluent flow, and effluent concentration as
input. Mean stream flow and mean effluent flow are
provided by the E-FAST2 Main Facility File. The stream
flow coefficient of variation is estimated using the mean
stream flow, low stream flow (7Q10, also available in the
E-FAST2 Main Facility File) and empirically derived
coefficients specific to each subbasin that are available in
the Basin Coefficient Statistical File. The coefficients of
variation of effluent flow and concentration are assumed to
be 0.24 and 0.85, respectively.
Following entry of chemical loading and the COC, the
probability of exceedance is calculated by PDM using the
Di Toro algorithm. PDM does not estimate exceedances for
chemicals discharged to still waters, such as lakes, bays, or
estuaries. For these water bodies, the days of exceedance is
Page 26 of 205
-------
assumed to be zero unless the predicted surface water
concentration exceeds the COC.
Additional details of the E-FAST model are found at:
https://www.eDa. gov/sites/oroduction/files/2015-
04/docum ents/efast2m an .odf
EPA did not have enough chronic toxicity data available to
make a statistical comparison of the chronic COC with the
estimated surface water concentration using confidence
intervals. Instead, in scenarios with limited data, EPA picks
the lowest (most sensitive) toxicity value and divides by an
assessment factor of 10 to determine the chronic COC.
There is uncertainty associated with the use of assessment
factors (e.g., what value is large enough to be protective)
which has been added as a discussion in Section 4.3.4.
RQs ~
SACC
SACC COMMENTS:
EPA has revised the risk evaluation and re-calculated RQs
corresponding to the updated surface water estimates from
2015 and 2018 TRI data (Table 4-2 and Table 4-3). All acute
and chronic RQs < 1 except for the POTW (Rock Creek
STP) in Hillsboro, Oregon, where the RQ was 1.1. The RQ
increased from the Draft RE and reflects the combination of
the two facilities (Intel - Aloha Campus and Intel - Ronler
Acres Campus) that discharge to the same POTW, as was
suggested by the committee.
Because frequency and duration of exposure also affects the
potential for adverse effects in aquatic organisms, the
number of days that the chronic COC was exceeded was also
calculated using E-FAST as described in Section 2.3.2.
Facilities with an RQ > 1 for the acute risk scenario or an
RQ > 1 and 20 days or more of exceedance for the chronic
risk scenario would suggest the potential for environmental
Several Committee members noted that the RQ for the
chronic (environmental) risk scenario (NMP Risk
Evaluation, Table 4-2, p. 209) is very close to one. RQs
approaching 1 are likely to be very sensitive to small
changes in estimates of maximum exposure
concentrations and/or COCs.
Recommendation: Flag RQs close to 1 for further
evaluation because they are sensitive to small changes in
the estimates of exposure concentrations and/or COCs.
Page 27 of 205
-------
risks posed by NMP. The 20-day exceedance time frame was
derived from partial life cycle tests (e.g., daphnid chronic
and fish early life stage tests) that typically range from 21 to
28 days in duration. Because the surface water concentration
at Hillsboro was predicted to exceed the chronic COC for 2
days per year, risk from chronic exposure was not indicated
for aquatic receptors.
Exposure data used for ecological risk characterization
32
PUBLIC COMMENTS:
EPA used its E-FAST model to predict surface water
concentrations at the TRI/DMR facilities based on
facility-specific emissions and wastewater treatment
removal. The Probabilistic Dilution Model (PDM) was
used to predict the number of days a stream
concentration may exceed the designated concentration
of concern. It is unclear whether EPA used the dilution
factor for the site-specific receiving water body or the
national 7Q10 dilution factor, which is equivalent to 1.0.
EPA should clarify the reasoning for using the 7Q10
value for the facility-specific receiving water body
associated with the facilities discharge, rather than the E-
FAST PDM 7Q10 for dilution.
EPA uses the 7Q10 hydrologically based low flow in several
program offices, including the NPDES permit writing
program. The 7Q10 is the lowest 7-day average flow that
occurs (on average) once every 10 years. The hydrologically
based low flow is computed using the single lowest flow
event from each year of record, followed by application of
distributional models (typically the Log Pearson Type III
distribution is assumed) to infer the low flow value. National
Pollutant Discharge Elimination System (NPDES) permit
writers often need to calculate low flow statistics for
reasonable potential analyses and water quality-based
effluent limitation (WQBEL) calculations or to confirm
estimates provided by the permittee during the NPDES
permit development process. The EPA E-FAST model also
uses the accepted 7Q10 low flow to estimate site-specific
NMP stream concentrations at facilities reporting NMP
releases.
The Probabilistic Dilution Model (PDM) is used for
predicting downstream chemical concentrations from an
industrial discharge. It calculates the probability that a given
target stream concentration will be exceeded, and the
number of days per year the exceedance condition will exist.
The calculation of probability assumes that receiving stream
flow, effluent flow, and effluent concentration are log-
Page 28 of 205
-------
normally distributed. The statistics involve both the
arithmetic and logarithmic forms of the mean and coefficient
of variation (., standard deviation/mean) for the flow and
concentration of both the stream and the effluent. PDM can
predict frequency of exceedance of the concern
concentration in streams that have a record of flow data from
gaging stations as well as in streams without gaging stations.
Detailed information on the E-FAST model is publicly
available at:
httos:/Av ww.eoa.gov/sites/Droduction/files/2015-
04/docum ents/efast2m an .odf
54
PUBLIC COMMENTS:
EPA applied a number of conservative estimates in
evaluating environmental exposures in the draft risk
evaluation, particularly regarding surface water
concentrations (e.g., 12- and 250-day acute and chronic
release scenarios). These conservative assumptions may
be suitable for a "first-tier" exposure assessment, but
EPA should more clearly articulate what approach(es) it
would use for higher tier environmental exposure
estimates and justify assumptions regarding release
scenarios. Further, since EPA is estimating exposures at
specific facilities using this approach, it should
understand and incorporate those facility-specific
conditions into its assessment.
The "first-tier" or conservative screening level analyses for
risks to aquatic organisms did not identify risks for any of
the direct or indirect dischargers of NMP reporting to the
TRI in 2015 or 2018. These results do not indicate the need
for any further detailed environmental risk analyses.
54, 46
PUBLIC COMMENTS:
EPA needs to be more transparent about how it will
analyze environmental risks that are already subject to
EPA regulation under other environmental laws when
conducting risk evaluations. It is recommended that EPA
more clearly explain its approach, consider whether its
approach is consistent across its TSCA risk evaluations,
and if not, to explain why not. For example, EPA should
As described in Section 1.4.2, EPA believes it is both
reasonable and prudent to tailor TSCA risk evaluations when
other EPA offices have expertise and experience to address
specific environmental media, rather than attempt to evaluate
and regulate potential exposures and risks from those media
under TSCA. EPA believes that coordinated action on
exposure pathways and risks addressed by other EPA-
administered statutes and regulatory programs is consistent
Page 29 of 205
-------
address when TRI estimates are adequate to predict
concentrations in air, water, and land, and when they are
not, in addition to what constitutes adequate "regulation"
under other environmental laws and regulation within
EPA's purview, providing justification when EPA will
either not analyze it further or not review it at all in a risk
evaluation. Differences in treatment of NMP and
methylene chloride (MC) illustrate the problem. MC is a
regulated priority pollutant under the Clean Water Act
(CWA) with CWA monitoring data and CWA
technology-based standards, but EPA decided to analyze
the ambient water pathway anyway. For NMP, however,
which is not regulated under CWA and has no
monitoring data, EPA dropped the ambient water
pathway based on estimated TRI release data.
with statutory text and legislative history, particularly as
they pertain to TSCA's function as a "gap-filling" statute,
and also furthers EPA aims to efficiently use Agency
resources, avoid duplicating efforts taken pursuant to other
Agency programs, and meet the statutory deadline for
completing risk evaluations. EPA has therefore tailored the
scope of the risk evaluation for NMP using authorities in
TSCA Sections 6(b) and 9(b)(1).
During problem formulation, EPA performed a first-tier
screening analysis of risks from ambient air, ambient water,
sediment, and land-applied biosolids. EPA did not identify
risks from human or environmental exposures that may
result from these pathways. In the final risk evaluation, EPA
updated the evaluation of risks to aquatic life and the general
population from ambient water exposures using more recent
TRI release data. As described in Section 4.1, EPA did not
identify risks to environmental receptors or the general
population from ambient water.
Page 30 of 205
-------
4. Occupational and Consumer Exposure
Occupalional :mtl Consumer Kxposure
Charge Question 3.1: Please comment on the reasonableness of the characterization of occupational exposure for workers and
occupations non-users. What other additional information, if any, should be considered.
Charge Question 3.2: Please comment on the transparency of EPAs approach and the assumptions EPA used to characterize
exposure for ONUs.
Charge Question 3.3: Please comment on the approaches and assumptions used and provide any specific suggestions or
recommendations for alternative approaches, models or information that should be considered by the Agency for improving the
workplace exposure assessment. More specifically, if other sources of monitoring data are available to estimate air concentrations for
worker exposures, please provide specific citations.
Charge Question 3.4: Please comment on assumptions used in the absence of specific exposure information (e.g., dermal surface
area assumptions: high-end values, which represents two full hands in contact with a liquid: 890 cm2 (mean for females), 1070 cm2
(mean for males); central tendency values, which is half of two full hands (equivalent to one full hand) in contact with a liquid and
represents only the palm-side of both hands exposed to a liquid: 445 cm2 (females), 535 cm2 (males)).
Charge Question 3.5: Please comment on EPAs approach to characterizing the strengths, limitations and overall confidence for each
occupational exposure scenarios presented in Section 2.4.1. Please comment on the appropriateness of these confidence ratings for
each scenario. Please also comment on EPAs approach to characterizing the uncertainties summarized in Section 2.4.1.4.
Charge Question 3.6: Please comment on the approach used and provide any specific suggestions or recommendations for
alternative approaches, models or information that should be considered by the Agency for improving its assessment of consumer
inhalation exposure, including specific citations of data sources characterizing consumer emission profiles of NMP-based products.
Charge Question 3.7: Please comment on EPAs approach to characterizing the strengths, limitations and overall confidence for each
consumer exposure scenarios presented in Section 2.4.2. Please comment on the appropriateness of these confidence ratings for each
scenario Please also comment on EPAs annioach to chaiacteiizinu the uncertainties summarized in Section 2 4 2 6
#
Summary of Commcnls for Specific Issues Related lo
Charge Question 3
KPA Response
Dermal absorption and exposure from direct skin contact with liquids
SACC
SACC COMMENTS:
Recommendation: Revisit dermal absorption modeling
and adopt a matrix of standardized approaches based on
conditions of use, the physical-chemical properties of the
agent of interest, and its vehicle (if any).
Modeling of dermal absorption of NMP is captured in the
PBPK model and differs from dermal absorption modeling
applied in other TSCA risk evaluations. The NMP risk
evaluation does not cite Frasch and Bunge (2015) or Frasch
(2011). This comment appears directed at the dermal
modeling approach applied in TSCA risk evaluations for
Page 31 of 205
-------
In the draft risk evaluation, the Frasch and Bunge (2015)
model is cited and used rather than the Frasch et al.
(2011) model that was used in previous risk evaluations.
The Frasch and Bunge (2015) model deals with
disposition of skin depot left behind after skin
decontamination but does not address disposition of total
applied dose. This is considered as misuse of the Frasch
and Bunge (2015) model.
The draft risk evaluation links to CEM 2.1, which
includes 4 dermal models, while previous evaluations
linked to CEM 2.0, which included 3 dermal models
(model identified only as "CEM" in the evaluations).
This may lead to confusion. The risk evaluation needs to
clearly identify the actual model used and consider
assigning version numbers to models in a consistent and
standardized manner.
other chemicals. EPA is currently reevaluating its approach to
dermal exposure modeling for TSCA risk evaluations. The
NMP PBPK models incorporate reasonably available NMP-
specific data on dermal absorption. For example, reasonably
available data demonstrate that dermal permeability of neat
NMP is higher than permeability of 50% NMP in water. The
PBPK model used to model human exposures in the final risk
evaluation adjusts dermal permeability based on the weight
fraction of NMP in products associated with each exposure
scenario. While the solvents present in NMP containing-
products may influence dermal absorption, EPA does not
have data on the impact of the specific solvents and product
formulations relevant for each condition of use on the dermal
permeability of NMP.
EPA used CEM 2.1 and checked that the links are
appropriate.
SACC
SACC COMMENTS:
Recommendation: Exposure factors should more
adequately reflect uncertainty and avoid use of excessive
significant digits.
EPA used hand area as a surrogate for exposed skin. One
Committee member noted that information from
occupational agriculture is supportive of the idea that
hands are disproportionately exposed as a result of
normal human behavior and that hand area was therefore
a reasonable starting point. It was thought, however, that
EPA estimates to three significant digits were
unrealistically precise. It was suggested that hand area
represents an exposure factor for which use of distributed
values rather than point estimates would be relatively
easy to implement. It was also noted that surveyed
EPA uses three or more significant digits only when reporting
or using values reported in literature sources and in its
guidance documentation. For example, EPA's modeling
guidance documents specify surface area input values that
have three significant digits, and EPA reports these values
consistently in its first 10 risk evaluations. EPA presented all
occupational PBPK modeling results with two significant
figures in Section 2.4.1.3.
EPA clarified in Section 2.4.1.1 that EPA has no reasonably
available information on actual surface area of contact with
liquid and that the assumed values represent adequate
surrogates for most uses' central tendency and high-end
surface areas of contact with liquid that may sometimes
include exposures to much of the hands and also beyond the
hands, such as wrists, forearms, neck, or other parts of the
Page 32 of 205
-------
graffiti removers self-reported exposure to skin other
than hands.
body. EPA accounts for distributed values using the central
tendency and high-end assumptions for surface areas.
32, 49,
52, 54,
56,31,
64
PUBLIC COMMENTS:
In the draft risk evaluation, EPA's default assumption is
that the total skin surface area of hands is in prolonged
contact with the liquid product. This assumption is
inaccurate, does not reflect the actual work activities, and
does not take into consideration information provided to
EPA detailing engineering controls and chemical
handling procedures that explicitly prevent dermal
contact with liquid NMP or other forms of residual
NMP. The low NMP concentrations at semiconductor
facilities during routine or maintenance tasks are not
indicative of the presence of liquid NMP, and are
therefore inconsistent with EPA's assumption of
extensive dermal contact. While an Intel employee may
periodically touch liquid NMP while wearing personal
protection equipment (PPE), such contact would be brief
and the employee's hands would never be immersed in
liquid. Accordingly, the chronic exposure scenario used
in the draft risk evaluation is not reflective of Intel's
work practices and exposure potential for conditions of
use in the semiconductor industry. This assumption
results in exposure scenarios driven by dermal contact
with the liquid. For example, in the electronics industry,
in most scenarios presented, 100% of the area under the
curve (AUC) {i.e., internal dose) is due to dermal
contact, including tasks such as maintenance, truck
unloading, and fabrication). Justification for these
exposure assumptions for these occupational scenarios is
needed.
Immersion of one or two hands in concentrated or neat
NMP solvent for prolonged periods is implausible, as it
EPA has improved and clarified dermal input parameter
assumptions in Section 2.4.1.1. EPA clarified in Section
2.4.1.1 that the exposure duration assumptions of full-shifts
for high-ends account for the possibility of repeated contact
with NMP such that NMP does not fully volatilize from the
skin before the next contact event, potentially resulting in
prolonged exposure.
EPA has expanded the range of contact durations for OESs
where values of both shift durations and task durations were
reasonably available.
EPA clarified in Section 2.4.1.1 that EPA has no reasonably
available information on actual surface area of contact with
liquid and that the assumed values represent adequate
surrogates for most uses' central tendency and high-end
surface areas of contact with liquid that may sometimes
include exposures to much of the hands and also beyond the
hands, such as wrists, forearms, neck, or other parts of the
body.
EPA clarified in Section 2.4 that non-immersive dermal
contact with liquid films is evaluated.
In the Electronics Manufacturing OES, EPA includes 6
worker activities within semiconductor manufacturing. EPA
added several PBPK model runs using semiconductor
industry-proposed input values and data including assumed
contact durations. EPA has not found reasonably available
data on actual contact durations or contact surface area for
workers in the semiconductor industry and most other OESs.
Page 33 of 205
-------
is a skin irritant that can cause dermatitis, blistering, or
cracking (E.U. SCCS, 2011) that would be difficult to
tolerate for prolonged periods of time (one or two hands
immersed in solvent for 30 or 60 hours/week,
respectively).
The semiconductor worker scenarios are characterized
by margins of exposure (MOEs) >30 when dermal liquid
contact is assumed to be negligible. Thus, the draft
agency "unreasonable risk determination" for
semiconductor workers is highly sensitive to the
unsubstantiated assumption of extensive and immersive
skin contact with liquid NMP.
o An assumed condition of use with immersive and
prolonged contact with NMP is inconsistent with the
statement in the 2016 peer-reviewed publication that
"human exposures to NMP will be primarily via the
inhalation route" (Poet et al., 2016, Sup. Al, p. 5).
EPA has not provided a transparent substantiated
analysis in the 2019 draft risk evaluation explaining
the inconsistency in the stated contribution of liquid
contact between the peer-reviewed paper and the
draft TSCA evaluation,
o The unexpected dominant contribution of NMP
contact with the skin to internal exposure should
have resulted in additional steps by the Agency to
characterize uncertainty and refine model
assumptions.
EPA must include appropriate justification of dermal
exposure assumptions for occupational scenarios in the
draft risk evaluation and better represent the occupational
exposure scenarios experienced by workers for both
central tendency and high-end scenarios. Table 4-49
indicates that 52 of the 58 exposure calculations were
EPA added discussion in Section 2.4.1.1 regarding the
relative contributions of each exposure pathway to total
exposures, which vary according to parameter values for
NMP weight fraction in the liquid product contacted, skin
surface areas in contact with the liquid product and with
vapor, durations of dermal contact with liquid product and
with vapor, air concentration for inhalation and vapor-
through-skin exposure, body weight of the exposed person,
and glove protection factor and respirator assigned protection
factor (if applicable). In scenarios where the three parameters
involving dermal contact with liquid product (NMP weight
fraction in the liquid product contacted, skin surface areas in
contact with the liquid product and with vapor, durations of
dermal contact with liquid product) have relatively high
values, this route can be the dominant route for worker
exposures.
To illustrate the contribution of inhalation and vapor-through
skin versus dermal contact with liquids, the male worker and
male ONU AUC values can be compared for the same work
activity for an OES because the PBPK inputs for both
workers and ONUs utilize the same NMP air concentration,
while the worker PBPK inputs include parameters for dermal
contact with liquid and the ONU PBPK inputs assume no
dermal contact with liquid.
For example, for the OES Laboratory Use, the central
tendency scenario (PF = 1) results are a male worker AUC of
77 hr-mg/L and male ONU AUC of 0.023 hr-mg/L. These
results indicate a 0.03% contribution from inhalation and
vapor-through-skin exposure and a 99.97% contribution from
dermal contact with liquid for the worker. For the same OES,
the central tendency scenario (PF = 20) results are a male
Page 34 of 205
-------
driven entirely by the dermal exposure levels (>88%
NMP exposure resulting from the dermal route). Because
the dermal route has an outsized effect on the overall
exposure - and consequently the risk determination -
EPA should ensure that these values are as accurate as
possible, rather than relying on overly conservative
assumptions.
worker AUC of 3.4 hr-mg/L and male ONU AUC of 0.023
hr-mg/L (unchanged because no ONU dermal exposure).
These results indicate a 0.68% contribution from inhalation
and vapor-through-skin exposure and a 99.32% contribution
from dermal contact with liquid for the worker. These results
show that, with decreasing dermal exposure to liquid,
inhalation and vapor-through-skin exposure have an
increasing contribution to exposure results.
For the same Laboratory Use OES, the high-end scenario (PF
= 1) results are a male worker AUC of 400 hr-mg/L and male
ONU AUC of 0.86 hr-mg/L. These results indicate a 0.22%
contribution from inhalation and vapor-through-skin exposure
and a 99.78% contribution from dermal contact with liquid.
Compared with the central tendency (PF =1) scenario, the
NMP air concentration increased by >4000% and contact
duration and hand surface area increased by 100%. The
exposure results between the central tendency and high-end
scenarios show higher contributions from inhalation and
vapor-through-skin exposure; however, the increase in the
contributions of these pathways is not proportional to increase
in air concentration. For this OES, regardless of central
tendency or high-end and PF, dermal contact with liquid is
the dominant pathway for workers.
EPA believes that engineering controls would not impact
contact duration with liquids but would generally reduce air
concentrations. Such reductions would be reflected in air
monitoring data. EPA considers chemical handling practices
by reflecting different worker activities in each OES to the
extent that these activities are known.
Page 35 of 205
-------
EPA does not expect that NMP air concentrations correlate to
dermal contact, which is indicated by worker activities.
EPA accounts for potential glove use by applying a range of
glove protection factors for every worker activity modeled as
indicated in Section 2.4.1.1. The modeling results for each
OES central tendency and high-end scenario are presented in
Table 2-77 and for all scenarios for all OESs in the
Supplemental Excel File on Occupational Risk Calculations.
EPA does not have reasonably available data or information
to inform specific durations of contact and associated
concentrations and formulations that would be implausible or
cause toleration issues.
SACC,
31,49,
52 53,
56, 54,
64
SACC COMMENTS:
Assumptions regarding work shift duration, specifically
the assumption that central tendency exposures involve
durations <8 hours seem unrealistic and should be
reassessed.
The draft risk evaluation should consider assuming
longer work-days for some workers since 12-hour shifts
were noted in the literature for degreasing of optical
lenses (Xiaofei et al., 2000) and as reported in the
sampling data from SI A (2019).
Recommendation: Revisit shift duration assumptions or
explain why results are not sensitive to that parameter.
PUBLIC COMMENTS:
Table 2-32, p. 102 in the draft risk evaluation and Table
2-42, p. 84 of the Supplemental Information of the NMP
Draft Risk Evaluation report exposure durations for
several tasks that are incorrect. For example, EPA made
an incorrect assumption that semiconductor industry
workers could be exposed to NMP throughout their
EPA clarified in Section 2.4.1.1 that shift durations are
assumed to be 8 (standard) or 12 hours, depending upon
available data, and that durations of contact with liquids are
based on fractions of shift-durations or other assumptions
(e.g., task durations). EPA does not assume any shift
durations < 8 hours. EPA assumes 12-hour shifts for several
subgroupings of the Electronics Industry OES where air
monitoring data indicates such durations. Therefore, EPA
revised shift durations based on reasonably available data.
EPA does not have reasonably available data or information
that shows assumed exposure durations for dermal contact
with liquids to be incorrect for any tasks. EPA added several
PBPK model runs using semiconductor industry-proposed
input values and data including their assumed contact
durations. EPA has not found reasonably available data on
actual contact durations or methods for measuring these
durations for workers in any industry, including the
semiconductor industry.
Page 36 of 205
-------
entire work shift (8-12 hours), rather than episodically.
However, no individuals in semiconductor
manufacturing handles containers for 6-12 hours/day.
Although Intel factory shifts range up to 12 hours and a
given maintenance activity may take several hours, the
actual time any worker would come into contact with
NMP (always while wearing PPE) would be only a small
fraction of this time. For most maintenance tasks, contact
would be short; potential for chemical contact would
typically be between 15 and 60 minutes (except in very
unusual circumstances, such as if cleaning up a small
spill contained within complex equipment).
Page 128: it is incorrectly assumed that truck unloading
at semiconductor sites is a 4- or 8-hour/day task. Data
submitted by SIA (2019) indicate that the task takes 2-4
hours and is performed no more than 2 times each year.
In the lithium ion battery industry, employees prepare
batches 1-2 times/day, 3-4 times/week, with a duration of
2.5 hours per batch.
Page 132, Table 2-66: The task durations for
semiconductor applications are inaccurate.
The frequency and duration assumptions in the draft risk
evaluation did not consider the duration and frequency of
use data provided to EPA from semiconductor fab
facilities including worker exposure monitoring for NMP
conducted at 14 facilities with a total of 118 samples and
is not reflective of the actual work activities. It is
requested that EPA reconsider their estimated duration of
potential exposure and take into consideration the data
provided to them. Many tasks involve episodic handling
of NMP and task duration is short.
The current dermal liquid contact exposure assumptions
are based primarily on a policy rather than a "best
EPA clarified in Section 2.4.1.1 that the contact duration
assumptions of full-shifts for high-ends account for the
possibility of repeated contact with NMP such that NMP does
not fully volatilize from the skin before the next contact
event, potentially resulting in prolonged exposure. In this
section EPA also clarified that where available, EPA utilized
exposure durations from the available task-based inhalation
monitoring data for generating what-if type exposure
scenarios assuming that the workers were contacting NMP-
containing liquids over only the monitoring duration {i.e., the
entire task duration). Task-based duration estimates do not
account for either liquid remaining on the skin after the task
is completed or for workers performing a task multiple times
during their shift. EPA expanded PBPK runs using both shift-
based and task-based duration estimates for many OESs.
EPA did not assume that truck unloading at semiconductor
sites is a 4- or 8-hour/day task but assumed shift-based
durations for central tendency and high-end estimates and
task-based durations for what-if estimates. EPA removed the
truck unloading from chronic estimates since this task is not
performed 4 or 5 days per week.
EPA used the most recent industry-provided task duration
estimates in some PBPK runs for the lithium ion battery
industry and for the semiconductor applications, including fab
facilities.
EPA's current dermal liquid contact exposure assumptions
are based on the "best available science" approach and have
considered detailed information supplied by the assessed
industry, including industry-proposed parameter values as
Page 37 of 205
-------
available science" approach that considered detailed
information supplied by the assessed industry. The
equations used by EPA imply immersion for prolonged
periods of time. Rather than a generic assumption of
immersion in NMP-containing liquid, the dermal
exposure chapter of the American Industrial Hygiene
Association (AIHA) reference text "Mathematical
Models for Estimating Occupational Exposures to
Chemicals, 2nd Edition" advises that scenario-specific
liquid loading, surface area, and contact time should be
determined based on the conditions of use. The chapter
notes that "a far more realistic scenario is to consider a
finite volume of chemical deposited on the skin that is
subsequently removed by one or more mechanisms, such
as washing or evaporation" (Sahmel et al., 2009, p. 119).
It is implausible that hand surface area of liquid NMP
contact and fraction of the shift exposed to liquid be the
same in dissimilar industries such as paint, coatings,
adhesives, and semiconductor manufacturing, which
EPA grouped together.
Many tasks involve the use of NMP in well-ventilated
spaces with conditions favoring the evaporation of
incidentally generated solvent residual. As described in
Sahmel et al. (2009), the consideration of evaporation of
volatile or semi-volatile chemicals from the skin is an
important determinant of dermal exposure potential.
At least one peer-reviewed approach capable of using
scenario-specific factors is available for dermal liquid
exposure assessment. The IH SkinPerm model is freely
available from AIHA (https://www.aiha.org/public-
resources/consumer-resources/topics-of-interest/ih-apps-
tools) and presented in the peer-reviewed literature in
Tibaldi et al. (2014). This model allows for consideration
well as additional parameter values that consider more
factors, such as repeated contact with liquids during the
workers' shifts and time for NMP-containing liquids to
evaporate. EPA's approach is consistent with the dermal
exposure chapter of the American Industrial Hygiene
Association (AIHA) reference by using scenario-specific
surface area and contact time that are determined based on the
conditions of use. The liquid loading aspect covered in
AIHA's dermal chapter and in AIHA's IH SkinPerm model is
handled differently because PBPK modeling for internal dose
does not use a liquid loading parameter as do the more
simplistic potential dose models covered by the AIHA
reference. EPA clarified in Section 2.4 that non-immersive
dermal contact with liquid films is evaluated. EPA does not
have reasonably available data to indicate dissimilarity of
industries grouped into OESs. EPA did not group paint,
coatings, adhesives, and semiconductor manufacturing into an
OES.
Regarding AIHA's IH SkinPerm model contact time (h)
based on a consideration of evaporation, this model's
treatment does not account for repeated contacts during a
worker's shift, task duration, nor worker activities for
particular NMP OESs and COUs. Therefore, the evaporation-
based contact times estimated by this AIHA model are less
useful for EPA's risk evaluation.
EPA considers evaporation of volatile or semi-volatile
chemicals from the skin as a determinant of dermal exposure
potential by using contact duration. This evaporation is only
one of many determinants toward contact duration.
Page 38 of 205
-------
of realistic exposure scenario factors including skin
surface loading (mg/cm2) and contact time (h) based on a
consideration of evaporation. EPA should explore
application of this model to the NMP risk evaluation.
EPA does not specify the loading of NMP on the skin or
gloves because a scenario equivalent to skin immersion
in solvent was assumed. Research has shown that the
amount of substance deposited on the skin or gloves can
vary by activity (Sahmel et al., 2009). The AIHA dermal
exposure assessment chapter suggests default dermal
loading of 0.7-2.1 mg/cm2 for incidental contact with
liquids (Sahmel et al., 2009). The tasks described by SIA
(2019a) indicated very limited contact opportunities of
NMP with skin or gloves. Thus, it is reasonable to
assume that the maximum daily loading of NMP -
containing liquid during a work-shift is approximately
0.7-2.1 mg/cm2.
Cardno ChemRisk used IH SkinPerm (Tibaldi et al.,
2014) to determine reasonable contact times for NMP.
The model indicated times to complete evaporation of 20
and 60 minutes when the loading was 0.7 and 2.1
mg/cm2, respectively. This conclusion was insensitive to
surface area over the range of 10-1,000 cm2. Cardno
ChemRisk confirmed that the dermal permeability
constant used by EPA of 4.78xl0"4 was similar to the
value of 3.66xl0"4 predicted by the algorithm of IH
SkinPerm; thus, predictions of dermal absorption were
similar in both methods.
Scenario-specific factors available for dermal liquid exposure
assessment in the IH SkinPerm model are not specific enough
to specifically determine surface area of contact, the number
of repeated contacts during a worker's shift, task duration, or
worker activities for particular NMP OESs and COUs.
Therefore, parameters estimated by this AIHA model are less
useful for EPA's risk evaluation.
To address dynamic loading on the skin (i.e., where
deposition is defined by an amount/area/time deposited) or
exposure to very thin films would require significant revision
of the PBPK model. It is likely that for a film on exposed skin
on the order of microns of thickness (1.02 mg/cm2 is
equivalent to a layer 10 [^m thick), evaporation will become a
significant factor, with that evaporation being temperature
dependent. A film on exposed skin will be simultaneously
warmed by body heat and cooled by evaporation. The U.S.
EPA is not aware of a PBPK model of dermal exposure that
accounts for the complex interplay of these factors; i.e., such
a model is not in the realm of available science. On the other
hand, if NMP penetrates under a protective glove, that film
would not be subject to evaporation and EPA is not aware of
science to indicate that absorption would vary as a function of
the film thickness, as long as it was present. Therefore, EPA
considered two options: making the best possible use of the
Poet et al. PBPK model (with minor corrections) or
performing the risk assessment without a PBPK model. The
use of the PBPK model under the assumption that the
exposed skin is effectively immersed in NMP was considered
the preferred option, making use of the best available science,
despite its limitations.
Page 39 of 205
-------
Unlike EPA estimates of contact durations, the Cardno
ChemRisk analysis equates evaporation time to contact
durations, which does not account for extended, continued
contact or repeated contacts over a shift.
SACC
SACC COMMENTS:
SACC expressed concern that glove protection factor
(GPF) assumptions made by EPA were overly optimistic
and unsubstantiated. The adequacy of the ECETOC TRA
model was questioned. The references cited do not
support an assumption that workers consistently wear
appropriate chemical-resistant gloves. It is questionable
whether worker training on glove use is routine and
adequate. Some workers fail to use gloves even after
training. Some scenarios (e.g., soldering) could affect the
integrity of gloves. Chemically resistant gloves degrade
with age, even over the course of hours, and may not be
changed out appropriately. Because improper glove use
can make exposures worse (e.g., Rawson et al., 2005),
GPFs <1 should be considered at the lower bound.
Recommendations: Consider reducing the assumed GPFs
used in the NMP draft risk evaluation, given uncertainty
regarding worker training and glove material selection.
Provide greater justification for adoption of specific GPF
values. Adopt language such as "No unreasonable risk is
found if a GPF of X is achieved," leaving room for
uncertainty as to whether that outcome can be achieved
in practice.
Regarding the numerical values of glove PFs, EPA has not
found reasonably available data or methods to improve upon
the ECETOC TRA model. Therefore, EPA retains this model
and its method and values. In Section 2, EPA has removed
assignments/ assumptions of specific glove PFs to apply to
each OES. To the extent that scenario-specific information on
glove use is available, it is described in Section 2.4.1.2. Table
2-77 has been updated to include worker exposures for all
glove PFs for all OESs.
EPA agrees that improper glove use can make exposures
worse. However, EPA clarified in Section 2.4.1.1 that its
approach uses glove PFs to reduce skin surface area.
Therefore, using PF < 1 would increase surface area above
EPA's assumed values, which is not the likely effect of
improper glove use. Also, EPA found no reasonably available
approaches or method toward quantifying PF < 1. Improper
glove use would create conditions closer to occlusion and
would be more likely to increase contact duration as noted in
the Supplemental File on Occupational Exposure Assessment.
Assuming longer contact durations would be a better
approach for improper glove use.
For the purpose of this risk evaluation, EPA makes
assumptions about potential personal protective equipment
(PPE) use based on reasonably available information and
expert judgment. EPA considers each condition of use and
constructs exposure scenarios with and without engineering
controls and /or PPE that may be applicable to particular
Page 40 of 205
-------
worker tasks on a case-specific basis for a given chemical.
Again, while EPA has evaluated worker risk with and without
PPE, as a matter of policy, EPA does not believe it should
assume that workers are unprotected by PPE where such PPE
might be necessary to meet federal regulations, unless it has
evidence that workers are unprotected. For the purposes of
determining whether or not a condition of use presents
unreasonable risks, EPA incorporates PPE use based on
information underlying the exposure scenarios. These
assumptions are described in the unreasonable risk
determination for each condition of use, in Section 5.2.
Additionally, in consideration of the uncertainties and
variabilities in PPE usage, including the duration of PPE
usage, EPA uses the high-end exposure value when making
its unreasonable risk determinations in order to address those
uncertainties. EPA has also outlined its PPE assumptions in
Section 5.1 and EPA's assumptions are described in the
unreasonable risk determination for each condition of use, in
Section 5.2.
SACC,
54, 56,
57
SACC COMMENTS:
Describe how default GPFs are assigned when the type
and appropriateness of glove type and proper use or other
PPE use are not known.
Use of empirical glove permeability data could
strengthen the risk evaluation relative to use of
hypothetical GPF. Crook and Simpson (2007) published
the results of an NMP glove permeability study that was
not cited in the draft risk evaluation.
Committee members questioned the decision to apply the
same GPF to all scenarios. Industry-specific protections
factors should be considered.
The Committee recommended that EPA lower the
assumed GPF to 5 for the following exposure scenarios:
In Chapter 2 and in the Supplemental File on Occupational
Exposure Assessment, EPA has removed assignments/
assumptions of specific glove PFs to apply to each OES.
Table 2-77 in Section 2.4.1.3 has been updated to include
worker exposures for all glove PFs for all OESs.
EPA states in Section 2.4.1.1 that, as indicated in Table 2-3,
use of PFs above 1 is recommended only for glove materials
that have been tested for and shown to be effective for
preventing permeation of the NMP-containing liquids
associated with the condition of use. Therefore, EPA has
included consideration of permeation/ efficacy/ effectiveness
by considering use of PFs of 5, 10, and 20.
Page 41 of 205
-------
Section 2.4.1.2.1 Manufacturing; 2.4.1.2.2 Repackaging;
2.4.1.2.3 Chemical processing, excluding formulation;
2.4.1.2.4 Incorporation into Formulation, mixture or
reaction product; and 2.4.1.2.8. Electronic parts
manufacturing.
Assumptions underlying assigned GPFs were difficult to
follow for individual scenarios, in both the draft risk
evaluation and Supplemental File.
Recommendation: State (and display in tabular form)
expected glove use and GPF assumed for each condition
of use scenario and include an assessment of associated
uncertainty.
PUBLIC COMMENTS:
EPA should incorporate NMP-specific data on glove
permeation into its risk assessment to provide a more
accurate characterization of their impact on internal dose
and risk.
EPA was provided with information about the efficacy of
different glove materials for reducing potential hazards
from NMP-containing paint strippers in a July 2015
report: "Assessment of the Efficacy of Different Glove
Materials for Reducing Potential Hazards Associated
with NMP Containing Paint Strippers." It was not
apparently considered for the draft risk evaluation, nor
was it put into the public docket as the Group requested.
To ensure that this important information is available and
included in the final risk evaluation, the report will soon
be published as open access in the Journal of Exposure
Science and Environmental Epidemiology.
In Appendix E, EPA presents information gathered in support
of understanding glove use for handling pure NMP and for
paint and coatings removal using NMP formulations. EPA
states in Section 2.4.1.1 that this information in Appendix E
may be generally useful for a broader range of uses of NMP
and is presented for illustrative purposes. EPA has
incorporated NMP-specific data on glove permeation,
including information and data from the Crook and Simpson
(2007) studv. in this appendix.
In Section 4.2.2 of the risk evaluation, EPA presents risk
estimates for occupational exposures both with and without
glove use (glove PFs 1, 5, 10, and 20) for each occupational
exposure scenario. Table 4-54 presents risk estimates with
and without gloves for all conditions of use.
61
PUBLIC COMMENTS:
Notably, "EPA has not found information that would
indicate specific activity training (e.g., procedure for
glove removal and disposal) for tasks where dermal
EPA includes high-end scenarios with PF = 1 to account for
scenarios in which workers are not provided protective
gloves, are provided inadequate gloves, or are not adequately
trained.
Page 42 of 205
-------
exposure can be expected to occur in a majority of sites
..." EPA must therefore consider the foreseeable
exposure scenarios in which employees are not provided
protective gloves, or, worse, are provided inadequate
gloves or are not adequately trained and thus face even
greater dermal exposures due to glove contamination and
occlusion of NMP close to the skin.
EPA assumes that any NMP on the skin is "removed by
cleaning at the end of the work period." EPA offers no
evidence that all workers actually do clean their hands
and other exposed body parts following each shift, nor
that facilities are available for them to do so. In the
absence of such cleaning, dermal exposure durations -
and associated risks - will be greater than those
estimated by EPA.
EPA must consider the fact that clothing can absorb
NMP liquids and/or vapors. As many workers return
home in the same clothes they were wearing at work, this
absorption creates that potential for additional exposures
that EPA has not addressed in either of its draft risk
evaluations.
EPA has clarified in Section 2.4.1.1 that it is assumed that
workers usually clean their exposed skin following each shift.
EPA did not find reasonably available information on
prevalence of or facilities for cleaning or that dermal contact
with liquids will exceed EPA estimates.
The frequency and magnitude of take-home exposure is
dependent on several factors, including personal hygiene and
visibility of the chemical on skin or clothing. EPA does not
have methods to reliably predict take-home exposures
associated with returning home from work in potentially
contaminated clothing.
Dermal exposure - Vapor-through-skin
SACC
SACC COMMENTS:
Discussion of dermal exposures via NMP vapor that
penetrates clothing fabrics, and direct skin contact with
clothing saturated with NMP vapor, along with
associated uncertainties, should be included in the
evaluation.
EPA has included discussion in Uncertainties Sections 2.4.1.4
and 4.3 that dermal exposures to NMP vapor that may
penetrate clothing fabrics and the potential for associated
direct skin contact with clothing saturated with NMP vapor
are not included in quantifying exposures. The discussion
further notes that these uncertainties could potentially result
in underestimates of exposures.
33, 57
PUBLIC COMMENTS:
With respect to the dermal vapor pathway, there is clear
evidence that this pathway is important in humans since
the combined contributions from inhalation and dermal
EPA has included the dermal vapor pathway, which EPA
refers to as the vapor-through-skin route. The PBPK model
accounts for inhalation exposure, dermal exposure to liquid,
and dermal exposure to vapor. EPA discusses vapor-through-
Page 43 of 205
-------
absorption of vapor (when wearing trousers and short-
sleeved shirts) to the internal dose were 1.5- to 1.7-fold
higher than that from inhalation alone (Bader et al.,
2008).
A fuller characterization of exposure pathways should be
conducted for whole-body exposures to humans {i.e.,
dermal exposure to NMP vapor is explicitly included).
skin exposure in Sections 2.4.1, 3.2.2, and 3.2.5.5 of the risk
evaluation. EPA includes discussion of Bader et al. (2008) in
Section 3.2.5.5. EPA considers the current characterization of
pathways to be clear and adequate.
54, 56
PUBLIC COMMENTS:
For the vapor-through-skin route of exposure, EPA
assumed that workers wore short-sleeved shirts and long
pants. EPA assumed that the head, arms, and hands are
entirely exposed unless PPE is worn. Together, the
fractional skin area exposed to vapor (SAVC) is 25% of
the total skin surface area in the absence of PPE or liquid
dermal contact (lines 4774-4781). Information submitted
to EPA shows that the practice in industrial settings, such
as electronic part manufacturing, is for complete
coverage of head, torso, legs, arms, and hands.
Assumptions regarding skin exposure for the vapor-
through-skin route of exposure should reflect actual
industry practice in the use of PPE.
The PBPK model accounts for reduction in skin surface area
for vapor-through-skin exposure based on PPE usage. EPA
included additional PBPK runs for semiconductor fab
workers assuming 98% skin coverage to supplement runs that
assume 75% skin coverage. EPA has included discussion in
Uncertainties Sections 2.4.1.4 and 4.3 that dermal exposures
to NMP vapor that may penetrate clothing fabrics and the
potential for associated direct skin contact with clothing
saturated with NMP vapor are not included in quantifying
exposures. The discussion further notes that these
uncertainties could potentially result in underestimates of
exposures.
Inhalation exposure
SACC
SACC COMMENTS:
In the draft risk evaluation, three pathways (vapor
inhalation, dermal absorption from liquid, and dermal
absorption from vapor) were assessed. One Committee
member suggested that aerosol inhalation should be
investigated as well.
EPA modeled exposures to NMP in aerosols in the
commercial automotive servicing OES. Also, EPA states in
Section 2.4.1 that inhaled vapor/mist/dust will not be
considered as an inhalation exposure because EPA does not
have reasonably available data or methods to fractionate the
total NMP inhaled into the amount of NMP that deposits in
the upper respiratory system and the amount of NMP that
enters the lung. EPA considers aerosol to be essentially
equivalent to mist.
SACC
SACC COMMENTS:
EPA has added to 2.4.1.1 that EPA has modeled inhalation air
concentrations for workers in 11 of 16 OESs and far-field
Page 44 of 205
-------
Provide justification for not exploring a range of
available inhalation exposure models for estimating
occupational exposures.
inhalation air concentrations for ONUs in 1 of 16 OESs. EPA
has exhausted all modeling opportunities with the data that
are reasonably available and therefore was unable to model
inhalation air concentrations for workers in the remaining 5
OESs and far-field inhalation air concentrations for ONUs in
the remaining 15 OESs.
A^rc^Jitc exposure
SACC
SACC COMMENTS:
Recommendation: Estimate aggregate exposures, since
some occupational workers may have both dermal and
inhalation exposures that are non-negligible, and discuss
the relative contributions of each exposure pathway to
total exposures.
PBPK modeling for workers estimates aggregate exposure
and accounts for dermal, inhalation, and vapor-through-skin
routes.
In Section 2.4.1.1, EPA added discussion that the relative
contributions of each exposure pathway to total exposures
varies according to parameter values for NMP weight fraction
in the liquid product contacted, skin surface areas in contact
with the liquid product and with vapor, durations of dermal
contact with liquid product and with vapor, air concentration
for inhalation and vapor-through-skin exposure, body weight
of the exposed person, and glove protection factor and
respirator assigned protection factor (if applicable). In
scenarios where the three parameters involving dermal
contact with liquid product (NMP weight fraction in the
liquid product contacted, skin surface areas in contact with
the liquid product and with vapor, durations of dermal contact
with liquid product) have relatively high values, this route can
be the dominant route for worker exposures.
Exposure monitoring data
51,34,
49, 61
PUBLIC COMMENTS:
EPA acknowledged it "only found inhalation monitoring
data for the use of NMP in semiconductor
manufacturing" and had no data at all regarding the use
of NMP in manufacturing lithium ion batteries (NMP
Risk Evaluation, p. 100) or any "inhalation monitoring
data specifically related to the use of NMP-based
EPA revised the occupational exposure assessment in the risk
evaluation to separately assess occupational exposure
scenarios associated with three categories of electronic part
manufacturing: lithium ion battery manufacturing
(2.4.1.2.15); Other electronics manufacturing, including
capacitor, resistor, coil, transformer, and other inductor
manufacturing (2.4.1.2.9); and semiconductor manufacturing
Page 45 of 205
-------
soldering products." EPA cannot validly make an
"unreasonable risk" determination about the use of NMP
in lithium ion battery manufacturing, given the Agency's
very limited knowledge about how NMP is used in
lithium ion batteries and the substantial differences
between lithium ion battery manufacturing and
semiconductor manufacturing. With regard to soldering,
EPA assumed, without any supporting data, that "most
NMP may be destroyed in the soldering process,
mitigating the potential for significant inhalation
exposures."
(2.4.1.2.10). In these separate OESs, EPA revised and
expanded PBPK runs for industry-specific work activities
using industry-specific air concentration data sets provided in
public comments for the lithium ion battery manufacturing
industry, for the semiconductor manufacturing industry, and
from the OSHA data set for capacitor, resistor, coil,
transformer, and other inductor manufacturing (LICM. 2020a;
Semiconductor Industry Association. 2020. 2019b, c; OSHA..
2017).
EPA revised the Soldering OES in Section 2.4.1.2.12 of the
risk evaluation to assess potential inhalation and vapor-
through-skin exposures to NMP during soldering by
including air monitoring data form a surrogate activity.
Conditions of use - Formulating
50
PUBLIC COMMENTS:
Fujifilm Holdings America Corporation submitted
worker exposure data, including air monitoring and
manufacturing/handling data on NMP (see below) and
requested that EPA consider that, under these existing
industrial user conditions, risk to the worker has been
minimized, and therefore, handling and use of NMP does
not present an unreasonable risk and does not warrant
use restrictions that would prevent these successful
operations from continuing.
Fujifilm is regulated by multiple federal, state, and local
agencies for compliance with both worker and
environmental safety. In addition, Fujifilm utilizes Best
Management Practices, including PPE as needed to
protect the operator against exposure via dermal, oral, or
respiratory routes as a required practice. Mandatory
worker protection has long been in place and the wearing
EPA revised the air concentration inputs in the Formulation
OES in Section 2.4.1.2.4 to incorporate these data provided
bv Fujifilm (2020).
EPA revised the worker activities Section 2.4.2 of the
Supplemental File on Occupational Exposure Assessment to
include information from Fujifilm that their workers are
required to have chemical hygiene training prior to handling
NMP. EPA also included additional PPE and engineering
control information from Fujifilm in this same section of this
supplemental file.
Page 46 of 205
-------
thereof is a condition of employment when handling
such chemicals.
o No employee is permitted to handle or be exposed to
these chemicals before adequate chemical hygiene
training.
o In addition, training and rigorous adherence to the
procedures is required at all times,
o Fujifilm requires the use of long-sleeve rubber
gloves, which are impervious to the chemicals. This
means that there is no transmittance of any chemical
through the gloves to the skin,
o For further protection, the purpose of the extended
glove length is for incidental contact protection.
Note: for reference, the gloves in use are Showa
brand #3416: neoprene-coated, 15-mil thickness, and
14-inch gauntlet cuff with interlock knit lining,
o The manufacturing and use of products that contain
NMP does not pose an unreasonable risk - inhalation
hazards are minimal due to the chemical's low vapor
pressure (0.29-0.32 mm/Hg @ 68F) and it is not
processed at elevated temperatures,
o In addition, for skin and eye exposure, employees do
not have direct contact with the chemical when
processing and packaging,
o In the case of liquid splashing or spilling, skin and
eye exposure and contact is prevented by the required
use of safety glasses and impermeable nitrile gloves.
Shower rooms and uniform changes are available if
needed.
o Respirator use will vary from operation to operation
dependent upon the type ventilation systems the
customers have employed. The laminating clears and
adhesives are used primarily for credit card
Page 47 of 205
-------
production industry and NMP would not be present
in the end product.
For open process batch blending: Maximum batch size of
300 gallons containing a concentration of approximately
10-25% NMP. Annual usage is approximately 21,000
lbs. During this process:
o PPE: Single use impervious nitrile gloves and safety
glasses to prevent dermal exposure,
o Ventilation: Facility uses localized ventilation to
minimize inhalation exposure,
o Industrial hygiene monitoring (air sampling):
Because of the chemical's low vapor
pressure/volatility and the localized ventilation
employed, air monitoring has not been necessary as
we do not anticipate detectable amounts. This is
additionally based on data from testing at another
Fujifilm facility - Carrollton, Texas. Occupational
Safety and Health Administration (OSHA)
permissible exposure limits (PELs)/time-weighted
average (TWA)/short term exposure limits (STELs)
are not established for NMP.
o Suppliers: Supplied by U.S.-based vendors in 55-
gallon drums as 100% NMP or in a resin mixture,
o Waste disposal: Waste residue from cleaning
equipment and off-specification product is disposed
of at a licensed treatment, storage, and disposal
facility where it is used as a fuel supplement for
energy recovery.
Conditions of use - Soldering
34, 61
PUBLIC COMMENTS:
EPA proposes a finding that an estimated 4 million
workers exposed to NMP from soldering face no
unreasonable risks. This finding, based largely on the
EPA revised the Soldering OES in Section 2.4.1.2.12 of the
risk evaluation to assess potential inhalation and vapor-
through-skin exposures to NMP during soldering by
including air monitoring data from a surrogate activity.
Page 48 of 205
-------
erroneous assumption that all of those workers would
have access to and use PPE, is unwarranted.
For the purpose of this risk evaluation, EPA makes
assumptions about potential PPE use based on reasonably
available information and expert judgment. EPA considers
each condition of use and constructs exposure scenarios with
and without engineering controls and /or PPE that may be
applicable to particular worker tasks on a case-specific basis
for a given chemical. Again, while EPA has evaluated worker
risk with and without PPE, as a matter of policy, EPA does
not believe it should assume that workers are unprotected by
PPE where such PPE might be necessary to meet federal
regulations, unless it has evidence that workers are
unprotected. For the purposes of determining whether or not a
condition of use presents unreasonable risks, EPA
incorporates assumptions regarding PPE use based on
information and judgment underlying the exposure scenarios.
These assumptions are described in the unreasonable risk
determination for each condition of use, in Section 5.2.
Additionally, in consideration of the uncertainties and
variabilities in PPE usage, including the duration of PPE
usage, EPA uses the high-end exposure value when making
its unreasonable risk determinations in order to address those
uncertainties. EPA has also outlined its PPE assumptions in
Section 5.1 and EPA's assumptions are described in the
unreasonable risk determination for each condition of use, in
Section 5.2
Conditions of use - Lithium ion battery industry
49, 56
PUBLIC COMMENTS:
EPA grouped lithium ion battery manufacturing and
lithium ion cell production under the much broader
category of "Use in Electronic Equipment, Appliance,
and Component Manufacturing," which includes the use
of NMP for "cleaning of electronic parts, coating of
EPA revised the occupational exposure assessment in the risk
evaluation to separately assess occupational exposure
scenarios associated with three categories of electronic part
manufacturing: lithium ion battery manufacturing
(2.4.1.2.15); Other electronics manufacturing, including
capacitor, resistor, coil, transformer, and other inductor
Page 49 of 205
-------
electronic parts, including magnet wire coatings, and
photoresist and solder mask stripping," in addition to
"lithium ion battery manufacturing." In so doing, EPA
has incorrectly assumed that exposure, engineering and
workplace controls, job descriptions, and uses of NMP
are substantially similar in all types of manufacturing in
this category. In fact, there are few similarities between
the use of NMP in manufacturing lithium ion batteries
and the other uses of NMP covered by this category (see
details below). EPA evaluated little, if any, information
about the specific use of NMP in the manufacturing of
lithium ion batteries. The exposure scenarios in the draft
risk evaluation are based solely on information that EPA
gathered about semiconductor manufacturing, which are
not applicable to lithium ion battery manufacturing.
EPA cannot validly apply generic exposure assumptions
and make an "unreasonable risk" determination about the
use of NMP in "lithium ion battery manufacturing,"
given the Agency's very limited knowledge about, and
understanding of, how NMP is used in lithium ion
batteries and the substantial differences between
potential exposure scenarios in lithium ion battery
manufacturing and semiconductor manufacturing.
Lithium ion battery manufacturing should be considered
as a separate condition of use.
EaglePicher disagrees with EPA's determination that the
use of NMP in manufacturing lithium ion batteries
presents an unreasonable risk of injury to workers.
EPA's assessment must include consideration of
engineering controls (described below). If risks are
preliminarily identified, EPA then must consider whether
the risk is mitigated by the use of PPE, and if so, no
finding of unreasonable risk is warranted.
manufacturing (2.4.1.2.9); and semiconductor manufacturing
(2.4.1.2.10). In these separate OESs and where feasible, EPA
revised and expanded PBPK runs for industry-specific work
activities using some industry-specific PBPK input data and
information provided in public comments for the lithium ion
battery manufacturing industry (EaglePicher Technologies.
2020j. b; i ซA \ t 2020a. b, c) and for the semiconductor
manufacturing industry ("Intel Corporation. 2020;
Semiconductor Industry Association. 2020; Intel Corporation.
2019; Semiconductor Industry Association. 2019a. b, c).
EPA's risk conclusions take PPE use into account. In risk
characterization, EPA calculated risks for each occupational
worker COU with and without various levels of PPE
protection. The tables in Section 4 show the extent to which
protective gloves and masks mitigate risks.
Page 50 of 205
-------
49, 56
PUBLIC COMMENTS:
The semiconductor-related NMP activities described
above as the basis of EPA's risk assumption are very
different from those occurring in the lithium ion
cell/battery industry.
o This industry does not engage in the unloading of
trucks containing virgin NMP; it only receives sealed
containers of virgin NMP and pre-mixed binder-
NMP solution, with no associated exposure risks,
o This industry does not load trucks with waste NMP,
thus eliminating any related risks, or use NMP for
"coating of electronic parts or "stripping" of any
kind.
o At no time do any employees come into direct dermal
contact with NMP, and the duration of contact is
considerably less.
ฆ The handling of NMP in small containers in cell
manufacturing facilities is limited to infrequent
use in the laboratory or small-scale operations
where they are opened only in ventilated hood
areas with personnel equipped with extensive
PPE (e.g., Figure 8) for no more than 30 minutes
a shift (even in these operations, mixing and
further processing takes place in fully enclosed
systems).
o Employees wear PPE with an assigned protection
factor of 1,000 that precludes inhalation or dermal
contact.
o The total volume of NMP used by EaglePicher
Technologies, LLC, is small. Their Joplin facility
uses <1,100 kg annually. At the East Greenwich
facility, the annual volume is only 800 kg.
EPA assessed lithium ion cell manufacturing work activities
indicated by the information in public comments
(EaelePicher Technologies. 2020b; 1 2020a. b). These
activities include: Container handling, small containers;
Container handling, drums; Cathode coating; Cathode
mixing; Research and development; and, Miscellaneous.
To supplement shift-based contact duration estimates, EPA
used the task duration estimates in the public comments to
assess what-if (task duration-based) PBPK runs for the
lithium ion cell manufacturing.
EPA included summaries of relevant details provided in the
public comments, particularly process description and PPE
information specific to the lithium ion cell manufacturing
industry, in Section 2.15 (Electronics Part Manufacturing
OES) of the Supplemental Information on Occupational
Exposure Assessment document.
EPA thanks the commenter for providing information on the
EPA-issued consent order and Significant New Use Rule
(SNUR) for the use of cathode powders and cathode mixing
in lithium ion cell manufacturing. These consent orders and
SNURs were issued for chemicals that are not NMP.
Page 51 of 205
-------
Lithium ion cells are produced in a tightly controlled
manufacturing environment and closed pipe systems are
used for NMP transfer.
o The engineering controls and PPE the industry
employs are expressly designed to prevent worker
exposure to NMP.
ฆ Because of purity concerns alone, nowhere in a
commercial lithium ion cell manufacturing
process are workers expected to immerse their
hands in NMP or NMP-based slurries - with or
without proper PPE. Non-routine operations such
as maintenance activities or recovery from
process upsets require the use of PPE because, in
the absence of PPE, they could put workers in
contact with NMP outside of established
engineering controls.
ฆ Access to cathode mixing, coating, and drying
areas where NMP is used is tightly controlled.
o Personnel working in mixing and coating areas
receive extensive training regarding the processes
and proper PPE. Standard Operating Procedures
(SOPs) are utilized for routine and non-routine tasks
and specify training and required PPE. Personnel
entering these areas for routine work must undergo a
gowning procedure for quality and safety purposes
that includes donning in-process safety shoes, Tyvek
suits, nitrile gloves, safety glasses, hairnet, and mask.
These PPE are not intended for operations involving
intentional contact with NMP or NMP-based slurries,
o Additional PPE is provided for work that will involve
contact or potential contact with NMP and includes
chemical resistant suits, respirators, and chemical-
resistant gloves, depending on the task performed.
Page 52 of 205
-------
These workers are required to wear full-body
chemical resistant suits with booties/shoe covers. The
equipment includes a PAPR and hood with
organic/acid gas + high-efficiency particulate air
cartridge coverage. Gloves are required, typically
double latex for limited contact with NMP. Butyl
gloves are required when contact with NMP is
expected.
o For the purposes of ensuring worker health and
safety, exposure risk assessments are conducted to
verify the efficacy of engineering and administrative
controls.
o In all cases, lithium ion cell manufacturers strive to
control every material in the NMP pathway,
including the metals and other materials used in
piping, valves, and mixing and coating equipment.
Any contact with workers and their PPE is avoided
whenever possible,
o NMP recovery systems are fully automated, closed
systems. Only maintenance workers with prescribed
PPE interact with these systems. Maintenance
procedures are conducted only on de-pressurized
systems. This means for large operations, shipments
of virgin NMP are less frequent compared to smaller
production facilities and other industries. Where
these shipments occur, and in the case of condensed
NMP liquid shipments and shipments for disposal of
degraded NMP and distillation bottoms, workers are
fully protected from potential inhalation and dermal
exposures through the use of PPE.
Finally, EPA has issued several consent orders and
associated significant new use rules for the use of
cathode powders that already require the extensive use of
Page 53 of 205
-------
PPE in cathode mixing operations during lithium ion cell
manufacturing, such as 40 CFR ง 721.11027. Therefore,
EPA-imposed PPE is already required and should be
taken into consideration in this risk evaluation.
Conditions of use - Magnet wire industrial processing
47
PUBLIC COMMENTS:
The magnet wire industry has long utilized NMP as a
solvent/ diluent in high-performance magnet wire
enamels, thinners, and cleaners. In the magnet wire
industrial process, a copper or aluminum wire is routed
through an applicator of solvent-based enamel coating.
The size of applicator may vary throughout the industry,
but most contain V2-I gallon of coating, which contains,
at most, 80-85% concentration of NMP. NMP does not
react with the other ingredients in this coating but is
simply mixed in to facilitate the smooth application of
the enamel. NMP's role here is critical since rough
application of the enamel would result in a blistered film
and ultimately cause failure of the magnet wire to create
the electro-magnetic field. After leaving the applicator,
the wet-coated wire passes through a curing oven where
the NMP evaporates from the mixture, leaving a thin
film of cured polymer on the wire. Magnet wire usually
gets several coats of enamel, each followed by a pass
through the curing oven. Once finished, there is no NMP
exposure risk to the end-user of which the National
Electrical Manufacturers Association (NEMA) is aware
under normal conditions of use.
0 The applicator used for solvent-based enamel coating
contains, at most, an 80-85% concentration of NMP.
0 The curing process occurs in ovens that are
completely enclosed, and there is no human exposure
to NMP during this process.
EPA updated the process description information and PPE
information for the magnet wire coating process in Section
2.9 (Other Electronics Manufacturing OES) of the
Supplemental Information on Occupational Exposure
Assessment document.
As described in Section 1.4.2 of the risk evaluation, EPA did
not evaluate exposures via the drinking water pathway or
NMP land releases to underground injection, RCRA Subtitle
C hazardous waste landfills, or RCRA Subtitle D municipal
solid waste (MSW) landfills in this risk evaluation. These
exposure pathways fall under the jurisdiction of other EPA-
administered statutes and associated regulatory programs.
During problem formulation, EPA performed a first-tier
screening analysis of risks from ambient air, ambient water,
sediment, and land-applied biosolids. EPA did not identify
risks from human or environmental exposures that may result
from these pathways, including inhalation of outdoor air
containing NMP released from industrial and commercial
facilities.
Page 54 of 205
-------
o Any vapor emitted during application moves directly
into the curing oven, wherein at least 90% of the
NMP combusts,
o The finished product contains only trace amounts of
NMP due to the curing process previously described.
Another facet to magnet wire manufacturing is
maintenance cleaning. Due to regular, widespread
industry application of preventive safety measures
(described below), use of NMP as a solvent for
cleaning/degreasing operations in magnet wire facilities
DOES NOT present an 'unreasonable risk' to workers
under these conditions of use, as per TSCA Section
6(b)(4)(A).
o Enameling equipment is bathed in agitated tanks of
NMP. These tanks range in size but are commonly
around 50 gallons,
o This process is completely enclosed while equipment
is cleaned. When the cycle completes, the operator
retrieves the equipment by opening the tank lid,
which prompts the basket to rise up and drains the
NMP back into the tank,
o Emissions consist of evaporation from the NMP bath
and from the cleaned parts removed from the bath,
o Human exposure to NMP is controlled through the
use of PPE such as gloves, aprons, and goggles, as
well as engineering controls.
NMP losses to the environment are limited by strict
controls on air emissions through the combustion process
mentioned prior. To be sure, some NMP vapors may be
emitted, for example, during equipment cleaning. EPA
has estimated annual emissions from a magnet wire
operating line without an incinerator at 84 Mg/year.
Almost all operating lines now have an incinerator, so
Page 55 of 205
-------
the actual amount is expected to be much less.
Any liquid waste NMP and/or solid waste wet with NMP
(paper, plastic, rags, etc.) are handled in compliance with
the Resource Conservation and Recovery Act (RCRA).
In the event that small amounts of NMP do leak into the
environment, EPA recognized that NMP has low hazard
for ecological receptors and low persistence if released
into aquatic or terrestrial environments.
Conditions of use - Semiconductor industry
52, 64,
31
PUBLIC COMMENTS:
NMP is used in semiconductor manufacturing as a
solvent or to remove residue from product wafers.
Semiconductor manufacturing involves the fabrication of
circuits that are typically <100 nanometers in dimension.
The process of manufacturing advanced semiconductors
takes place in highly advanced and complex fabrication
plants ("fabs") and requires exceptionally precise and
controlled manufacturing equipment and processes. Such
processes occur within equipment, which is, by design,
intended to isolate the manufacturing process and
chemicals from workers. Modern semiconductor
manufacturing tools are enclosed, ventilated, and
automated, thus preventing worker exposure. Under
these near pristine and highly controlled conditions, there
are no unreasonable risks to workers attributable to
exposures to NMP.
SIA has provided extensive information to EPA on the
industry's practices and procedures for handling NMP,
including meeting with EPA officials in November 2017
to summarize the conditions of use of NMP at
semiconductor fabs and hosting a group of EPA officials
in February 2019 to tour a semiconductor fab of a
member company to provide a first-hand understanding
EPA updated the process description and PPE information for
semiconductor manufacturing in Section 2.10 (Semiconductor
Manufacturing OES) of the Supplemental Information on
Occupational Exposure Assessment document.
To supplement shift-based contact duration estimates, EPA
updated the what-if (task duration-based) work activities for
semiconductor manufacturing based on the task duration
estimates provided by SIA in public comments
(Semiconductor Industry Association. 2020, 2019a).
EPA updated the central tendency and high-end NMP weight
fractions for the semiconductor work activities using values
provided by the SIA in these public comments.
EPA added several PBPK runs for semiconductor fab workers
assuming 98% skin coverage to supplement runs that assume
75% skin coverage. These runs are available in the
Supplemental Excel File on Occupational Risk Calculations.
For any particular male Fab worker or Fab ONU activity, the
differences in AUC internal concentrations obtained by
varying only the assumed whole-body skin coverage between
75%) and 98%> but no other parameter variation was found to
be less than 1% in EPA's anecdotal comparison. Therefore,
Page 56 of 205
-------
of the use and handling of chemicals at a fab. SIA
submitted details on task durations and frequencies,
which showed that task durations are short and human
exposures are accordingly time-limited. SIA has
described its engineering controls and chemical handling
procedures to EPA in presentations and in written
documentation submitted to EPA. These procedures are
designed explicitly to prevent any dermal contact with
liquid NMP or other potential forms of residual NMP.
Several submissions to EPA describe the risk
management measures implemented at fabs, including
depictions and descriptions of PPE worn by workers to
minimize the potential that they might come in contact
with NMP (e.g., specific documented procedures for
selecting the proper gloves for a particular chemical and
task and for donning and removing the gloves), as well
as information concerning the structure and operations of
fab facilities, which are designed to largely eliminate
opportunities for any human contact with wafers and the
chemicals used within semiconductor manufacturing
equipment. This information clearly demonstrates that
fab workers have minimal opportunities for direct
exposures to NMP.
Workplace practices at Intel (listed below), and those
that are common in the semiconductor industry,
successfully mitigate the risk of worker exposures to
NMP.
o Intel requires work controls, such as flushing filters
prior to filter changes, draining of NMP-containing
sinks, use of tools to retrieve parts, and use of wipes
in addition to chemical-resistant gloves, to minimize
contact with NMP during maintenance.
the skin coverage assumption does not appear to significantly
impact the AUC internal concentration estimates.
Page 57 of 205
-------
o Container changeouts are performed in a dedicated
bulk chemical room and dip tubes are utilized for
easy container changes. Minimal chemical is exposed
during the dip tube change process (potential for a
few drops cleaned with a wipe),
o Employees who perform work in areas or during
tasks where they may be exposed to NMP are
required to use chemical resistant PPE. The current
PPE required includes MAPA Trionic 514+ (or
equivalent) chemical-resistant gloves, chemical
resistant gowns, and eye and face protection. PPE
assessments are performed and documented by
Intel's Environmental Health & Safety professionals
prior to use of NMP.
o In addition, semiconductor fab workers wear long-
sleeved coveralls with hoods and boots as well as
gloves and safety glasses that provide >98% skin
coverage. PPE such as chemical-resistant aprons and
gloves, face shields, and respiratory protection are
used when necessary to further reduce worker
exposure. Clothing and PPE provide >96% skin
coverage for workers performing NMP-related tasks
outside the fab.
o Employees working with NMP are required to take
documented safety training to ensure that they are
qualified to perform tasks and can don and doff and
dispose of PPE safely. In addition to the PPE
training, equipment-specific training and chemical
safety training is required for employees who work
with NMP and/or in areas where NMP is used,
o Intel uses multiple engineering and administrative
controls to successfully reduce the risk of exposure to
NMP and other chemicals in the workplace.
Page 58 of 205
-------
Engineering controls include the use of bulk
chemical delivery to reduce manual handling of
chemical containers, lockout/tagout to turn off bulk
chemical supplies prior to equipment maintenance,
flushing of filters and housing prior to work,
integrated local ventilation exhaust on manufacturing
equipment, liquid leak detection systems, use of
ventilated parts clean sinks or hoods for parts
cleaning, and emergency machine off systems.
Administrative controls include the use of tools, parts
clean baskets, and wipes to handle any chemical or
chemical-contaminated parts in order to reduce
contact with chemicals (chemical-resistant PPE is
still required), immediate bagging of contaminated
parts and waste, and a prohibition on immersion of
gloves in liquid chemicals.
52,31,
64
PUBLIC COMMENTS:
Monitoring data (described below) submitted by SIA to
EPA confirm that worker exposure at semiconductor fabs
is minimal and presents no unreasonable risks to human
health. SIA submitted contemporary information on
worker exposure at semiconductor fabs collected by
member companies (see SIA's 2019 workplace exposure
study at Appendix A). These data included 118 air
sampling datasets and details regarding the durations and
frequencies of tasks undertaken by workers in fab
facilities.
o Only 5 of the 118 personal air samples that SIA
member companies collected showed concentrations
of NMP above the limits of detection (LODs). Three
of the five samples (0.01, 0.02, and 0.07 ppm) were
for fab maintenance tasks. Two of the five were for
waste truck load/virgin NMP truck offload - tasks
EPA updated the air concentrations for semiconductor
manufacturing work activities and did not adjust the duration-
adjusted air concentrations to normalize to contact duration
estimates due to the high number of non-detect values. The
air concentration values used by EPA are very similar to
those proposed by SIA in the public comment
(Semiconductor Industry Association. 2020) with their
proposed input values for PBPK runs. Frequency of truck
unloading is accounted for in the analysis by modeling only
acute and not chronic exposures for this work activity.
Page 59 of 205
-------
that occur at many industrial sites and that are not
specific to semiconductor manufacturing where
measured exposures were <0.4 and 1.2 ppm.
o Of the 5 measured samples that did have NMP
concentrations above LOD, the highest 8-hour TWA
concentration was 1.18 ppm for tanker truck
offloading. The virgin NMP truck offload task is
conducted once per year and corrective actions have
been identified to reduce potential exposures. The
measured exposure in this instance was only 0.18
ppm above the Cal/OSHA 1.0 ppm 8-hour TWA,
more than a factor of 3 less than the 3.5 ppm ECHA
limit, and 10 times lower than AIHA's 10 ppm 8-
hour Workplace Environmental Exposure Limit.
Intel began exposure sampling in the 1990's to evaluate
and address potential workplace exposures to NMP. As
no OSHA federal PEL exists, Intel adopted the CAL
OSHA limit of 1 ppm (8-hour TWA) in 2006. Because
Intel employees work 12-hour shifts, Intel employs a
reduced exposure limit of 0.76 ppm (extrapolated from
CAL OSHA's 8-hour PEL of 1 ppm). In 2018, with risk
assessments being performed in Europe by the ECHA for
purposes of Registration, Evaluation, Authorization and
Restriction of Chemicals (REACH) and in the United
States under TSCA, Intel conducted another thorough
review of employee use and exposure to NMP. In its
2018 updated analysis, Intel collected data from its large
manufacturing sites in the United States, Ireland, and
Israel. The 2018 data were provided to EPA as a subset
of the observations reported in the submission made by
the SIA.
o All 2018 breathing zone samples collected at Intel
were orders of magnitude below the CAL OSHA
Page 60 of 205
-------
exposure limit, and 43 out of the 44 samples
collected were below the detection limit for the
validated National Institute for Occupational Safety
and Health (NIOSH)/OSHA methods (analyses were
performed using either NIOSH 1302 or OSHA
PV2043 SOP-5, with method detection limits of 3 or
5 jag, respectively),
o The analysis included all job categories where
employees could potentially be exposed to NMP and
included the following: equipment maintenance
(working on equipment that contains NMP where
there is potential for contact with NMP), routine
operations (working in the vicinity of NMP but no
physical contact with NMP), and container
changeouts (changing bulk NMP containers).
Statistical analysis of the 2018 and past data using
the AIHAIHSTAT statistical modeling package
indicated a very low probability of exceeding Intel's
adopted exposure limit of 1 ppm (i.e., at a 95%
confidence level, the likelihood of exceeding the
CAL OSHA OEL is 3.4xl0"8).
52,31,
64
PUBLIC COMMENTS:
In the draft risk evaluation, semiconductor
manufacturing was inappropriately grouped with "Paint
additives and coating additives not described by other
codes" (p. 315) and "Solvents (for cleaning or
degreasing): Use in electrical equipment, appliance and
component manufacturing" (p. 316-317). The
semiconductor industry should be assessed separately,
rather than as part of a broader "electronics parts
manufacturing" sector.
The semiconductor manufacturing is considerably
different from these other industrial operations. EPA's
EPA revised the occupational exposure assessment in the risk
evaluation to separately assess occupational exposure
scenarios associated with three categories of electronic part
manufacturing: Lithium ion battery manufacturing
(2.4.1.2.15); Other electronics manufacturing, including
capacitor, resistor, coil, transformer, and other inductor
manufacturing (2.4.1.2.9); and Semiconductor manufacturing
(2.4.1.2.10). In these separate OESs and where feasible, EPA
used some industry-specific PBPK input data and information
provided in public comments for the lithium ion battery
manufacturing industry (EaglePicher Technologies. 2020a, b;
LICM. 2020a, b, c) and for the semiconductor manufacturing
Page 61 of 205
-------
draft risk evaluation improperly grouped semiconductor
manufacturing along with other industrial activities that
have differing conditions of use. EPA's assumption that
the practices in the semiconductor manufacturing
industry are similar to other electronics manufacturing
operations is not accurate and is inconsistent with the
Agency's risk evaluation rules. The regulations at 40
CFR 702.41(a)(5) require that risk evaluations rely on
analyses that are "suited for their intended purpose" and
"well-tailored" to enable a "technically sound
determination" concerning the conditions of use. In
evaluating the conditions of use in semiconductor
operations within the same category as other industry
sectors with different operations and conditions of use,
EPA relied on estimates of dermal exposures that greatly
exaggerated the conditions documented in the exposure
study SIA provided to EPA and incorrectly applied the
same assumptions and drew the same conclusions for all
conditions of use in this broad range of categories.
It is inappropriate and unnecessary to group the
semiconductor industry's conditions of use of NMP with
other industrial sectors when EPA had available the
information needed to better understand and more
reasonably evaluate the potential for semiconductor
workers to be exposed to NMP under the conditions of
use unique to semiconductor fabrication facilities.
Unfortunately, it appears that much of the information
and data that SIA provided were not incorporated in the
draft risk evaluation docket and may not have been
thoroughly reviewed or were only partially considered by
EPA personnel when preparing the draft risk evaluation.
This reflects a deficiency that should be corrected before
the final risk evaluation is prepared and this must be
industry (Intel Corporation. 2020; Semiconductor Industry
Association. 2020; Intel Corporation. 2019; Semiconductor
Industry Association. 2019a. b, c).
EPA reviewed all information in the public comments
provided by SIA and updated the PBPK inputs for the
semiconductor manufacturing OES work activities, including
NMP weight fractions, NMP air concentration, and task
duration (for what-if, task duration-based work activities).
In Chapter 2, EPA has removed assignments/ assumptions of
specific glove PFs to apply to each OES. Table 2-77 has been
updated to include worker exposures for all glove PFs for all
OESs. Table 2-77 in Section 2.4.1.3 has been updated to
include worker exposures for all glove PFs for all OESs.
EPA clarified in Section 2.4.1.1 that EPA has no reasonably
available information on actual surface area of contact with
liquid and that the assumed values represent adequate
surrogates for most uses' central tendency and high-end
surface areas of contact with liquid that may sometimes
include exposures to much of the hands and also beyond the
hands, such as wrists, forearms, neck, or other parts of the
body. EPA accounts for distributed values using the central
tendency and high-end assumptions for surface areas. EPA
clarified in Section 2.4 that non-immersive dermal contact
with liquid films is evaluated.
Page 62 of 205
-------
accomplished if the Agency is to meet its obligations
under Section 26 of the amended statute to consider
information that is readily available and apply a weight-
of-the-evidence approach when assessing risks. By
ignoring or undervaluing the SIA data, EPA failed to rely
on the best available information and therefore did not
apply a weight-of-the-evidence approach.
Based on the data submitted by SIA, EPA should
consider the conditions of use in the semiconductor
manufacturing sector separately from those in other
industrial activities and sectors. Specifically, EPA should
take into account:
o Duration and frequency of tasks during which
exposures to NMP can occur (e.g., truck unloading),
o PPE used during such operations,
o Engineering controls employed to minimize
exposure, and
o Sampling data indicating the extremely low potential
exposure to NMP when used in fab operations, which
take place in a controlled environment inside
manufacturing equipment and in maintenance tasks.
Task descriptions provided by SIA (2019a) show that
there are generally limited opportunities for skin contact
with NMP-containing liquid based on the work
descriptions. If EPA chooses to consider dermal
exposure, EPA should use the information provided to
reassess and refine its dermal exposure estimates
specifically for the conditions of use in semiconductor
manufacturing operations. In particular, EPA should: (1)
assign the highest level protection factors (PFs) in
modeling of the level of dermal protection provided by
gloves used in semiconductor manufacturing operations;
(2) reduce the estimated duration of potential dermal
Page 63 of 205
-------
exposures during semiconductor manufacturing
operations to no greater than 2 hours/day per individual;
and (3) employ a surface area for dermal exposure that is
substantially less than immersion of full hands.
Conditions of use - Small scale operations
45
PUBLIC COMMENTS:
EPA should consider evaluating small-scale conditions
of use separately from the bulk loading from drums and
from large-scale laboratory use. Hach manufactures
chemical reagents and instruments for water quality
analysis. NMP is an ingredient in the Hach product,
Silver 2 Reagent Solution Pillows. This laboratory
reagent is sold only in unit-dose packaging containing <5
mL solution per test. Exposures are limited by both the
small amount handled in the unit-dose package and the
required PPE. In this industry, small-scale incorporation
of NMP into a mixture occurs, in which NMP is
transferred from hand-held containers, such as a 1-gallon
container, to the mixing vessel. Fewer than 10
gallons/day of NMP are handled, and workers are
required to wear PPE. This is very different than the EPA
assumption that workers unload bulk NMP from 20
drums per hour.
EPA updated the process description in Section 2.14
(Laboratory Use OES) of the Supplemental Information on
Occupational Exposure Assessment based on this
information. For the Laboratory Use OES, EPA did not find
reasonably available data to distinguish separate PBPK input
parameters based on scale.
Conditions of use - Disposal
34, 55,
51, 61
PUBLIC COMMENTS:
The NMP draft risk evaluation does not assess exposure
or evaluate the risks associated with disposal-related
releases, including spills and accidents, as it is required
to do under TSCA. EPA states that "disposal of NMP via
underground injection is not likely to result in
environmental and general population exposures"
because such injection is regulated under SDWA and
RCRA. This is untrue (NMP is not regulated as
As described in Section 1.4.2 of the risk evaluation, EPA
believes it is both reasonable and prudent to tailor TSCA risk
evaluations when other EPA offices have expertise and
experience to address specific environmental media, rather
than attempt to evaluate and regulate potential exposures and
risks from those media under TSCA. EPA has therefore
tailored the scope of the risk evaluation for NMP using
authorities in TSCA Sections 6(b) and 9(b)(1).
Page 64 of 205
-------
hazardous waste under RCRA) and in violation of
TSCA, which expressly defines disposal as a condition
of use, requiring EPA to evaluate risks associated with
disposal, as opposed to merely assuming that those risks
will be adequately managed under other laws. EPA
provides no evidence that exposure and risk are
insignificant for NMP releases to underground injection
wells and hazardous waste landfills or that existing
regulations adequately control these pathways for
environmental release.
As noted in the NMP Problem Formulation, "NMP has
been detected in industrial landfill leachate (Danish EPA,
2015). Although it is not currently subject to any
proposed or promulgated water regulations, NMP has
been detected in wastewater (WHO, 2001) and is
included on EPA's Drinking Water Contaminant
Candidate Lists (CCL) 3 and 4 because it is a suspected
contaminant in public water systems that may require
regulation under SDWA." EPA has acknowledged that
"no [landfill] liner can be expected to remain impervious
forever," and that "even with stringent waste
management standards, waste management units may
fail, accidents may occur during transport and handling,
and ... chemicals may continue to be released and build
up in the environment." EPA must assess the
concentrations found in air and water (surface and
ground) near injection (and other disposal) facilities.
In the Problem Formulation for NMP, EPA
acknowledged that it should identify "Other groups of
individuals within the general population who may
experience greater exposures due to their proximity to
conditions of use identified in Section 2.2 that result in
releases to the environment and subsequent exposures
While NMP is not classified as RCRA hazardous waste,
RCRA Subtitle C hazardous waste landfills and RCRA
Subtitle D municipal solid waste (MSW) landfills where
NMP may be disposed are subject to regulation under
RCRA. These methods of disposal fall under the jurisdiction
of and are addressed by other EPA-administered statutes and
associated regulatory programs. Environmental disposal of
NMP via injection into Class I wells is covered under the
jurisdiction of SDWA and disposal of NMP via underground
injection is not likely to result in environmental and general
population exposures. NMP is one of 109 contaminants listed
on EPA's fourth CCL. Because the drinking water exposure
pathway for NMP is being addressed under the regular
analytical processes used to identify and evaluate drinking
water contaminants of potential regulatory concern for public
water systems under SDWA, EPA has not included this
pathway in the risk evaluation for NMP under TSCA. As a
result, EPA did not evaluate exposures via the drinking water
pathway or on-site NMP land releases that go to RCRA
Subtitle C hazardous waste landfills or RCRA Subtitle D
municipal solid waste (MSW) landfills, or associated
exposures to the general population or terrestrial species, in
the risk evaluation.
The comment recommends identifying fenceline
communities as PESS and urges EPA to evaluate risk of
cumulative exposures. Populations exposed through
pathways excluded from the risk evaluation were not
identified as PESS. EPA disagrees with public comments on
the draft risk evaluation that suggest fenceline
subpopulations should be identified as PESS. TSCA provides
EPA with the discretion to identify the PESS that are relevant
to the chemical-specific risk evaluation [TSCA Section
6(b)(4)(A)], General population exposure through air, surface
Page 65 of 205
-------
(e.g., individuals who live or work near manufacturing,
processing, use or disposal sites)." This might include
tribal communities, workers, and other members of the
public who might spend time on the site, live near the
site, and consume food and water from near the site.
EPA should identify all populations living near disposal
and other waste management sites as potentially exposed
subpopulations. The multiple exposure scenarios
associated with proximity to unlined disposal site
releases to environmental media, including air, water,
and waste pathways excluded from the NMP draft risk
evaluation, must be analyzed for both individual and
cumulative exposures.
water, sediment, and land-applied biosolids were evaluated
based on fate properties of NMP and screening level
analysis. As described in Section 4.6.1.3, EPA did not
identify risks to the general population through these
pathways. As described in Section 1.4.2, general population
exposures through drinking water and disposal are beyond
the scope of the risk evaluation.
EPA considered the reasonably available information and
used the best available science to determine whether to
consider aggregate or sentinel exposures for a particular
chemical. EPA evaluated aggregate risks across exposure
routes for each condition of use but concluded that there is
insufficient information about likely co-exposures to support
analysis of aggregate exposure across multiple conditions of
use.
Spills and leaks generally are not included within the scope of
a TSCA risk evaluation because, in general, they are not
considered to be circumstances under which a chemical
substance is intended, known, or reasonably foreseen to be
manufactured, processed, distributed, used, or disposed of. To
the extent there may be potential exposure from spills and
leaks, EPA is also declining to evaluate environmental
exposure pathways addressed by other EPA-administered
statutes and associated regulatory programs.
First, EPA does not identify NMP spills or leaks as
"conditions of use." EPA does not consider NMP spills or
leaks to constitute circumstances under which NMP is
manufactured, processed, distributed, used, or disposed of,
within TSCA's definition of "conditions of use." Congress
specifically listed discrete, routine chemical lifecycle stages
within the statutory definition of "conditions of use" and EPA
does not believe it is reasonable to interpret "circumstances"
under which NMP is manufactured, processed, distributed,
Page 66 of 205
-------
used, or disposed of to include uncommon and unconfined
spills or leaks for purposes of the statutory definition. Further,
EPA does not generally consider spills and leaks to constitute
"disposal" of a chemical for purposes of identifying a COU in
the conduct of a risk evaluation.
In addition, even if spills or leaks of NMP could be
considered part of the listed lifecycle stages of NMP, EPA
has "determined" that spills and leaks are not circumstances
under which NMP is intended, known or reasonably foreseen
to be manufactured, processed, distributed, used, or disposed
of, as provided by TSCA's definition of "conditions of use,"
and EPA is therefore exercising its discretionary authority
under TSCA Section 3(4) to exclude NMP spills and leaks
from the scope of the NMP risk evaluation. The exercise of
that authority is informed by EPA's experience in developing
scoping documents and risk evaluations, and on various
TSCA provisions indicating the intent for EPA to have some
discretion on how best to address the demands associated
with implementation of the full TSCA risk evaluation
process. Specifically, since the publication of the Risk
Evaluation Rule, EPA has gained experience by conducting
ten risk evaluations and designating forty chemical
substances as low- and high-priority substances. These
processes have required EPA to determine whether the case-
specific facts and the reasonably available information justify
identifying a particular activity as a "condition of use." With
the experience EPA has gained, it is better situated to discern
circumstances that are appropriately considered to be outside
the bounds of "circumstances... under which a chemical
substance is intended, known, or reasonably foreseen to be
manufactured, processed, distributed in commerce, used, or
disposed of' and to thereby meaningfully limit circumstances
subject to evaluation. Because of the expansive and
Page 67 of 205
-------
potentially boundless impacts that could result from including
spills and leaks as part of the risk evaluation (e.g., due to the
unpredictable and irregular scenarios that would need to be
accounted for, including variability in volume, frequency, and
geographic location of spills and leaks; potential application
across multiple exposure routes and pathways affecting
myriad ecological and human receptors; and far-reaching
analyses that would be needed to support assessments that
account for uncertainties but are based on best available
science), which could make the conduct of the risk evaluation
untenable within the applicable deadlines, spills and leaks are
determined not to be circumstances under which NMP is
intended, known or reasonably foreseen to be manufactured,
processed, distributed, used, or disposed of, as provided by
TSCA's definition of "conditions of use."
Exercising the discretion to not identify spills and leaks of
NMP as a COU is consistent with the discretion Congress
provided in a variety of provisions to manage the challenges
presented in implementing TSCA risk evaluation. See e.g.,
TSCA Sections 3(4), 3(12), 6(b)(4)(D), 6(b)(4)(F). In
particular, TSCA Section 6(b)(4)(F)(iv) instructs EPA to
factor into TSCA risk evaluations "the likely duration,
intensity, frequency, and number of exposures under the
conditions of use...," suggesting that activities for which
duration, intensity, frequency, and number of exposures
cannot be accurately predicted or calculated based on
reasonably available information, including spills and leaks,
were not intended to be the focus of TSCA risk evaluations.
And, as noted in the preamble to the Risk Evaluation Rule,
EPA believes that Congress intended there to be some
reasonable limitation on TSCA risk evaluations, expressly
indicated by the direction in TSCA Section 2(c) to "carry out
[TSCA] in a reasonable and prudent manner."
Page 68 of 205
-------
For these reasons, EPA is exercising this discretion to not
consider spills and leaks of NMP to be COUs.
Second, even if NMP spills or leaks could be identified as
exposures from a COU in some cases, these are generally not
forms of exposure that EPA expects to consider in risk
evaluation. TSCA Section 6(b)(4)(D) requires EPA, in
developing the scope of a risk evaluation, to identify the
hazards, exposures, conditions of use, and potentially exposed
or susceptible subpopulations the Agency "expects to
consider" in a risk evaluation. This language suggests that
EPA is not required to consider all conditions of use, hazards,
or exposure pathways in risk evaluations. EPA has chosen to
tailor the scope of the risk evaluation to exclude spills and
leaks in order to focus analytical efforts on those exposures
that present the greatest potential for risk.
In the problem formulation documents for many of the first
10 chemicals undergoing risk evaluation, EPA applied the
same authority and rationale to certain exposure pathways,
explaining that "EPA is planning to exercise its discretion
under TSCA 6(b)(4)(D) to focus its analytical efforts on
exposures that are likely to present the greatest concern and
consequently merit a risk evaluation under TSCA...." This
approach is informed by the legislative history of the
amended TSCA, which supports the Agency's exercise of
discretion to focus the risk evaluation on areas that raise the
greatest potential for risk. See June 7, 2016 Cong. Rec.,
S3519-S3520.
In addition to TSCA Section 6(b)(4)(D), the Agency also has
discretionary authority under the first sentence of TSCA
Page 69 of 205
-------
Section 9(b)(1) to "coordinate actions taken under [TSCA]
with actions taken under other Federal laws administered in
whole or in part by the Administrator." TSCA Section 9(b)(1)
provides EPA authority to coordinate actions with other EPA
offices, including coordination on tailoring the scope of
TSCA risk evaluations to focus on areas of greatest concern
rather than exposure pathways addressed by other EPA-
administered statutes and regulatory programs, which does
not involve a risk determination or public interest finding
under TSCA Section 9(b)(2). EPA has already tailored the
scope of this risk evaluation using such discretionary
authorities with respect to exposure pathways covered under
the jurisdiction of other EPA-administered statutes and
associated regulatory programs (see Section 1.4.3).
34
PUBLIC COMMENTS:
NMP is present in bio-solids from wastewater treatment,
which are then applied to land as fertilizer. EPA excludes
biosolids from the NMP draft risk evaluation because
"NMP concentrations in surface water resulting from
land application of biosolids are expected to be much
less than those associated with direct release of
wastewater treatment plant effluents to surface water."
EPA offers no support for that statement, and it does not
evaluate the combined effects of NMP from direct
discharges and biosolids application on the same
waterbodies.
As described in Section 2.1.1 of the risk evaluation, EPA
considered exposures from biosolids during problem
formulation and concluded that no further analysis of this
pathway was needed. In the NMP Problem Formulation, EPA
explains that "NMP exhibits high water solubility (1000 g/L)
and limited
potential for adsorption to organic matter (estimated log Koc =
0.9); therefore, land releases will ultimately partition to the
aqueous phase (i.e., biosolids associated waste water and soil
pore water) upon release into the environment. Because NMP
readily biodegrades in environments with active microbial
populations, NMP residues that remain following waste water
treatment are not expected to persist. NMP concentrations in
biosolids-associated water are expected to decrease, primarily
via aerobic degradation, during transport, processing
(including dewatering), handling, and land application of
biosolids (which may include spraying)."
Page 70 of 205
-------
53, 55
PUBLIC COMMENTS:
Legacy use of products containing NMP was not
considered in this draft risk evaluation. Per decision of
the Ninth Circuit Court of Appeals, EPA can no longer
exclude "legacy" chemical uses from a risk evaluation,
nor can it exclude any conditions of use from
consideration. The Court also affirmed that "TSCA's
definition of'conditions of use' clearly includes uses and
future disposals of chemicals." EPA does not have the
discretion to pick-and-choose conditions of use for
inclusion in a risk evaluation.
EPA did not identify any "legacy uses" or "associated
disposals" of NMP, as those terms are described in EPA's
Risk Evaluation Rule, 82 FR 33726 (July 20, 2017).
Therefore, no such uses or disposals were added to the scope
of the risk evaluation for NMP following the issuance of the
opinion in Safer Chemicals, Healthy Families v. EPA, 943
F.3d 397 (9th Cir. 2019).
The uses of NMP in the past are not "legacy" uses. As
described in EPA's Risk Evaluation Rule (82 FR 33726 (July
20, 2017)), a legacy use is an ongoing use of a chemical
substance in a particular application where the chemical
substance is no longer being manufactured, processed, or
distributed in commerce for that application. The example
provided in the Rule is insulation, which may be present in
buildings after a chemical substance component is no longer
being made for that use. EPA is not aware of legacy NMP
uses.
Conditions of use - Other
SACC COMMENTS:
Recommendation: Include NMP as an agrochemical
formulant as an occupational use in the draft risk
evaluation and discuss implications for exposure and
risk.
Only active pesticidal ingredients are regulated under the
Federal Insecticide, Fungicide, and Rodenticide Act
(FIFRA), not formulants (adjuvants). Therefore, this use
does not fall within the purview of any other office of the
Agency.
NMP is used as an inert ingredient in pesticide products.
Agricultural chemical inert ingredients are considered from a
risk perspective. Such agents are often included in pesticide
formulations for a number of reasons. EPA understands this
use as an ingredient in pesticides would be regulated under
the Federal Insecticide, Fungicide, and Rodenticide Act
(FIFRA), and is therefore outside the definition of chemical
substance as regulated by TSCA.
63
PUBLIC COMMENTS:
EPA should carefully consider the ways in which
conditions of use can vary across different segments of
industry, and consider the actual, rather than
EPA appreciates the commenters' suggestion. TSCA (U.S.C.
ง 3(4)) defines the conditions of use as "the circumstances,
as determined by the Administrator, under which a chemical
substance is intended, known, or reasonably foreseen to be
manufactured, processed, distributed in commerce, used, or
Page 71 of 205
-------
hypothetical, conditions of use in specific industry
sectors within the scope of the risk evaluation.
disposed of." EPA carefully considered the best approach to
describe the conditions of use across and within sectors using
NMP, and relied on communications with companies,
industry groups, environmental organizations, and public
comments to supplement the use information. In the
occupational assessment, EPA presents variations across
different segments of industry when reasonably available data
are found. For example, EPA revised the occupational
exposure assessment in the risk evaluation to separately
assess occupational exposure scenarios associated with three
categories of electronic part manufacturing: lithium ion
battery manufacturing (2.4.1.2.15); Other electronics
manufacturing, including capacitor, resistor, coil, transformer,
and other inductor manufacturing (2.4.1.2.9); and
semiconductor manufacturing (2.4.1.2.10). In these separate
OESs and where feasible, EPA revised and expanded PBPK
runs for industry-specific work activities using industry-
specific PBPK input data and information provided in public
comments for the lithium ion battery manufacturing industry
(1 j\lePicher Technologies. 2020;i, h; 1 ICM. 2020a, b. c) and
for the semiconductor manufacturing industry (Intel
Corporation. 2020; Semiconductor Indus sociation.
2020; Intel Corporation. 2019; Semiconductor Industry
Association. 2019a, b. c).
31, 52
PUBLIC COMMENTS:
NMP concentrations in specific conditions of use should
be considered in the overall risk evaluation. SIA
provided data on NMP weight percentage in chemical
formulations and waste as part of its 2019 study.
EPA used NMP concentration data specific to the conditions
of use being assessed, including weight fractions provided by
SIA.
Section 2.4.1.1 of the draft risk evaluation details the
parameters considered for the PBPK model. To support the
draft risk evaluation, EPA determined the weight fraction of
NMP in various products through information provided in the
available literature, Safety Data Sheets, previous risk
assessments and the 2017 NMP Market Profile. Where a data
Page 72 of 205
-------
point was provided as a range of NMP concentrations for a
certain product (e.g., paints and coatings), EPA utilized the
mid-range (middle) and high-end (maximum) weight
fractions to estimate potential exposures. Where multiple data
points for a given type of product (e.g., paints and coatings)
were available, EPA estimated exposures using the central
tendency (50th percentile) and high-end (95th percentile)
NMP concentrations. The SIA-provided weight fraction data
were used in the Semiconductor Manufacturing OES in
Section 2.4.1.2.10.
31
PUBLIC COMMENTS:
On the basis of the draft risk evaluation and the materials
made public in the supplemental assessments, SIA
reviewers have not been able to reproduce EPA's
statistical analysis when interpreting the air sampling
results. This lack of transparency makes it difficult for
other scientists to fully assess whether EPA is making
use of the best available science.
EPA updated the air concentrations for work activities
assessed for semiconductor manufacturing. EPA provided
additional explanation of the analysis of SIA's air sampling
results in the Supplemental Information on Occupational
Exposure Assessment.
51
PUBLIC COMMENTS:
NMP is not regulated as a Hazardous Air Pollutant
(HAP) under the Clean Air Act (CAA) so there are no
applicable federal emission limits and no reason to
expect that EPA will use the CAA to evaluate the risks of
NMP air emissions and take action to reduce this source
of exposure.
NMP is not regulated as a hazardous air pollutant. As
described in Section 4.6.2.3, EPA performed a first-tier
screening analysis for risks from ambient air exposures
during problem formulation. EPA did not identify risks from
human exposures that may result from inhalation of outdoor
air containing NMP released from industrial and commercial
facilities.
Exposure for Occupational Non-Users (ONUs)
SACC
SACC COMMENTS:
Very limited data informs ONU exposure assumptions.
In particular, the apparent paucity of actual data
underlying the ONU air concentration assumption was
noted.
Several Committee members expressed concern that the
assumption that ONUs were always exposed to cleaner
Table 4-47 shows that EPA found no reasonably available
NMP air concentration data for ONUs, including relevant
area monitoring. EPA updated Section 2.4.1.1 to indicate that
EPA does not have reasonably available parameters,
including proximity of ONUs to workers or to emission
sources, to estimate near-field/ far-field NMP air
concentrations for both workers and ONUs except for the
Page 73 of 205
-------
(unfiltered) air than users was questionable, given that
both near field/far field separation distances and facility
air handling equipment were likely to be highly variable
both across and within industries.
One Committee member pointed out ONUs in home
enterprises might be children in very close proximity to
users.
One Committee member felt that ONU and worker
exposures were not sufficiently distinguished and/or
contrasted.
Recommendation: Develop specific ONU exposure
scenarios for each condition of use, and tabulate
expected exposures based on the information that has
already been collected.
Recommendation: Discuss the potential for modeling
ONU air exposures as a function of proximity to active
use, including why proximity may be more important for
some conditions of use than others.
Commercial Automotive Servicing OES. EPA used this
model to estimate near-field/ far-field NMP air concentrations
for the Commercial Automotive Servicing OES, as described
in the Supplemental Information on Occupational Exposure
Assessment. EPA also updated Section 2.4.1.1 to note that
proximity may be more important for some conditions of use
where ONUs are in close proximity to workers or to emission
sources. Also, Section 2.4.1.1 states that "EPA expects that
ONUs are exposed to lower air concentrations than workers
since they may be further from the emission source than
workers," which is similar to the comment (that ONUs were
always exposed to cleaner (unfiltered) air compared to
workers). EPA agrees that both near field/far field separation
distances and facility air handling equipment are likely to be
highly variable both across and within industries but has not
found reasonably available data that refutes EPA's
expectation.
ONU modeling is for adults only. Children are covered as
bystanders in consumer modeling (see Section 2.4.2.5).
In each subsection of 2.4.1.2, EPA describes whether or not
ONU-specific monitoring data or modeling is available for
each OES.
EPA developed specific PBPK runs for ONUs for each OES
using NMP air concentration estimates for workers except for
the Commercial Automotive Servicing OES that uses
estimates of near-field and far-field NMP air concentrations.
These PBPK runs are presented in the Supplemental Excel
File on Occupational Risk Calculations. Table 2-77 presents
the results of the ONU PBPK runs for all OESs.
34, 61
PUBLIC COMMENTS:
The range of workers defined by EPA as ONUs - which
include "supervisors, managers, engineers, and other
personnel in nearby production areas" - is too broad to
EPA has not found additional reasonably available
information or data to explore different categories of ONUs
beyond the ONU categories presented in this risk evaluation.
EPA presents all reasonably available information on the job
Page 74 of 205
-------
warrant a single categorization. Supervisors and
managers have very different exposure patterns than
skilled trade workers and other "shop floor" ONUs, yet
all of them are assumed to face similar risks under EPA's
categorization. Separate ONU worker categories should
be considered.
descriptions and functions of ONUs in NMP workplaces in
each subsection of 2.4.1.2 and in the Supplemental
Information on Occupational Exposure Assessment.
51, 61
PUBLIC COMMENTS:
EPA's determination that there are no unreasonable risks
to ONUs is unsupportable (pp. 27-28). The draft risk
evaluation provides few details on the job functions of
ONUs in NMP workplaces, number of ONUs exposed to
NMP, and the nature and duration of this exposure.
It is assumed that all ONUs lack dermal contact with
NMP, but this assumption is implausible. Cleaning and
maintenance of NMP-contaminated equipment would
unavoidably result in dermal contact, as would sampling
and testing of NMP-containing process streams or
products for quality control purposes; spills and
equipment leaks would also likely result in dermal
contact. ONUs are also less likely to be provided PPE,
and therefore, exposure may be as great, or greater than,
those of other workers. Removing dermal exposure
entirely from EPA's determination of risks to ONUs
severely skews EPA's risk estimates and ignores
exposure scenarios that are highly likely in real-world
use and handling of NMP.
EPA also claims "ONU inhalation exposures are
expected to be lower than inhalation exposures for
workers directly handling the chemical substance" (p.
303). To account for this assumed difference in
inhalation exposure, EPA bases its unreasonable risk
determinations for ONUs on "central tendency risk
estimates" rather than high-end inhalation exposure
EPA added all reasonably available information on the job
functions of ONUs in NMP workplaces in Section 2.4.1.2
subsections. In Table 2-4, EPA presents numbers of ONUs in
each OES where exposure to NMP may occur. In Section
2.4.1.1, EPA clarified that the nature and duration of ONU
exposure is inhalation and vapor-through-skin and occurs
over the same duration, whether based on task or shift
duration, as estimated for the worker.
EPA considers the activities of cleaning and maintenance of
NMP-contaminated equipment, and sampling and testing of
NMP-containing process streams or products for quality
control purposes, to be worker activities because they may
result in dermal contact with liquids. EPA generally defines
ONUs in Section 2.4.1 of the RE as "supervisors, managers,
and other employees that may be in the production areas but
do not perform tasks that result in direct dermal contact with
liquids." Based on this definition, EPA does not expect ONUs
to have dermal exposures to liquids.
Where EPA had monitoring or modeled data specific to
ONUs, unreasonable risk determinations were made based on
high-end exposures. For conditions of use where the data did
not distinguish between worker and OINU inhalation
exposures, there was uncertainty regarding ONU exposure.
Page 75 of 205
-------
levels. Assuming that there is never high-end inhalation
exposure by ONUs is unsupportable, since there are
undoubtedly some ONU inhalation exposure scenarios
that are similar in magnitude and duration to those of
workers involved in direct NMP operations. For
example, workers in shared work areas close to
equipment emitting NMP vapors could have nearly the
same level of inhalation exposure as workers using this
equipment. The Agency itself acknowledges that,
"[w]hen EPA does not have ONU-specific exposure
data, EPA's assumption that 50th percentile air
concentrations predicted for workers in these activities
are a good approximation of exposure is uncertain."
The draft risk evaluation does not include workplace
monitoring data to show exposure levels for ONUs
specifically.
EPA must obtain more information about real-world
ONU exposure scenarios or base its risk determinations
on more plausible default assumptions that reflect likely
conditions in NMP workplaces.
In each subsection of 2.4.1.2, EPA describes whether or not
ONU-specific monitoring data or modeling is available for
each OES.
EPA does not have reasonably available data or information
to develop more plausible default assumptions for ONUs.
EPA had sufficient reasonably available information to
complete the NMP risk evaluation using a weight of the
scientific evidence approach based on the best available
science. EPA selected the first 10 chemicals for risk
evaluation based in part on its assessment that these
chemicals could be assessed without the need for regulatory
information collection or development. When preparing this
risk evaluation, EPA obtained and considered reasonably
available information, defined as information that EPA
possesses, or can reasonably obtain and synthesize for use in
risk evaluations, considering the deadlines for completing the
evaluation. In some cases, when information available to EPA
was limited, the Agency relied on models; the use of modeled
data is in line with EPA's final Risk Evaluation Rule and
EPA's risk assessment guidelines.
Environmental pathways of human exposure
51,55,
61
PUBLIC COMMENTS:
The NMP draft risk evaluation departs from - in the
SACC's words - basic "risk assessment principles" by
excluding "well-known exposure routes" for this
chemical and failing to provide an "overall assessment of
risks." Contrary to SACC's explicit advice, EPA's draft
risk evaluation excludes all human exposures from
environmental releases of NMP, resulting in the absence
of any consideration of environmental pathways that
contribute to overall human exposure and risk. EPA
As described in Section 1.4.2, EPA believes it is both
reasonable and prudent to tailor TSCA risk evaluations when
other EPA offices have expertise and experience to address
specific environmental media, rather than attempt to evaluate
and regulate potential exposures and risks from those media
under TSCA. EPA believes that coordinated action on
exposure pathways and risks addressed by other EPA-
administered statutes and regulatory programs is consistent
with statutory text and legislative history, particularly as they
pertain to TSCA's function as a "gap-filling" statute, and also
Page 76 of 205
-------
excluded all risks that the general population faces from
exposures due to releases of NMP to land, air, and water,
based on the assumption that other statutes adequately
address these exposures. No analyses or data have been
presented, however, to show that other statutes are
protective of the general population. For example, NMP
is not currently regulated under the SDWA. It is also not
listed as a HAP under the CAA. That means that there is
no federal limit for NMP level in drinking water and that
it is subject to only limited CAA regulations, which
(unlike TSCA) do not encompass all known sources of
the chemical and do not require the elimination of
unreasonable risk.
This unlawful interpretation of TSCA has twice been
rejected by the SACC and overlooks the widespread
presence of NMP in environmental media to which
millions of people are exposed. The exclusion of known
exposure pathways violates both the intent and letter of
TSCA. EPA is required to evaluate all such risks under
TSCA, regardless of whether other statutes regulate
them. In a recent decision on EPA's TSCA risk
evaluation rule, the Ninth Circuit Court of Appeals ruled
that EPA is unambiguously not granted the discretion "to
exclude conditions of use, or their associated exposures
and risks, from TSCA risk evaluations." If any of NMP's
conditions of use results in air emissions or releases to
water, these exposures are an essential part of the risk
evaluation and must be considered by EPA, regardless of
whether or not they might be addressed under other laws.
The releases and exposures that EPA is ignoring are far
from trivial. EPA data on releases of NMP to air, water,
and land for years 2015-2017 in the draft risk evaluation
(p. 30) show that >9,500,000 lbs/year are released into
furthers EPA aims to efficiently use Agency resources, avoid
duplicating efforts taken pursuant to other Agency programs,
and meet the statutory deadline for completing risk
evaluations. EPA has therefore tailored the scope of the risk
evaluation for NMP using authorities in TSCA Sections 6(b)
and 9(b)(1).
EPA did not include exposures via the drinking water
pathway or disposal to underground injection, RCRA
Subtitle C hazardous waste landfills, or RCRA Subtitle D
municipal solid waste (MSW) landfills in this risk evaluation,
as these exposure pathways fall under the jurisdiction of
other EPA-administered statutes and associated regulatory
programs.
The recent Ninth Circuit Court of Appeals decision in Safer
Chemicals, Healthy Families v. EPA, 943 F.3d 397 (9th Cir.
2019) was limited to review of EPA rulemaking, Procedures
for Chemical Risk Evaluation Under the Amended Toxic
Substances Control Act, 82 FR 33726 (July 20, 2017),
commonly referred to as the TSCA Risk Evaluation Rule. As
such, the Ninth Circuit decision did not opine on EPA's
statutory authority under TSCA, as discussed above, to
exclude conditions of use, or associated exposures and risks
in TSCA risk evaluations.
As described in Section 4.6.2.3 of the risk evaluation, EPA
evaluated potential exposures and risks to the general
population through ambient water, land-applied biosolids, and
ambient air during problem formulation. Based on
environmental fate properties of NMP and first-tier screening
level analyses, EPA did not identify risk to the general
population from these pathways. Specifically, EPA did not
identify risks from human exposures that may result from
contact with or ingestion of surface water, land releases of
Page 77 of 205
-------
the environment. EPA's approach effectively reduces
this quantity to zero. Excluding releases to air is also
contrary to the EPA Problem Formulation, which said
"[ijnhalation is expected to be a relevant route of
exposure for the general population due to the propensity
for NMP air releases from ongoing commercial and
industrial activities." The most recent round of reporting
for the TRI showed total NMP air emissions from 280
facilities of 1,532,507 million pounds in 2017 (an
underestimate of total releases because emissions from
small commercial operators below the TRI reporting
thresholds are not captured).
EPA's exclusion of environmental releases that may be
subject to other laws ignores the comprehensive multi-
media scope of TSCA as framed by Congress. Residents
of areas with a concentration of manufacturing and use
facilities may also work at these facilities and be exposed
to NMP both on the job and during non-work activities.
Consumers who use NMP-based products may also be
exposed to NMP air emissions, particularly if they live
near emitting facilities, and may also be exposed to NMP
through drinking water or proximity to waste sites.
Similarly, workers exposed to NMP at their places of
employment may also inhale NMP from ambient air or
have dermal contact with NMP-containing products used
in their homes, adding to their overall exposure.
Determination of overall risk requires assessing exposure
by all of these pathways in combination.
NMP (including those that may result from land-application
of biosolids), or inhalation of outdoor air containing NMP
released from industrial and commercial facilities. As the
commenter notes, on p.36-37 of the problem formulation,
EPA identified oral, dermal and inhalation exposures relevant
to general population exposures. Based on subsequent
screening level analysis in the problem formulation, EPA
concluded (p47-48) that exposures through air, surface water,
sediment, and land-applied biosolids do not require further
analysis in the risk evaluation because exposure through these
pathways is unlikely to present a risk concern.
The final risk evaluation includes an updated screening level
analysis of risks to the general population for incidental
ingestion or dermal contact with NMP in surface water. EPA
found no unreasonable risks to the general population from
NMP under the conditions of use within the scope of the risk
evaluation.
EPA considered the reasonably available information and
used the best available science to determine whether to
consider aggregate or sentinel exposures for a particular
chemical. EPA concluded that there is insufficient
information about likely co-exposures to support analysis of
aggregate exposure across multiple conditions of use. In
Section 4.5 of the risk evaluation, EPA acknowledges that the
decision not to aggregate risk across conditions of use could
result in an underestimate of risk.
Consumer exposures
SACC
SACC COMMENTS:
EPA appreciates the SACC's comments. Descriptions and
justification on the consumer conditions of use can be found
Page 78 of 205
-------
The origin of the various consumer conditions of use is
insufficiently justified in the draft risk evaluation and
would benefit from further explanatory text.
Recommendation: Expand the justifications for each
consumer condition of use, citing primary sources
whenever possible, and conducting a sensitivity analysis
on exposure factors.
in the documents preceding this risk evaluation, including the
Scope of the Risk Evaluation for NMP and the NMP Problem
Formulation. In the NMP Problem formulation EPA
explained consumer conditions of use were identified through
"extensive research and outreach, including review of
published literature and online databases, SDSs, company
websites and various databases. EPA met with environmental
groups, chemical users, states, industry groups, companies
and other stakeholders to identify consumer COUs."
SACC
SACC COMMENTS:
Consumer conditions of use in the draft risk evaluation
assume limited frequency of use and a distinction
between near-field (zone 1) and far-field (zone 2)
exposures. Some activities, such as simultaneous
hobby/craft work and childcare might lead to "child
bystander" inhalation exposure to zone 1 air. Some small
residences might also feature rapid mixing of air among
rooms. The Committee was particularly concerned about
households in which hobby/craft work is routine (e.g.,
internet sellers deriving a significant portion of their
income from "hobby/craft" activities). In such
households, evaluation of chronic as well as acute
exposures might be appropriate.
Recommendation: Treat enterprises co-located with
residences as a distinct consumer condition of use and
consider chronic exposures in that case.
EPA did not identify reasonably available information on
chronic exposures to NMP that may occur in the home. EPA,
however, did consider high-intensity users which would be
consumers that use craft products in a greater amount for a
longer duration. EPA also considered potential acute
exposures to both adult and child bystanders resulting from
acute exposures. This analysis of risks to bystanders is
summarized in tables 4-49 and 4-50 in the risk evaluation.
51, 55
PUBLIC COMMENTS:
The draft risk evaluation only addresses developmental
(fetal mortality) risks to consumers, ignoring potential
effects on fertility on the grounds that "consumer
exposure is not chronic in nature" (p. 160). The rationale
for this approach is EPA's assumption that consumer
exposure is limited to "a single use event which may
As stated above, EPA did not identify reasonably available
information indicating chronic consumer exposures to NMP.
EPA used product-specific data and consumer survey data to
characterize the consumer activity patterns and use and
exposure scenarios. The consumer survey data provides
statistical range of data on the use of certain categories of
products, including paint removers, adhesives, stains and
Page 79 of 205
-------
occur over a 24-hour period" and that a "consumer uses a
single product or product type" (p. 159). These
assumptions, which EPA acknowledges may
underestimate NMP exposures, disregard use scenarios
for consumer products that could result in repeated NMP
exposure over time. The draft risk evaluation identifies
12 separate categories of NMP-containing consumer
products, representing 52 discrete products, (p. 140).
Some of these products (adhesives, adhesive removers,
paint removers, paints,, sealants, stains, and varnishes)
would be expected to be used regularly by hobbyists,
artists who work at home, or home renovators. Others
(engine cleaners and degreasers and auto interior
cleaners) would likely be used frequently by consumers
who maintain and repair their own or friends' vehicles.
Given the many different household functions performed
by NMP-containing products, it is likely that many
consumers use multiple products either simultaneously,
resulting in greater acute exposures than addressed by
EPA, or over time, resulting in chronic exposures that
put them at risk of reproductive harm.
varnishes, degreasers and paints as outlined in the risk
evaluation. EPA selected not only median but also high-end
input parameters in order to develop a high-intensity use
scenario to capture the exposures to those consumers who
would use the products in greater quantities and for a longer
duration such as arts and crafts hobbyists, do-it-yourself
home renovators, or consumer auto repair hobbyists.
SACC
SACC COMMENTS:
The draft risk evaluation assumes that consumer products
are less concentrated than industrial formulations.
However, at least one Committee member objected,
noting that consumers can acquire industrial
formulations in many cases.
EPA identified consumer product concentrations based on
information on SDSs for products available to consumers
whether or not those products were formulated for
consumers. For example, if a product was labeled 'for
commercial use only' but could be purchased in a retail store
or online retail platforms, it was included as a consumer
product. However, some industrial formulations are not
available through means normally available to consumers and
can only be purchased through direct wholesale distributors.
These formulations tend to be more concentrated than those
sold in the retail space. For industrial products that are sold in
retail space accessible to typical consumers, product labeling
Page 80 of 205
-------
may be geared to the industrial or commercial use. Therefore,
it is not always apparent the product would also meet the
needs of the consumer and may decrease the likelihood of a
consumer purchase.
SACC
SACC COMMENTS:
The current NMP risk evaluation may overestimate the
number of minutes that children engage in mouthing
behavior. EPA is referred specifically to Table 1 in
Babich et al. (2004), which summarizes prior reports and
regulatory defaults.
Table 1 of Babich et al., (2004) lists mouthing time/duration
(in minutes per day) estimates from a variety of sources for
PVC based teethers and toys. Since NMP data was found in
children's blankets, EPA used the Consumer Exposure Model
(CEM) exposure scenario for textile articles that are mouthed
by children for mouthing durations cited in CPSC. These
mouthing durations are within the ranges cited in Babich et
al. C2004Y
SACC
SACC COMMENTS:
Recommendation: Cite primary, rather than secondary,
data sources whenever possible, and reference the work
by Akesson and Jonsson (2000), Anundi et.al (1993), and
Beaulieu and Schmerber (1991); they are not listed in the
references section although exposure levels used in the
evaluation seem to derive from these sources.
Where possible, EPA has revised the final risk evaluation to
cite primary sources rather than secondary sources. For
example, four primary sources cited in a secondary source
have been added to Table 2-31 and to the text below this table
for Paint and Coating Removers. Regarding the three specific
sources suggested in this comment, the first source listed,
Akesson and Jonsson (2000), is a primary source that is not
reasonably available to EPA and is therefore cited in Table 2-
31 and in the associated footnote. The second source listed,
Anundi et.al (1993), was reviewed but not used for graffiti
remover data. Instead, a robust data set in Anundi et al.
(2000) was used for graffiti removers because the single,
short-term data point in Anundi et.al (1993) was older and
fell into the range of the newer data. The third source listed,
Beaulieu and Schmerber (1991), contained data for the
microelectronics industry are over 30 years old and were not
used due to availability of a robust and more relevant recent
data set less than 10 years old.
51, 55
PUBLIC COMMENTS:
EPA's Problem Formulation for NMP cites evidence
from two Canadian studies that use of NMP-containing
The risk evaluation accounts for consumer exposure to NMP
through air. EPA used the Consumer Exposure Model to
estimate air concentrations associated with each consumer
Page 81 of 205
-------
products in homes and buildings results in elevated
levels of NMP in indoor air (p. 33). Although the
Problem Formulation commits (p. 58) to further
"[ejvaluate the indoor exposure pathways based on
available data," the draft risk evaluation itself makes no
mention of NMP levels in indoor air. Elevated NMP
concentrations in indoor air would represent another
contributor to chronic consumer exposure, adding to
direct dermal and inhalation exposure from product use.
exposure scenario. Those concentrations were then used as
inputs in the PBPK model used to predict total internal
exposures (blood concentrations). Exposures predicted for
bystanders of consumer users are based on air concentrations
alone since no dermal contact is anticipated.
EPA did cite NMP air concentration monitoring data after
Daint removal use fNIOSH. 1993), com oared that to modeled
NMP air concentration and found the two were relatively
similar and not a risk to non-users.
33, 57
PUBLIC COMMENTS:
Many exposure assumptions overestimate actual
exposures.
To ensure that the range of expected exposures to NMP
by consumers is accurately characterized, EPA should
account for the most likely exposure scenarios, which
involve product use in outdoor and/or garage settings
(ABT, 1992).
For the small, unventilated room scenarios, two
additional options should be included to account for
higher air change rates associated with "Window Open"
and "Exhaust Fan On."
For consumer and worker exposure scenarios that are
inconsistent with product labeling instructions, these
should be presented separately as "Product Misuse
Scenarios" so that risk management options for these
scenarios can be addressed independently from "Product
Use Scenarios."
EPA modeled medium-intensity use of each of the consumer
product scenarios to provide information for most prevalent
exposures. Table 2-78 in the risk evaluation lists the various
locations where the consumer product was modeled to be
used. For example, EPA modeled a deck adhesive product to
be used outdoors, though adhesives could be used anywhere
inside a home as well and modeled engine degreasing use in
the garage.
EPA is also concerned about high-intensity consumer use as
they represent a fraction of the population that would be more
highly exposed, such as do-it-yourself paint removal or
engine repair where longer duration of use and a greater
amount of the product are anticipated. Given that there are
plenty of reasons for a consumer not to open windows (cold
or very hot weather) or use an exhaust fan (not available in
most rooms other than bathrooms), EPA chose not to include
these scenarios in the current risk evaluation.
For consumer exposure, there were no "product misuse
scenarios" considered.
33, 57
PUBLIC COMMENTS:
EPA conducted quality control of all Consumer Exposure
Model inputs and outputs.
Page 82 of 205
-------
EPA should conduct a strong quality assurance/quality
control check to ensure that any errors on Csat and other
parameters are corrected and implemented appropriately
in the Consumer Exposure Model (CEM).
33, 57
PUBLIC COMMENTS:
The saturation of air with NMP is dependent upon
humidity, and Csat is negatively correlated with
humidity. A relative humidity of 50% would correspond
to a Csat value of approximately 525 mg/m3, which is
considerably lower than the value assumed by EPA.
The distribution of relative humidity in indoor air in the
United States (e.g., low, average, and high humidity
values of-15%, 50%, and 85%) should be
characterized based on recent surveys (e.g., USHUD,
2010) and incorporated into the risk assessment to
characterize a distribution of Csat values based on the
relationships described (e.g., 1,030, 525, and 132 mg/m3)
for indoor air modeling.
Separate assumptions for relative humidity should be
made for the use of aqueous vs. non-aqueous NMP
products separately, to avoid an unrealistic assumption of
low humidity following application of aqueous NMP
products.
EPA's assessment does not consider condensation and
aerosol droplet formation for concentrations of NMP
vapor exceeding 470 mg/m3 (or 116 ppm).
In EPA's 2015 assessment, a Csat value of 640 mg/m3
was considered, but this value is not considered in the
current risk evaluation; no explanation is provided.
Since publication of the 2015 Paint Remover Risk
Assessment, EPA changed the NMP vapor pressure from
0.190 to 0.345 mm Hg which has affected the Csat value. The
estimated Csat calculated from CEM is now 1.84E+3 mg/m3,
using a vapor pressure of 0.345 mm Hg. In this instance, the
high intensity use Engine Degreaser scenario would not meet
Csat.
EPA varied the three most sensitive variables used in the
Consumer Exposure Model to predict consumer exposures:
weight fraction, mass of product and duration of product use.
From these exposure model results, EPA presents in the risk
evaluation, the medium-intensity and high-intensity use
scenarios. The high-intensity use scenario captures the
highest duration of use, highest mass and highest average
weight fraction and represents an upper end of the expected
consumer exposures to NMP even with other variables such
as humidity constant.
For EPA's NMP consumer exposure estimates, the PBPK
model was used to estimate total blood concentration from
dermal contact, vapor-through-skin, and inhalation of NMP
during product use. EPA modeled exposure to NMP during
consumer product use (with limited duration for most
products) and then inhalation exposures within the rest of the
house for 24 hours. EPA anticipates that if there is already
significant dermal contact with the liquid, the amount of
aerosol would likely only add a small increment of exposure.
Page 83 of 205
-------
57
PUBLIC COMMENTS:
EPA should carefully review the inconsistency of the
values from the exposure models used for CEMs versus
those used for worker exposures [Chemical Screening
Tool for Exposures and Environmental Releases
(ChemSTEER)] as there seem to be fundamental
differences in the underlying assumptions.
EPA uses different models for different exposure scenarios.
The assumptions for ChemSTEER are specific to
occupational scenarios, whereas CEM is for consumer
exposure scenarios. For example, for consumer exposure,
there is no assumption of glove use, there are specific
parameters associated with indoor homes such as air
exchange rate, activity patterns (time spent using the
consumer product as well as time inside the house and outside
of the house) that are inherently different from occupational
scenarios. For this risk evaluation, the only parameter values
in common for both consumer and occupational scenarios are
adult body weights and hand surface areas, and these values
are consistent.
57, 33
PUBLIC COMMENTS:
The relative magnitude of the air concentrations
modeled for consumer exposures (e.g., peak
concentrations up to 1,300 mg/m3; 24-hour TWA up to
103 mg/m3) compared to worker exposures (e.g., 8-hour
TWA up to 64 mg/m3) is counterintuitive.
EPA should identify differences in key assumptions and
where possible, consider that the application of the more
reasonable predictions of the ChemSTEER model be
adopted for consumer scenarios that are comparatively
similar.
There are assumptions in consumer exposure scenarios that
are inherently different from those of occupational scenarios.
For example, there may be occupational scenarios that
include industrial-sized facilities with ventilation, which is
not an assumption in consumer scenarios. Thus, given the
specific consumer product, it's duration of use and room of
use, it is reasonable to have a higher air concentration than
that found in an occupational setting.
EPA uses different models for different exposure scenarios.
The assumptions for ChemSTEER are specific to
occupational scenarios, whereas CEM is for consumer
exposure scenarios. For example, for consumer exposure,
there are specific parameters associated with indoor homes
such as air exchange rate, activity patterns (time spent using
the consumer product as well as time inside the house and
outside of the house) that are inherently different from
occupational scenarios. For this risk evaluation, the only
parameter values in common for both consumer and
Page 84 of 205
-------
occupational scenarios are adult body weights and hand
surface areas, and these values are consistent.
PPE
SACC
SACC COMMENTS:
Recommendation: EPA should use the authority
provided under the new TSCA rules to obtain better
information whenever data in the open literature are
found to be sparse (e.g., worker and ONU practices and
exposures from manufacturers and processors, empirical
data to support the specific condition of use modeling
results and increase support for occupational exposure
estimates, consumer exposure data, data on NMP content
in consumer products, etc.).
Recommendation: Contact the State of California Air
Resources Board for their data on NMP content in
consumer products.
It was noted that the State of California has established a
PEL for NMP, so some data should be available there.
The literature underpinning the American Conference of
Governmental Industrial Hygienists Biological Exposure
Indices for NMP should also be examined to see if any
useful data exist there.
One Committee member suggested that exposure to skin
beyond hands was likely for many occupational
conditions of use and that some useful data with respect
to skin exposure might be gleaned from studies
sponsored by the Agricultural Handlers Exposure Task
Force.
EPA obtained and considered reasonably available
information, defined as "information that EPA possesses, or
can reasonably obtain and synthesize for use in risk
evaluations, considering the deadlines ... for completing
such evaluation." Given the timeframe for conducting risk
evaluations on the first 10 chemicals, use of TSCA data
gathering authorities has been limited in scope. In general,
EPA intends to utilize TSCA data gathering authorities more
routinely for the next 20 risk evaluations.
EPA reviewed the California Air Resources Board (CARB)
and CalOSHA websites for additional NMP occupational air
monitoring data and did not find relevant data.
EPA reviewed the //-Methyl-2-Pyrrolidone: BEI(R) 8th
Edition Documentation published by the American
Conference of Governmental Industrial Hygienists and found
no new data. This documentation did refer to data in one field
study, Anundi et al, 2000, which is cited in the risk evaluation
and its data used.
EPA reviewed the Agricultural Handlers Exposure Task
Force website for data related to NMP and did not find any
relevant data. The information from this source includes a
database of doses, provided in mg or ug of exposure per
pound of pesticide applied. The information also includes
PPE assumptions and frequency of application. The source
does not provide PBPK inputs needed for NMP assessment,
such as air concentrations, exposure duration, or surface area
exposed.
51,34,
49, 61
PUBLIC COMMENTS:
EPA lacks sufficient exposure/monitoring data to support
proposed findings of no unreasonable risk. EPA's
EPA obtained and considered reasonably available
information, defined as "information that EPA possesses, or
can reasonably obtain and synthesize for use in risk
Page 85 of 205
-------
evaluation of workplace risks from NMP exposure is
flawed because it relies on limited worker exposure data.
EPA has substantial authority under TSCA sections 4, 8,
and 11 to require the submission of existing exposure
information, as well as additional monitoring or testing
to fill data gaps. Thus far, however, EPA has not
exercised that authority for any of its risk evaluations. It
has also failed to ask employers to share the workplace
monitoring data that they are required to preserve under
OSHA regulations, or asked OSHA and other state and
federal agencies to provide access to the extensive
exposure information in their direct possession. In
finalizing the NMP risk evaluation, EPA should make
every effort to obtain additional workplace monitoring
data from OSHA, state agencies, and industry.
evaluations, considering the deadlines ... for completing
such evaluation." EPA selected the first 10 chemicals for risk
evaluation based in part on its assessment that these
chemicals could be assessed without the need for regulatory
information collection or development. Given the timeframe
for conducting risk evaluations on the first 10 chemicals, use
of TSCA data gathering authorities has been limited in scope.
In general, EPA intends to utilize TSCA data gathering
authorities more routinely for the next 20 risk evaluations.
In the 2017 Procedures for Chemical Risk Evaluation Under
the Amended Toxic Substances Control Act (82 FR 33726,
July 20, 2017), EPA committed to, by codifying, interagency
collaboration to give the public confidence that EPA will
work with other agencies to gain appropriate information on
chemical substances. This is an ongoing deliberative process
and EPA is not obligated to provide descriptions of pre-
decisional and deliberative discussions or consultations with
other federal agencies. In the interest of continuing to have
open and candid discussions with our interagency partners,
EPA is not intending to include the content of those
discussions in the risk evaluation.
For the NMP risk evaluation, EPA reviewed and integrated
NMP monitoring data from the OSHA Chemical Exposure
Health Data (CEHD) database (OSHA. 2017). specifically for
the following OESs: Other electronics (capacitor, resistor,
coil, transformer, and other inductor) manufacturing (Section
2.4.1.2.9), printing (Section 2.4.1.2.11), and spray/ wipe
cleaning (Section 2.4.1.2.16). Additionally, EPA requested
NMP monitoring data from OSHA and did not receive any
additional data to supplement the CEHD data.
EPA also requested and received NMP monitoring data from
the DoD, specifically for paint removal and spray application
of paint containing NMP. EPA did not integrate these data
Page 86 of 205
-------
into the assessment due to lack of metadata resulting in a
confidence rating below other monitoring data used in the
assessment. These data are included and discussed in
Appendix A of the Supplemental File on Occupational
Exposure Assessment.
53
PUBLIC COMMENTS:
EPA should gather additional data or provide further
explanation related to identified issues in the final risk
evaluation, particularly related to air concentrations and
exposure times during formulation of products with
NMP, application, and use of relevant products for both
consumers and workers and laboratory use.
EPA reviewed additional sources recommended by the SACC
and public comments. For occupational exposures, EPA has
added and integrated the relevant information into the
assessment. This occupational exposure data and information
includes air concentration data and durations of tasks and
shifts for formulation, printing, electronics manufacturing and
cleaning. EPA also reviewed data for consumer products and
made adjustments to the final risk evaluation. For example,
EPA reviewed paint remover formulations and revised the
high-intensity use scenario to include paint removers with
60% NMP, instead of 50% NMP.
SACC
SACC COMMENTS:
Discuss the efforts made (if any) to obtain sample-
relevant data when not reported (generally designated as
unknown in tables) by the sources of the data.
EPA may seek critical sample-relevant metadata on a case-
by-case basis if its provision would significantly impact the
occupational exposure assessment. For NMP, EPA was not
aware of data in need of additional sample-relevant metadata
where its provision would significantly impact the
occupational exposure assessment.
SACC
SACC COMMENTS:
One Committee member found the description of the
actual amount of data available for specific occupational
exposure scenarios inadequate. Estimates of central and
high-level exposures based on just two samples are
highly uncertain, even if the analytical quality of the data
is high.
The draft risk evaluation does not describe a systematic
and robust protocol for determining when the available
measurements are sufficient to derive estimates of
exposure for an occupational exposure standard with
Table 4-47 provides a summary of actual amounts of air
monitoring data and of model estimates for each OES. Table
4-48 provides a summary of dermal parameter data and
assumptions for each OES.
For occupational exposure estimation, EPA uses data sets of
most sizes due to the prevalence of data scarcity. EPA uses
modeled estimates according to the data integration strategy
described in Appendix C of the Supplemental File on
Occupational Exposure Assessment. This appendix indicates
that EPA uses data with the highest quality ratings and may
Page 87 of 205
-------
reasonable reliability. Clarify how EPA decides when
there is enough good quality data to derive a sufficiently
reliable estimate of exposures for specific scenarios. The
risk evaluation should discuss why and where very few
data were used instead of modeling estimates.
Procedures used in estimating air concentrations when
data are very sparse should be clarified and standardized.
Recommendation: Describe the protocol used for dealing
with parameter estimation when few actual exposure
measurement data points are available and describe the
protocol for deciding to use modeled values rather than
available measurements.
supplement the highest quality data with data of lower quality
when warranted. This appendix also discusses the factors,
including dealing with few measurements or samples or using
modeled values or assumptions, for assigning confidence
ratings for PBPK input parameter sets.
Clarification/rationale/justification needed
SACC
SACC COMMENTS:
Clarity on how central and upper exposure tendencies
were derived is needed. The risk evaluation needs to be
more explicit about uncertainties in these estimates since
the risk evaluation assumes that workers and ONUs have
similar central tendency inhalation exposures.
In Section 2.4.1.1, EPA explained for occupational exposures
how EPA selected grouped sets of individual input parameter
values intended to represent central tendency and high-end
occupational exposure scenarios.
EPA clarified in all OES subsections of 2.4.1.2 that EPA
assigns the same confidence level for PBPK inputs for both
workers and ONUs because lower surface areas for liquid
contact for ONUs have higher certainty but air concentrations
experienced by ONUs have lower certainty. These factors
offset one another in determining ONU confidence level
using worker confidence level as a starting point.
In Section 2.4.1.4, EPA is explicit in the discussions of
uncertainties of individual parameters. Also, EPA modeled
exposures for workers and ONUs using both central tendency
and high-end air concentrations.
SACC
SACC COMMENTS:
Material sourced from the paint strippers work plan
document should have been sourced from original
literature for the sake of clarity.
EPA added the citations to the original reference sources
from the paint strippers work plan document in Section
2.4.1.2.8 and in the Supplemental File on Occupational
Exposure Assessment.
Page 88 of 205
-------
SACC
SACC COMMENTS:
Provide the rationale for assigning medium reliability for
exposure estimates assuming PPE use when the Agency
acknowledges that information on glove use is lacking.
In Chapter 2 and in the Supplemental File on Occupational
Exposure Assessment, EPA has removed assignments/
assumptions of specific glove PFs to apply to each OES.
Table 2-77 has been updated to include worker exposures for
all glove PFs for all OESs. EPA clarified the uncertainties of
PF values in the overall confidence rating discussion at the
end of each OES subsection in Section 2.4.1.2. Additionally,
EPA added Appendix C of the Supplemental File on
Occupational Exposure Assessment, which describes how
various factors including the glove PFs impact the confidence
of the PBPK inputs.
SACC
SACC COMMENTS:
Provide clarification on how specific scenarios were
mapped to subcategories and categories of use across
multiple industries (NMP risk evaluation, Sections 1.4.1
and 2.4).
In Section 2.4, EPA clarified how the conditions of use listed
in Table 1-6 were crosswalked to the occupational and
consumer exposure scenarios assessed in this report provided
in Table 2 2: EPA crosswalked/mapped the exposure
scenarios to conditions of use using professional judgment
based on reasonably available data and information.
Consumer confidence ratings
SACC
SACC COMMENTS:
Some committee members questioned why all consumer
scenarios have the same level of overall confidence. The
extent to which this single input is driving the level of
confidence was questioned and it was suggested that
some type of sensitivity analysis might better quantify
confidence.
As described in Section 2.4.2.6, there is an absence of direct
measurement and monitoring of consumer exposures to NMP.
Exposure estimates for consumers are therefore all based on
modeling approaches that rely on similar sets assumptions
with similar sources of uncertainty. This results in similar
levels of overall confidence across all consumer exposure
scenarios.
Occupational confidence ratings
SACC
SACC COMMENTS:
Recommendation: Tabulate specific condition of use
factors influencing confidence levels in a manner
analogous to tables of input parameters (NMP risk
evaluation, Table 2-66) and exposure results (NMP risk
evaluation, Table 2-67), in order to increase
transparency.
EPA clarified in the introductory paragraphs of Section
2.4.1.2 that key strengths and limitations of each PBPK input
parameter set are listed and used to determine qualitative
overall confidence ratings, and these lists and ratings are
provided at the end of each OES subsection as well as
clarifying that the occupational exposure data integration
strategy and factors impacting the overall confidence ratings
Page 89 of 205
-------
are available in Appendix C of the Supplemental Information
on Occupational Exposure Assessment. EPA added
explanations of the factors impacting qualitative confidence
ratings and their directional influence in this appendix. EPA
does not have a feasible design for such a tabular summary of
factors and directional influence for each OES, which would
essentially duplicate the paragraphs at the end of each OES
subsection in 2.4.1.2.
Uncertainty assessment
SACC
SACC COMMENTS:
Some Committee members expressed dissatisfaction
with the qualitative treatment of uncertainty in the draft
risk evaluation, preferring a more formal probabilistic
approach.
Recommendation: Perform more extensive uncertainty
assessment, including clearly discussing the uncertainties
related to the assumptions used in the draft risk
evaluation. These uncertainties should be discussed in
context of choice of the MOE.
EPA considers the uncertainties associated with each
condition of use, and how the uncertainties may result in a
risk estimate that overestimates or underestimates the risk.
Based on such analysis in part, EPA determines whether or
not the identified risks are unreasonable.
To the extent possible, EPA has characterized uncertainties
related to exposure, hazard, and risk. EPA lacks reasonably
available information to quantify all sources of uncertainty.
Sources of uncertainty that can be quantified are described
quantitatively, but those for which information is lacking are
described qualitatively. EPA has inserted additional
discussion of the sources of uncertainty related to PBPK
modeling of exposure estimates in Section 4.3.
Regarding occupational inputs to PBPK exposure modeling,
EPA has fully explained uncertainties in Section 2.4.1.4.
While most uncertainties in occupational inputs can't be
quantified, this section does address directional impacts of
uncertainties to the extent feasible. EPA also varied
occupational inputs to the extent feasible. EPA added a new
qualitative sensitivity discussion on modeling to Appendix B
of the Supplemental Information on Occupational Exposure
Assessment.
Benchmark MOEs are selected based on the total uncertainty
factors associated with the hazard POD and they do not
Page 90 of 205
-------
account for uncertainty around potential human exposure
estimates. Uncertainty and variability related to human
exposure are captured in the various assumptions used to
derive exposure estimates.
Page 91 of 205
-------
5. Human Health Effects
11 iiiii;hi Mcallli KITects
Charge Question 4.1: Please comment on the reasonableness of the evaluation of human health hazards. Are there any additional
NMP specific data and/or other information that should be considered?
Charge Question 4.2: Please comment on the conclusions presented regarding the genotoxic and carcinogenic potential of NMP.
Charge Question 4.3: Please comment on the validity of endpoints selected as the basis for PODs and their relevance to the
evaluation of human health risks across lifestages.
Charge Question 4.4: Please comment on the strength of evidence for, and general applicability of fetal mortality (resorptions) for
evaluating the human health risks associated with acute exposure to NMP.
Charge Question 4.5: Please comment on the strength of evidence for, and the general applicability of decreased fetal/pup body
weight or decreased fertility for evaluating the human health risks associated with chronic exposure to NMP.
Charge Question 4.6: Please comment on whether the document adequately identified uncertainties, assumptions, and data gaps
associated with the selected PODs and whether the analysis addressed tiem sufficiently
Summary of Comments lor Specific Issues Related lo
Charge Question 4
KPA Response
Hazard identification - Additional information to include or exclude
SACC
SACC COMMENTS:
Recommendation: Include more detailed summaries and
describe the context of all studies that were only briefly
summarized in the risk assessment studies referenced in
the current NMP risk evaluation.
Recommendation: Clearly state how summaries of
studies where full text is unavailable are being used in
health hazard identification and risk characterization.
The draft risk evaluation cites past draft risk assessments
of NMP performed by EPA (2015) and the Organisation
for Economic Cooperation and Development (OECD,
2007). Some of these previous documents contained
summaries of studies to which the Committee did not
have full access to. Several Committee members
concluded that these data should not be considered (e.g.,
The studies that are the basis for quantitative analysis in
hazard characterization are available to EPA and the public.
Other studies were not fully available to EPA. EPA
acknowledged results presented in summaries of these
studies, but did not rely on results summarized by secondary
sources for dose-response information. Where possible, EPA
has revised the final risk evaluation to cite primary sources
rather than secondary sources and to be more transparent
about which studies are available to EPA and have gone
through data quality review. For example, in the genotoxicity
section EPA has inserted data quality ratings for studies that
were available to EPA. While EPA acknowledges and has
reviewed summaries of additional studies, conclusions on
genotoxicity are based on information in the studies that were
fully available to EPA.
Page 92 of 205
-------
be given 0 weight in the weight of evidence [WOE]),
since the ability to peer review the risk evaluation
requires independent analysis of all underlying data, and
that is not possible with data from summary-only
sources.
Genotoxicity and cancer studies were available only as
summaries for review. The committee was unable to
validate the interpretation of cancer data provided in
Table 3-6 because the data were not provided or
available in supplemental files
The cancer studies summarized in Table 3-6 (Mallev et al
2001; Lee et al 1987) were available to EPA as unpublished
study reports but results of these studies are also available to
EPA and the public as peer-reviewed papers published in
scientific journals. Footnotes to the table cite the unpublished
studies that correspond to these published studies.
SACC
SACC COMMENTS:
Recommendation: Consider using information not
otherwise useful for POD derivation for qualitative
supporting evidence in the evidence integration phase of
the risk evaluation.
EPA includes several studies in the evidence integration
portion of the assessment that provide qualitative evidence
even though they are not able to inform dose-response
analysis. These include studies that report developmental
neurotoxicity and developmental toxicity but only evaluate
effects of single high doses (Hass et al 1995; Hass et al
1994), the case report of acute maternal effects and pregnancy
loss at 31 weeks following an accidental occupational
exposure to NMP (Solomon et al 1996), and a studv with rat
whole embryo cultures that provides additional evidence of
embrvotoxicitv (Flick et al 2009). While these studies do not
provide quantitative information sufficient to support
derivation of a POD, they are discussed in Section 3.2.4
(Weight of Scientific Evidence) of the risk evaluation and
contribute to the overall weight of the scientific evidence.
SACC
SACC COMMENTS:
Recommendation: Include human data from the worker
studies, including Haufroid (2014).
With respect to adult neurotoxicity, it is important to
include the results from the Haufroid (2014) worker
study that identifies neurotoxicity as an impacted health
endpoint.
In response to this comment, EPA reconsidered human
evidence from the Haufroid et al. (2014) paper as well as the
Nishimura (2009) studv, both of which were evaluated using
EPAs systematic review data quality criteria. The Haufroid
2014 study was primarily focused on evaluating the efficacy
of biomarkers of exposure to NMP. Evaluation of
neurotoxicity in this study was limited to self-reported
Page 93 of 205
-------
symptoms in response to a survey. In addition, the sample
size for currently exposed individuals was very small (n=8)
while other study participants (n=22) were recalling past
exposures and symptoms. The study did not find any
association between NMP exposure and any of the health
effects evaluated, including neurotoxicity. As stated by the
authors, "the number of symptomatic (work-related or not)
cases was mostly low (below 10 or even 5 in each cell)
making interpretation quite difficult and consideration of
possible confounders impossible."
The Nishimura (2009) study also has a limited sample size
(n=14 exposed workers). It included some clinical
neurobehavioral endpoints in addition to survey responses.
The study found no significant association between
occupational exposure and any of the clinical endpoints. The
mean score on the self-rated depression scale was
significantly lower in exposed workers relative to controls,
but the study reports that NMP exposure did not contribute to
SDS in multiple regression analysis.
Overall, EPA concluded that the human health hazard
information provided by these small occupational studies is
limited and difficult to interpret. In response to this comment,
EPA inserted a brief reference to these studies in the hazard
identification portion of the risk evaluation: "Two cross-
sectional occupational epidemiology studies report no
significant association between NMP exposure and
neurobehavioral endpoints, but very small sample sizes and
limitations in study design (including reliance on self-
reported effects for many endpoints) make it difficult to
interpret these results (Haufroid et at.. 2014; Nishimura et at..
20091"
Page 94 of 205
-------
Hazard identification - Irritation and sensitization
SACC,
32
SACC COMMENTS:
Recommendation: Provide better characterization of
dermal exposure studies and reevaluate if available data
are sufficient to determine if NMP is an irritant or
sensitizer.
Recommendation: Identify the inefficient assessment of
NMP-induced irritation/sensitization in the available
literature as a data gap.
The Committee indicated concern with the inaccurate
categorization of study results to "Irritation and
Sensitization" and "Immune Toxicity." These are terms
which could be misleading, adding confusion to the risk
evaluation.
Most studies employed oral or inhalation routes of
exposure, but one of the major routes of exposure for
humans is dermal. Therefore, the Evaluation should
discuss whether absorption of NMP via oral, inhalation
and dermal routes results in similar
PUBLIC COMMENTS:
EPA should conduct a more thorough evaluation of the
evidence, including consideration of data quality and
human relevance, to better characterize potential irritant
and sensory effects.
EPA has slightly modified the narrative around irritation to be
more specific. The final risk evaluation makes clear that there
is not enough data to determine whether NMP is an irritant or
skin sensitizer, stating, "Limited data from secondary sources
suggesting that NMP is not a sensitizer (RIVM. 2013; Lee et
al.. 1987) are insufficient to suDDort conclusions on
sensitization with a high degree of confidence." EPA
identifies the limited data on sensitization as a source of
uncertainty in Section 3.2.6.
As described in Section 3.2.2 on toxicokinetics and Section
3.2.5.5 on derivation of internal doses, the available
toxicokinetic studies demonstrate that NMP readily enters
systemic circulation following inhalation, dermal and oral
exposures. EPA assumed that once NMP enters systemic
circulation, all routes of exposure result in similar
distribution. The revised risk evaluation includes this
additional description of evidence for distribution to tissues
from systemic circulation: "In rats administered a single
intravenous dose, NMP was distributed to all major organs
with the highest concentrations detected in the liver and
intestines (Wells and Digenis. 1988)." In the PBPK models
EPA used to establish PODs and estimate human exposures,
distribution of NMP to tissues is assumed to be flow-limited.
Hazard identification - Developmental neurotoxicity
SACC,
38
SACC COMMENTS:
Recommendation: Integrate the Hass et al. (1994) study
findings in the WOE discussion on developmental
toxicity.
PUBLIC COMMENTS:
The Hass et al., 1994 study is included in the WOE
discussion on developmental toxicity and was considered in
the dose-response portion of the risk evaluation (e.g., see
Table 3-8 and Figure 3-4). EPA has inserted a statement
about the neurodevelopmental effects reported in Hass et al
1994 into the Weight of Evidence section for developmental
toxicity: "Hass et al. ( 1) also reported
Page 95 of 205
-------
EPA should regulate NMP as a developmental
neurotoxic agent, with potential lasting adverse effects
on neurological functioning.
neurodevelopmental effects following inhalation exposure
during gestation. The effect was evaluated at a single dose
and has not been evaluated in other studies, resulting in a lack
of information about potential neurodevelopmental effects at
lower exposure concentrations."
As stated in the risk evaluation, "Effects on postnatal
neurological behavior were reported following whole-body
inhalation exposure to 151 ppm (612 mg/m3) NMP during
gestation (Hass et al.. 1994). However, because behavioral
effects were only evaluated at this single exposure level, no
NOAEL has been identified for developmental neurotoxicity
and dose-response for this endpoint cannot be characterized."
In the absence of other data evaluating potential
neurodevelopmental effects of lower levels of NMP, EPA is
unable to further evaluate this endpoint or incorporate it into
dose-response assessment, but it does contribute to the overall
weight of the scientific evidence. EPA acknowledges the
absence of other data on developmental neurotoxicity as a
source of uncertainty in the risk evaluation.
Hazard identification - Scope and data gaps
51,34,
38
PUBLIC COMMENTS:
EPA should address all NMP-related health endpoints -
neurotoxicity, liver and kidney effects, immunotoxicity,
and developmental and reproductive harm - in its final
risk evaluation or provide a detailed science-based
justification for retaining its current narrow approach.
EPA is obligated under TSCA to obtain and assess the
information necessary to determine whether health
effects that are now poorly characterized present
unreasonable risks of injury. The proper time to take
these steps is before EPA initiates a risk evaluation.
EPA evaluated the reasonably available information from
animal toxicology studies and narrowed the scope of its
hazard identification based on the available evidence. While
EPA summarized reasonably available evidence for other
health endpoints, the systematic review process and weight of
the scientific evidence discussion are focused around
endpoints that have been identified previously as primary
targets of NMP. While EPA agrees that there is limited
information on some endpoints, EPA considers the database
adequate for risk evaluation without the need to separately
address immune effects on their own.
SACC,
51
SACC COMMENTS:
The Committee wondered if NMP is an immunotoxicant.
EPA evaluated the reasonably available information from
animal toxicology studies. While EPA agrees that there is
Page 96 of 205
-------
The Committee concluded that available data are
unsuitable for determining dose-response for
immunotoxicity because of use of outcomes that may or
may not derive from an immune reaction cascade and
study timeframes representing acute, not chronic,
exposures.
PUBLIC COMMENTS:
The insufficiency of immunotoxicity data should be
considered a data gap for NMP. EPA should use its
information gathering authorities to fill this data gap.
limited information on immunotoxicity, EPA considers the
database adequate for risk evaluation without the need to
separately address immune effects on their own. EPA
identifies the limited data on immunotoxicity as a source of
uncertainty in Section 3.2.6.
SACC
SACC COMMENTS:
Recognize and directly state that some endpoints such as
cardiometabolic and endocrine are not able to be
adequately assessed given the WOE.
EPA considers the database adequate for risk evaluation
without the need to separately address cardiometabolic and
endocrine effects on their own. EPA revised Section 3.2.6 and
Section 4.3.5 on Human Health Hazard Assumptions and
Uncertainties to specifically identify the limited data on
sensitization, immunotoxicity, cardiometabolic effects,
endocrine effects, and developmental neurotoxicity as sources
of uncertainty that may result in an underestimate of risk.
Hazard Identification - Maternal toxicity
SACC
SACC COMMENTS:
Recommendation: Evaluate maternal systemic toxicity as
a separate and distinct endpoint from fetal toxicity.
Committee members felt that there was evidence
supporting both maternal systemic toxicity as well as
fetal toxicity in the studies provided, but that it is
difficult to distinguish between developmental and
maternal toxicity given the data presented.
An integrated assessment of developmental and maternal
toxicity was conducted, in accordance with Agency policy
( ). Since EPA and OECD guideline prenatal
developmental and reproductive toxicology studies are
designed to include doses that are maternally toxic, this is not
an uncommon issue. For studies in which developmental
outcomes are observed at doses that are not excessively toxic
in the dam, current information is inadequate to assume that
developmental effects are the result only of maternal toxicity.
When the developmental LOAEL is the same for the adults
and offspring, it might be because both the adult and
developing organisms are sensitive at that dose level. Even if
the effect in the developing organisms are secondary to
Page 97 of 205
-------
maternal toxicity, the effects might be reversible in adults
(e.g., body weight deficits) yet permanent or of greater
severity in the offspring (e.g., death). Furthermore, no
additional information, e.g., as recommended in Beyer et al.
( ), was identified that would further elucidate the
contribution of maternal toxicity to developmental outcomes
for the available studies.
Interestingly, the OECD had reviewed available
developmental and reproductive toxicology studies for NMP
in the course of implementation of the HPV Programme,
noting that maternal toxicity had been observed at treatment
levels that were maternally toxic. Notably, the OECD NMP
report specifically concluded that "developmental effects are
not considered secondary to maternal toxicity" (OECD.
2007Y
SACC
SACC COMMENTS:
Recommendation: Include a discussion of the
uncertainties and assumptions relating to differentiating
maternal and fetal toxicity and with the choice of fetal
mortality or resorption as the endpoint for acute
exposure.
Section 4.3.5 includes a discussion of the uncertainties related
to reported effects on maternal body weight in some of the
developmental studies considered in hazard characterization.
54
PUBLIC COMMENTS:
Given that doses in the Saillenfait et al. (2002, 2003)
studies resulting in statistically significant increases in
fetal mortality were the same or higher than doses
causing maternal toxicity, EPA should consider re-
evaluating the fetal mortality POD. At the very least,
EPA should provide additional discussion of this issue.
As described in Section 4.3.5, "The maternal effect reported
in the Saillenfait (2003) inhalation study (transient decrease
in body weight gain and food consumption) has been cited as
a confounding factor by some study authors. EPA does not
concur with this assertion, specifically as it relates to the
observed decrease in maternal body weight gain on GD 6-21
(minus gravid uterine weight). Although a decrease in
maternal body weight gain was observed, it is not statistically
significant. Dams weighed roughly 235 g at GD 0, and
whereas the controls gained approximately 32 grams, the high
dose dams gained slightly less, roughly 26 grams. Given the
lack of significant change in maternal body weight gain, it is
Page 98 of 205
-------
unlikely that the observed decreases in fetal and pup body
weights reflect a secondary effect of maternal toxicity. In
other key and supporting studies, including an inhalation
studv (Solomon et al 199\ 1' ! Huoont De Nemours & Co.
1990), and an oral savage studv (Saillenfait et al 2002),
similar decreases in pup body weight were observed at
similar exposure levels, in the absence of any effects on
maternal body weight."
Hazard Identification - Carcinogenicity
SACC,
48,51,
34
SACC COMMENTS:
Recommendation: Although not considered a most
sensitive health effect associated with NMP exposures, a
WOE carcinogenesis summary should be added to the
risk evaluation.
Recommendation: Recognize that there is evidence that
NMP has mitogenic properties while acknowledging that
there are not enough data to conclude whether NMP is a
promotor.
The Committee was split on whether there are enough
data to draw the conclusion that there is no carcinogenic
potential from NMP exposures. Several Committee
members noted that there is evidence of effects in the
current data, including neoplasms at multiple stages and
evidence of stimulatory proliferation, and others
suggested that additional testing must be done in at least
one other species besides the rat to be confident in
making this determination. Some Committee members
indicated that while these data are not strongly
suggestive of carcinogenicity, they do have some
positive findings that should not be disregarded and
should be noted in the report.
Some data suggest that a dose-response may be nonlinear
In response to these comments, EPA has inserted a summary
or conclusions in Section 3.2.3.2.2 and the following Weight
of Evidence for Carcinogenesis summary to the evaluation
(Section 3.2.4.2):
"The reasonably available scientific information does not
provide strong evidence for carcinogenicity. Inhalation
exposure studies are more relevant to human exposure
scenarios than oral exposure studies. The inhalation cancer
bioassav (Lee et al 1987) reported a significant increase in
pituitary adenocarcinoma incidence in rats at the middle dose
after 18 months of exposure, but no significant effect after 24
months of exposure and no effect at the highest dose. The
lack of dose-response relationship makes it difficult to
determine that effects are related to exposure and prevents
quantitative dose-response analysis. In oral dietary studies
(Mallev et al 2001), there was no significant association
between NMP exposure and increased tumor incidence in
rats. There was a small but significant increase in liver tumor
incidence in male, but not female mice. While some evidence
is suggestive of a potential cancer risk at maximally tolerated
doses, the data are inconsistent and do not demonstrate a clear
dose-response relationship. In addition, available in vivo and
in vitro studies report no evidence of genotoxicity. The
reasonably available data is insufficient to support a
Page 99 of 205
-------
with a threshold (Nohmi, 2018). The potential for a non-
monotonic dose response is not discussed in the risk
evaluation.
PUBLIC COMMENTS:
There is no apparent reason for disregarding the Malley
et al. (2001) findings. This study should have been
considered evidence supporting the determination that
NMP poses a risk of cancer. Toxicological evidence of
cancer should not be dismissed on the basis that it occurs
only in the high dose group, unless it is accompanied by
evidence of excessive toxicity.
The final NMP evaluation must fully address the
evidence of NMP carcinogenicity and make a
determination of unreasonable risk for this endpoint
using a linear low-dose extrapolation unless it can
provide convincing evidence of an MOA that is not
relevant to humans (i.e., peroxisome proliferation).
quantitative evaluation of cancer risks from NMP and EPA
did not further evaluate cancer risks in the dose-response
assessment or risk characterization."
POD derivation - Clarification of doses used
SACC
SACC COMMENTS:
Recommendation: Clarify whether maternal or fetal
doses were used for risk characterization of adverse
developmental effects.
Section 4.5 of the draft risk evaluation is not clear on
whether fetal endpoints are referenced to fetal or
maternal doses. Section 4.3.5 (NMP Risk Evaluation, p.
277, lines 6474-6476) indicates that there was a lack of
significant changes in maternal body weight for the
corresponding decreases in observed fetal and pup body
weights. These data support the concept that the fetal
dose may be more appropriate for fetal endpoints than
maternal dose, thus buttressing the use of fetal AUC and
Cmax, as opposed to maternal values.
While fetal blood concentrations would provide the most
accurate estimate of doses achieved at the target site, this type
of information is rarely available for dose-response analysis.
Dose-response relationships in developmental studies are
often based on maternal oral or inhalation doses with no
information about internal doses. For the NMP risk
evaluation, PBPK models allow EPA to evaluate exposure,
hazard, and risk in terms of internal doses. The PBPK model
does not model fetal blood concentrations. EPA assumes that
the average concentration reaching the fetus will be
proportional to maternal exposure and used maternal blood
concentrations as the metric to evaluate dose-response
relationships in developmental exposure studies. EPA has
Page 100 of 205
-------
slightly revised Sections 3.2.5.2 and 3.2.5.6 to further clarify
this point.
Acute POD - Endpoint selection and dose-response analysis
SACC,
57
SACC COMMENTS:
Recommendation: Discuss why fetal mortality/resorption
was chosen as the critical endpoint when several
endpoints, including reduced body weight and reduced
perinatal survival, were affected at similar exposure
concentrations.
Most of the Committee concluded that fetal mortality
was an appropriate endpoint for evaluating acute
exposure to NMP. It was noted, however, that several
endpoints, including reduced body weight and reduced
perinatal survival were affected at similar exposure
concentrations. The Committee did not understand why
resorption was chosen over these other measurements.
PUBLIC COMMENTS:
EPA's default assumption that developmental effects
could arise as a result of a single exposure (i.e., duration
equivalence of 15- and 1-day exposures) is a
conservative one. The NMP dose for a single-day
exposure is predicted to be approximately 2.3-fold higher
than that for a 15-day exposure to yield an equivalent
rate of fetal resorptions. Based upon this consideration,
EPA's acute POD value of 214 mg/L based on 15-day
exposures should be adjusted to a value that is 2.3-fold
higher (492 mg/L) for a single-day exposure.
In the revised risk evaluation, EPA used "post-implantation
loss" (combining resorptions and fetal mortality) as the
critical endpoint for acute exposures. Resorptions are
generally considered to be appropriate endpoints for
evaluation of acute effects while other endpoints such as
reduced fetal body weight are generally considered to be
more appropriate for evaluating risks from repeated
exposures. As discussed in the risk evaluation, "Resorptions
can occur following a single exposure during a sensitive
developmental stage and as such, resorptions and fetal
mortality are considered a relevant endpoint for acute effects
(van Raaii et al 2003)." This was demonstrated in an
analysis that compared the potency (NOAELs and LOAELs)
of developmental toxicity reported in repeated dose studies
and single dose studies (van Raaii et al 2003). Van Raaii et
al. found that there is a relatively small difference between
repeated and single dose studies in the NOAELs and
LOAELs reported for embryonic and fetal resorptions. While
the difference in potency of single and repeated doses varied
across chemicals, for some chemicals the potencies of single
and repeated doses were equal. The study authors concluded
that "resorptions observed in standard guideline-based
developmental toxicity studies are considered to be relevant
endpoints for setting limits for acute exposure."
In contrast, while reduced fetal body weight is a sensitive
endpoint that is considered a marker for fetal growth
restriction which is often assumed to be representative of
repeated dose rather than acute exposures (van Raaii et al..
2003).
Page 101 of 205
-------
EPA also considered postnatal survival as a potential
candidate for acute risk characterization but concluded that
this outcome was not consistently observed across studies and
when increased post-natal mortality was observed, the
NOAELs were within the same range as other sensitive
endpoints, such as reduced fetal body weight. Additionally,
postnatal deaths occurred following multiple in utero and
postnatal exposures, thus these outcomes were less likely to
have resulted from a single developmental insult.
54, 32,
57
PUBLIC COMMENTS:
Considerable uncertainties are associated with derivation
of risk estimates based on fetal resorptions and could be
better analyzed and documented.
Importantly, there are NMP-specific data in Schmidt
(1976) for which the issue of exposure duration can be
addressed. This study assessed the effect of multiple
exposure periods for mice exposed to NMP via
intraperitoneal injection.
Schmidt ( ) demonstrates a 2- to 3-fold increase in post-
implantation loss, as compared to untreated controls,
following a single 166 mg/kg i.p. injection of NMP in mice
on GD 7, 9, or 11. This supports the acute developmental
toxicity of NMP, and it is consistent with the premise that a
single-dose of NMP could result in embryo/fetal death. The
effect and potency of a single exposure is likely to vary
across specific days in development. As the Schmidt (1976)
paper does not test the potency of single doses at every
potentially critical period of gestation, it does not provide
sufficient evidence to conclude that there are no specific
periods of gestation during which a single exposure could be
as potent as repeated dose exposures. The Schmidt (1976)
paper does however provide additional evidence that the
difference in potency between single doses and repeated
doses of NMP is relatively small. This is consistent with the
finding of the Van Raaij analysis described above. EPA
assumed that single exposures to NMP could be as potent as
repeated dose exposures during critical periods of
development. The uncertainties and assumptions made
around selection of post-implantation loss as an acute
endpoint are described in Section 3.2.5.1.
SACC
SACC COMMENTS:
In the final risk evaluation, EPA used "post-implantation
loss" as the critical health effect for acute exposures. This
Page 102 of 205
-------
Recommendation: Distinguish between embryonic
resorption and fetal mortality in tables and text
throughout the draft risk evaluation.
Fetal mortality and fetal resorption are two different
endpoints but are considered as one in the draft risk
evaluation, with confusion evident in both tables and
text.
endpoint integrates early embryonic loss and fetal mortality in
a single endpoint. In the developmental toxicity study
protocols that evaluate these outcomes, biomarkers of
embryo/fetal death (e.g., empty implantation sites, or early
and late resorptions) encompass both embryo and fetal
developmental stages. EPA selected this combined endpoint
because it can be modeled as a dichotomous endpoint. In
addition, both stages of pregnancy loss are reported in several
studies following NMP exposure. Considering embryonic
resorptions and fetal mortality independently could
underestimate the total impact of NMP on offspring survival
through gestation.
SACC,
57
SACC COMMENTS:
Recommendation: Verify that fetal mortality data have
been analyzed on a per-litter basis, re-analyze
appropriately otherwise, and if this is not possible,
discuss the reasoning for making conclusions based on
fetal mortality data that are not litter adjusted.
PUBLIC COMMENTS:
Dichotomous data for developmental toxicity studies are
best assessed using a nested BMD model to account for
potential litter effects (i.e., effects that are not randomly
distributed across litters); however, this would require
access to the raw study data, which to date are not
available.
In the final risk evaluation, EPA modeled post-implantation
losses (including both resorptions and fetal mortality) based
on reported mean losses per litter. As stated in the revised
BMD modeling supplemental file:
"To perform this analysis incidence of post-implantation loss
from the reported litter means were modeled with BMDS
standard dichotomous models after adjusting for litter effects
using a Rao-Scott transformation. Normally, individual
animal data are necessary in order to account for intralitter
correlation present in nested developmental toxicity data (i.e.,
the observation that pups from one litter are more likely to
respond alike one another compared to pups from another
litter). In this situation, study authors were unable to provide
litter level data and instead an approximate approach was
used. Briefly, the numbers of total implantations and total
fetal mortality (dead fetuses plus resorptions) were scaled by
a design effect in order to approximate the true variance of
the clustered data. This transformation is called the Rao-Scott
transformation and has been shown to reasonably
approximate the variance due to clustering and intralitter
correlation in developmental toxicity data (Fox et al 2016)."
Page 103 of 205
-------
SACC
SACC COMMENTS:
One Committee member recommended that the study by
Bartsch et al. (1976) that reports on acute toxicity in rats
and mice to a range of solvents should be included in this
discussion as it adds to the WOE and may address
uncertainties.
Bartsch ) reported LD50s for a range of solvents
following acute exposures in rats and mice. The risk
evaluation cites a range of LD50s in rodents across different
endpoints (including LD50 data from Bartsch (1976) in the
discussion of acute toxicity in Section 3.2.3.1. The acute POD
selection is based on much more sensitive acute
developmental endpoints, which were reported at doses well
below the LD50s reported in the reasonably available data.
Chronic POD - Transparency of data
SACC,
33, 54
SACC COMMENTS:
The Committee discussed reproductive toxicity in terms
of male and female fertility and found it difficult to come
to any conclusions given the complexity, and sometimes
lack of transparency, in the data and analysis provided in
the Exxon (1991) study. It was noted that the draft risk
evaluation disagreed with the conclusions of the study
regarding male and female fertility effects.
It should be noted that methods and data from all of the
NMP Producers Group 1999 study (both rat and mouse
experiments), which also examined reproductive
endpoints, were also not available to the Committee for
this risk evaluation.
PUBLIC COMMENTS:
EPA indicated that the NMP reproductive studies were
not included because they did not have access to the
study reports. It is speculated that EPA made an
unnecessarily restrictive and interpretation of TSCA
Section 14.
EPA's independent analysis of the data in the Exxon 1991
study identified statistically and biologically significant
effects on fertility in both males and females at all doses
tested. While the study authors describe the effect as close to
the range of historical controls, EPA concluded that
concurrent controls in the study are the more appropriate
basis for comparison.
The NMP Producers Group 1999 studies were not shared
with the committee because they were not available to EPA at
the time of the meeting. In the draft risk evaluation, EPA
discussed the evidence presented in summaries of the studies
but was unable to evaluate study quality and did not rely on
quantitative dose-response information from the studies.
EPA has subsequently obtained access to the full studies.
EPA evaluated the studies using the systematic review data
quality criteria, performed dose-response analysis for
developmental endpoints reported in the studies, and
incorporated results of the studies into hazard identification,
weight of the scientific evidence and dose-response analysis.
Chronic POD - Critiques of the Exxon (1991) key study
SACC,
60, 33,
57, 33
SACC COMMENTS:
The Committee members noted that controls for the
contemporaneous experiments are typically the better
The Exxon 1991 study was inadvertently not made available
to SACC members prior to the peer review meeting.
However, it had been shared with SACC members before the
second day of the meeting.
Page 104 of 205
-------
comparison group unless contemporaneous control level
are at the extreme ends of the historical control
distribution. The data used in these analyses are not
available; hence, neither EPA nor the Committee were
able to examine contemporaneous or historical control
distributions, nor validate that conclusions were derived
from correct statistical tests.
PUBLIC COMMENTS:
Exxon (1991) should not be considered a high-quality
study and should not be the basis for EPA's human
health risk assessment.
The international regulatory authorities rated the Exxon
study (reliability score of 2) inferior in quality to those
conducted by the NMP Producers Group (reliability
scores of 1) (OECD, 2007).
Exxon (1991) failed to adjust the concentrations of NMP
in feed to reflect changes in food consumption that occur
during pregnancy and lactation.
Other issues include increased probability of
brother:sister matings, fertility problems at the specific
Charles River site, reduction of females available for
mating at the start of the P2 generation, and mating
confirmation missed by laboratory personnel within the
fertility and fecundity indices.
The reproductive effects noted in the Exxon (1991) study
are likely related to the sensitivity of the testing strain's
particular breeding background and the use of the 500
mg/kg/day dose level for the breedings to produce the
Fib, F2a, and F2b litters. It is believed that the animals
tested in the Huntingdon (1999) and BASF (1999)
studies more accurately represent the true sensitivity of
rats to NMP-related effects on fertility/fecundity.
EPA's independent analysis of the data in the Exxon 1991
study identified statistically and biologically significant
effects on fertility in both males and females at all doses
tested. While the study authors describe the effect as close to
the range of historical controls, EPA, in agreement with this
SACC comment, concluded that concurrent controls in the
study are the more appropriate basis for comparison.
Based upon the study report submitted to EPA, the Exxon
(1991) study was rated high-quality by EPA. Study quality
ratings from other organizations lack transparency and thus
may not be particularly informative. Study quality issues
raised in public comments for the most part reflect choices in
study design (e.g., not adjusting the concentration of NMP in
the feed), non-critical reporting issues, and speculation about
the adequacy of the study animals. For example, NMP
Producers Group submitted a report by Dr. Willem Faber
which concluded that the Huntingdon (1999) and BASF
(1999) (i.e., the two NMP Producers Group studies) "more
accurately represent the true sensitivity of rats to NMP-
related effects on fertility/fecundity." This conclusion was
based on Dr. Faber's review of the Exxon (1991) study, as
well as on the conclusions of two additional documents
pertaining to that study, i.e., 1) a report of a "Good
Laboratory Practices Audit" of the Exxon (1991) study
written by an independent consultant, Linda Calisti, dated 22-
Feb-2001 and 2) a review of the Exxon (1991) study (and
other related correspondence) by Dr. Mildred Christian
(Argus International, Inc.) dated 22-Jul-1999. The report by
L. Calisti concluded that the Exxon (1991) report should be
classified as non-GLP; however, the study report states that
the study had been conducted under FDA GLP regulations,
and the audit was conducted after the study data records had
already been destroyed. Although the audit identified a
Page 105 of 205
-------
For fertility and fecundity, EPA should consider placing
the data of Exxon (1991) within the context of data
collected from other reproductive toxicity studies (NMP
Producers Group, 1999a, 1999b; Solomon et al., 1995).
EPA should consider all of the fertility and fecundity
data within an overall data quality context.
number of procedural and record-keeping errors, they did not
appear to invalidate the study GLP status or invalidate the
study findings. The report by Dr. Christian agreed in principle
with information submitted to California EPA (OEHHA) and
concluded that EPA did not have sufficient information to
identify the 500 mg/kg/day dose of NMP as toxic to
reproduction. This conclusion was based on an assumption
that the Exxon (1991) study rats might have been carriers of
genetically-mediated testicular abnormalities and decreased
fertility known to occur in some Charles River rats and to be
sporadically expressed. The report indicates that this genetic
anomaly was identified in the Charles River Laboratory
Raleigh Production Room 1, the Raleigh Facility Room RIO,
and the Kingston Facility Room K83. However, according to
the GLP audit by L. Calisti, the source of the Exxon study
animals was the Kingston Facility Area K97. Thus, there is
no evidence that the Exxon (1991) study animals were
carriers of this genetic variant; it is merely speculative.
Additionally, regarding the conclusion that the NMP
Producers Group studies "more accurately represent the true
sensitivity of rats to NMP-related effects on
fertility/fecundity": first, concurrent controls were used for
statistical comparison to the treated groups in the Exxon 1991
study, not controls from the NMP Producers studies, and the
P and F1 control males did not exhibit these effects.
Secondly, it is possible that the fertility response in the Exxon
1991 study is more representative of the human population,
and thus a better predictor of the potential effects of NMP for
human health risk assessment. Infertility has been reported to
affect approximately 15% of couples globally; males are
solely responsible for 20-30% of infertility cases and
contribute to 50% of cases overall (Agarwal et al., 2015;
doi:10.1186/s 12958-015-0032-1). In the US, 9% of men aged
Page 106 of 205
-------
25-44 years of age reported consulting a physician on
infertility issues during their lifetime
(https://www.cdc.gov/reproductivehealth/infertility/).
The weight of the scientific evidence for male reproductive
toxicity in the risk evaluation (Section 3.2.4.2.) includes all
reasonably available data that inform the issue (Sitarek and
Stetkiewicz. 2008; NMP Producers Grot ?a. b; Exxon
Biomedical. 1991) and studies that provide information and
support for the mechanistic plausibility of male reproductive
toxicity following NMP exposures.
Chronic POD - Endpoint selection
33
PUBLIC COMMENTS:
An assumption of early-life susceptibility for this
endpoint is inconsistent with the exposure scenarios to
which the assessment is applied. EPA should adopt a
different endpoint that is applicable to adult worker
exposures (e.g., effects on fetal/pup body weight) or
revisit the assumption of early-life susceptibility for the
assessment of fertility effects.
EPA selected a chronic POD based on reduced male fertility
and female fecundity in rats following exposures throughout
gestation, lactation, development, and prior to mating.
Relying solely on the available studies, and without
additional targeted testing, there is no way to determine
which periods of exposure contributed most to this effect.
EPA assumes that this endpoint may be relevant for sensitive
phases of human reproductive development, including
pubertal development that may be ongoing in young workers.
This POD is also assumed to be protective of other endpoints
for which data are not reasonably available but which may be
relevant to workers.
Areas that lack clarity
SACC
SACC COMMENTS:
There is a lack of clarity on how the draft risk evaluation
discriminates between positive vs. negative results. For
example, Table 3-8 (NMP risk evaluation, pp. 187-8)
shows positive, negative, and N/A results for effects of
NMP exposures on various developmental endpoints.
How these results factor into the overall WOE is not
clear since the draft risk evaluation does not provide
Differences in outcomes across studies may be due to
differences in study design (exposure timing and duration,
timing of outcome evaluation, etc), or other unknown
confounding biological factors (strain sensitivity, metabolic
changes, etc). While statistical power is a relevant
consideration, these biological aspects of study design that
could influence outcome are also important contributors to
study outcomes.
Page 107 of 205
-------
In selecting endpoints as the basis for PODs, EPA
qualitatively considers the weight of the scientific evidence
for specific outcomes of interest based on the quality of the
studies, consistency of effects, relevance of effects for human
health, coherence of the spectrum of reproductive and
developmental effects observed and biological plausibility of
the observed effects.
EPA did not simply select the most sensitive study to be the
basis for the POD. As described in Section 3.2.5.1, EPA
considered the reduced fertility reported in several studies to
be a robust, biologically plausible endpoint that is highly
relevant to humans and that is consistent with the continuum
of reproductive and developmental endpoints reported across
available studies.
EPA did not put greater weight on studies from specific
exposure pathways. The PBPK model facilitates evaluation
NMP toxicity based on internal blood concentrations
regardless of exposure pathway. While the conditions of
inhalation studies may be more relevant to human exposures
than oral studies, they introduce more uncertainties around
the level of exposure achieved. Whole body inhalation
exposures may result in simultaneous oral exposures due to
grooming behavior, resulting in a potential underestimate of
total dose achieved. In addition, due to the hygroscopic nature
of NMP, condensation reported in whole body inhalation
studies may decrease the level of exposure to NMP achieved
through inhalation while simultaneously increasing the
amount of oral exposure to NMP following deposition on fur.
For these reasons, oral exposure studies provide more reliable
Page 108 of 205
weights for individual health outcomes examined in each
study. The risk evaluation could discuss the statistical
power associated with the test for each study/outcome as
one way of indicating which results are less likely to be a
false negative or false positive.
Suggestions:
o Consider providing a rationale for selecting the most
applicable toxic endpoint for the assessment that is
based on the available science, not simply choosing
the most sensitive, especially when it is not
corroborative with other data from other studies,
o Consider using greater weights for data collected
from the more applicable exposure pathways and
accompany the quality review score of a study with
the associated weights for data on health outcomes
addressed in that study.
-------
dose-response information, though exposure conditions are
less directly relevant to human exposures.
SACC
SACC COMMENTS:
Recommendation: Specify which reproductive effect is
particularly sensitive and consistent between studies.
In the draft risk evaluation (p. 178, lines 4371-74), the
following sentence is unclear and confusing as written:
"While reproductive effects are less consistently reported
across studies than developmental effects, reduced
fertility following exposure throughout gestation,
lactation, growth, puberty, and prior to mating is a
particularly sensitive endpoint."
EPA has revised the paragraph referenced by reviewers for
clarity: "Several studies are available to assess the
reproductive effects of NMP exposure. Reproductive effects
are less consistently reported across studies than
developmental effects, but significant reductions in fertility
were reported in three studies. The reduced male fertility and
female fecundity observed in the second generation of the
Exxon studv (1991) are particularly sensitive endooints.
These significant reductions in male fertility and female
fecundity occurred in the second generation following
exposure throughout gestation, lactation, growth, puberty, and
prior to mating. Other studies with shorter exposure periods
limited to the weeks prior to mating, also reported reduced
fertility in male and female rats (Sitarek et al 2012; Sitarek
and Stetkiewicz. 2008), although NOAELs in these studies
were higher than the LOAEL for reproductive effects
identified in the Exxon study."
Other
SACC
SACC COMMENTS:
Recommendation: Include air odor thresholds for NMP
and note that it also has poor chemosensory warning
properties.
One Committee member recommended that the risk
evaluation include and discuss the fact that the air odor
thresholds recommended for NMP at 4 ppm low to 10
ppm high are significantly above the 1 ppm PEL.
EPA has inserted the reported air odor thresholds for NMP
for reference. OSHA has not established a PEL for NMP. The
California PEL of lppm does not serve as the basis for any of
EPA's risk conclusions in this risk evaluation.
Page 109 of 205
-------
6. Dose-Response Assessments
Dose-Response Assessments
Charge Question 5.1: Please comment on EPA's use of the PBPK model used to derive internal dose estimates (Poet et al. 2010,
2016). Please comment on whether the model is clearly and transparently described and technically and scientifically adequate for
supporting the NMP draft risk evaluation. Specifically, please address the structure of the PBPK model, parameter calibration and
model predictions of the available in vivo data. Please comment on the dose metrics selected for acute (Cmax) and chronic (AUC)
PODs.
Charge Question 5.2: Please comment on the BMD analysis conducted on the endpoints identified from the key studies. Please
specify whether the BMD calculations were appropriately conducted and documented and whether the BMRs applied for each endpoint
are appropriate.
#
Summary orCommenls lor Specific Issues Related to
Charge Question 5
KPA Response
PBPK modeling - Model transparency
SACC
SACC COMMENTS:
Recommendations: Describe which version of the
PBPK model was used, and clearly describe
modifications made to the model since it was
published. Articulate clearly the meaning of "peer-
reviewed" with respect to the PBPK model used.
The Committee was unable to determine whether the
EPA 2015 peer review of the Poet PBPK models
reviewed the same models used in the draft risk
evaluation. Appendix I indicates that the PBPK models
had experienced extensive revisions since the 2015
review and might be considered a new model. Any
PBPK model employed in an evaluation should have a
recent complete peer review. The description of the
PBPK model as "peer reviewed" could be misleading.
EPA's PBPK model is based on a model that was published in
the peer-reviewed literature bv Poet et al (2010). EPA
modified the model for use in the 2015 risk assessment. There
are some differences between EPA's model and the 2016
model published by Poet et al. Appendix I (now Appendix J)
describes the EPA PBPK model used for this risk evaluation.
The Appendix has been revised to include a more complete
description of all parameters used in the model and all
modifications made to the model since the Poet et al 2010
publication.
While there are differences in the details, the model is in most
aspects that which has been described in the peer review
literature bv Poet et al. (2010). The model went through
additional review as part of EPA's 2015 risk assessment of
NMP.
SACC,
56, 33,
31, 54
SACC COMMENTS:
Recommendation: The final PBPK model code and
model parameters used in assessing each scenario
EPA shared the PBPK model code publicly immediately
following the peer review meeting. The model itself has not
been substantively changed since the 2015 risk assessment;
Page 110 of 205
-------
should be fully described and made available for
external peer review {i.e., PBPK model code files
should be provided in the docket as part of the
supplemental materials for the NMP draft risk
evaluation in a form that would allow the
informed/scientific/modeling public to review and run
to validate findings).
There was concern that the PBPK model used had been
modified since the EPA 2015 review and a full review
of the final code used had not been performed. Without
the final code, validation by members of the
Committee was not possible. The complete and
documented PBPK model code used in the draft risk
evaluation was not available to the Committee for its
experts to fully review. Code is shown for various
aspects of the model but not for all components and not
in a systematic manner.
PUBLIC COMMENTS:
Access to the full code, the model input parameters,
and tabular outputs are requested.
Other information needed includes weight fraction in
the liquid product, skin surface area, GPF, dermal
exposure duration, air concentration, and worker body
weight.
only model inputs have been modified. A summary of model
parameters, corrections made since the 2010 model and
additional minor corrections made in response to this public
comment process are described in the revised model code and
Appendix J. Scenario-specific exposure model inputs
(including weight fraction, body weights, exposure duration,
skin surface area, etc), PBPK exposure estimates and risk
calculations are all included in the occupational exposure and
consumer exposure supplemental excel files.
SACC,
54
SACC COMMENTS:
Recommendation: Explain and clarify the structure of
the PBPK model as used to derive internal doses of
NMP in this risk evaluation. Some on the Committee
remarked that the information provided in the draft risk
evaluation (Appendix I) is insufficient to fully
document the modeling approach used (for example,
no diagram is provided describing the rat or human
To increase transparency of the model, EPA has expanded its
discussion of the PBPK model in Appendix I (now Appendix
J). EPA inserted a figure outlining the structure of the rat and
human PBPK models (as described by Poet et al (2010) and
added tables summarizing all partition coefficients and
parameters used in the model.
Appendix J also documents optimization of specific
Page 111 of 205
-------
PBPK model structure, with all tissue compartments
described, etc.).
PUBLIC COMMENTS:
EPA should expand its discussion of this specific
PBPK modeling exercise, including its overall
objective and any conclusions that can be drawn based
on the calculated applied doses across candidate
studies.
parameters that was performed based on data from available
studies, including a discussion of the purpose and conclusions
of these parameter optimization exercises.
SACC
SACC COMMENTS:
Recommendation: Poet et al. (2016) should be
referenced in the risk evaluation and details of the two
human dermal studies clearly described in the risk
evaluation.
An inhalation study in adult humans was used in the
draft risk evaluation (Bader and Vanthriel, 2006). The
description seems to differ from what was reported in
Poet et al. (2016). In the study, dermal penetration was
estimated for volunteers wearing shorts and tee shirts
and volunteers breathed filtered air, so there was no
exposure to the vapor. Details about the clothing worn
by the inhalation group in the other study are missing
and should be provided.
As described above, the PBPK model was modified from the
Poet (2010) publication for use in the 2015 risk assessment.
Since then, some additional modifications have been made in
response to public comment. Poet et al (2016) is not cited in
the risk evaluation because it is not consistent with the PBPK
model used by EPA. Several parameters differ in the two
versions of the model. Appendix J is the appropriate reference
for details of the PBPK model. To increase model
transparency, the appendix has been modified to provide
additional details of the PBPK model structure and parameters.
As provided for the Akesson dermal study, details about
clothing worn in the Bader and Van Thriel inhalation study are
included in Appendix J: "Volunteers wore slacks and T shirts
and thus had arms exposed to vapor."
SACC,
54
SACC COMMENTS:
Recommendation: Describe and tabulate the settings
for PBPK model internal parameters used for each
condition of use scenario and make available in a
supplemental document.
Values for internal parameters (body weight, volumes
of tissues, respiration rate, flow rates through and
among tissues, and other physiological rate parameters,
etc.) used to model the internal dose metrics were not
readily available. These are critical to understanding
The standard input parameters and partition coefficients used
in the PBPK model are now summarized in tables in Appendix
J. The input parameters that are exposure-scenario specific can
be found in the supplemental occupational and consumer risk
calculator excel files. The risk calculator files contain all
scenario-specific PBPK model inputs as well as the PBPK
outputs {i.e., exposure levels predicted by the PBPK model)
and the final risk calculations based on those predicted
exposures. EPA's PBPK model code is publicly available on
the EPA website.
Page 112 of 205
-------
the adequacy of the PBPK model in occupational
settings.
PUBLIC COMMENTS:
EPA should also provide an example calculation with
selected model inputs and provide a citation to the
exact location of the PBPK files where the code and
output specifically for the applied dose calculations are
located.
PBPK modeling - Assumptions, uncertainties, and sensitivity analysis
SACC,
54
SACC COMMENTS:
Recommendation: Explicitly state assumptions and
uncertainties associated with use of the physiologically
based pharmacokinetic (PBPK) model to estimate
internal dose metrics, particularly route-to-route
extrapolation. Provide estimates of the magnitude of
influence of each assumption on the final estimate of
the POD.
PBPK model calibration represents multi-variable
fitting, which can result in apparently good fits without
being physiologically logical or "correct." This
uncertainty is not discussed in the draft risk evaluation.
One Committee member suggested that assumptions
used to construct the model, and uncertainties in the
model's parameterization, might affect the PODs.
The Committee agreed there should be more
information provided on the model and on how central
and high-end tendencies were derived.
PUBLIC COMMENTS:
EPA should provide additional discussion of
uncertainties and consider providing sensitivity
analyses using alternative PBPK parameters.
EPA has inserted additional discussion of the assumptions and
sources of uncertainty related to the PBPK models used to
derive PODs from animal data and to estimate human
exposures for each COU in Section 4. To the extent possible,
EPA describes the potential magnitude of each source of
uncertainty. For several sources of uncertainty, EPA lacks the
quantitative information that would be necessary to
characterize uncertainty for some parameters.
For many human exposure scenarios, EPA relies on
assumptions about specific exposure parameters for which
there is a lack of reliable data. Where plausible alternate
assumptions have been proposed by stakeholders, EPA has
considered 'what if scenarios, estimating exposure based on
several alternate assumptions. These alternate exposure
estimates provide an indication of the magnitude of
uncertainty associated with EPA's assumptions. EPA believes
that the qualitative and quantitative analyses included in the
risk evaluation provide sufficient information to support risk
conclusions.
A comprehensive quantitative uncertainty analysis for all
parameters used in the rat and human PBPK models would
require substantial additional analysis that is beyond the
current capacity of the agency within the time available for
completion of risk evaluation. For example, one source of
Page 113 of 205
-------
uncertainty is that the model only uses a single compartment
to describe the skin vs. multiple skin layers or a partial
differential equation. EPA is not able to quantify the
uncertainty from this assumption without building the more
elaborate model forms, calibrating new parameters and/or re-
fitting new ones, and seeing how that impacts model results.
PBPK modeling - General model design and assumptions
SACC
SACC COMMENTS:
Recommendation: Verify that there is no concern for
buildup of NMP during prolonged/repeated exposures
by applying the PBPK model to an appropriately
selected scenario.
A Committee member indicated that more support is
needed for the statement made in footnote 1 of Table 4-
3 (NMP Risk Evaluation, p. 212), which states: "It is
assumed that there is no substantial buildup of NMP in
the body between exposure events due to NMP's short
biological half-life (-2.5 hrs.)" Support for this
statement can be provided by applying the PBPK
model to a suitable scenario (e.g., one that would be
expected to show build up, such as from long
exposures at high concentrations).
The meaning of the phrase "substantial buildup" in this
context needs definition.
To provide a quantitative basis for the assertion that there is a
lack of "substantial buildup", EPA has added additional
analysis demonstrating that there is no buildup of NMP week
to week. For workplace exposures, seven days are simulated,
with the first five being exposure days, and average or peak
concentration is calculated over this time. Simulations were
performed starting on either day 0 or at 8 months of pregnancy
to assure that the largest internal dose that might occur from a
work week during pregnancy is calculated, but the difference
between those two points of pregnancy was minimal. As is
shown by the example plots now included in Appendix J, the
blood concentration returns essentially to zero by the end of
the 7th day; i.e., after 2 (weekend) days without exposure,
even for the highest exposure scenario.
For evaluation of consumer exposures, it is assumed that these
are single-day exposures associated with home projects that
are not performed repeatedly over multiple days. So,
accumulation cannot occur for such scenarios.
SACC
SACC COMMENTS:
Recommendations: Justify running the PBPK model
for the first day or week of pregnancy instead of for the
entire pregnancy.
The text states that the PBPK model was run for the
first day or week of pregnancy when physiological
changes are minimal; however, the point of using a
pregnancy model is to be able to predict dose metrics
For rat developmental studies, simulations were run for the
entire portion of pregnancy during which exposure occurred
and peak and average blood concentrations were calculated.
When calibrating the model against human PK collected from
non-pregnant adults, it is appropriate to use simulations when
there are minimal changes vs. later in pregnancy.
For human pregnancy, simulations were run to estimate the
internal doses at several points in time during pregnancy, from
Page 114 of 205
-------
for the entire pregnancy to capture the impact of the
physiological changes during pregnancy. This requires
more computational time, but it should be done even if
only for one case to demonstrate that EPA's
assumptions would result in the most conservative
estimate.
beginning to end. It was found that the highest internal dose
was predicted for early pregnancy because respiration rate
increases as BW0.75 and dermal contact is assumed to be
constant, while BW increases, indicating that internal
dose/BW decreases as pregnancy progresses. Therefore, the
early pregnancy simulations provide the most health-
protective estimates.
SACC
SACC COMMENTS:
Recommendation: Describe the process used to
develop the pregnant human female PBPK model and
how it was used along with the pregnant female rat
PBPK model to inform chronic toxicity endpoints.
It appears that no data informing processes in pregnant
human females were used in developing the pregnant
human female PBPK model. The approach used was
not clearly described.
One Committee member explained that, during
pregnancy, induced enzyme activities in humans could
be variable and increase up to 30%. It was unclear to
the Committee the extent to which the PBPK model
accounts for induced enzyme activity and if it does, the
extent to which this source of variability is accounted
for. This needs clarification.
The PBPK models for pregnancy accounted for pregnancy-
related changes in blood flow, respiration and body weight.
Changes in enzyme activity that occur during pregnancy are
not captured in the PBPK model. There is evidence that
Cytochrome P450 2E1 (CYP2E1) contributes to NMP
metabolism (ligocka et al. 2003) and that CYP2E1
expression decreases during pregnancy in rats and mice, but
there is insufficient information about the overall impact of
these changes on NMP metabolism. There is also insufficient
information about other metabolic pathways that contribute to
metabolism and how they may change during pregnancy. EPA
concluded that there is not sufficient quantitative information
about pregnancy-related changes in NMP metabolism to
incorporate this into the model. In the absence of more
quantitative information, EPA assumed the interindividual
uncertainty factor of 10 (with a factor of 3 designated for
toxicokinetic differences across individuals) is sufficient for
addressing metabolic differences associated with pregnancy.
SACC
SACC COMMENTS:
Recommendation: Demonstrate that modeling a single
exposure on each day of gestation is comparable to
running repeated dosing and computing the average
AUC.
If the repeated exposures are going to be used to
calculate internal dose metrics that are then used to
assess acute exposure, then modeling a single exposure
As described above, for workplace exposures, seven days are
simulated, with the first five being exposure days, and average
or peak concentration is calculated over this time. Simulations
were performed starting on either day 0 or at 8 months of
pregnancy to assure that the largest internal dose that might
occur from a work week during pregnancy is calculated, but
the difference between those two points of pregnancy was
minimal. As is shown by the example plots now included in
Page 115 of 205
-------
on each day of gestation must be demonstrated to be
comparable to running repeated dosing and computing
the average AUC. To this end, it could be useful to
have a table comparing the single dose and repeated
dose studies mentioned in the draft risk evaluation (p.
195, Section 3.2.5.1, lines 4547-4553) that were used
to justify this approach.
Appendix J, the blood concentration returns essentially to zero
by the end of the 7th day; i.e., after 2 (weekend) days without
exposure, even for the highest exposure scenario. This
demonstrates the buildup is not expected from repeated dose
exposures during pregnancy.
SACC
SACC COMMENTS:
Recommendation: Explain how the co-exposure of
NMP with 2-pyrrolidinone impacts the
pharmacokinetics of NMP.
The oral data used for calibration of the rat model was
for NMP co-exposure with 2-pyrrolidinone. There is no
discussion in the publication's supplemental material
as to how the co-exposure could have affected the
pharmacokinetics of NMP. These same data were used
to justify the addition of dual absorption. While the
authors also cite other publications to support this
modification, it is unclear if any of these publications
are NMP-specific.
Clarify how oral data by Midgley et al. (1992) used to
fit a 2-compartment stomach in a study that co-
administered NMP with another chemical, 2-
pyrrolidinone, was acceptable for describing the
absorption pharmacokinetics of NMP.
Since absorption of NMP is not known to be transporter-
mediated {i.e., occurs by simple chemical diffusion) it is
unlikely that the presence of 2-pyrrolidinone would have an
effect. Further, the same absorption rate constant adequately
fits the Ghantous data from female rats, though it over-predicts
the Ghantous male rat data, which were obtained with no co-
exposure. If anything, the co-exposure would be predicted to
reduce absorption though some form of competition. Thus, in
combination with these other data, there is no indication that
absorption is under-predicted (which would lead to an under-
prediction of risk). With the focus of the assessment being on
developmental risk, the fact that the male rats in the Ghantous
study appeared to absorb NMP more slowly than the females
is not a significant issue.
SACC
SACC COMMENTS:
Recommendation: Comment on the impact of not
including dermal uptake of NMP vapors or oral uptake
of NMP during grooming on the internal doses in the
rat PBPK model and how this impacts the estimates of
and associated confidence in the final POD value.
Air exposures in the PBPK model were parameterized
based on a nose-only exposure study (Ghantous, 1995),
Separate model code was developed for rats vs. humans. The
rat code was not changed to include vapor-through skin
absorption. Vapor-dermal uptake and grooming were not
included, but it is presumed that these contribute little to
overall uptake (< 20%). Because those routes are not included,
the estimated inhalation dose from inhalation studies in rats
could under-predict the total exposure levels that result in
reproductive toxicity effects in animal. While this source of
Page 116 of 205
-------
and then used to predict whole-body exposures
described in Saillenfait et al. (2002; 2003). Therefore,
in rats, model predictions for whole-body exposures for
developmental toxicity studies do not include any
contributions from dermal uptake of vapors or oral
uptake via grooming. Estimated internal doses could be
lower than the actual internal doses in these studies,
resulting in an underestimate of the POD.
uncertainty could result in an underestimate of the dose at the
POD, this would not contribute to an underestimate of risk
because the primary studies used for quantitative analyses for
the PODs following acute and chronic exposures would not be
impacted by this source of uncertainty.
SACC
SACC COMMENTS:
The PBPK model models skin as a Continuously
Stirred Tank Reactor (CSTR). Traditionally,
experimental permeation data are evaluated using
mathematics representing steady state diffusive flux
through a membrane. Diffusive transport in a
membrane and loading/release in a CSTR are
fundamentally different. Matching of traditionally
determined experimental permeability coefficients and
permeability coefficients backfitted by PBPK modeling
should not be expected or viewed as confirmatory.
While the U.S. EPA clearly will correct model features that
are considered to be critically deficient, even if a model has
been accepted for peer-reviewed publication, EPA does defer
to the journal peer review process, particularly for aspects of a
model that do not critically impact model predictions. This
aspect of the model structure was clearly present in the model
as originally described by Poet et al.( ). The fit to the rat
dermal exposure data of Pavan et al. (2003) is adequate and is
not critical since the rat bioassays were by oral or inhalation
exposure. The fit to the human dermal PK data of Akesson et
al. (2004) was also considered adequate. In particular, the peak
concentration after exposure to neat or 50% NMP was not
under-predicted, and while the initial rate of absorption from
50% NMP was somewhat over-predicted, the AUC from that
exposure was not under-predicted. So, use of this model
feature as accented for the Poet et al. (2010) publication does
not appear to under-predict human dosimetry and hence risk,
and therefore it is appropriate for the EPA to defer to the
journal peer review acceptance.
The comment also implies that fitting of the permeability
constant to the PK data as occurred for this model should not
be acceptable. EPA respectfully disagrees, as it is typically the
case that PBPK parameters are empirically fitted to
observational data, as occurs for the oral absorption and
metabolic elimination constants for this model. By analogy,
Page 117 of 205
-------
inhalation uptake for some gases is less than predicted by the
alveolar equilibration equation because a process referred to as
"wash-in/wash-out" occurs in the upper respiratory tract and
conducting airways. This effect is typically accounted for by
empirically fitting an absorption fraction, since inclusion of a
mechanistic model for absorption and desorption in the
conducting airways would increase model complexity by a
large extent and is presumed not to improve the overall model
applicability for prediction of systemic dosimetry. Rejection of
empirically fitted absorption constants would reduce the
number of PBPK models considered acceptable for use, when
these are otherwise the preferred means of estimating human
dosimetry and is not viewed as an action that would
significantly improve EPA's ability to accurately predict
human risk.
SACC
SACC COMMENTS:
Permeability coefficients are vehicle specific.
Permeability coefficients obtained from neat
compound experiments (e.g., Payan et al., 2003) should
not be assumed to apply when NMP: water mixtures are
used at varying NMP weight fractions (e.g., see
Marquet et al., 2015).
EPA appreciates this comment as distinct permeability values
were estimated for neat vs. dilute NMP in humans but these
were not applied consistently for all analyses.
As noted in response to a previous comment, since most of the
rat bioassays did not involve dermal exposure, the fact that rat
dermal permeability was only measured with neat NMP is not
an issue with regard to rat model application.
Akesson et al. (2004) did measure dermal absorption in
humans exposed to neat or 50% NMP and separate permeation
constants were fit to each concentration. Although a wider
range of concentrations wasn't tested, the decision on how
best to use these results is described in response to a similar
comment, below.
SACC
SACC COMMENTS:
There is a disparity between predicted dermal vapor
uptake and the Bader et al. (2008) data. The impact of
this disparity is unknown, meaning that it is uncertain
whether the PBPK model over- or under-predicts the
As described in Section 3.2.5.5 of the risk evaluation, "The
discrepancy between the Bader et al. (2008) data and the
current model predictions could be because the subjects in
Bader and van Thriel (2006), on which this model is based,
wore long-sleeved shirts, thereby reducing dermal absorption
Page 118 of 205
-------
dose metrics.
or due to the use of an idealized model of inhalation uptake
which could over-predict uptake by that route." In the absence
of scenario-specific information on clothing, EPA assumes
that the surface area of skin exposed to vapor through skin is
25%, corresponding to the face, neck, arms and hands. While
there is some uncertainty around the impact of that
assumption, EPA makes scenario-specific adjustments to the
surface area of skin exposed to vapor where EPA has
information to indicate that workers are more fully protected
by PPE.
SACC
SACC COMMENTS:
Recommendation: Modify the discussion in Appendix I
on the PBPK models to address the issues identified by
the Committee with the rat and human PBPK models
listed below:
a) Based on the figures in the Poet et al. (2016)
publication, the rat model underpredicts the rate of the
metabolite into the plasma for the higher dose and
overpredicts it for the lower dose. This could have an
impact on predicted dose metrics for NMP (i.e., if the
under- and overprediction of metabolite data is due to
metabolism parameters).
b) The rat model calculates changes in tissue volume for
maternal fat, liver, uterus, and mammary tissue, but
only adjusts blood flow to these tissues, based on the
volume increases, for fat, uterus, and mammary tissue.
There is no explanation of why the liver blood flow is
not adjusted. This may have been an oversight and
could affect dose metric predictions.
c) In the rat model, urinary excretion is subtracted from
the arterial blood compartment equation but the
equation for urinary excretion uses a volume for
venous blood. In the submodel for the metabolite, this
Appendix J (previously Appendix I) has been updated to
provide summary information about model structure and
parameters. In response to specific points raised by the SACC:
a) Dr. Poet was not in complete agreement with changes that
the U.S. EPA made to her model and she implemented
some additional changes and modified parameter values
prior to the 2016 paper. While Dr. Paul Schlosser agreed to
be a co-author on that paper, the U.S. EPA modeling does
not use the parameters and fits shown by Poet et al. (2016),
but those shown in the PBPK appendix for the assessment.
In particular, EPA chose to focus the human model
calibration on the low-concentration data, as these are
considered most relevant to low-concentration exposures.
As shown by the simulations vs. the human data, these
parameter estimates are likely to over-predict human
dosimetry and risk at higher exposures but should not
substantially under-predict dose/risk.
b) This is an oversight in the documentation. Liver blood
flow is also increased in proportion to liver volume.
c) We appreciate the SACC noting the error in the urinary
excretion equation, but since the excretion constant (KLN)
is fitted, it can be corrected without effecting the
quantitative predictions. Defining KLNnew =
Page 119 of 205
-------
is reversed with urinary excretion being subtracted
from the venous blood compartment (using a total
blood volume as there is no arterial blood compartment
in the submodel) but using the arterial blood volume in
the equation for urinary excretion.
d) The rat model code in the Poet et al. (2016) publication
had the model variable named DNN defined as a state
variable {i.e., the integration of a rate variable and then
again with an arithmetic equation in a DISCRETE
block). Appendix I describes some model changes
made with regards to liquid dermal exposure and the
parameter, DNN. It was not possible to verify that
these changes were implemented correctly.; In the
INITIAL section of the model the equation for DDNX
is defined as ConcL*VLiq. In the DISCRETE block
called REAP, it is defined as ConcL/VLiq*FAD.
Given that this parameter is used in conjunction with
DNN, this may have been corrected but the Committee
cannot tell without EPA's code listing.
e) The M script file included in the supplemental material
for the 2016 publication indicated the parameter
NumFet is set to 0.01. This parameter is described as
the number of fetuses in the litter and setting NumFet
to this value has an impact on several other equations
such as maternal body weight. There is no explanation
of why this is done or if it is changed for all runs
conducted for this draft risk evaluation.
f) The constant for VSkC, described as fractional skin
volume, is set the same in the human model as it is set
in the rat model (as found in the M scripts). While the
Committee did not search the literature for fractional
skin volumes for rats and humans, the value being the
same for both seems unlikely.
KLN*VV/VA, then KLNnew*CA*VA = KLN*CA*VV.
Since venous blood volume is assumed to be 75% of the
total blood volume and arterial blood is 25%, setting
KLNnew = 3 *KLN and correcting the equation in the
model results in identical predictions. These changes have
been made.
d) The U.S. EPA version of the model does not include
variable DNN. EPA does apply the factor, FAD, to
calculate a corrected liquid concentration (CONCL2 =
CONCL*FAD), to account for < 100% absorption.
CONCL2 is then used as the initial concentration in the
mass balance for NMP in the liquid layer on top of the
skin.
e) Since the PK data used for rat model calibration were in
non-pregnant animals, NUMFET = 0.01 for those
simulations, to minimize the impact of the fetal
compartment. (Setting NUMFET = 0 would lead to divide-
by-zero errors.) For developmental bioassay simulations,
NUMFET =14. The EPA model workspace uses a
'ratparam.m' script to set default parameters for non-
pregnant rats, including NUMFET = 0.01, and
'pregratparam.m' to set pregnancy-specific parameters
for those simulations.
f) For the rat VSKC is set to 0.19 in the initial block of the
EPA model code and is not reset in the parameter scripts.
For the human it is set to 0.051 in the 'human_params.m'
script. These are the values as listed in Table 1 of Poet et
al. (2010). cited from Brown et al. (1997).
g) During the time when liquid is present on part of the
human skin, it is reasonable to assume that vapor does not
simultaneously contact that portion of the skin. It is
assumed that if vapor is present in the air and PPE is not
being worn, the exposed skin is otherwise exposed to
Page 120 of 205
-------
g) How the human PBPK model defines the skin
compartments for liquid and vapor exposure assumes
that the two do not overlap. It is questionable how
accurate a description this is for the actual exposures.
h) The model defines the variable named DDN as
(ConcL-1.0)*VLiq0*FAD. Appendix I states that this
was done to avoid potential division by zero; however,
it appears this could cause a larger issue by potentially
running the model with a negative amount in liquid if
ConcL is zero. There are better ways to avoid division
by zero.
i) The model defines QPlal in the INITIAL section in
terms of VPlal before VPlal is defined. The INITIAL
section is not sorted when the model is compiled so
when the model calculates QPlal, it does not yet know
the value for VPlal and, therefore, will use a value of
5.55e+33. The value of QPlal will be quite large at the
first time point and may affect AUC and Cmax
estimates. It was not mentioned in Appendix I in the
modifications to the model that this was corrected.
j) QSlw is calculated in the model code by subtracting
QPlal. The pregnancy code is reported to be based on
Gentry et al. (2002), but that model does not subtract
QPlal from QSlw.
k) PDose, PDose2, and PDose3 are labeled in the
INITIAL section as being in units of mg/kg, but when
ODosel, ODose2, and ODose3 are calculated by
multiplying these by body weight (in kg) and FracOr
(unitless), they are labeled as being in (iinoles. Given
the equations later in the model, ODosel, ODose2, and
ODose3 should be labeled as being in mg.
1) The SCHEDULE statement, IF (ON3) SCHEDULE
OND3.AT.S3, uses the parameter ON3 as a logical
vapor. PPE is assumed to occlude the skin from vapor
exposure when worn. It is possible that a portion of the
skin which is assumed to be only exposed to vapor is
actually exposed to liquid. This would be an error in the
exposure assessment and model input parameters; i.e., the
area exposed to vapor should be reduced and the area
exposed to liquid should be increased to reflect the actual
exposure. But otherwise EPA believes it is reasonable to
assume that a given cm2 of skin cannot be simultaneously
exposed to vapor and liquid. Since the absorption from
liquid far exceeds the absorption from vapor, the key
factor is that liquid exposure should not be under-
estimated. EPA believes that its estimation of liquid
exposure area is reasonable.
h) DDN: The multiplier FAD is used to force this term to
zero outside of periods with explicit liquid-dermal
exposure. With a density of 1.02xl0A6 mg/L, the
concentration in even a 10% solution is 5 orders of
magnitude greater than 1. For none of the liquid-dermal
exposure simulations conducted does the concentration
even approach 1. So, for practical purposes this is a non-
issue. While other means of avoiding divide-by-zero may
exist, EPA believes that the equation as it exists is
sufficient for the current model application.
i) In the U.S. EPA human model code, VPlal is defined on
line 217 and QPlal is defined on line 234.
j) The change in the calculation of QSlw from the Gentry et
al. (2002) model is described in the PBPK appendix
section on "Tissue and Blood-Flow Mass Balances." This
and other changes noted were made to provide balanced
flow volumes.
k) The U.S. EPA model code only uses PDOSE and
ODOSEl. The comments on the same line as the equations
Page 121 of 205
-------
variable. 0N3 has not been declared as a logical
variable and is defined with a value of 1.0. Logical
variables must have an integer value of 1 or 0 or, in the
CSL file with ".TRUE.", ".T.", ".FALSE." or ".F." The
Committee is unable to tell if this was corrected in the
models used, and this was not mentioned in the
changes in Appendix I.
m) The DISCRETE blocks named DOSE1, DOSE2, and
DOSE3 use parameters TIME, TIME2, TIME3, and
REPTM. While these parameters are set in the M script
file, they are not set or defined as a constant in the
model definition file. This could result in inaccurate
predictions if the values are not set prior to running the
model as the model does not have default values,
n) The model is coded to start the simulation after
pregnancy has started {i.e., increases in tissues volumes
calculated in INITIAL block as the non-pregnant
fractional value plus an increase based on the point in
the pregnancy at which the simulation is to start), and
the initial body weight is set at the default pre-
pregnancy weight. The body weight needs to be
updated after these initial tissue volumes are calculated
in the INITIAL section to account for these larger than
pre-pregnant tissue compartments. Then in the
DERIVATIVE section, the model currently adds the
increase from the initial tissue volumes {i.e., VFat-
VFatl) to the initial body weight at the beginning of the
simulation for fat, mammary tissue and uterus but not
for fetal and placental weight. This needs to be
corrected as the body weight (once modified as
necessary in the INITIAL section) already includes a
value for fetal and placental weight and only needs the
additional increase included.
correctly identify it as having units of mg. A stray
comment re. converting to |imole (presumed to be a legacy
comment) has now been deleted.
1) While ON3 has not been defined as a logical variable,
examination of model simulation outputs shows that the
scheduling statements have worked correctly, as the value
is either set to 0 or 1 in various scripts. This can be seen in
the plots of the Bader and van Thiel exposures (ON3 = 1)
vs. Akesson and Paulsson (ON3=0). The .csl file has now
been updated to declare it as a logical variable. But the
reviewer is mistaken that one must use .TRUE., etc., in
setting its value. Logical variables can also be set to 0 or 1
as the equivalent of FALSE and TRUE, respectively.
Computer logic is ultimately performed using binary
numbers, not symbols, and the use of .TRUE., etc., is
simply a convenience introduced in acslX.
m) The EPA version of the model does not use the DOSE1,
etc., discrete blocks,
n) Because the initial masses of the placenta and fetus (as
defined by the model equations) are 6.8e-5 and 4.9e-9 kg,
respectively, the error from not subtracting their initial
values in the overall mass balance is minimal. Therefore,
the model code has not been updated as suggested,
o) RADVL, which is the rate of vapor absorption on areas of
skin which are exposed to liquid during times when liquid
isn't present, is another component of the EPA model with
which Dr. Poet disagreed, hence does not appear in the
code for the Poet et al. (2016) publication. It is defined on
line 427 of EPA's file, HumPregRev2.clean.csl. The
equations for RADL (absorption when liquid is present)
and RADVL are as follows:
RADL = (PVL*SAL/1000.0)*(CSURF -
(CSKL/PSKL))*czone*BRUSH
Page 122 of 205
-------
o) Appendix I mentions an equation for RADVL which is
not in the code listing in the supplemental material for
the 2016 publication. RADVL is mentioned in some
comments in the model code but there is no equation.
Without EPA's code listing, the Committee is unable
to tell if this was added,
p) PregTime is set in the M script file to 0.0001. This
seems like an odd choice.
! Net rate of delivery to "L" skin from liquid, when liquid
is there
ADLL = integ(RADL, 0.0)
RADVL = (PV* SAL/1000.0)*(CI - (CSKL/PSKA))*(1.0-
Czone*BRUSH)
! Net rate of delivery to "L" skin from air, when liquid not
present
p) PregTime is not in the EPA model and the primary input
time constant is GDStart, which is the day of gestation on
which exposure simulation starts; it is either set to 0 or 240
days.
34
PUBLIC COMMENTS:
EPA should provide the blood:air partition coefficient;
this is the key parameter that an inhalation toxicologist
needs to understand respiratory tract absorption
The blood-air partition coefficient for NMP is 450. This is
now included in the PBPK Appendix (Appendix J) in the risk
evaluation.
54, 56
PUBLIC COMMENTS:
EPA should use dermal permeability constants in the
PBPK model that accurately reflect the variability of
skin thickness on the hand.
The comment is suggesting that the surface area of the skin be
divided into sub-regions, each with its own thickness. Areas
with greater thickness will have less absorption, areas with
lower thickness will have greater absorption. Since absorption
is proportional to the difference between concentration on the
surface and concentration in blood, the result will be the same
as using a single surface area with a weighted-average
thickness, where the weighting is by the area of skin with each
thickness. EPA is not aware of detailed information on the
variation in skin thickness across the hands, data that would be
required for such an elaboration of the model. If there are data
to indicate that the average thickness of skin on the palms and
under side fingers is different from that used in experiments
from which the dermal penetration constant was measured (for
example, that skin is on average one-half the thickness of skin
on the upper arm), then the model could easily be adjusted to
reflect this difference without sub-dividing the liquid-exposed
Page 123 of 205
-------
skin compartment.
57, 33
PUBLIC COMMENTS:
The parameter values used in the PBPK model are
overly conservative. EPA should consider using human
parameter values that are appropriate for the
concentration range of interest {i.e., near the POD
values), rather than those that were specifically
optimized for low concentrations of NMP.
As described in Appendix J, the mechanism for nonlinearity of
NMP concentrations in blood is unclear. EPA applied
parameter values derived from low exposure levels in order to
avoid creating a model that underestimates blood
concentrations for low level exposures relevant to conditions
of use considered in this risk evaluation.
57
PUBLIC COMMENTS:
Because future decisions made for NMP under TSCA
are expected to be driven in part by the risks associated
with high concentrations NMP in air, it is critical that
the human PBPK model be appropriately
parameterized for these exposure conditions.
EPA should either utilize all of the data from Bader
and van Thriel (2006) to estimate a single set of
parameters for describing NMP metabolism in humans
or estimate two sets of metabolism parameters: one for
low intensity exposures [<2.5 ppm, continuing to
utilize the 2.5 ppm data from Bader and van Thriel
(2006) alone], and another for high intensity exposures
[>2.5 ppm, utilizing the 10 and 20 ppm data from
Bader and van Thriel (2006)].
Considerable thought was put into why the human internal
concentrations were relatively reduced at higher vs. lower
concentrations from this study. Given the interval between
exposures, metabolic induction did not seem likely. The other
possibility was that the subjects were reducing their activity
level and hence respiration at the higher exposures, perhaps
because of the NMP odor. Without respiration data, such
concentration-dependent behavior cannot be verified, and
reduced respiration is not something that a person involved in
physical labor would be able to achieve. Hence, EPA does not
believe it appropriate to extrapolate this empirical
concentration-dependence to predictions for workplace or
residential user exposures; i.e., in the absence of knowing the
underlying mechanism. In the absence of such a correction, the
over-prediction of internal doses for some of the Bader and
van Thiel (2006) subjects is less than a factor of 3.
To follow the suggestion exactly as given would lead to a
discontinuity in the predicted internal dose between
predictions for individuals exposed to 2.4 ppm and 2.6 ppm,
for example, where the individual exposed to 2.6 ppm would
be predicted to have a lower internal dose than the person
exposed to 2.4 ppm. This is clearly unrealistic. A more
realistic approach would be to determine Vmax as a function
of exposure concentration (or blood or liver concentration)
such that there is a continuous, increasing relationship between
Page 124 of 205
-------
exposure and internal dose. But such a model would require
scientific review. Since the current model is health-protective
(does not under-predict internal dosimetry) and the suggested
revision is beyond the scope of the current assessment, EPA
will consider such a revision in the future.
PBPK modeling - Application of the PBPK model to estimate rat exposures in key studies
SACC,
57, 33,
54
SACC COMMENTS:
Recommendation: Justify the decision to model a 50 g
post-weanling rat instead of examining internal dose
for the period of newborn rat to mature mating rat.
One Committee member questioned why the model
was run for post-weanling rats at body weight of 50g
rather than for the period of new-born rat to mature
mating rat. Given that this simulation is to get an
internal dose metric for a 2-generation study, running
only this body weight is potentially losing any changes
in predicted internal dose due to the growth of the rat.
Also, running the model from newborn rat to mating
adult continuously better accounts for what the actual
internal dose metric would be at mating.
In Section 3.2.5.2 (NMP risk evaluation, p. 197, lines
4615-4617), it states that the internal dose metrics for
young post-weanling rats are the lowest. It would be
useful to see a table to demonstrate this.
PUBLIC COMMENTS:
EPA should use a duration-weighted average dose to
NMP over the total exposure period (pre-mating,
mating, gestation, lactation) as the basis for BMD
modeling and also revisit the assumption of early-life
susceptibility for the assessment of fertility effects, or
adopt a different endpoint that is more directly
EPA is not able to model internal blood concentrations of
NMP in newborn rats because the PBPK model does not
include lactation and cannot predict NMP concentrations in
milk. This means that it is not possible to calculate average
exposures over the entire exposure period. Instead, EPA
modeled internal exposures for post-weaning rats consuming
known levels of NMP through food. Of the exposures that
could be predicted in the PBPK model, the exposures
predicted during the post-weaning life stage are the lowest that
occurred throughout the exposure period. Because metabolism
is assumed to scale allometrically (a standard assumption that
is hard coded into the PBPK model), the metabolism/kg BW is
higher for smaller animals, leading to a lower internal dose.
Internal doses simulated for the Exxon 1991 study are
included in Section 5.2 of Risk Evaluation for N-
Methylpyrrolidone (NMP), Benchmark Dose Modeling
Supplemental File. Docket EPA-HQ-OPPT-2019-0236 (11 S
i). A reference to this documentation is now
included in the risk evaluation.
It is not possible to determine which phase(s) of exposure
were most sensitive to NMP and most contributed to the
reduced fertility in the study. It is also unknown how sensitive
phases in rat reproductive development translate to specific
phases of sensitivity in humans. EPA assumed that any phase
of the exposure could have been responsible for the effect and
therefore selected the lowest dose predicted in the PBPK
Page 125 of 205
-------
applicable to adult worker and consumer exposures.
EPA did not adequately consider rat size and
associated internal dose within the context of its
assumption of early-life susceptibility, which results in
a nearly 2-fold increase in the estimated NMP potency.
Because male rats in the Exxon (1991) study grew to
>700 g (i.e., well above the maximum weight of 450 g
included in the table), Table 4-1 should be expanded to
include doses for larger rats.
Inspection of the BMD modeling results in Table 4-4
of EPA's Benchmark Dose Modeling Supplemental
File demonstrate a consistent pattern between points of
departure for P2/F2A vs. P2/F2B litters for endpoints
in both male and female animals (Table 5). This pattern
is consistent with a duration effect, not an early life-
stage effect.
model over the course of the exposure period as the basis for
BMD modeling.
SACC
SACC COMMENTS:
Recommendation: Clarify how the PBPK model
accounts for exposure during lactation.
It is not clear how the PBPK model accounts for
exposure during lactation when there is no lactational
component in the PBPK model. Exposure during this
time period and its effect on internal doses are lost.
The PBPK model does not account for offspring exposure
through lactation. The model estimates maternal blood
concentrations but cannot predict milk concentrations or fetal
exposures. As stated above, this is one reason why doses are
based on 50 g post-weaning rats rather than rats exposed
through lactation.
SACC
SACC COMMENTS:
A figure showing rat dermal PBPK predicted versus
observed pharmacokinetic data should be included for
completeness.
This is shown in Figure 4 in Appendix J.
PBPK modeling - Application of the PBPK model to estimate human occupational exposures
SACC
SACC COMMENTS:
Recommendation: Justify the assumption of no
EPA assumes that dermal NMP exposure is constant. Rather
than assuming that a specific amount of NMP is applied just
Page 126 of 205
-------
decrease in coverage of NMP chemical on the surface
of the skin over time due to absorption or evaporation.
once in a shift and that a set fraction of that will be absorbed,
EPA assumes that over the course of a work shift, additional
exposures may continue to occur (replacing what might be lost
to absorption or evaporation). Evaporation from skin would
only be applicable in scenarios where gloves are not used. For
all COUs where gloves are worn, the dermal load of NMP that
comes in contact with skin under gloves is only reduced by
absorption (as evaporation is prevented by gloves). Glove
protection factors already adjust for the reduction in exposure
that may be provided by glove use in these scenarios. For all
of the COUs, there is insufficient data on the frequency of new
exposures that occur over the course of a shift. EPA's
assumption of continuous exposure could contribute to an
overestimate of risk for some COUs if dermal exposure occurs
less frequently, but would not underestimate risk.
SACC
SACC COMMENTS:
Clarify whether the data referenced in the
Semiconductor Industry comments at the Committee
meeting are useful in reducing uncertainties in PBPK
model inputs on exposure.
Clarify why the permeability coefficient (Kp) for
dermal absorption was refit to the in vivo data, 4.7xl0"3
versus 4.6xl0"3.
Clarify why the sex difference in rats observed in
plasma NMP levels over time, as seen in the data by
Ghantous (1995), is not reflected in the PBPK model
simulated trend presented in Figure_Apx_I4.
Regarding PBPK occupational inputs, EPA has included data
and assumptions provided in semiconductor industry
comments in many PBPK runs for occupational exposures.
While weight fraction data provided in semiconductor industry
comments reduce uncertainties, EPA cannot determine
whether uncertainties in PBPK model inputs on exposure are
reduced by using assumptions provided in semiconductor
industry comments because EPA has no data to determine
whether the proposed industry assumptions are more accurate
than the assumptions applied by EPA.
As described in the "Dermal Model & Data" section of
Appendix J, an error was found in the Poet et al. (2010) model
equations for dermal absorption (they did not account for
bidirectional diffusion). Therefore, the parameter was re-fit to
the dermal data after the equation was corrected.
To the extent that sex differences in physiological parameters
are known for rats, the rat parameters were set to those for
female rats, given the focus on developmental exposures.
Page 127 of 205
-------
Brown et al. (1997), a standard source for physiological
parameters, does not report sex-specific differences for any
tissues in rats other than those for gonads. Thus, the only
physiological difference between males and females in model
inputs is BW, which has a relatively small effect on predicted
dose, as shown. While it would be possible to fit sex-specific
metabolic rates to those data, given the focus on
developmental effects, the model fits focused on female rats
were considered adequate for dose-response assessment.
31
PUBLIC COMMENTS:
Tables 2-66 and 2-67, pp. 131-132: EPA should
explain how acute exposure peak blood concentrations
(in mg/L) and chronic exposure AUC values (in hour-
mg/L) were determined for fab workers and
maintenance.
EPA has clarified in the introduction to Section 2.4.1.3 that
Table 2-77 (had been Table 2-67 in the draft risk evaluation)
PBPK exposure results include acute exposures, which are
peak blood concentrations (Cmax in mg/L), and chronic
exposures, which are area under the curve (AUC in hr mg/L).
EPA has provided updates to Tables 2-76 and 2-77 in the risk
evaluation that include updated PBPK input parameter sets
and results, including for fab workers and maintenance in the
semiconductor industry. Input selection for the PBPK model
that estimates the acute and chronic exposures values are
shown in Section 2.4.1.2 for all OESs, and the semiconductor
industry is covered in 2.4.1.2.10.
52
PUBLIC COMMENTS:
While many aspects of the 2019 EPA NMP PBPK
model are adequately supported by primary and
secondary peer reviewed literature, the use of the
model to assess dermal liquid exposures lacked
reference to sufficient peer reviewed or scientific
consensus information.
While there are differences in the details, the model is in most
aspects that which has been described in the peer review
literature bv Poet et al. (2010), which includes predictions for
dermal exposures. Recognizing that while peer reviewed
scientific publications are an indicator of quality, but are not
entirely sufficient to determine a model's suitability for
regulatory application, the U.S. EPA looks to and expects that
its external review process, this SACC review in particular,
will provide a sufficient level of peer review for the entirety of
a risk analysis, including but not limited to any PBPK
modeling used. However, per EPA policy, consensus among
the peer review panel is not required.
Page 128 of 205
-------
52
PUBLIC COMMENTS:
EPA stated an intention in the draft risk evaluation to
use a higher dermal permeability constant for neat
NMP exposures, but appears to have used the lower
dilute NMP permeability coefficient irrespective of
weight fraction (EPA, 2019a).
Use of the higher permeability constant did not impact
the conclusion of this assessment. It is noted that the
increased permeability associated with neat NMP skin
contact is unlikely to occur in the semiconductor
industry because potential exposure events are
transient, and likely do not occur every shift.
EPA appreciates the thoroughness of the review and the fact
that this mistake was identified. In the expectation that
absorption (permeability, PVL) will be a continuous function
of the concentration of NMP in an aqueous solution - it would
not change suddenly to a large extent when a solution changes
by 1% weight fraction (WF) - EPA believes that setting
permeability to be a continuous function of WF fraction is the
most realistic approach. Since data are only available for two
weight fractions, 50% and 100%, EPA believes the best
resolution is to assume PVL is constant at the value estimated
from the 50% WF data for concentrations below 50% and that
the PVL increases linearly with concentration between 50%
and 100%), as indicated by the lines in the graph. Model
simulations and margin of exposure calculations were revised
based on this assumed relationship.
The 2nd comment is not entirely correct. There is no evidence
that the duration or frequency of contact would have an impact
on permeability. Hence a higher value of PVL will be used for
exposure to concentrations over 50%, as described above,
whether the exposure is transient and does not occur daily or is
expected to occur for longer periods on a daily basis.
52
PUBLIC COMMENTS:
The density of NMP is not defined as static variable in
the .m file of the 2019 EPA code. Thus, if the model is
executed in a new acslX workspace, it is necessary to
explicitly define density in the script file.
Since the density of NMP is set in the initial block of the .csl
file, and is a physical constant that doesn't change, it does not
need to be defined in a script. If a new acslX workspace was
created with this CONSTANT statement in the initial block of
the csl, the value will be correctly set. However, since acslX
software has been discontinued, EPA does not expect such an
event to occur. Future versions of the model would need to be
created in a different software.
52
PUBLIC COMMENTS:
A correction to the skin:blood partition coefficient had
not been documented in the supplemental materials of
the Poet et al. (2016) human model code.
As described above, the EPA PBPK model differs from the
PBPK model published by Poet et al (2016). See Appendix J
of the risk evaluation for skin:blood partition coefficient used
in the current model and a description of changes made to the
Page 129 of 205
-------
dermal model.
BMD modeling - Model fit
SACC,
57
SACC COMMENTS:
Recommendation: Reexamine the dose-response fit to
the combined Saillenfait et al. (2003, 2002) data on
resorptions in rat. Fit a dose-response to each dataset
separately and compare the final POD estimate to that
obtained for maternal toxicity for corroboration.
The placement (hence the calculation) of the
benchmark dose (BMD) and benchmark dose level
(BMDL) is not intuitive and appears far to the right of
an apparent threshold (see p. 23, EPA, 2019f). The
final model fit could be due to an artifact of combining
results from two studies. The highest modeled dose
appears to be driving this model. For some Committee
members, these findings raised questions of the utility
of the BMDL for this endpoint. Some Committee
members questioned whether this effect is real (due to
a threshold for maternal toxicity), an artifact from
issues associated with study design and/or quality, or
due to random variability (Type II error; false
positive). The Committee suggests that EPA consider
modeling the data for each study separately, after
reevaluating the studies for confidence in the results
(Saillenfait et al., 2002, 2003).
PUBLIC COMMENTS:
In Table 3-3-2 in Section 3.1 of the draft assessment
report (EPA, 2019a), EPA acknowledges that none of
the models had an acceptable p-value of >0.1.
EPA has revised BMD modeling for post-implantation loss
(resorptions and fetal mortality) and reduced fetal body weight
to ensure appropriate model fits. The revised BMD modeling
resulted in a new acute POD of 437 mg/L Cmax. Revised
models are summarized in Section 3.2.5.6 of the risk
evaluation and described in detail in the revised BMD
supplemental file. EPA revised the dose-response and risk
characterization portions of the risk evaluation to reflect
changes in the BMD analysis, the resulting PODs, and revised
risk estimates.
Specifically, for the acute POD, EPA considered the post-
implantation loss data from the Saillenfait et al. (2002) and
(2003) studies both independently and as a combined dataset.
EPA analyzed results from the Saillenfait et al. (2003)
inhalation study and, consistent with the study authors, found
no statistically significant effect at any of the doses tested.
Because there was not a statistically significant dose-response
observed in the Saillenfait et al. (2003) post-implantation loss
dataset, EPA did not consider using BMDs or BMDLs derived
from this dataset alone as an acute POD. The internal doses
achieved in the Saillenfait et al. (2003) inhalation studv were
all below the internal doses achieved in the Saillenfait et al.
(2002) oral studv. In order to provide a more robust dataset
that includes dose-response information at the lower end of the
dose response curve, EPA performed BMD modeling on the
combined dose-response data from the oral and inhalation
studies. EPA also performed BMD modeling on the Saillenfait
et al. (2002) oral studv alone. The combined model
incorporates a more complete set of dose-response data,
maximizing the power of the model and reducing the risk of
Page 130 of 205
-------
overinterpreting statistical noise at the lower end of the dose-
response curve. Prior to conducting BMD modeling on the
post-implantation data, EPA applied a Rao-Scott
transformation to adjust for litter effects in the absence of
access to individual animal data. Using this approach, good
BMD model fits were achieved for both the combined dataset
and for the 2002 oral exposure data alone. Ultimately the
BMDL identified in the combined dataset was identical to the
BMDL identified in the Saillenfait et al. (2002) studv alone
{i.e., a BMDL of 437 mg/L Cmax was identified for both
datasets). This is the POD EPA ultimately selected as the basis
for risk characterization for acute exposures.
The maternal effects reported in these studies do not negate
the observed embryonic and fetal toxicity. Without additional
information, it is difficult to determine if the maternal toxicity
contributes to or is independent of the developmental effects.
In addition, maternal toxicity in this study would be
considered indicative of repeated dose effects relevant for
derivation of the chronic POD rather than the acute POD. In
the Saillenfait 2002 study, significant effects on fetal body
weight (which are considered relevant for the chronic POD)
occur below doses at which significant effects on maternal
body weight were reported.
BMD modeling - BMR selection
SACC,
33, 57,
63, 54
SACC COMMENTS:
Recommendation: Consider revising the BMR for fetal
resorptions to a value >1% to avoid a BMDL that
extrapolates far outside the range of observed
responses.
The BMR of 1% results in a BMDL that is well below
the level where a statistical effect was observed. Other
BMRs tested (0.5 and 1 standard deviation [SD]) may
be more appropriate. EPA guidance indicates that a
EPA selected a benchmark response (BMR) of 1% for post-
implantation loss (resorptions and fetal mortality) to reflect
the severity of the endpoint. This approach is consistent with
the general principles for BMR selection that are outlined in
EPA's BMD modeling guidance. For quantal data, the
guidance describes a BMR of 10% extra risk as standard, but
notes that "biological considerations may warrant the use of a
BMR of 5% or lower for some types of effects (e.g., frank
effects)." Selection of a BMR of 1% for post-implantation
Page 131 of 205
-------
BMDLio is always provided for comparison but was
not observed for resorptions.
PUBLIC COMMENTS:
The decision to adopt a BMR of 1% relative deviation
(RD) in determining the POD (BMDL1RD) value for
fetal resorptions is not supported by the data for NMP
and is inconsistent with standard practice and EPA
guidelines.
Summit Toxicology conducted supplemental BMD
analyses for the acute toxicity of NMP. Based on these
analyses, it is recommended that EPA adopt a BMR of
0.5 SD in dose-response analysis for the acute toxicity
of NMP (acceptable model fit was achieved).
Even if one ignores the fact that the goodness of fit for
all EPA model runs of the combined data is
unacceptable per EPA BMD guidelines {i.e., <0.0001
when it should be >0.1), a conservative BMR of 0.5
SD more than doubles the BMDL {i.e., from 216 to
514).
loss is also consistent with IRIS assessments for TCE and
1,2,3-trichloropropane which applied 1% BMRs for prenatal
loss, cardiac malformations, and reduced numbers of live
pups/litter.
EPA selected the BMR based on the severity of the endpoint
(pregnancy loss) and not based on the statistical power of key
studies. As EPAs BMR guidance states, "It is important to
recognize that the BMR need not correspond to a response
that the study could detect as statistically significantly
different from the control response provided that the response
is considered biologically significant." (page 20 of EPA's
Benchmark Dose Technical Guidance).
The available NMP studies do not have sufficient power to
detect a 1% difference in post-implantation loss. EPA
concluded that the lack of power in the key studies introduces
uncertainty into the analysis but does not require application
of a less cautious BMR.
BMD modeling - Software versions
SACC,
57
SACC COMMENTS:
At least two different versions of EPA's Benchmark
Dose Software (BMDS) appear to have been used
(versions 2.7 and 3.1.1)
The Committee found it disconcerting that some nested
dichotomous models examined are not included in
BMDS 3.1.1, and EPA has provided no justification for
this exclusion.
The Committee wondered if analysis results would be
the same performed using BMDS 2.7 as obtained using
BMDS 3.1.1 and the EPA-preferred NLogistic nested
dichotomous model.
As detailed in the updated BMD Modeling Supplemental File,
EPA used BMD software (BMDS) package versions
(released 07/31/2019), 3.1.2 (released 11/08/2019), or 3.2
(released 08/20/2020) to model post-implantation loss,
resorptions, fetal and pup body weight changes, male fertility
and femal fecundity, and absolute testes weight datasets. For
these endpoints, choice of BMDS was dictated by software
availability at the time of BMD modeling. As each BMDS
release provides updates, fixes, and enhancements to BMDS
version 3, EPA chose to use the most up-to-date BMDS
version available when BMD modeling was originally
conducted. Because there weren't major updates {i.e., updates
Page 132 of 205
-------
The Committee questions why different versions of
BMDS were used for the acute BMDs than for the
chronic BMDs. The text states why two different
versions were used for the chronic BMDs but not why
a third version was used for the acute BMDs.
PUBLIC COMMENTS:
It is recommended that EPA rely upon the latest
version of BMDS in their TSCA risk assessment for
NMP.
In addition, since variance does not appear to vary in a
systematic manner as a function of dose, an assumption
of constant variance should be considered.
that would significantly alter BMD results) made to the
various releases of BMDS version 3 used in the NMP risk
evaluation, EPA chose not to rerun all BMD modeling with
the most up-to-date software {i.e., BMDS 3.2). A complete
history of updates made to various BMDS 3 releases is
available here.
The pup death and stillbirth endpoints were modeled using
BMDS version 2 J (released 08/18/2017) because it contains a
larger suite of nested dichotomous models compared to
BMDS version 3, and nested dichotomous models are
preferred for these endpoints because they contain an intra-
litter correlation coefficient for the assessment of litter-
specific responses. Both BMDS 2.7 and BMDS version 3
contain the same NLogististic model, which is preferred
because it has received more extensive QA testing and is
deemed to be the most reliable nested model. In contrast, the
NCTR and RaiVR nested dichotomous models are included in
BMDS 2.7, but not BMDS version 3. Because BMDS 2.7 and
BMDS version 3 contain the same version of the preferred
NLogistic model, and because BMDS 2.7 contains the
additional nested dichotomous models, BMDS 2.7 is preferred
over BMDS version 3 to model pup death and stillbirth
endpoints.
For BMD modeling, EPA first assumed responses to be
normally distributed with constant variance across dose
groups. If no model achieved adequate fit to response means
and response variances under those assumptions, models that
assume normal distribution with modelled {i.e., non-constant)
variance were applied.
Dose-response analysis - Integration of data across exposure routes
SACC
SACC COMMENTS:
EPA did not evaluate risks from NMP inhalation exposure
independently. For NMP, EPA used the PBPK models to
evaluate risks based on internal blood concentrations resulting
Page 133 of 205
-------
Recommendation: Justify why oral data are used to
estimate inhalation risk when adequate inhalation data
are available.
The Committee questioned the use of oral data to
estimate inhalation risk when good inhalation data are
available. Added to this is the statement that reports
that excreted amounts of NMP metabolites are higher
after inhalation dosing than after oral dosing (NMP risk
evaluation, Section 3.2.5.5, p. 199, lines 4727-4729).
These differences in excretion also support the
Committee's conclusion that combining the Saillenfait
oral and inhalation study data might not be appropriate.
The same reason given in Section 3.2.5.6 (NMP risk
evaluation, p. 205, lines 4916-4921) for why these
studies were not combined for decreased fetal body
weight might also apply as a reason for not combining
these study data for the "acute" endpoints.
from combined exposures across all exposure routes. Rather
than identifying separate PODs for each route of exposure,
EPA derived a single set of acute and chronic PODs in terms
of internal blood concentrations. These PODs are designed to
evaluate risks from the internal NMP concentrations resulting
from simultaneous inhalation, dermal, and vapor through skin
exposure routes for each condition of use. Differences in
absorption associated with each route are accounted for in the
PBPK model. The PBPK model also takes into account
various elimination routes, including exhalation, metabolism,
urinary elimination, and desorption from open skin,
irrespective of the route of exposure. For the acute endpoint,
inclusion of data from both inhalation and oral studies
allowed EPA to make use of a more complete dose-response
dataset that covers a wider range of internal doses. The
inhalation study alone is either too low in dose-range or too
limited in statistical power to identify significant effects on
the acute endpoint, post-implantation loss. EPA included the
2003 data in the combined dataset to ensure that BMD
modeling included data from lower dose range, but the 2003
data on its own could not serve as the basis for a POD.
Dose-response analysis - Assumptions and uncertainties
54
PUBLIC COMMENTS:
Overall, EPA should provide a more robust discussion
of the assumptions and uncertainties inherent in the
dose-response assessment in the revised risk
evaluation. In addition, EPA should consider whether
alternative PODs and/or BMRs are warranted. A
statement regarding the uncertainty in the appropriate
dose metric and outcome response rate for resorptions
should be added to the table.
In the final risk evaluation, EPA considers a range of PODs in
Section 3.2.5.6 and provides a rationale for selection of each
POD over the other PODs considered. In Section 3.2.6, EPA
discusses uncertainties around dose metrics and endpoint
selection as part of the overall confidence in the final PODs.
A more detailed discussion of the rationale for specific BMD
approaches, BMRs and model selection for each endpoint is
included in the BMD supplemental file.
Page 134 of 205
-------
Dose-response analysis - Other suggestions
SACC
SACC COMMENTS:
SACC suggests the following:
o Add the numbers for increased fetal mortality to
Figures 3-2 or 3-3.
o The phrase "small difference" (NMP risk
evaluation, Section 3.2.5.1, p. 195, line 4550)
should be clarified; "small" is a relative term.
In Section 3.2.5.5 (NMP risk evaluation, p. 199, lines
4707-4710), it states that the PBPK model was run for
humans to estimate internal doses, which were then
compared to the rat internal doses. This sounds
backwards from what is normally done. It should be
explained.
The word "similar" (NMP risk evaluation, Section
3.2.5.6, p. 201, line 4828) could use more information
(e.g., show the values that are similar in parentheses
after the statement).
More details should be given in the main text (NMP
risk evaluation, Tables 3-10 and 3-11) as to how the
PBPK model runs were conducted for the rat BMD
modeling (i.e., run as pregnant rat, non-pregnant, etc.).
In Table 3-10 (NMP risk evaluation, Section 3.2.5.6, p.
202), the lines for the Becci study and the Sitarek study
(line for NOAEL) do not state whether the internal
dose is Cmax or AUC.
Section 3.2.5.6 (NMP risk evaluation, p. 203, lines
4861-4865) states that a UF was included to account
for human variability. A Monte Carlo analysis could be
used (like with the MC assessment) to account for
human toxicokinetic variability.
Section 3.2.5.6 (NMP risk evaluation, p. 205, lines
4903-4914) discusses how data may possibly be
Thank you for the suggestions.
The exposure-response arrays (now figures 3-2, 3-3, 3-4
and 3-5) have been expanded to include additional
reproductive and developmental endpoints for easier
comparison across endpoints.
As described above, EPA added a sentence with additional
quantitative information from the van Raaij study to
support the phrase "small difference."
The first paragraph in Section 3.2.5.5 was revised for
clarity.
In Section 3.2.5.6, the word 'similar' is now followed by
the specific values being compared in parentheses.
More details have been added to the narrative and to table
footnotes to clarify the specific dose metrics used as the
basis for BMD modeling for each endpoint.
Table 3-10 has been modified to specify that NOAELs
from the Becci and Sitarek studies are in terms of Cmax.
EPA did not perform Monte Carlo analysis because there
is insufficient quantitative information on the range of
human variability to support such an analysis.
Whole body inhalation studies were considered as part of
the weight of the scientific evidence and some endpoints
were further evaluated in dose-response analysis. While
there is uncertainty around specific levels of exposure
achieved in these studies, the studies demonstrate that
developmental effects are consistently observed across
exposure routes. Data from these inhalation studies
provide supporting evidence, but they did not provide the
quantitative basis for PODs. The modeling for the acute
POD incorporated data from the Saillenfait 2003 study but
Page 135 of 205
-------
compromised due to condensation of NMP in whole
body inhalation experiments and subsequent oral
consumption by the exposed rats. Should this then
eliminate whole body inhalation studies from
consideration or is there sufficient information to show
that it was not an issue for any inhalation studies used
in the risk evaluation? The rationale for inclusion
should be presented.
is really driven by the oral exposure data (demonstrated by
the fact that the BMDL for the combined analysis is the
same as the BMDL for the Saillenfait 2002 study alone).
The chronic POD is based on effects on fertility in a two-
generation dietary study.
Page 136 of 205
-------
7. Risk Characterization
Risk Characterization
Charge Question 6.1: Please comment on whether the information presented to the panel supports the conclusions outlined in the draft
risk characterization section concerning NMP. If not, please suggest alternative approaches or information that could be used to further
develop a risk estimates within the context of the requirements stated in EPA's Final Rule, Procedures for Chemical Risk Evaluation
Under the Amended Toxic Substances Control Act (82 FR 33726).
Charge Question 6.2: Please comment on the validity of specific confidence summaries presented in Sections 4.2 and 4.3.
Charge Question 6.3: Please comment on any other aspect of the human health risk characterization that has not been mentioned above.
Charge Question 6.4: Please comment on the assumptions, strengths and weaknesses of the MOE approaches used to estimate the acute
and chronic risks associated with occupational and consumer use of NMP-containing products, including the MOEs calculated with
PBPK-derived internal doses. Please comment on the selection of composite uncertainty factors that were used to derive benchmark
MOEs risk estimation.
Charge Question 6.5: Please comment on this approach to evaluating the relative contribution of each exposure route to aggregate risk.
Charge Question 6.6: Please comment on whether the risk evaluation has adequately addressed potentially exposed or susceptible
subpopulations.
Charge Question 6.7: Please comment on whether the risk evaluation document has adequately described the uncertainties and data
limitations associated with the methodologies used to assess the human health risks. Please comment on whether this information is
presented in a clear and transparent manner.
Charge Question 6.8: Please comment on whether EPA has adequately, clearly, and appropriately presented the reasoning, approach,
#
Summary ol*Comments lor Specific Issues Related lo
Charge Question 6
KPA Response
Ecological Risk
SACC
SACC COMMENTS:
Recommendation: Provide more detail on the selection of
environmental receptors and better justify not considering
risks to terrestrial organisms via surface waters.
The Committee agreed that the risk characterization section
does not support clearly the conclusion of no unreasonable
risk for aquatic organisms, and recommends further
justification of the estimates used, adding UFs where
During problem formulation, EPA performed a first-tier
screening analysis of risks from ambient air, ambient water,
sediment, and land-applied biosolids. EPA did not identify
risks from environmental exposures that may result from
these pathways. In the final risk evaluation, EPA updated
the evaluation of risks to aquatic life from ambient water
exposures using more recent TRI release data. As described
in Section 4.1, EPA did not identify risks to environmental
receptors or the general population from ambient water.
Page 137 of 205
-------
needed, and discussing further those scenarios where risk
quotients are close to 1.
Occupational Risk - ON Us
SACC
SACC COMMENTS:
Lack of exposure data for ONUs is a major issue.
Recommendation: Explore and expand the use of modeling
approaches to exposure estimation when exposure
measurement data are insufficient or lacking (e.g., workers
and ONUs for NMP).
EPA has added to 2.4.1.1 that EPA has modeled inhalation
air concentrations for workers in 11 of 17 OESs and far-
field inhalation air concentrations for ONUs in 1 of 17
OESs. EPA has exhausted all modeling opportunities with
the data that are reasonably available and therefore was
unable to model inhalation air concentrations for workers in
6 OESs and far-field inhalation air concentrations for ONUs
in the 16 OESs. However, air monitoring data for workers
was available and used for workers and ONUs when
modeling was not possible.
Occupational risk - Self-employed and small businesses
55
PUBLIC COMMENTS:
A risk determination for workers and ONUs, both self-
employed and in small businesses, that incorporates OSHA's
exemptions and practical exceptions should be added to the
risk evaluation.
EPA does not have reasonably available data or information
to distinguish how OSHA's exemptions and practical
exceptions would cause exposure differences for workers
and ONUs, both self-employed and in small businesses.
Benchmark MOEs - Uncertainty factors
SACC,
33, 57
SACC COMMENTS:
Recommendation: Consider deriving data-driven
extrapolation factors (DDEFs) as an alternative to default
UFs.
PUBLIC COMMENTS:
Based on PBPK modeling of individual data from human
volunteers exposed via inhalation (reflects uptake via
inhalation and dermal absorption of vapor) to three
concentrations of NMP vapor (Bader and van Thriel, 2006),
peak NMP levels in blood (mg/L) for each exposure
concentration were determined to have a coefficient of
variation (CV) of approximately 0.21, while AUC blood
EPA expects that DDEFs would result in very similar total
UFs as the default UFs that are already applied for NMP. In
addition, the analysis presented in public comments is based
on a small sample of healthy workers (n=7) described in
Bader and van Thriel (2006). EPA concluded that this small
dataset is not sufficient as the basis for defining the
potential range of toxicokinetic variability across the
population. In the absence of data that adequately captures
the true variability in the population, EPA applied default
uncertainty factors for interindividual variability. Therefore,
rather than identifying benchmark MOEs of 20 and 21 as
suggested by commenters, EPA identified a benchmark
MOE of 30 for acute and chronic exposures.
Page 138 of 205
-------
levels (mg/L x h) for each exposure concentration had a CV
of approximately 0.28 (Poet et al., 2016).
Using the CV for peak blood levels and assuming a normal
distribution in a healthy worker population, a tk DDEF of
2.0 (1.21 x 1.645, rounded to two significant figures) is
judged sufficient to protect 95% of a healthy worker
population and yields a net UFH of 6.3 [2.0 (tk) x 3.16 (td)].
Using the CV for AUC data similarly, a DDEF of 2.1 (1.28 x
1.645) and a net UFH of 6.6 [2.0 (tk) x 3.16 (td)] can be
calculated. When these net UFH values are combined with
the UFA value of 3.16, the composite UFs for acute (peak)
and chronic (AUC) exposures to workers are 20 (6.3 x 3.16)
and 21 (6.6 x 3.16), respectively.
For occupational scenarios, the default MOE of 30 should be
replaced by data-driven MOEs of 20 (acute exposures) and
21 (chronic exposures)
SACC,
34, 38,
51,59,
61
SACC COMMENTS:
Recommendation: Consider utilizing UFs for database
completeness and quality in TSCA chemical evaluations.
PUBLIC COMMENTS:
EPA's benchmark MOEs for acute and chronic effects
should include a further UF of 10 for database uncertainty
(data gaps include immune system, neurotoxicity,
reproductive, endocrine, and/or developmental endpoints).
EPA should be using its information gathering authorities to
fill the data gaps or, at a minimum, should apply an
additional 10X UF to account for these database
deficiencies.
Fetal death is a severe endpoint, not a sensitive one, and
EPA must acknowledge with appropriate UFs that many
adverse effects will occur at lower doses.
There is no universal list of hazard data required when
evaluating chemical risks under TSCA. Furthermore, for
NMP, EPA has sufficient, reasonably available hazard
information to conduct a risk evaluation and support the use
of the chosen hazard endpoints. Therefore, EPA did not use
a database uncertainty factor in the NMP risk evaluation.
EPA acknowledges the severity of the post-implantation
loss endpoint and applied a BMR of 1% in benchmark dose
modeling.
Page 139 of 205
-------
34,51,
38
PUBLIC COMMENTS:
EPA should provide justification for its deviations from its
standard inter- and intra-species UFs, or utilize the default
value.
It is recommended that EPA apply a UF of at least 20X for
intraspecies variability to account for the known
susceptibility of some subpopulations toNMP's
developmental and reproductive effects.
To provide protection for developmental effects that occur at
doses below those causing death, a UF beyond the default
intraspecies 10X factor should be applied, as EPA has
previously done for other susceptible groups such as infants
and children.
EPA typically considers the interspecies uncertainty factor
of 10 to be comprised of a factor of 3 to account for
toxicokinetic differences and a factor of 3 to account for
toxicodynamic differences. As described in Section 3.2.5.6
of the final risk evaluation, toxicokinetic differences
between rats and humans are already accounted for in
PBPK modeling. EPA therefore applied a total interspecies
uncertainty factor of 3 to account for toxicodynamic
differences for both acute and chronic PODs. Consistent
with standard practice, EPA also applied the default
intraspecies uncertainty factor of 10 to both acute and
chronic PODs to account for sensitive subpopulations.
38, 51
PUBLIC COMMENTS:
EPA should apply a UF of 10 to reflect the absence of a
NOAEL for NMP's reproductive effects. The full UF of 10
for LOAEL-to-NOAEL extrapolation should be applied
since there is no basis for reducing it.
EPA applies uncertainty factors for LOAEL to NOAEL
extrapolation when a LOAEL serves as the point of
departure. To derive the chronic POD for NMP, EPA used
BMD modeling to identify a BMDL as the point of
departure. EPA does not apply LOAEL to NOAEL
uncertainty factors when PODs are derived from BMD
modeling, because BMD modeling addresses many of the
limitations associated with the NOAEL/LOAEL approach.
MOE determinations based on updated PBPK modeling (Cardno ChemRisk)
52
PUBLIC COMMENTS:
Based on Cardno ChemRisk analysis, there is not an
unreasonable risk in the unlikely event that neat NMP were
to contact the palm side of two gloved hands (plausible acute
worst case).
EPA conclusions regarding unreasonable risk for the use of
NMP in semiconductor manufacturing reflected the lack of
refinement and use of incorrect assumptions in the screening.
The Cardno ChemRisk analysis indicates a differentiation on
exposure potential between jobs, with some functions having
EPA does not have reasonably available information and
data to verify parameter assumptions in the Cardno
ChemRisk analysis.
EPA also did not identify reasonably available information
or data indicating that any of the assumptions EPA used in
the analysis are incorrect. For the Semiconductor
manufacturing OES (Section 2.4.1.2.10), EPA revised and
expanded PBPK runs for industry-specific work activities
using industry-specific sets provided in public comments
Page 140 of 205
-------
no opportunity for dermal direct contact. The resultant acute
and chronic MOEs were >30, indicating support for a
conclusion that use of NMP in the semiconductor industry
does not present an unreasonable risk.
for the semiconductor manufacturing industry.
As demonstrated by various PBPK parameter sets, the EPA
analysis indicates a differentiation on exposure potential
between jobs as does the Cardno ChemRisk analysis. EPA's
analysis uses ONUs to indicate functions having no
opportunity for dermal direct contact.
PPE assumptions - General
SACC
SACC COMMENTS:
Recommendation: Provide a more balanced description of
exposures and MOE estimates with and without PPE in the
text of Section 4 to reflect the estimates provided in the
tables.
Recommendation: Provide a balanced description of risk
with and without PPE use in the text of Section 4.
The risk characterization section (Section 4) does not
provide a clear summary of the rationale, approach,
assumptions, and uncertainties about PPE, which are more
clearly described in other sections of the draft risk
evaluation.
The text referring to the values in the tables of Section 4
occasionally neglects to discuss MOEs when PPE use is not
assumed, resulting in an unbalanced discussion of the
information presented in the tables.
In Section 4, risks for occupational exposures are calculated
and presented both with and without PPE to provide risk
managers and the public with complete information about
how PPE use could influence risk. In Section 4.2.2 of the
risk evaluation, EPA presents risk estimates for
occupational exposures both with and without glove use
(glove PFs 1, 5, 10, and 20) for each occupational exposure
scenario. Table 4-54 presents risk estimates with and
without gloves and respirators for all conditions of use. The
narrative in Section 4.6.2.1 describes risk with and without
glove use.
61
PUBLIC COMMENTS:
EPA's assumption of PPE use violates TSCA's requirement
to "use scientific ... methods, protocols, [and]
methodologies ... in a manner consistent with the best
available science." The best available science for
occupational risk assessment requires the measurement of
worker exposures and risks without PPE.
These non-PPE measurements permit OSHA and other
regulatory agencies to determine whether risks can be
For the purpose of this risk evaluation, EPA makes
assumptions about potential PPE use based on reasonably
available information and expert judgment. EPA considers
each condition of use and constructs exposure scenarios
with and without engineering controls and /or PPE that may
be applicable to particular worker tasks on a case-specific
basis for a given chemical. While EPA has evaluated
worker risk with and without PPE, as a matter of policy,
EPA will not assume that workers are unprotected by PPE
Page 141 of 205
-------
eliminated through the use of engineering controls and
hazard elimination before the consideration of PPE,
consistent with the well-established occupational hierarchy
of controls.
where such PPE might be necessary to meet federal
regulations, unless the Agency has evidence that workers
are unprotected. For the purposes of determining whether or
not a condition of use presents unreasonable risks, EPA
incorporates assumptions regarding PPE use based on
information and judgement underlying the exposure
scenarios. These assumptions are described in the
unreasonable risk determination for each condition of use,
in Section 5.2. Additionally, in consideration of the
uncertainties and variabilities in PPE usage, including the
duration of PPE usage, EPA uses the high-end exposure
value when making its unreasonable risk determinations in
order to address those uncertainties. EPA has also outlined
its PPE assumptions in Section 5.1 and EPA's assumptions
are described in the unreasonable risk determination for
each condition of use, in Section 5.2.
The OSHA regulations at 29 CFR 1910.132 require
employers to assess a workplace to determine if hazards are
present or likely to be present which necessitate the use of
PPE. If the employer determines hazards are present or
likely to be present, the employer must select the types of
PPE that will protect against the identified hazards, require
employees to use that PPE, communicate the selection
decisions to each affected employee, and select PPE that
properly fits each affected employee.
SACC
SACC COMMENTS:
Recommendation: Describe assumptions and uncertainties
regarding PPE use, including how they differentially affect
conclusions about risks for multiple occupational exposure
scenarios.
There should be cross-referencing to Section 2.4.1.1 (Glove
Use, NMP risk evaluation, pp. 68-69 and Table 2-3, p. 70),
In Section 2, EPA has removed assignments/ assumptions
of specific glove PFs to apply to each OES. Table 2-77 has
been updated to include worker exposures for all glove PFs
for all OESs. To the extent that information is available on
PPE use for specific occupational exposure scenarios, it is
described in Section 2. In Risk Characterization (Section 4)
EPA presents risks for all occupational exposure scenarios
Page 142 of 205
-------
where PPE is discussed in detail. Tables 4-5 through 4-36
should indicate the category of glove use, for example, by
adding footnotes linking to Table 2-3. (SACC provided a
table of scenarios that are sensitive to glove use assumptions
in their report)
The Evaluation should explicitly recognize that the assumed
glove use and the GPF assignment have a major effect on
MOEs for risk characterization in multiple occupational use
scenarios, and describe the extent to which these
assumptions and uncertainties about PPE use influence
conclusions about risk.
both with and without glove use and with and without
respirator use. In Section 4.2.2, in the discussion of
strengths and limitations of risk estimates calculated for
each OES, EPA describes glove use assumptions as a
source of uncertainty. Table 4-54 demonstrates the extent to
which gloves and respirators influence risk estimates for
each condition of use.
31,33,
56, 57
PUBLIC COMMENTS:
EPA assumptions of no PPE or ineffective PPE for the high-
end exposure scenario should be justified.
The statement indicating that "high-end occupational
exposure estimates for NMP use in cleaning, metal finishing,
electronic parts manufacturing, automotive servicing, and
use in (or removal of) paints, coatings, adhesives and
sealants show risks that are not mitigated via glove use" is
inaccurate. The SI A data submitted to EPA indicate that
semiconductor risks ARE mitigated via glove use.
EPA's assumptions on worker exposure controls for "Use in
Electrical Equipment, Appliance and Component
Manufacturing" are not aligned with ongoing workplace
practices and need to be re-evaluated.
In Section 2, EPA has removed assignments/ assumptions
of specific glove PFs to apply to each OES. Table 2-77 has
been updated to include worker exposures for all glove PFs
for all OESs. EPA revised and expanded PBPK runs for
industry-specific work activities using industry-specific
data and information provided in public comments for the
semiconductor manufacturing industry. To the extent that
information is reasonably available on PPE use for specific
occupational exposure scenarios, it is described in Section
2.
In Section 4, EPA presents risks for all occupational
exposure scenarios both with and without glove use and
with and without respirator use. Table 4-54 demonstrates
the extent to which gloves and respirators influence risk
estimates for each condition of use.
PPE assumptions - Glove protection factors
59, 57,
52, 49
PUBLIC COMMENTS:
EPA should utilize chemical-specific data for PFs of gloves
that are likely to be used as PPE in industrial settings with
NMP (Zellers and Sulewski, 1993; Stull et al., 2002; Crook
and Simpson, 2007). Chemical-specific GPFs (Kirman,
EPA appreciates the commenters' recommendations and
considered the information submitted concerning the use of
chemical specific gloves and protection factors (PF)
assigned to specific industries occupational exposure. In
Section 2.4.1.1 of the draft risk evaluation EPA discussed
the parameters and assumptions made about glove use and
Page 143 of 205
-------
2020; Attachment 1) were provided with this comment
(Docket 0057, comment 51).
The submitted report "Assessment of the Efficacy of
Different Glove Materials for Reducing Potential Hazards
Associated with NMP Containing Paint Strippers" was not
listed among the references used by EPA in this assessment.
Was this in error?
For lithium ion battery manufacturing and the semiconductor
industry, EPA should assign a higher GPF (e.g., PF 20) and
recalculate the MOEs.
The strict work rules in the semiconductor industry and
training programs support a GPF of 20 (95%).
Page 144 of 205
associated protection factors (PFs) based on information
including worker training and NMP chemical-resistant
gloves. EPA recommends the commenters also refer to
table 2-3 in the risk evaluation to review EPA's glove
protection factors for different dermal protection strategies.
For the purpose of this risk evaluation, EPA makes
assumptions about potential PPE use based on reasonably
available information and expert judgment. EPA considers
each condition of use and constructs exposure scenarios
with and without engineering controls and /or PPE that may
be applicable to particular worker tasks on a case-specific
basis for a given chemical. Again, while EPA has evaluated
worker risk with and without PPE, as a matter of policy,
EPA does not believe it should assume that workers are
unprotected by PPE where such PPE might be necessary to
meet federal regulations, unless it has evidence that workers
are unprotected. For the purposes of determining whether or
not a condition of use presents unreasonable risks, EPA
incorporates assumptions regarding PPE use based on
information and judgement underlying the exposure
scenarios. These assumptions are described in the
unreasonable risk determination for each condition of use,
in Section 5.2. Additionally, in consideration of the
uncertainties and variabilities in PPE usage, including the
duration of PPE usage, EPA uses the high-end exposure
value when making its unreasonable risk determinations in
order to address those uncertainties. EPA has also outlined
its PPE assumptions in Section 5.1 and EPA's assumptions
are described in the unreasonable risk determination for
each condition of use, in Section 5.2.
The chemical-specific glove PFs referred to in this
comment were generated using a limited approach (in
Kirman, 2020) using laboratory-generated data on
-------
permeation testing of NMP-containing liquids. Actual glove
PFs holistically take into consideration other factors beyond
permeation, including several listed in Section 2.4.1.1:
NMP-based product that penetrates the gloves, including
potential seepage through the cuff from improper donning
of the gloves and the gloves may occlude the evaporation of
NMP from the skin. EPA recognizes that the references
"Zellers and Sulewski, 1993; Stull et al., 2002; Crook and
Simpson, 2007" are covered by Kirman, 2020, and several
of these references are cited in Appendix F of the risk
evaluation.
The report "Assessment of the Efficacy of Different Glove
Materials for Reducing Potential Hazards Associated with
NMP Containing Paint Strippers" noted in docket reference
EPA-HQ-OPPT-2016-0231-0910, is apparently a pre-cursor
to Kirman, 2020. This report uses a similar limited
approach to Kirman, 2020, and data and PFs from these
sources are not used in the risk evaluation and are therefore
not listed among the references.
In Chapter 2, EPA has removed assignments/ assumptions
of specific glove PFs to apply to each OES. Table 2-77 has
been updated to include worker exposures for all glove PFs
for all OESs.
59, 52
PUBLIC COMMENTS:
For several conditions of use, EPA asserts that it used lower
PFs than it actually relies on in its risk determinations, and
there are contradictions within the text that are inconsistent
with information that it relied on in its risk estimates table
(Table 4-50) and risk determination table (Table 5-1). The
PF used in the risk estimates should be clarified and stated
consistently in text and tables.
EPA appreciates the commenters' suggestion and has
worked to add clarity to the use of glove protection factors
(PF) in the risk evaluation. Section 5.2 describes the PPE
assumptions for each condition of use and how those
assumptions contribute to the unreasonable risk
determination. In other sections of the risk evaluation, EPA
provides additional information on each occupational and
consumer exposure scenario, shows risk estimates with a
range of glove PF for workers, and describes the
assumption of no PPE for consumers. EPA has outlined its
Page 145 of 205
-------
PPE assumptions in Section 5.1 and has supplemented some
sources and information in Section 2.4.1.1. of the risk
evaluation and Appendix E Occupational Exposures. EPA
has also added tables in Section 4.2.2 to clarify the PPE
assumptions made for each COU. Additionally, in
consideration of the uncertainties and variabilities in PPE
usage, including the duration of PPE usage, EPA uses the
high-end exposure value when making its unreasonable risk
determinations in order to address those uncertainties.
In Section 4.2.2 of the risk evaluation, EPA presents risk
estimates for occupational exposures both with and without
glove use (glove PFs 1, 5, 10, and 20) for each occupational
exposure scenario. Table 4-54 (formerly table 4-50)
presents risk estimates with and without gloves and
respirators for all conditions of use.
61
PUBLIC COMMENTS:
Without data on which gloves are provided to which
employees, EPA has no basis for assuming specific GPFs in
its draft risk evaluation.
EPA appreciates the comments. Whenever EPA had data on
PPE and glove use for specific uses, we incorporated the
information into the risk evaluation. EPA has outlined its
PPE assumptions in Section 5.1 and has supplemented some
sources and information on respirator use in Section 2.4.1.1.
of the risk evaluation and Section 1.4.6 of the Supplemental
Information on Releases and Occupational Exposure
Assessment. EPA has also added a table in Section 4.2.2.1
to make the PPE assumptions made for each COU clearer.
Additionally, in consideration of the uncertainties and
variabilities in PPE usage, including the duration of PPE
usage, EPA uses the high-end exposure value when making
its unreasonable risk determinations in order to address
those uncertainties.
PPE assumptions - Glove use assumptions, OSHA guidance and SDS information
42, 55,
51,34,
PUBLIC COMMENTS:
EPA assumes that proper gloves will always be used for
NMP handling, and that workers are trained in proper glove
In Chapter 2, EPA no longer uses professional judgment to
predict the likelihood of the use of glove and has removed
Page 146 of 205
-------
use. No data or other evidence are presented to support this
assumption.
In the NMP draft risk evaluation, EPA states that it "used
professional judgment to predict the likelihood of the use of
gloves ... and the associated [protectiveness factors] are
presented as what-if scenarios."
o It is not clear what this judgment is based on, given that
EPA earlier states that it "does not know the likelihood
that workers wear gloves of the proper type and have
training on the proper usage of gloves," and that the
"assumed glove protection factor values are highly
uncertain."
o EPA provides no evidence that industry has conducted
testing to identify the best glove materials for each of the
many NMP-containing products and mixtures to which
workers are exposed.
For several work scenarios (construction trades, small
businesses, self-employed workers, etc.), widespread glove
use is unlikely.
The unfounded assumptions that persons will always use
gloves and that they will always effectively reduce risk have
no evidentiary support in the administrative record. EPA
cannot rely on OSHA regulations and use of gloves to
reduce risk. TSCA requires the assessment of risk to workers
in the absence of PPE, and if risks are identified, it can then
be considered whether the risks would or would not be
mitigated by PPE.
If glove use is non-existent or limited, risks to exposed
workers would be unreasonable according to EPA's risk
benchmarks and worker protections would be required under
TSCA section 6(a). In its final risk evaluation, EPA should
adhere to its own unreasonable risk criteria and not
Page 147 of 205
assignments/ assumptions of specific glove PFs to apply to
each OES. Table 2-77 has been updated to include worker
exposures for all glove PFs for all OESs.
To the extent that information is reasonably available on
PPE use for specific occupational exposure scenarios, it is
described in Section 2.
In Section 4, risks for occupational exposures are calculated
and presented both with and without PPE to provide risk
managers and the public with complete information about
how PPE use could influence risk. In Section 4.2.2 of the
risk evaluation, EPA presents risk estimates for
occupational exposures both with and without glove use
(glove PFs 1, 5, 10, and 20) for each occupational exposure
scenario. Table 4-54 presents risk estimates with and
without gloves and respirators for all conditions of use.
For the purpose of this risk evaluation, EPA makes
assumptions about potential PPE use based on reasonably
available information and expert judgment. EPA considers
each condition of use and constructs exposure scenarios
with and without engineering controls and /or PPE that may
be applicable to particular worker tasks on a case-specific
basis for a given chemical. Again, while EPA has evaluated
worker risk with and without PPE, as a matter of policy,
EPA does not believe it should assume that workers are
unprotected by PPE where such PPE might be necessary to
meet federal regulations, unless it has evidence that workers
are unprotected. For the purposes of determining whether or
not a condition of use presents unreasonable risks, EPA
incorporates assumptions regarding PPE use based on
information underlying the exposure scenarios. These
assumptions are described in the unreasonable risk
determination for each condition of use, in Section 5.2.
Additionally, in consideration of the uncertainties and
-------
recharacterize risks that meet these criteria as "reasonable"
based on "uncertainty."
OSHA cannot cite an employer for failing to follow
manufacturer recommendations in a Safety Data Sheet
(SDS). In the absence of such a requirement, there is no
basis for EPA's assumption that the Hazard Communication
Standard will result in the uniform use of appropriate PPE.
variabilities in PPE usage, including the duration of PPE
usage, EPA uses the high-end exposure value when making
its unreasonable risk determinations in order to address
those uncertainties. EPA has also outlined its PPE
assumptions in Section 5.1 and EPA's assumptions are
described in the unreasonable risk determination for each
condition of use, in Section 5.2.
33, 57
PUBLIC COMMENTS:
OSHA data support the position that workers use appropriate
gloves and other PPE. It is therefore inappropriate for EPA
to assess scenarios in which no gloves or the wrong glove
types are used.
While EPA has evaluated worker risk with and without
PPE, as a matter of policy, EPA does not believe it should
assume that workers are unprotected by PPE where such
PPE might be necessary to meet federal regulations, unless
it has evidence that workers are unprotected. In Section
2.4.1.1, EPA presented an example from a published report
on a scenario, graffiti removal, in which no gloves or the
wrong glove types are used.
51,59,
61
PUBLIC COMMENTS:
For scenarios that involve only industrial sites, EPA
inappropriately assumes that "SDS recommendations are
followed and that workers are likely to wear protective
gloves and have specialized training on the proper usage of
these gloves, corresponding to a protection factor of 20" (p.
213).
EPA points out that SDSs for NMP and NMP-containing
products recommend gloves (p. 68) and the NMP draft risk
evaluation states that "EPA expects ... [that] OSHA
regulations for worker protection and hazard communication
will result in use of appropriate PPE consistent with the
applicable SDSs in a manner adequate to protect them" (p.
335).
o No evidence is provided indicating that these SDSs are
read by most employers, let alone shared with workers,
or that their recommendations are consistently
implemented.
Information reasonably available to EPA, including data
submitted by chemical manufacturers and processors,
indicates that PPE is generally used. EPA does not assume
that the inclusion of PPE on SDSs is sufficient to ensure
PPE use. While EPA considers the information on SDSs,
EPA does not make PPE use assumptions based solely on
SDSs.
EPA generally assumes compliance with OSHA
requirements for protection of workers, including the
implementation of the hierarchy of controls. In support of
this assumption, EPA used reasonably available information
indicating that some employers, particularly in the
industrial setting, are providing appropriate engineering, or
administrative controls, or PPE to their employees
consistent with OSHA requirements.
EPA does not have reasonably available information to
support this assumption for each condition of use; however,
EPA does not believe that the Agency must presume, in the
Page 148 of 205
-------
o SDSs are often inaccurate, incomplete, and too technical
for many workers to understand.
In its proposed 2017 rule to ban MC and NMP paint
removers, EPA concluded that enhanced warnings and
directions for use would not be effective because
"consumers and professionals do not consistently pay
attention to labels for hazardous substances; consumers,
particularly those with lower literacy levels, often do not
understand label information... "
A comprehensive survey of SDSs identified "a number of
common themes ... regarding inaccuracies, incompleteness,
[and] incomprehensibility" and cautioned that "there are
serious problems with the use of SDSs as hazard
communication tools."
Even if gloves were worn by certain workers, EPA has little
to no information about the types of gloves worn. Without
data on which gloves are provided to which employees, EPA
has no basis for assuming specific GPFs in its draft risk
evaluation.
absence of such information, a lack of compliance with
existing regulatory programs and practices. Rather, EPA
assumes there is compliance with worker protection
standards unless case-specific facts indicate otherwise, and
therefore existing OSHA regulations for worker protection
and hazard communication will result in use of appropriate
PPE in a manner that achieves the stated APF or PF. EPA's
decisions for unreasonable risk to workers are based on
high-end exposure estimates, in order to account for the
uncertainties related to whether or not workers are using
PPE. EPA believes this is a reasonable and appropriate
approach that accounts for reasonably available information
and professional judgement related to worker protection
practices, and addresses uncertainties regarding availability
and use of PPE.
In Sections 2 and 4, the risk evaluation presents exposure
and risk estimates for each occupational exposure scenario
both with and without glove use. To the extent that EPA has
information about glove use in a specific industry, it is
described in Section 2.4.1.2. In Section 4.2.2, EPA
acknowledges that glove protection factors for each
exposure scenario are a source of uncertainty for exposure
and risk estimates.
PPE assumptions - Respirators
42
PUBLIC COMMENTS:
The assumption that respirators are used by all exposed
workers, under all conditions of chemical use and are 100%
effective is erroneous. Respirators are considered to be less
effective in protecting employees than other engineering and
administrative controls {i.e., NIOSH hierarchy of controls).
Employers are highly unlikely to institute a respiratory
protection program without an express requirement {i.e., no
OSHA PEL).
EPA states in Section 2.4.1.1 that "Few literature sources
indicate the use of respirators for reducing worker
exposures to NMP by inhalation. Therefore, EPA central
tendency and high-end scenarios do not incorporate
protection factors for respirator use." PBPK results shown
in Table 2-77 do not include reductions for respirator use.
In Section 4, EPA presents risk estimates for each COU
with and without gloves and respirators. As demonstrated
by the risk estimates summarized in Table 4-54, respirator
Page 149 of 205
-------
The 8-hour use of PPE should not be used in the risk
characterization.
use has minimal impact on overall risk estimates for
workers exposed to NMP.
Potentially exposed or susceptible subpopulations
SACC,
38,51,
55, 54,
61
SACC COMMENTS:
Recommendation: Revise Section 4.4 to describe PESS
consistent with how it is described in other sections of the
draft risk evaluation and to be as specific as the descriptions
used elsewhere in the document.
The statement "Due to limited information on the degree that
humans of varying gender, age, health status, or genetic
makeup might vary in the disposition of, or response to NMP
a factor of 10 was applied" bears repeating in Section 4.4.
It would be helpful if the draft risk evaluation listed the
categories of PESS described throughout the draft risk
evaluation and use this list as a guide for the summary
presented in Section 4.4 to assure completeness and
consistency.
Section 4 should emphasize more the most critical
population of concern (i.e., pregnant women or women who
could become pregnant).
PUBLIC COMMENTS:
EPA must consider and analyze each of the following types
of subpopulations: infants, children, pregnant women,
workers, and the elderly.
Due to the developmental neurotoxicity risks, pregnant
women, fetuses, and children should all be specifically
included.
Due to the reproductive risks, reproductive aged men and
women should be included.
Due to the risks for neurotoxicity, immunotoxicity, and other
risks, elders and people with health conditions should be
included.
In an effort to be concise and avoid repetition, Section 4.4
does not repeat everything stated in previous sections, but
does provide references and links back to slightly more
detailed discussions earlier in the document. All discussions
of PESS are consistent, though previous sections go into
slightly more detail. Section 4.4 lists the general categories
of PESS considered in the risk evaluation.
Consistent with the statement cited by the SACC, Section
4.4 has been slightly modified to state, "There is
insufficient information to support a quantitative analysis of
interindividual variability in other potentially susceptible
populations. An uncertainty factor of 10 was applied to
account for interindividual variability across gender, age,
health status, genetic makeup, or other factors, but the
actual effect of various factors contributing to biological
susceptibility on overall risk is unknown."
Although EPA agrees with the SACC that pregnant women
and women who could become pregnant are a critical
population to consider for developmental effects, they are
not the only population that needs to be emphasized in this
section. As described throughout Section 3, male exposure
to NMP prior to conception may reduce reproductive
success (by reducing male fertility and offspring survival).
Children and adolescents may also be sensitive to
reproductive effects, though the specific phases of
development during which exposure may have the most
impact is not known. Consistent with public comments,
Section 4 highlights the potential susceptibility of pregnant
women as well as both males and females of reproductive
age, children and adolescents, and people with health
Page 150 of 205
-------
Individuals with chronic liver or kidney disease or other
systemic ailments may be at particularly high risk for organ
damage or other systemic adverse effects. Thus, workers or
consumers at risk for developmental and reproductive effects
may well also be at risk for liver and kidney damage, and
vice-versa.
Page 151 of 205
conditions or metabolic differences. Section 4.4 states, "The
developmental effects identified as a critical human health
endpoint for acute exposures in this draft risk evaluation are
a major concern for pregnant women, the developing fetus,
and women who may become pregnant. The reproductive
effects identified as a critical human health endpoint for
chronic exposures may be of concern for all people of
reproductive age as well as for infants, children and
adolescents whose reproductive systems are still
developing.
Other populations that may be more sensitive to the hazards
of NMP exposure include people with pre-existing
conditions, and people with lower metabolic capacity due to
life stage, genetic variation, or impaired liver function. The
magnitude of the effect of each of these factors alone or in
combination on overall risk is unknown."
While EPA does not have specific information on
susceptibility of elderly people to NMP, they are more
likely to have pre-existing conditions. Potential
susceptibility to genetic variation and preexisting conditions
related to neurotoxicity and immunotoxicity and other risks
are specifically discussed in Section 3.2.5.4:"Genetic
variations or pre-existing conditions that increase
susceptibility of the reproductive system, the hepatic, renal,
nervous, immune, and other systems targeted by NMP
could also make some individuals more susceptible to
adverse health outcomes following consumer or workplace
exposures."
Finally, EPA notes that TSCA Section 3(12) lists examples
of human receptors that may be considered PESS but
provides for EPA to identify the relevant subpopulations for
each chemical substance. Infants, children, pregnant
women, workers, and the elderly are examples of human
-------
receptors that may be identified as PESS in individual
chemical risk evaluations.
SACC
SACC COMMENTS:
Recommendation: Consider formally calculating risk for the
CYP2E1 susceptible subpopulation.
Biologically susceptible subpopulations, including variations
in CYP2E1, other unidentified genetic variations, preexisting
conditions in the liver and other organ systems, and other
chemical co-exposures are discussed in adequate detail. This
information should be more completely summarized in
Section 4.4.
Many on the Committee concluded that there seems to be
enough information to attempt calculating risks for the
CYP2E1 susceptible subpopulation.
There is evidence that CYP2E1 contributes to NMP
metabolism CLigocka et al 2003), suggesting genetic
variation in CYP2E1 may increase susceptibility to NMP
by altering metabolism. While there is quantitative
information about variation in CYP2E1 expression across
the population, there is insufficient information to predict
the impact of that variation on risk. EPA does not have
sufficient quantitative information about the magnitude of
differences in NMP metabolism caused by genetic variation
in CYP2E1. There is also insufficient quantitative
information about variability of other metabolic pathways
that may contribute to NMP metabolism. EPA concluded
that there is not sufficient quantitative information about
variation in NMP metabolism to incorporate this into the
model. In the absence of more quantitative information,
EPA assumes that an interindividual uncertainty factor of
10 (with a factor of 3 designated for toxicokinetic
differences across individuals) is sufficient for addressing
metabolic differences within the population. The potential
role of CYP2E1 activity in increasing susceptibility is
discussed in Section 3.2.5.3.
SACC
SACC COMMENTS:
Recommendation: Expand the discussion on PESS to include
potential risks of developmental neurotoxicity in infants and
impacts of lower CYP2E1 activity in fetuses/infants.
The discussion on PESS in the risk characterization section
should include the following:
o Infants: Postnatal endpoints are not directly addressed in
the draft risk evaluation, but discussion of associated
endpoints and risks could be incorporated if
developmental neurotoxicity were considered.
EPA acknowledges the substantial uncertainty around the
increased susceptibility to young infants due to differences
in CYP2E1 expression in Section 3.2.5.3: "The variability
in CYP2E1 in pregnant women could affect how much
NMP reaches the fetus, which typically does not express
CYP2E1 (Hines. 2007). Newborns and verv voung infants
are particularly susceptible to NMP exposure because they
are metabolically immature. CYP2E1 is not fully expressed
in children until about 90-davs of age (Johnsrud et al
2003). The variability in CYP2E1 was identified as an
Page 152 of 205
-------
o Fetuses and infants: The draft risk evaluation does
discuss concerns with enzyme CYP2E1 activity
including a known mutation that can put fetuses and
infants at risk, but risk characterization was not provided
for this subpopulation. If this risk factor were considered,
risk calculations would change with the result that the
safety of this vulnerable population would become a
concern.
important uncertainty that was reflected in the calculation
of the intraspecies uncertainty factor (human variability)."
As discussed in response to the previous comment, EPA
does not have sufficient information to provide a
quantitative analysis of the impact of this important
metabolic difference on risk. To specifically address
uncertainty related to developmental neurotoxicity and
metabolic differences in the risk characterization section,
EPA added the following statement to Section 4.4: "there is
uncertainty around the impact of metabolic differences in
young infants on susceptibility. There is also uncertainty
around susceptibility of infants and young children to
neurodevelopmental effects of NMP which have been
documented in animals at high doses but have not been
characterized at lower doses."
PESS - Tribal communities
55
PUBLIC COMMENTS:
The multiple exposure scenarios associated with proximity
to unlined disposal site releases to environmental media
must be analyzed for both individual exposures and the
cumulative exposures that tribal members face from their
customary and traditional tribal lifeways (inhalation, dermal,
ingestion).
As part of this analysis, EPA should identify all populations
living near disposal and other waste management sites as
potentially exposed subpopulations. Groups living near
National Priority List (NPL) sites and proposed NPL sites
should be included as well.
EPA believes it is both reasonable and prudent to tailor
TSCA risk evaluations when other EPA offices have
expertise and experience to address specific environmental
media, rather than attempt to evaluate and regulate
potential exposures and risks from those media under
TSCA. EPA has therefore tailored the scope of the risk
evaluation for NMP using authorities in TSCA Sections
6(b) and 9(b)(1). As described in Section 1.4.2 of the risk
evaluation, EPA did not include exposures via the drinking
water pathway or disposal to underground injection, RCRA
Subtitle C hazardous waste landfills, or RCRA Subtitle D
municipal solid waste (MSW) landfills in this risk
evaluation. These exposure pathways fall under the
jurisdiction of other EPA-administered statutes and
associated regulatory programs.
Page 153 of 205
-------
Populations exposed through pathways excluded from the
risk evaluation were not identified as PESS. EPA disagrees
with public comments on the draft risk evaluation that
suggest fenceline subpopulations should be identified as
PESS. TSCA provides EPA with the discretion to identify
the PESS that are relevant to the chemical-specific risk
evaluation [ TSCA Section 6(b)(4)(A)], General population
exposure through air, surface water, sediment, and land-
applied biosolids were evaluated based on fate properties of
NMP and screening level analysis. As described in Section
4.6.1.3, EPA did not identify risks to the general population
through these pathways. As described in Section 1.4.2,
general population exposures through drinking water and
disposal are beyond the scope of the risk evaluation.
Regarding cumulative exposures, EPA considered the
reasonably available information and used the best
available science to determine whether to consider
aggregate or sentinel exposures for a particular chemical.
EPA used PBPK modeling to evaluate total risks from
combined inhalation, dermal, and vapor through skin
exposures for each COU. EPA concluded that there is
insufficient information to support analysis of aggregate
exposure across multiple conditions of use. EPA
acknowledges that the decision not to aggregate risk across
conditions of use could result in an underestimate of risk.
55
PUBLIC COMMENTS:
Not considering legacy use, and the risks it poses,
disproportionately affects tribes' exposures, in this case due
to the unique disposal circumstances on tribal lands and in
tribal communities. EPA must consider the impacts of legacy
use of NMP on tribal populations.
EPA did not identify, based on reasonably available
information, any "legacy uses" or "associated disposals" of
NMP, as those terms are described in EPA's Risk
Evaluation Rule, 82 FR 33726 (July 20, 2017). Therefore,
no such uses or disposals were added to the scope of the
risk evaluation for NMP following the issuance of the
Page 154 of 205
-------
opinion in Safer Chemicals, Healthy Families v. EPA, 943
F.3d 397 (9th Cir. 2019).
55
PUBLIC COMMENTS:
Tribes must be considered as a potentially exposed
subpopulation under TSCA due to higher consumption of
locally caught fish and other aquatic life {i.e., subsistence
fishing), substandard housing, lack of worker safety
protocols, unique water uses, and other lifeways {i.e.,
environmental activities for dietary sustenance, socio-
cultural activities, ceremonial and spiritual purposes,
recreation, and general well-being).
EPA disagrees with public and scientific advisory
committee comments on the draft risk evaluation that
suggest tribal communities should be identified as PESS.
Populations exposed through pathways excluded from the
risk evaluation were not identified as PESS. TSCA provides
EPA with the discretion to identify the PESS that are
relevant to the chemical-specific risk evaluation [ TSCA
Section 6(b)(4)(A)], As described in Section 1.4.2, EPA
determined that general population exposures through
drinking water and disposal pathways are under the
jurisdiction of other EPA-administered statutes and fall
outside the scope of the risk evaluation. As described in
Sections 2.1.1 and 4.6.2, EPA concluded based on first-tier
analysis and environmental fate properties considered
during problem formulation that general population
exposure through ambient air, ambient water, sediment, and
land-applied biosolids do not pose a human health risk and
did not require further analysis in the risk evaluation.
Commenters note that the HBCD risk evaluation identified
subsistence fishermen as PESS; however, HBCD is
classified as a persistent bioaccumulative toxic (PBT)
compound and expected to bioaccumulate through the food
chain. NMP is not a PBT and has low bioaccumulation
potential. Therefore, NMP is not a significant concern for
communities with elevated fish ingestion and the
consumption of fish along with other trophic transfer
pathways were not included in the scope of the risk
evaluation.
55
PUBLIC COMMENTS:
Under the SDWA and CWA, multiple tribes use individual
groundwater well systems that are not regulated or
As described in the Problem Formulation, EPA did not
evaluate risks from NMP exposure through drinking water.
Exposures to the general population via drinking water,
Page 155 of 205
-------
monitored and have members on remote systems that are not
POTWs due to the system size.
which includes finished surface and ground water are
covered under SDWA.
EPA performed an additional analysis of potential
exposures through ambient water regulated under the Clean
Water Act. While EPA does not have sufficient information
about the specific levels of exposure experienced by tribes,
the ambient water analysis includes a consideration of
potential high-end exposures to NMP through surface
water. EPA did not identify risks from incidental ingestion
of or dermal contact with NMP in surface water.
55
PUBLIC COMMENTS:
Small tribal businesses and self-employed tribal workers are
exempt from OSHA, and in rural Alaska non-hub
communities, where the majority of Alaska's federally
recognized tribes live, OSHA will only provide assistance
and compliance visits if three separate entities request them.
Most of rural Alaska's communities do not have three
entities to which the workplace exposures discussed in the
draft risk evaluation would be relevant.
EPA evaluated occupational risk according to exposures
and hazards identified through reasonably available
information. Risk conclusions do not rely on OSHA
standards.
EPA considers each condition of use and uses exposure
scenarios with and without PPE that may be applicable to
particular worker tasks on a case-specific basis for a given
chemical. For the purposes of determining whether or not a
condition of use presents unreasonable risks, EPA
incorporates assumptions regarding PPE use based on this
information and judgement underlying the exposure
scenarios. These assumptions are described in the
unreasonable risk determination for each condition of use,
in Section 5.2. While EPA has evaluated worker risk with
and without PPE, as a matter of policy, EPA does not
believe it should assume that workers are unprotected by
PPE where such PPE might be necessary to meet federal
regulations, unless it has evidence that workers are
unprotected. In the case of tribal businesses and/or small
business that may not be subject to OSHA regulations,
consistency of PPE use is a source of uncertainty in the risk
evaluation. In consideration of the uncertainties and
Page 156 of 205
-------
variabilities in PPE usage, EPA uses the high-end exposure
value when making its unreasonable risk determinations in
order to address those uncertainties as well as to capture
exposures for PESS. EPA has also outlined its PPE
assumptions in Section 5.1.
55
PUBLIC COMMENTS:
In order to make an accurate risk characterization of tribal
communities, EPA needs to consider releases of NMP from
landfills as well as any other disposal facility, such as a
transfer station or recovery facility.
As described in the Problem Formulation and in Section
1.4.2 of the risk evaluation, EPA did not analyze risks of
NMP releases from landfills because they are under the
jurisdiction of other statutes.
EPA did not include releases to land from RCRA Subtitle D
municipal solid waste (MSW) landfills or exposures to the
general population from such releases in the risk evaluation.
As NMP is not classified as a RCRA hazardous waste,
NMP containing solid waste may be sent to RCRA Subtitle
D municipal solid waste (MSW) landfills. While permitted
and managed by individual states, MSW landfills
established after 1989 are required by federal regulations to
implement some of the same requirements as Subtitle C
landfills. MSW landfills must have a liner system with
leachate collection and conduct ground water monitoring
and corrective action when releases are detected. MSW
landfills are also subject to closure and post-closure care
requirements, as well as providing financial assurance for
funding of any needed corrective actions. MSW landfills
have been designed to allow for the small amounts of
hazardous waste generated by households and very small
quantity waste generators (< 220 pounds per month). Bulk
liquids, such as free solvent, may not be disposed of in
MSW landfills. See 40 CFR part 258.
NMP containing solid wastes are not expected to be sent to
Subtitle C incinerators because NMP is not a hazardous
waste and due to higher cost of such incineration as
compared with MSW or other incinerators. Emissions from
Page 157 of 205
-------
hazardous waste incinerators were not evaluated. However,
it is possible that NMP containing solid wastes could be
sent to subtitle C incinerators due to other characteristics of
an NMP-containing solid waste mixture.
55
PUBLIC COMMENTS:
It is recommended that "the context of the assessment would
be improved by including a graphic" ... "that illustrates
exposure routes for potentially sensitive or highly exposed
populations."
EPA appreciates this suggestion. While EPA is unable to
develop this in the timeframe available to complete the
NMP risk evaluation, EPA is developing graphics to
illustrate potential exposure routes and receptors for future
risk evaluations.
exposure
SACC
SACC COMMENTS:
Recommendation: Explain the rationale for deciding not to
consider risks from combined occupational exposures
through multiple tasks. Discuss this decision as a source of
uncertainty in the characterization of risk to the population.
Workers may perform multiple work activities within an
occupational exposure scenario, especially when sub-
scenarios within the use scenario have maximum exposure
duration of less than a full shift. This is especially relevant to
work activities (sub-scenarios) of short duration in many use
scenarios in manufacturing, repackaging, chemical
processing-excluding formulation, printing & writing, and
recycling & disposal (NMP risk evaluation, Table 2-66).
For most work activities, EPA calculated potential risks
based on the assumption that the exposure from that
particular work activity is continuous over either a task
duration, if available, and portions of a shift, half a shift (for
central tendency exposures) or over the whole shift (for
high end exposures). While some workers may be exposed
to NMP through multiple work activities in a shift, EPA
assumes that workers generally perform one activity at a
time. The portions of a shift account for multiple contacts,
which may be for one or more activities. EPA expects that
the risks calculated from a full shift of exposure through
single work activities will approximate potential risks for
workers who split their shifts across multiple activities that
are each performed for shorter durations.
SACC,
51,55,
34, 48,
61,46,
59
SACC COMMENTS:
Recommendation: Explain the rationale for deciding not to
consider aggregate risks from combined occupational and
consumer exposure pathways. At a minimum, the
implications of not aggregating should be specifically
described. Discuss this decision as a source of uncertainty in
the characterization of risk to the worker/consumer
population.
PUBLIC COMMENTS:
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. EPA defines aggregate
exposures 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).
EPA considered the reasonably available information and used
the best available science to determine whether to consider
Page 158 of 205
-------
-------
risk evaluations is consistent with statutory text in TSCA
Section 6(b)(4)(A), which instructs EPA to conduct risk
evaluations to determine whether a chemical substance
presents an unreasonable risk "under the condition of use."
Assumptions and uncertainties - General
SACC,
54
SACC COMMENTS:
Recommendation: Describe more appropriately the
assumptions and uncertainties in Section 4.3.6 of the draft
risk evaluation.
Committee members expressed concerns that Section 4.3.6
Risk Characterization Assumptions and Uncertainties (NMP
Risk Evaluation, p. 278) does not reflect appropriately what
the title of the section implies. A Committee member
suggested that the description of uncertainties in dermal
exposure parameters found in section 2.4.1.4. of the draft
risk evaluation (pp. 136-137, lines 2864-2884) provides a
clear description of sources of uncertainty and could be used
as a template for the discussion that summarizes assumptions
and uncertainties in the parameters of the risk
characterization.
PUBLIC COMMENTS:
EPA should consider providing a more detailed discussion of
uncertainty in the risk characterization section of the revised
risk evaluation and quantitatively demonstrate the impacts of
alternative plausible approaches on the calculated theoretical
risks.
Section 4.3.7 (formerly 4.3.6 as identified by the comment)
focuses on assumptions and uncertainties related to the
overall integration of exposure and hazard information via
the PBPK model.
Because specific assumptions and uncertainties vary for
each exposure scenario, and uncertainties related to
exposure and hazard characterization have been described
in previous sections, Section 4.3.7 does not provide the
same level of detail as previous discussions of uncertainty.
Scenario-specific strengths, limitations, and overall
confidence are presented along with risk estimates in
Section 4.2.2. To the extent that EPA is able to quantify
them, the effect of specific exposure assumptions and
parameters on exposure estimates are described in Section 2
and can be further explored in the supplemental risk
calculator excel files. Similarly, to the extent that EPA is
able to quantify them, the impact of various assumptions
and dose-response approaches is described in Section 3.2.6
and can be further explored in the BMD modeling
supplemental file. EPA added the new Section 4.3.6 to
discuss assumptions and uncertainties associated with the
PBPK models.
SACC
SACC COMMENTS:
Recommendation: Consider incorporating additional Monte
Carlo Analysis and/or sensitivity analysis to better
characterize uncertainties.
Characterization of uncertainties could be improved by
expanded use of global and specific sensitivity analysis, and
A meaningful Monte Carlo analysis would require
information about the variability of specific parameters. For
several important sources of uncertainty, EPA lacks the
quantitative information that would be needed to inform
Monte Carlo analysis. For example, EPA relies on
assumptions about specific dermal exposure parameters for
Page 160 of 205
-------
in some scenarios, Monte Carlo simulations. Several
Committee members suggested that Monte Carlo simulations
are worthwhile and be used across the board to the greatest
extent possible. This approach makes possible: 1) more
appropriate characterization of uncertainties due to, for
example, polymorphisms and enzyme activity; or 2) make
comparisons of risks under different assumptions of PPE use
and engineering controls.
Other members of the Committee suggested performing a
global and local sensitivity analysis. Performing limited
global sensitivity analysis followed by targeted sensitivity
analysis of key parameters would accomplish goals like
those that could be accomplished using Monte Carlo
simulations.
which there is a lack of reliable data.
Where possible, EPA has performed sensitivity analysis to
characterize the magnitude of uncertainty. For example,
where plausible alternate exposure assumptions have been
proposed by stakeholders, EPA has considered 'what if
scenarios, to estimate exposure based on alternate industry
assumptions. These alternate exposure estimates provide an
indication of the magnitude of uncertainty associated with
EPA's assumptions. EPA has inserted additional discussion
of sensitivity analyses in Section 2. Where possible, EPA
describes the magnitude of uncertainty and the potential
impact on risk estimates throughout Section 4.3.
SACC
SACC COMMENTS
Recommendation: Describe how uncertainties and biases
may be increased by creating and using condition of use
'categories' in exposure and hazard assessments and by
assigning estimated risks for the 'category' back to the
individual scenarios and specific conditions of use.
There is insufficient clarity about how and to what extent
uncertainties were considered, weighted, and propagated to
arrive at overall confidence about specific risks. For
example, decisions made in the process of 'grouping'
exposure scenarios and specific conditions of use into
'categories' assumed to have similar exposure and risks,
introduces uncertainties and if not performed carefully can
lead to biases in exposure and risk estimates. This is
particularly a concern when usable exposure measurement
data are available for only one or a couple of the specific
conditions of use in the condition of use 'category.' The
draft risk evaluation is not clear about this potential source
of uncertainty and bias and this should be described in more
Risk estimates for each condition of use are based on
reasonably available information and there is some
uncertainty around the representativeness of these estimates
across all exposure scenarios that a relevant for a given
COU. EPA provides risk estimates for both central
tendency and high end exposures in an effort to capture the
range of exposures that may be associated with each
condition of use. Where more detailed information is
reasonably available, separate risk estimates are shown for
several distinct tasks within a given COU (e.g., for lithium
ion manufacturing) to provide more specific information to
risk managers. In addition, EPA has separated the
electronics COU into two distinct COUs (semiconductor
manufacturing and other electronics manufacturing). Where
more detailed information is not reasonably available, the
representativeness of risk estimates for each COU across all
facilities and tasks is a source of uncertainty.
For occupational exposures, EPA added in Section 2.4.1.4
Page 161 of 205
-------
detail. Approaches used to account for these potential biases
in the final risk characterization need to be presented.
that the application of OESs and associated work activities
increases uncertainties in PBPK parameter inputs for OESs
that combine COUs and that the directional impacts due to
this application of either overestimating or underestimating
exposures estimated by PBPK modeling are not known.
Assumptions and uncertainties - Occupational exposure
SACC
SACC COMMENTS:
Recommendation: Clarify the uncertainty introduced by the
assumption of similar inhalation exposures for workers and
ONUs when estimating the relative contributions from
different exposure routes.
Regarding the approach to estimating total human exposures
from combined inhalation and dermal exposures using the
human PBPK model, the risk evaluation should more clearly
explain that this approach assumes the central tendency
estimates for inhalation exposures for workers, assigns
central tendency PBZ exposures to ONUs, and assumes
vapor through the skin exposures are similar for both
workers and ONUs. Worker exposure is via inhalation,
vapor through skin and dermal liquid contact. ONU exposure
is via only inhalation and vapor through skin exposure.
Under these assumptions, the only difference between
worker and ONU exposures is the amount of dermal liquid
contact of the worker. But if inhalation exposure for workers
is actually higher than for ONUs and assuming vapor
through the skin exposure estimates are similar, then the
difference between worker and ONU exposures now
includes both the difference in inhalation exposures and the
dermal exposure due to workers' contact with the liquid.
This adds uncertainty to the estimate of dermal exposure.
EPA added discussion in Section 2.4.1.1 regarding the
relative contributions of each exposure pathway to total
exposures, which vary according to parameter values for
NMP weight fraction in the liquid product contacted, skin
surface areas in contact with the liquid product and with
vapor, durations of dermal contact with liquid product and
with vapor, air concentration for inhalation and vapor-
through-skin exposure, body weight of the exposed person,
and glove protection factor and respirator assigned
protection factor (if applicable). EPA added clarifications to
include the uncertainty of this assumption in the overall
confidence discussions at the end of each OES subsection in
2.4.1.2. The uncertainty of this assumption is discussed in
Section 2.4.1.4. For most OESs, ONU-specific data and
modeling are not available; in these OESs, ONU inhalation
exposures are assumed to be lower than inhalation
exposures for workers directly handling the chemical
substance. To account for this uncertainty, EPA assumed
that the workers' central tendency air concentration
estimates (rather than the high-end estimates) are
appropriate for determining ONUs' air concentration
estimates. This assumption does not add uncertainty to the
estimate of dermal exposure since it has no impact on
dermal exposure.
SACC
SACC COMMENTS:
Recommendation: Include a discussion of potential dermal
exposures to NMP vapor penetrating through clothing, and
EPA has included discussion in Uncertainties Sections
2.4.1.4 and 4.3 that dermal exposures to NMP vapor that
may penetrate clothing fabrics and the potential for
Page 162 of 205
-------
from contact with NMP vapor-saturated clothing; describe
these additional exposure routes as sources of uncertainty in
the risk evaluation.
The Committee was concerned about the potential for NMP
vapor penetrating through regular (non-impervious fabric)
clothing, which could mean that the worker has >25%
exposed skin surface area assumed in the draft risk
evaluation, which would increase overall skin dermal
exposures.
In addition, it is likely that NMP vapor will adsorb/desorb to
and from fabric fibers until it equilibrates with airborne
NMP levels. This results in contact with NMP-impregnated
clothing becoming a source of ongoing dermal exposure
throughout the workday.
associated direct skin contact with clothing saturated with
NMP vapor are not included in quantifying exposures. The
discussion further notes that these uncertainties could
potentially result in underestimates of exposures.
SACC
SACC COMMENTS:
Recommendation: Include a summary of uncertainties about
PPE use.
To the extent that EPA has reasonably available
information about PPE use for specific occupational
exposure scenarios, it is summarized in Section 2.4.1.2.
Throughout Section 4.2.2, in describing the strengths,
limitations and overall confidence in risk estimates for each
exposure occupational scenario, the risk evaluation
describes glove protection factors as assumptions that are a
source of uncertainty. Section 4 presents risk estimates for
all occupational exposure scenarios for workers both with
and without gloves and respirators.
SACC
SACC COMMENTS:
Recommendation: Provide a description of the strengths,
limitations, and overall confidence for the risk
characterization for ONUs (NMP risk evaluation, Section 4.3
and in the summary).
Section 4.2.3 includes a discussion of the overall
confidence in risk estimates for ONUs. EPA clarified in
4.2.3 and in all OES subsections of 2.4.1.2 that EPA assigns
the same confidence level for PBPK inputs for both workers
and ONUs because lower surface areas for liquid contact
for ONUs have higher certainty but air concentrations
experienced by ONUs have lower certainty. These factors
offset one another in determining ONU confidence level
using worker confidence level as a starting point.
Page 163 of 205
-------
Assumptions and uncertainties - Consumer Exposure
SACC
SACC COMMENTS:
Recommendation: Emphasize that the CEM assumption that
bystanders remain in another room of the household while
the consumer uses NMP-containing products adds more
differential uncertainty to risk estimates for products that
could likely be applied in the presence of a bystander.
The Consumer Exposure Model (CEM) assumes that
bystanders remain in one room that is separated from the
consumer who is using the NMP-containing product in
another room. The draft risk evaluation discussion does not
emphasize that this assumption increases the uncertainty of
the risk characterization. This assumption may be more
appropriate for some products than others. For example, it
may not work for an NMP-containing adhesive that can be
used in any room of the house, compared to an NMP-
containing carbon remover that is typically used only in a
more open garage or outdoors.
EPA acknowledges that consumer bystanders were not
assumed to be exposed in same room as the users. EPA will
consider this refinement to the consumer modeling
approach for future evaluations.
However, for NMP, a disproportionate route of exposure is
dermal from direct handling. Recognizing that the most
sensitive receptor for acute exposures is a reproductive-age
woman, the risk evaluation assumes such a user and
estimates her total (dermal + inhalation) exposure will
always exceed any proximate bystander exposure,
regardless of their location.
Assumptions and uncertainties - Human health hazard
SACC,
57, 33
SACC COMMENTS:
Recommendation: Add details of the rat inhalation studies as
they relate to the total NMP dose (i.e., potential for oral
exposure), reproductive stages vis a vis exposure, and the
interpretation of maternal and fetal toxicity indicators.
There was insufficient attention to detail in the description of
the rodent studies used for risk evaluation. Some of these
details (i.e., the potential for additional oral ingestion of
NMP by preening, or the developmental stages in rodent
studies) could add to uncertainty, but they are not described
in the draft risk evaluation as sources of uncertainty.
PUBLIC COMMENTS:
EPA agrees that there is uncertainty around NMP doses
achieved in whole body inhalation exposure studies.
In Section 4.3.5, EPA has inserted additional discussion of
the uncertainty around doses achieved in inhalation studies
due to aerosolization and potential for oral ingestion due to
grooming behavior. This topic is also discussed in Section
3.2.6.
The PODs that were ultimately selected as the basis for risk
calculations were not reliant on results of inhalation studies.
EPA therefore concludes that this source of uncertainty is
unlikely to result in an over- or underestimate of risk.
Page 164 of 205
-------
It is estimated that the two pathways of vapor-to-dermal and
ingestion of vapors would combine to increase the total
NMP dose delivered to rats via whole-body exposures by a
factor of 3.3-fold compared to vapor inhalation alone. EPA
should include quantification of these two pathways (dermal
absorption of NMP vapors, ingestion of fur-adsorbed NMP
from grooming) when deriving POD values from rat studies
for inhalation exposures to NMP. The uncertainties
surrounding the potential dermal and oral uptake of NMP,
through grooming, in rats following whole-body inhalation
should be discussed in Section 4.3 of the risk assessment.
At a minimum, this important source of uncertainty should
be discussed in Section 4.3 of the risk assessment.
Risk Conclusions - Risk drivers and overall confidence in conclusions
SACC
SACC COMMENTS:
Recommendation: Indicate clearly the main driver(s) for the
risk estimates for each of the specific conditions of use.
Provide a more transparent justification for the overall levels
of confidence presented in the summaries.
The exposures, MOEs, and risk estimates need to be
examined carefully to ensure findings make sense across the
specific conditions of use and occupational exposure
scenarios in each condition of use category, acknowledging
the limits on the data that were available to derive the
estimates. A clearer justification is needed for the overall
level of confidence for each specific risk estimate.
It is not clear how limitations, strengths, and assessed
uncertainties translate into assigned low, medium, or high
levels of confidence. More information is needed, and the
decision process better defined, on how limitations and
strengths in risk estimates are considered and weighed in
assigning overall confidence to risk estimates. Some
Committee members suggested that quality scores could be
In the risk evaluation the main drivers for risk estimates
were identified in the unreasonable risk determination for
each condition of use. In the determinations in which
unreasonable risk was found, the term "unreasonable risk
driver" was used to label the risk drivers, and in
determinations where no unreasonable risk was found, the
exposure scenario with the highest risk estimate was
described. EPA thanks the SACC commenters for their
recommendations and has revised the risk evaluation to add
clarity on this issue.
EPA added discussion in Section 2.4.1.1 regarding the
relative contributions of each exposure pathway to total
occupational exposures, which vary according to parameter
values for NMP weight fraction in the liquid product
contacted, skin surface areas in contact with the liquid
product and with vapor, durations of dermal contact with
liquid product and with vapor, air concentration for
inhalation and vapor-through-skin exposure, body weight of
Page 165 of 205
-------
used, for example, when there is conflicting information,
such as in the toxicology data. In addition, qualifiers should
be removed from the risk evaluation when describing
uncertainty (e.g., "some uncertainty"). Numerical values
should be provided when uncertainties are quantifiable, and
or when describing levels of exposure (e.g., add the
percentile in addition to indicating "high end").
the exposed person, and glove protection factor and
respirator assigned protection factor (if applicable). In
scenarios where the three parameters involving dermal
contact with liquid product (NMP weight fraction in the
liquid product contacted, skin surface areas in contact with
the liquid product and with vapor, durations of dermal
contact with liquid product) have relatively high values, this
route can be the dominant route for worker exposures.
EPA assigned low, medium or high levels of overall
confidence qualitatively based on consideration of the
specific limitations, strengths, and uncertainties identified
for each exposure scenario and the hazard points of
departure.
SACC
SACC COMMENTS:
Recommendation: Remove qualifiers that do not provide
meaningful information when describing uncertainties and
provide numerical values whenever possible in conclusions
about level of confidence.
There was discussion about the description of risk as "not
unreasonable" instead of just saying "reasonable" or
considering the term "measurable risk."
EPA has revised the language describing uncertainties to
avoid unnecessary qualifiers. For many aspects of
uncertainty and confidence, EPA lacks sufficient
information to provide numerical values, but EPA has
provided quantitative information where possible.
SACC
SACC COMMENTS:
Recommendation: Address more specifically how
uncertainties impact MOEs when making "no unreasonable
risk" determinations and increase MOEs to accommodate
uncertainty.
There are uncertainties in the toxicity assessment and the
PBPK model. The extent of uncertainties appears too high to
support with confidence a determination of no unreasonable
risk.
The risk evaluation should only support "no unreasonable
risk" determinations when and where more solid evidence is
available, and uncertainties are judged to be truly small.
Section 5 of the risk evaluation describes how uncertainties
impact the determinations of unreasonable or no
unreasonable risk. The degree of uncertainty surrounding
the MOEs, cancer risk or RQs is a factor in determining
whether or not unreasonable risk is present. Where
uncertainty is low, and EPA has high confidence in the
hazard and exposure characterizations (for example, the
basis for the characterizations is measured or monitoring
data or a robust model and the hazards identified for risk
estimation are relevant for conditions of use), the Agency
has a higher degree of confidence in its risk determination.
EPA may also consider other risk factors, such as severity
Page 166 of 205
-------
Where uncertainties are large, MOEs need to be larger as
well, to reflect those larger uncertainties.
of endpoint, reversibility of effect, or exposure-related
considerations, such as magnitude or number of exposures,
in determining whether the risks are unreasonable under the
conditions of use.
SACC
SACC COMMENTS:
Recommendation: Provide a discussion of strengths,
weaknesses, and overall confidence for characterization of
risk to environmental receptors, following the pattern of
these descriptions for risks to human health (Sections 4.3.5
and 4.3.6).
Additional discussion has been added to Sections 4.1.2 and
4.3.4 of the final risk evaluation to discuss the strengths and
weaknesses of the environmental hazard and risk
characterization as well as identify the main sources of
uncertainty.
Risk conclusions - Other
34, 51
PUBLIC COMMENTS:
EPA repeatedly finds that risks that fall below the
benchmark MOE or that exceed EPA's cancer threshold are
reasonable, and need not be managed under TSCA.
In its final risk evaluation, EPA should adhere to its own
unreasonable risk criteria and not reclassify risks that meet
these criteria as "reasonable."
A calculated MOE for non-cancer risk that is less than the
benchmark MOE indicates the possibility of unreasonable
risk to human health. Whether there are unreasonable risks
will depend upon other risk-related factors, such as severity
of endpoint, reversibility of effect, exposure-related
considerations (e.g., duration, magnitude, frequency of
exposure, population exposed), and the confidence in the
information used to inform the hazard and exposure values.
If the calculated MOE is greater than the benchmark MOE,
generally it is less likely that there is unreasonable risk.
61,51,
53
PUBLIC COMMENTS:
While EPA's draft risk evaluations find that certain uses of
NMP pose unreasonable risks, EPA understates those risks
and thus violates TSCA's mandate to protect workers.
EPA miscalculates the severity of worker risks by
misconstruing OSHA requirements related to the use of
respirators and other PPE.
EPA has significantly understated NMP's risks because of
several omissions, indefensible assumptions and errors in its
risk evaluation methodology.. .Of most concern is EPA's
assumption that millions of exposed workers are "expected"
to wear protective gloves
In Sections 2 and 4, the risk evaluation presents exposure
and risk estimates for each occupational exposure scenario
both with and without glove use. To the extent that EPA has
reasonably available information about glove use in a
specific industry, it is described in Section 2.4.1.2. In
Section 4.2.2, EPA acknowledges that glove protection
factors for each exposure scenario are a source of
uncertainty for exposure and risk estimates.
For the purposes of determining whether or not a condition
of use presents unreasonable risks, EPA incorporates
assumptions regarding PPE use based on information and
judgement underlying the exposure scenarios. These
Page 167 of 205
-------
It is requested that EPA reach a conclusion of unreasonable
risk of injury to worker health, from chronic inhalation and
dermal exposure during drum unloading and loading into
shipping containers
assumptions are described in the unreasonable risk
determination for each condition of use, in Section 5.
Additionally, in consideration of the uncertainties and
variabilities in PPE usage, EPA uses the high-end exposure
value when making its unreasonable risk determinations in
order to address those uncertainties. EPA has also outlined
its PPE assumptions in Section 5.1. EPA agrees that there
are challenges associated with use of PPE; they are
described in Section 5. By providing risk estimates with use
of PPE, EPA is not recommending or requiring use of PPE.
Rather, these risk estimates are part of EPA's approach for
developing exposure assessments for workers that use the
reasonably available information and expert judgment.
When appropriate, in the risk evaluation, EPA will use
exposure scenarios both with and without engineering
controls and/or PPE that may be applicable to particular
worker tasks on a case-specific basis for a given chemical.
While EPA has evaluated worker risk with and without
PPE, it would be unrealistic for EPA to assume that workers
are typically unprotected by PPE.
56
PUBLIC COMMENTS:
EPA's draft finding of unreasonable risk for workers during
lithium ion cell manufacturing is based on erroneous
assumptions concerning how our industry handles NMP,
implements engineering controls, and protects workers.
Thank you for the comment. EPA worked with the lithium
ion cell manufacturing industry to incorporate substantiated
information into the Final Risk Evaluation.
EPA revised the occupational exposure assessment in the
risk evaluation to separately assess occupational exposure
scenarios associated with three categories of electronic part
manufacturing: Lithium ion battery manufacturing
(2.4.1.2.15); Other electronics manufacturing, including
capacitor, resistor, coil, transformer, and other inductor
manufacturing (2.4.1.2.9); and Semiconductor
manufacturing (2.4.1.2.10). In these separate OESs, EPA
revised and expanded PBPK runs for industry-specific work
Page 168 of 205
-------
activities using industry-specific air concentration data sets
provided in public comments for the lithium ion battery
manufacturing industry, for the semiconductor
manufacturing industry, and from the OSHA data set for
capacitor, resistor, coil, transformer, and other inductor
manufacturing (LICM. 2020a; Semiconductor Industry
Association. 2020, JOT Ik i \ )
52, 64
PUBLIC COMMENTS:
For purposes of the final risk evaluation, EPA must consider
the specific conditions of NMP use in the semiconductor
industry separately from other sectors with which they were
combined for purposes of the draft risk evaluation.
The Agency must update and enhance its final risk
assessment, and conclude, pursuant to 40 CFR Section
702.49(d) of EPA's risk evaluation rules, that the use of
NMP in the semiconductor industry does not present an
unreasonable risk. Moreover, EPA should determine that the
conditions of use in the sector require no further regulatory
scrutiny under Section 6 of TSCA.
Thank you for the comment. EPA worked with the
semiconductor industry to incorporate documented
assumptions and information into the Final Risk Evaluation.
EPA revised the occupational exposure assessment in the
risk evaluation to separately assess occupational exposure
scenarios associated with three categories of electronic part
manufacturing: Lithium ion battery manufacturing
(2.4.1.2.15); Other electronics manufacturing, including
capacitor, resistor, coil, transformer, and other inductor
manufacturing (2.4.1.2.9); and Semiconductor
manufacturing (2.4.1.2.10). In these separate OESs, EPA
revised and expanded PBPK runs for industry-specific work
activities using industry-specific air concentration data sets
provided in public comments for the lithium ion battery
manufacturing industry and for the semiconductor
manufacturing industry, and from the OSHA data set for
capacitor, resistor, coil, transformer, and other inductor
manufacturina CSIA. 2019b. c; SIA. 2020; LICM. 2020b;
OSHA. 2017Y
53
PUBLIC COMMENTS:
The high-end scenario may not represent actual exposure in
the workplace. EPA used monitoring data to develop an 8-
hour TWA for the high-end exposure scenario. Modeling for
this high-end scenario indicates that exposure does not meet
MOE threshold, indicating unreasonable risk. The 8-hour
period does not reflect actual exposure times of a worker
EPA differentiates between worker contact duration with
liquid and inhalation duration in Section 2.4.1.1 and
explains that PBPK inputs are adjusted to normalize these
durations.
Page 169 of 205
-------
during a shift. As a result, the conclusion of unreasonable
risk does not necessarily mean that workers are exposed to a
health risk in the workplace.
Adherence to TSCA - Effects on determination of risk
46, 56
PUBLIC COMMENTS:
The draft conclusions in this risk evaluation should not be
allowed to stand, consistent with TSCA, where best
available science has not been used.
EPA risk conclusions are based on best available science
and the weight of the scientific evidence.
51
PUBLIC COMMENTS:
The NMP draft risk evaluation is incomplete and inadequate
and does not comply with TSCA in the absence of sufficient
data to address whether all endpoints present an
unreasonable risk of injury. EPA's failure to develop risk
estimates for developmental neurotoxicity, immunotoxicity,
and endocrine effects is effectively a recognition that it
cannot make unreasonable risk determinations under TSCA
Section 6(b) for these endpoints using currently available
data.
EPA evaluated the reasonably available information from
animal toxicology studies. While EPA agrees that there is
limited information for some endpoints, EPA considers the
database adequate for risk evaluation without the need to
separately address immune effects on their own. In Section
3.2.6 and Section 4.3.5, EPA identifies the limited data for
developmental neurotoxicity, immunotoxicity and
endocrine effects as sources of uncertainty that could result
in an underestimate of risk.
59
PUBLIC COMMENTS:
EPA has failed to acknowledge that the requirements it relies
on derive from statutes (e.g., OSHA) that establish criteria
different than those under TSCA for establishing
requirements to address human and environmental health
risks. Many of these other statutes, for example, require EPA
or other agencies to consider factors such as cost and
feasibility when setting standards - factors that TSCA
explicitly forbids EPA from taking into account when
assessing risks.
EPA risk conclusions are based on exposure and hazard
characterization that EPA performed based on reasonably
available information. EPA evaluated exposure and hazard
without consideration of costs or other non-risk factors, and
without using other statutory risk assessment standards that
consider cost or non-risk factors. While EPA assumed PPE
use for some conditions of use based on the assumption of
compliance with OSHA standards, the risk associated with
a given level of exposure is based on EPA's independent
evaluation of reasonably available information.
Risk mitigation/management
53
PUBLIC COMMENTS:
EPA should use actual workplace exposure data or exposure
times for each conditions of use or recognize the limitations
of its approach when considering risk mitigation measures.
EPA uses all reasonably available actual workplace
exposure data for each condition of use. For duration of
dermal contact with liquids, EPA did not find reasonably
available actual workplace exposure durations and explains
Page 170 of 205
-------
assumptions as noted in Section 2.4.1.1. EPA explains the
uncertainties of these assumptions in Section 2.4.1.4.
53, 31
PUBLIC COMMENTS:
Since EPA will initiate risk mitigation in summer 2020, it is
recommended that EPA consider requiring PPE where it has
issued a finding of no unreasonable risk based on use of
PPE.
EPA should recognize limitations during risk mitigation so
as not to prescribe unnecessarily restrictive and unjustified
control measures, including banning NMP or setting
unreasonable de minimis values for the uses described.
As a result of lower exposure potential from significant
operational control, risk mitigation criteria should be specific
and less rigorous for semiconductor manufacturing.
There is concern that EPA will not be able to prescribe
adequate risk mitigation measures tailored to a condition of
use because of inconsistencies and vaguely supported
findings.
EPA can assist businesses by requiring PPE that mitigates
risk to "no unreasonable risk."
Thank you for the comment. For any conditions of use
found to present unreasonable risk in the final risk
evaluation, EPA will move immediately into risk
management. EPA will consider any comments related to
risk management at that time. EPA evaluated all conditions
of use of NMP under TSCA, including commercial and
industrial uses that result in occupational exposures. Risk
management activities are outside the scope of the risk
evaluation. As the commenter indicated, as appropriate for
any condition of use determined to present unreasonable
risk, EPA will consider feasibility in implementation of any
risk management actions that are proposed to address the
unreasonable risks that EPA has determined are presented.
In that context, EPA intends to analyze the applicability of
any PPE training, certification, and limited access
programs.
53
PUBLIC COMMENTS:
EPA's finding of unreasonable risk will be used to develop
risk mitigation measures after EPA finalizes its risk
evaluation in summer 2020. Yet, for several conditions of
use, EPA has not identified, to a high degree of certainty,
conditions causing risk.
There is concern that EPA's draft risk evaluation will
identify issues for further examination without clearly
identifying conditions leading to unreasonable risk to
workers and consumers. This in turn might result in EPA
developing unnecessary or flawed risk mitigation measures.
Thank you for the comment. Per the statute (see TSCA
Section 6(b)(4)(A)) and the implementing regulations for
risk evaluations (40 CFR part 702, subpart B), EPA must
make the unreasonable risk determination at the time of the
risk evaluation. Upon finding unreasonable risk, EPA will
apply risk management actions to the extent necessary so
that the chemical no longer presents such risk, in
accordance with TSCA Section 6(a). For any conditions of
use found to present unreasonable risk in the final risk
evaluation, EPA will move immediately into risk
management. EPA will consider any comments related to
risk management at that time. EPA evaluated all conditions
of use of NMP under TSCA, including commercial and
Page 171 of 205
-------
industrial uses that result in occupational exposures. Risk
management activities are outside the scope of the risk
evaluation. As the commenter indicated, as appropriate for
any condition of use determined to have unreasonable risk,
EPA will consider feasibility and implementation of any
risk management actions that are proposed to address the
unreasonable risks that EPA has determined are presented.
61
PUBLIC COMMENTS:
By assuming extensive use of PPE at the risk evaluation
stage, EPA conflates risk evaluation with risk management.
PPE is a risk management tool, albeit a poor one that may be
used only when preferable options are not available. As
such, PPE may only be considered, if at all, during the risk
management stage when it can be weighed against more
effective means of risk reduction.
Thank you for the comment. EPA has updated Table 2-77
to include worker exposures for all glove PFs for all OESs.
A hierarchy of controls is a method for eliminating
workplace hazards. While EPA has assessed the extent to
which certain exposure reduction tools that it assumes to be
in place may be reducing risks to workers, application of
the methodology of the hierarchy of controls is not relevant
to risk evaluations. EPA will manage unreasonable risks
presented by chemical substances when the Agency
undertakes regulatory action for COUs determined to have
unreasonable risk. Utilization of the hierarchy of controls
to recommend or require risk management actions in the
risk evaluation would be premature and inappropriate.
A hierarchy of controls is a method for eliminating
workplace hazards. While EPA has assessed the extent to
which certain exposure reduction tools that it assumes to be
in place may be reducing risks to workers, application of
the methodology of the hierarchy of controls is not relevant
to risk evaluations. EPA will manage unreasonable risks
presented by chemical substances when the Agency
undertakes regulatory action for COUs determined to have
unreasonable risk. Utilization of the hierarchy of controls to
recommend or require risk management actions in the risk
evaluation would be premature and inappropriate.
38, 48,
51, 61
PUBLIC COMMENTS:
EPA should immediately move forward to finalize a ban on
EPA appreciates the comment. As published in the
Methylene Chloride rule for consumer paint and coating
Page 172 of 205
-------
paint stripping uses as proposed in 2017.
removal, "EPA intends to incorporate NMP use in paint and
coating removal in the risk evaluation for NMP. EPA has
concluded that the Agency's assessment of the potential
risks from this widely used chemical will be more robust if
the potential risks from these conditions of use are
evaluated by applying standards and guidance under
amended TSCA. In particular, this includes ensuring the
evaluation is consistent with the scientific standards in
Section 26 of TSCA, including using best available science
and systematic review approaches." EPA evaluated all
conditions of use of NMP under TSCA, including
commercial and industrial uses that result in occupational
exposures. Risk management activities are outside the
scope of the risk evaluation.
As stated in the executive summary of the risk evaluation,
any proposed or final determination that a chemical
substance presents unreasonable risk under TSCA Section
6(b) is not the same as a finding that a chemical substance
is "imminently hazardous" under TSCA Section 7.
32, 54
PUBLIC COMMENTS:
EPA should revisit the clarity and organization of Table 5-1.
It is difficult to understand. EPA should boldface both the
"presents" and the "does not present" statements in Table 5-
1 to improve its risk communication about its risk
determinations.
In addition, EPA's "presents" and "does not present"
statements should cite to the Section 26 statutory and
regulatory requirements demonstrating that the
determinations are based upon best available science, weight
of the scientific evidence, and data quality.
A statement regarding the uncertainty in the appropriate dose
metric and outcome response rate for resorptions should be
added to the table.
Thank you for the comment. EPA added clarity to Table 5-1
and the risk determination language in Section 5 in the final
risk evaluation.
Page 173 of 205
-------
EPA should consider including a modified table that
represents the relevant endpoints and drivers, potentially
color-coding those that exceed benchmarks.
Page 174 of 205
-------
8. Content and Organization
('onlcnl suit! Organization
Charge Question 7.1: Please comment on the overall content, organization, and presentation of the NMP draft risk evaluation. Please
provide suggestions for improving the clarity of the information presented.
Charge Question 7.2: Please comment on the objectivity of the underlying data used to support the risk characterization and the
sensitivity of the agency's conclusions to analytic assumptions made.
Summary of Comments lor Specific Issues Related to
Charge Question 7
narge question /
Evaluation scope, methodology, and transparency
SACC I SACC COMMENTS:
EPA Response
SACC,
46,51,
55,46
Recommendation: Document and distribute to the
public a clear description of the following:
o The rationale for focusing TSCA evaluations only
on exposures to occupational users, non-
occupational users, and direct consumers of
chemical-containing products to the exclusion of
exposures to the broader public. [Note: the
conditions of use discussion in Section 1.4 is
inadequate for this purpose.]
o The timeline for combining risk assessments
conducted under TSCA, CAA, and CWA
regulations to produce a complete picture of risks
from TSCA listed chemicals,
o The rationale and approach for making
"unreasonable risk" determinations under TSCA.
[Section 5.1.1 of the NMP draft risk evaluation has
a good discussion of this but does not address the
PPE use issue that keeps arising in Committee
discussions.]
Readers of TSCA-focused evaluations receive only a
partial picture of risks to the chemical being assessed.
The Committee continues to express its concern that
EPA has added Section 1.4.2, which describes exposure
pathways and risks that fall under the jurisdiction of other EPA-
administered statutes or regulatory programs. As described in
Section 1.4.2, EPA believes it is both reasonable and prudent to
tailor TSCA risk evaluations when other EPA offices have
expertise and experience to address specific environmental
media, rather than attempt to evaluate and regulate potential
exposures and risks from those media under TSCA. EPA
believes that coordinated action on exposure pathways and risks
addressed by other EPA-administered statutes and regulatory
programs is consistent with statutory text and legislative history,
particularly as they pertain to TSCA's function as a "gap-
filling" statute, and also furthers EPA aims to efficiently use
Agency resources, avoid duplicating efforts taken pursuant to
other Agency programs, and meet the statutory deadline for
completing risk evaluations. EPA has therefore tailored the
scope of the risk evaluation for NMP using authorities in TSCA
Sections 6(b) and 9(b)(1).
As described in Section 4.6.2.3, some exposure pathways that
are not covered under other statutes were considered during
problem formulation, but not further analyzed in the risk
evaluation. During problem formulation, EPA evaluated
potentials exposures and risks to the general population through
Page 175 of 205
-------
this incomplete picture of risks may be used to promote
improper releases and disposal of chemicals that may
harm not only worker and ONU health, but also the
health of the general population.
The Committee encourages the Agency to rapidly
complete the assessment of all other releases and
general population exposures and associated hazard to
complete the risk picture for NMP.
PUBLIC COMMENTS:
EPA's draft risk evaluation excludes all human
exposures from environmental releases of NMP,
resulting in the absence of any consideration of
environmental pathways that contribute to overall
human risk exposure and risk. This approach is an
unlawful interpretation of TSCA, has twice been
rejected by the SACC, and overlooks the widespread
presence of NMP in environmental media to which
millions of people are exposed.
The air, water, and waste pathways excluded from the
NMP draft risk evaluation are significant contributors
to human exposure and should be included in risk
determinations.
EPA cannot evaluate, as it purports to do here, the
total, cumulative risk to public health and the
environment from these chemicals if it excludes
exposures through other pathways.
ambient water, land-applied biosolids, and ambient air. Based on
environmental fate properties of NMP and first-tier screening
level analyses, EPA did not identify risk to the general
population from these pathways. In the final risk evaluation,
EPA included an updated analysis of potential risks from
incidental ingestion and dermal contact with NMP in surface
water based on more recent TRI data.
EPA concluded that there is insufficient information to support
analysis of aggregate exposure across multiple conditions of use
for NMP. EPA acknowledges that the decision not to aggregate
risk across conditions of use could result in an underestimate of
risk.
Other data to consider
SACC
SACC COMMENTS:
One Committee member suggested studies (Babich et
al., 2004, Van Engelen et al., 2008) that provide better
information on children's mouthing behavior of toys.
EPA has included these references in the final risk evaluation.
SACC
SACC COMMENTS:
EPA held several meetings with States to gather chemical data
including a call specifically with Washington State Department
Page 176 of 205
-------
Several Committee members reiterated that EPA
should pursue additional data on use and use practices
from states (Washington, Vermont, Maine, and
Oregon) having chemical use permitting laws that
require reporting of these practices for targeted
chemicals. The Committee recognizes that such reports
may represent content or concentrations only (rather
than fully inform exposure) and are subject to
significant over-reporting rather than being based on
experimental data.
of Ecology in January 2017. EPA also routinely highlighted
NMP and the other "First 10 Chemicals" in meetings with the
Environmental Council of States (ECOS) and other EPA
quarterly calls on toxic chemicals. EPA held several public
comment periods encouraging the submittal of chemical use
information.
EPA carefully considered the best approach to capture the
conditions of use across and within sectors using NMP, and
relied on communications with companies, industry groups,
environmental organizations and public comments to
supplement the use information.
EPA had sufficient information to complete the NMP risk
evaluation using a weight of the scientific evidence approach.
EPA selected the first 10 chemicals for risk evaluation based in
part on its assessment that these chemicals could be assessed
without the need for regulatory information collection or
development. When preparing this risk evaluation, EPA
obtained and considered reasonably available information,
defined as information that EPA possesses, or can reasonably
obtain and synthesize for use in risk evaluations, considering the
deadlines for completing the evaluation. In some cases, when
information available to EPA was limited, the Agency relied on
models; the use of modeled data is in line with EPA's final Risk
Evaluation Rule and EPA's risk assessment guidelines.
However, EPA will continue to improve on its method and data
collection for the next round of chemicals to be assessed under
TSCA.
Editorial - Clarity/transparency
SACC
SACC COMMENTS:
EPA should use a standard format for citations,
ensuring that primary original data sources are cited
and reviewed (not simply subsequent papers or sources
using the original data), and including a sentence or
EPA has made revisions throughout the risk evaluation to cite
original sources and clarifying where only secondary sources
are available.
Page 177 of 205
-------
two stating the salient point when referencing
supporting documents.
SACC
SACC COMMENTS:
One Committee member suggested that using color in
tables to highlight significant findings rather than
simply holding values would improve clarity of
findings.
To improve clarity, EPA has modified table formatting to shade
cells to highlight significant findings (in addition to holding
values).
SACC
SACC COMMENTS:
SACC reported a number of editorial corrections to be
addressed in the revised risk evaluation.
EPA appreciates this feedback and has made editorial revisions
where possible to improve clarity.
31
PUBLIC COMMENTS:
Clarification/increased transparency is requested for the
following:
The basis for the Agency's handling of air sampling
data with observations falling below detection limits is
not transparent. EPA should provide a record of the
values used for the LOD in the Agency's statistical
adjustments.
EPA's statistical analysis when interpreting air
sampling results cannot be reproduced based on the
limited information provided. More details about the
PBPK model and its inputs are needed.
On page 176 of the Supplemental Information on
Occupational Exposure Assessment, why are line items
19 & 20 excluded? Is there confusion about the name
of the sample called "Fab Area samples?"
The Supplemental Information on Occupational Exposure
Assessment describes in Section 1.4.3.1 how EPA handles
datasets with results below detection limits. Specifically, for
datasets including exposure data that were reported as below the
limit of detection (LOD), EPA estimated the exposure
concentrations for these data, following EPA's Guidelines for
Statistical Analysis of Occupational Exposure Data (
1994). which recommends using the if the geometric
V 2
standard deviation of the data is less than 3.0 and if the
2
geometric standard deviation is 3.0 or greater.
Additional information on how EPA used datasets containing
results below detection limits for specific OES was included in
the appropriate subsections of the Supplemental Information on
Occupational Exposure Assessment (Section 2.8 Electronics
Parts Manufacturing and Section 2.9 Printing and Writing).
Line items 19 & 20 of page 176 of the Supplemental
Information on Occupational Exposure Assessment were
incorrectly labeled as personal breathing zone samples. This
was corrected to area samples. These data were excluded
because EPA used personal breathing zone monitoring data,
Page 178 of 205
-------
which is higher on the hierarchy in selecting data and
approaches for estimating air concentrations.
59
PUBLIC COMMENTS:
EPA has not provided public access to at least 12
sources that include health and safety information on
which EPA relies in its draft risk evaluation.
While they have entries in EPA's Heath and
Environmental Research Online (HERO) database,
those entries do not provide a means of accessing the
documents themselves. While some of the reference
list entries cite OTS, TSCATS, and other cross-
references, our searches for these failed to yield access
to the information sources.
While EPA claims not to have relied on two of these
studies, conflicting language elsewhere in the draft risk
evaluation requires clarification from EPA.
EPA appreciates this comment and generally expects to make
the information it uses for decision-making publicly available,
consistent with and subject to the requirements of TSCA
Section 14. The commenter identified several specific
references. EPA addressed access to each study as follows:
4 studies bv BASF (2001. 1989. 1986. 1983}
o 1983: Acute fi sh toxi city
o 1986: Acute fish toxicity
o 1989: Acute algae toxicity
o 2001: Chronic aquatic invertebrates
These are unpublished aquatic toxicity studies in the
HERO database.
DTI (2004): Survev of chemical substance in consumer
products. EPA is working to upload the document into
HERO.
Dupont (1990). A HERO link provided for this studv in the
draft risk evaluation linked to the wrong Dupont 1990
document. The link has been corrected in the final risk
evaluation. While the DuPont 1990 study is not publicly
available, it is consistently cited in tandem with Solomon,
1995, the publicly available version of the same study which
was published in the peer reviewed literature and is publicly
available. The data EPA relied on is publicly available
through Solomon, 1995 and the Dupont 1990 citation is
provided for reference and to be clear that both references
refer to the same study.
Huntingdon (1998). The information provided in this dermal
exposure study in rats indicates that certain solvents can
increase permeability of NMP. While EPA cites this study
Page 179 of 205
-------
for context, EPA does not have sufficient information about
each exposure scenario to incorporate solvent-specific
permeability estimates. The information in this study
provides useful context but does not provide the quantitative
analysis performed by EPA.
NMP Producers Group (1999a, b). EPA did not have access
to these two generation reproduction studies at the time the
draft risk evaluation was released. EPA has since obtained
access to both studies and posted them to the docket with
names of some study personnel redacted.
Prestige (2010): Safetv data sheet. The link to this SDS
appears to have become broken. In HERO, EPA replaced the
broken link with a link to the product web page and added a
PDF of the SDS as it appeared when it was originally found.
SI A (2019c): These exposure monitoring results are
available in the public docket.
59
PUBLIC COMMENTS:
In the draft risk evaluation, EPA stated that the NMP
Producers Group studies "contribute to overall weight
of evidence." However, an apparently contradictory
statement indicates that EPA did not use the results of
these studies in its analysis.
EPA needs to clarify that the studies will not be used in
its "overall weight of evidence" approach unless the
full studies are provided to EPA and made public along
with providing an opportunity for public comment on
them.
In the draft risk evaluation, EPA did not have access to either of
the NMP Producers Group two-generation reproductive toxicity
studies. EPA has since obtained access to both studies and
posted them to the docket [EPA-HQ-OPPT-2019-023 6], EPA
evaluated both studies using the systematic review data quality
criteria, performed dose-response analysis for developmental
endpoints reported in the studies, and incorporated results of the
studies into hazard identification, weight of the scientific
evidence and dose-response analysis.
54
PUBLIC COMMENTS:
EPA should more clearly explain its approach to
analyzing the potential risks of substances that are
already subject to other federal environmental
programs. EPA should also consider whether its
As described above, EPA has inserted Section 1.4.2, which
describes exposure pathways and risks that fall under the
jurisdiction of other EPA-administered statutes or regulatory
programs. EPA believes it is both reasonable and prudent to
tailor TSCA risk evaluations when other EPA offices have
Page 180 of 205
-------
approach is consistent across its TSCA risk
evaluations, and if not, to explain why not.
expertise and experience to address specific environmental
media, rather than attempt to evaluate and regulate potential
exposures and risks from those media under TSCA. EPA has
therefore tailored the scope of the risk evaluation for NMP
using authorities in TSCA Sections 6(b) and 9(b)(1).
Editorial - Accuracy
31
PUBLIC COMMENTS:
Comments on accuracy of the Supplemental Information
on Occupational Exposure Assessment:
The semiconductor manufacturing industry does not
use any open top tanks of NMP; the actual process in
the semiconductor industry is VERY different than that
described, and therefore, the descriptions are not
accurate.
The following statement is incorrect "data received
from European Semiconductor Industry (SIA) 2019."
The word "European" should be removed.
The high number of Semiconductor and Other
Electronic Component Manufacturing listed in Table 2-
39 is incorrect.
Based on process descriptions provided earlier in this
document, the last sentence on page 81 is not correct.
Semiconductor manufacturing has more controlled
NMP handling practices than other electronics parts
manufacturing.
The following statement is incorrect: "EPA did not find
data on exposure duration." SIA provided exposure
duration for each task and samples were taken for
duration of exposure.
Table 2-42 contains some information considered to be
inaccurate: No one in semiconductor manufacturing
handles containers for 6-12 hours/day; several of the
tasks in this table are also not completed on a daily
EPA revised the process description in Section 2.8.1
(Electronics Manufacturing) of the Supplemental Information
on Occupational Exposure Assessment. Specifically, EPA
included a process description specifically for semiconductor
manufacturing alone, which does not include the use of open
top tanks.
EPA made the suggested revision to remove "European" from
"data received from European Semiconductor Industry (SIA)
2019."
EPA revised Section 2.8.2.2 to include separate estimates of
number of sites and workers for each electronics manufacturing
OES. This revision resulted in fewer sites and workers
specifically applicable to semiconductor manufacturing.
EPA revised the occupational exposure assessment in the risk
evaluation to separately assess occupational exposure scenarios
associated with three categories of electronic part
manufacturing: Lithium ion battery manufacturing (2.4.1.2.15);
Other electronics manufacturing, including capacitor, resistor,
coil, transformer, and other inductor manufacturing (2.4.1.2.9);
and semiconductor manufacturing (2.4.1.2.10). In these separate
OESs, EPA revised and expanded PBPK runs for industry-
specific work activities using industry-specific air concentration
data sets provided in public comments for the lithium ion
battery manufacturing industry and for the semiconductor
Page 181 of 205
-------
manufacturing industry, and from the OSHA data set for
capacitor, resistor, coil, transformer, and other inductor
manufacturing (LICM. 2020a; Semiconductor Industry
Association. 2020. 2019b. c; OSHA. 2017). Therefore,
semiconductor manufacturing is no longer assumed to be
representative of all subcategories of the Electronic Part
Manufacturing OES, and the sentence that had been at the
bottom of page 81 was deleted.
EPA replaced the statement "EPA did not find data on exposure
duration" with "EPA did not find reasonably available data on
actual duration of dermal contact with liquids." EPA also
revised PBPK inputs for this OES to include "what-if' task
duration-based durations for liquid contact, which use tasks
durations provided from public comments, including from SIA
public comments (Semiconductor Industry Association. 2020.
2019b, c).
While EPA revised the assessment to include "what-if' task
duration-based PBPK inputs when available, EPA retains full-
shift and half-shift shift-based duration PBPK inputs for all
OESs due to uncertainty of task durations representing actual
durations of contact with liquids.
EPA revised the statement from page 86, Section 2.8.4, of the
draft risk evaluation to include a third exposure route, vapor-
through-skin exposure. The new sentence is in Section 2.9.4 due
to reorganizing of sections. The presence of a fraction, even a
majority, of non-detect values of air concentrations does not
remove the potential for exposure through any of these routes.
EPA deleted the statement from the draft risk evaluation on
page 172, "However, no other methods to address the reporting
Page 182 of 205
basis; and exposed skin surface area values do not
align with the percentage of exposed skin that SIA
provided when discussing the PPE used to handle
NMP.
On page 86, Section 2.8.4, the summary statement
seems inaccurate for the amount of data SIA provided,
most of which yielded non-detect results.
Regarding the statement on page 172, "However, no
other methods to address the reporting limit of
detection exist (EPA, 1994)," it was suggested that
EPA use the AIHA method defined in A Strategy for
Assessing and Managing Occupational Exposures and
the accompanying spreadsheet IHSTAT.
-------
limit of detection exist (EPA, 1994)." However, EPA retained
the approach used ( ) because it is consistent
with EPA's approach in other current chemical risk evaluations.
59
PUBLIC COMMENTS:
EPA's descriptions of its glove use assumptions and
analysis are contradictory within the text and are
inconsistent with the information that it relied on in its
risk estimates table (Table 4-50) and risk determination
table (Table 5-1).
In the risk characterization in Section 4, EPA presents risk
estimates for all occupational COUs both with and without PPE.
In the risk determination in Section 5, EPA makes unreasonable
risk determinations based on risk estimates and reasonably
available information on PPE use for each COU. EPA has
outlined the PPE assumptions in Section 5.1 and EPA's
assumptions are described in the unreasonable risk
determination for each condition of use, in Section 5.2.
57
PUBLIC COMMENTS:
In Table 3-3-1 of the BMD supplemental file, the
number of litters cited in the table for the 120 and 831
mg/L dose groups differs from the values used in the
BMD modeling and should be corrected.
This was an error. EPA has updated the number of litters from
21 to 22 for the 120 mg/L dose group, and from 5 to 25 for the
831 mg/L dose group to match the number of litters reported in
the original Saillenfait et al. (2002) publication. The undated
litter sizes appear in Table 2-2 of the updated BMD
Supplemental file.
Editorial - Suggested additions
SACC
SACC COMMENTS:
Recommendation: Consider adding a comprehensive
table/appendix for each occupational use scenario, sub-
scenario, and scenario characteristic (central tendency
or high-end) that, in addition to the information
provided in Table 2-66, also includes all of the acute
and chronic non-cancer risk estimates. This may be
helpful to industrial hygienists as they consider how
task duration, NMP concentration, hand exposure, and
PPE use modify risk.
In Section 2.4.1.3, EPA stated that the full range of this
modeling is presented in the spreadsheet Supplemental File on
Occupational Risk Calculations. This file covers all of the
aspects that are indicated in this comment.
SACC
SACC COMMENTS:
One Committee member suggested that a map of
facilities that use NMP or discharge NMP could be
helpful, especially in providing connections between
use, disposal, human health, and the environment. It
EPA thanks the Committee member for the suggestion to add a
map of facilities that use or discharge NMP. EPA will give
consideration to the feasibility of adding facility mapping to
future TSCA chemical risk evaluations, which will be weighed
against any security concerns.
Page 183 of 205
-------
was acknowledged that there are potential security
issues in producing such a map.
SACC
SACC COMMENTS:
One Committee member thought it would be helpful if
EPA was clear on which references in Table 1-5
actually include (NMP risk evaluation, p. 31, lines 481-
2) "information on conditions of use, hazards,
exposures and potentially exposed or susceptible
subpopulations," and where and how that information
was used in the risk evaluation.
EPA appreciates this suggestion. While this is not possible in
the time available to complete the NMP risk evaluation, EPA
will consider such a table for future risk evaluations.
SACC
SACC COMMENTS:
One Committee member suggested that a fact sheet on
dermal parameters should be a standard part of each
TSCA chemical risk evaluation. This fact sheet should
both include theoretical values and specifically provide
ranges based on physical-chemical properties and
experimental results where data are available in the
research literature.
EPA will consider this type of communication material for
future risk evaluations. Dermal exposure parameters are
described in the risk evaluation in Section 2.4.1.1,
2.4.2.3, and 3.2.5.5. Details of scenario-specific assumptions in
occupational exposure scenarios are available throughout
Section 2.4.1 and in the supplemental file Supplemental
Information on Occupational Exposure Assessment.
Systematic review - Limitations of guidelines studies
38
PUBLIC COMMENTS:
According to the TSCA Systematic Review, higher
quality studies are guideline studies or data collected
according to GLP requirements. Guideline studies are
most often designed to identify major toxic effects
(apical effects) like cancer, major organ weight gain or
loss, body weight gain or loss, skeletal malformations,
and clinical signs; they are not sufficiently sensitive to
reliably identify low-dose exposure, endocrine of
hormonal effects, or neurobehavioral effects that may
occur at low doses during critical windows of
development.
The TSCA Systematic Review guidelines result in
inappropriate favoring of industry studies, without
The TSCA risk evaluation strategies in some cases refer to
study guidelines along with professional judgement as helpful
guidance in determining the adequacy or appropriateness of
certain study designs or analytical methods. This should not be
construed to imply that academic research studies or other non-
guideline studies are automatically given lower confidence
ratings than guideline or Good Laboratory Practice (GLP)
studies typically conducted by industry. EPA considers
reasonably available, relevant data and information that
conform to the TSCA science standards when developing the
risk evaluations irrespective of whether they were conducted in
accordance with standardized methods (e.g., OECD test
guidelines or GLP standards).
Page 184 of 205
-------
assessing study quality. Studies conducted by industry,
as well as academic research studies, should be
systematically evaluated using a credible Systematic
Review method, such as used by the EPA Integrated
Risk Information System (IRIS) program and the
National Institute of Environmental Health Sciences.
EPA will publish a protocol document for the next TSCA risk
evaluations. To refine that protocol, EPA is reviewing existing
peer-reviewed systematic review approaches, consulting with
EPA's IRIS program, and considering feedback from the
NASEM TSCA Committee.
Systematic review - Information from ECHA dossiers
59
PUBLIC COMMENTS:
On p. 47 of its draft risk evaluation, EPA claims that
ECHA dossiers are existing chemical assessments
equivalent to EPA and Agency for Toxic Substances
and Disease Registry (ATSDR) governmental
assessments.
ECHA dossiers are not assessments and are not
government documents. They are compilations of
industry information submitted to ECHA that have not
been evaluated for quality or reliability by ECHA or
any other governmental entity. For EPA to equate them
with EPA and ATSDR assessments is simply wrong.
o At the bottom of each page of each dossier is a
statement that the information has not been reviewed or
verified by ECHA or any other authority. While some
chemicals do eventually undergo a "substance
evaluation" by government authorities under REACH,
NMP has not.
o EPA exacerbates the mischaracterization through its
text references to the industry's dossiers, typically cited
as "ECHA, [date]." Clicking on that link takes the
reader to EPA's entry for that source in its HERO data
system, in which the reference's author is prominently
listed as the "European Chemicals Agency." Such text
citations and HERO entries are misleading. All of these
Industry submitted ECHA dossiers have not gone through the
same level of review as government assessments. EPA removed
the reference to ECHA dossiers in the footnote offering
examples of previous assessments. Where possible, EPA now
cites original sources rather than summaries in ECHA dossiers.
Page 185 of 205
-------
documents were prepared by the industry registrants,
not ECHA.
59
PUBLIC COMMENTS:
EPA does not have access to the full studies for all of
the studies on which it relies in the draft risk
evaluation. The ECHA dossiers EPA has cited contain
only summaries of studies, not the studies themselves.
Even the best study summaries are incomplete
descriptions that do not allow for an independent
examination of study quality and conclusions reached
by authors.
Systematic review practices require access to full
studies, as details of study design and results are
necessary elements of consistently determining study
quality and ultimately evidence integration.
EPA needs to not only obtain copies of the full studies,
but also make full copies of all studies on which it
relies available to the public.
EPA needs to review the full study reports to confirm
that the information in the summaries meets the
scientific standards set forth in TSCA section 26.
Without access to full studies, EPA and the public will
be challenged or unable to assess and comment on the
quality of the studies.
EPA has obtained and reviewed reasonably available
genotoxicity studies using the systematic review data quality
criteria. However, some of the studies cited and summarized in
ECHA dossiers are not available to EPA. Where possible, EPA
has revised citations to refer to primary sources that are publicly
available rather than relying on secondary sources. EPA's
conclusions are based on data reasonably available from
primary sources. In some places, EPA notes the existence of
additional studies summarized in dossiers but for which EPA
does not have access to the full study. References to these study
summaries are only included to provide a complete picture of
reasonable available information, not to serve as the basis for
EPA decisions.
59
PUBLIC COMMENTS:
In the scoring sheet for the summary of the first of the
three degradation studies, EPA repeatedly noted it is a
"secondary source" that provides limited detail and that
the "primary source may have more detail." It is not
clear what the "primary source" actually is, and EPA
appears to rely on the industry-prepared summary
instead of the primary source. The lack of important
EPA has replaced the ECHA study summaries by their
respective primary sources: Shaven (1984) for the first Gerike
and Fischer (Vs and Krizek et al. (2015) for the second, and
U.S. EPA (:01:) (i.e., EPI Suite) for the third.
Page 186 of 205
-------
detail in the summary also calls into serious question
the medium ranking EPA assigned to it.
In the scoring sheet for the summary of the second of
the studies, EPA repeatedly refers to information that
was "not reported" or was "omitted." In a few places,
EPA adds without citing any basis that such omissions
"did not limit the interpretation of the results" or "were
not likely to have had a substantial impact on the study
results." It is exceedingly difficult to understand how
EPA can possibly draw such conclusions in the
absence of access to the full study (which is
unpublished). Nor are such flaws consistent with the
"high" quality ranking EPA assigned to the study. With
the study itself not made available, the public is left
with no ability to independently assess the validity of
such statements or EPA's reliance on the summary in
the draft risk evaluation.
The third study is apparently not an experimental study
but an estimation based on a quantitative structure-
activity relationship (QSAR) software model that EPA
developed. It is not at all clear why EPA did not run in
its own model rather than rely on an industry-prepared
summary of its run of the same model. Moreover, in
the summary in the ECHA dossier, the industry
registrant notes several major caveats regarding its
reliability. These statements by the industry registrant
are wholly at odds with EPA's assignment of a "high"
quality ranking, but these summaries were still used in
its analysis.
Systematic review - General
40
PUBLIC COMMENTS:
EPA published the
tor i\ivhJ: Supplemental Document to tne l Scope
ument in 2017 along with the scope document for NMP.
Page 187 of 205
-------
The protocol for the TSCA systematic review process
was not provided in the draft risk evaluation and does
not follow the general steps for a systematic review.
The TSCA method excludes the following steps:
protocol development, evidence identification,
evidence integration, and hazard identification. In
addition, the TSCA method uses a non-empirically
based 'scoring' system, includes metrics inside each
domain not relevant to study quality, and excludes
relevant studies.
This document outlined the literature search strategy and
title/abstract inclusion/exclusion criteria used for screening.
EPA subsequently published Aimlicaiion of Systematic Review
in TSCA Risk Evaluations that described the data aualitv criteria
used for each discipline and outlined data integration strategies
that will be further developed for the next risk evaluations.
Because the systematic review steps have been published and
are available to the public, EPA did not publish the protocols in
the risk evaluation documents.
EPA/OPPT's quality evaluation method was developed
following identification and review of various published
qualitative and quantitative scoring systems to inform our own
fit-for-purpose tool. The development process involved
reviewing various evaluation tools/frameworks (e.g., OHAT
Risk of Bias tool, CRED, etc.; see Appendix A of the
Application of Systematic Review in TSCA Risk Evaluations
document and references therein), as well as soliciting input
from scientists based on their expert knowledge about
evaluating various data/information sources specifically for risk
assessment.
51
PUBLIC COMMENTS:
The TSCA approach applies a rigid scoring system to
grade the "quality" of studies on chemicals. This
system could result in many studies being arbitrarily
classified as "poor" or "unacceptable" based on a small
number of reporting or methodology limitations that do
not negate their overall value for assessing health and
environmental risks. The consequence will be that
important evidence of public health impacts will be
either disregarded or given limited weight in risk
evaluations.
The updated criteria make it more difficult for
epidemiological studies to be scored as high quality,
EPA has comprehensively evaluated the reasonably available
human and animal studies forNMP.
The epidemiologic criteria were revised to more stringently
distinguish between High, Medium and Low studies. After
additional piloting of the criteria, EPA found that the initial
iteration of the epidemiological data quality criteria (as
published in the Application of Systematic Review in TSCA
Risk Evaluations) was inadvertently skewing quality scores
toward the tail ends of the scoring spectrum (High and
Unacceptable). In order for the criteria to represent a more
accurate depiction of the quality levels of the epi literature, the
criteria were revised. With the changes to the criteria, EPA
observed fewer studies with Unacceptable ratings and more
Page 188 of 205
-------
reflecting a consistent tendency by the TSCA program
to downplay the value of human evidence.
studies shifting from High to Medium, with only the highest
quality studies receiving a High overall rating.
EPA is in the process of revising the data quality criteria for
future assessments based on feedback from the NASEM TSCA
Committee.
34, 40,
51, 53
PUBLIC COMMENTS:
In EPA's "hierarchy of preferences," EPA does not
explain why some types of studies should receive
preference over others and on what basis these studies
should be assigned to a "higher level."
This hierarchy of preferences is not peer reviewed or
part of any documents on which EPA took public
comment.
There are no objective criteria for determining which
evidence to rely on and which to exclude, undermining
transparency and consistency and encouraging
subj ective judgments.
The hierarchy of preferences was used to exclude 39
relevant and acceptable sources. In the draft risk
evaluation, EPA recognizes that the quality of data in
excluded studies is acceptable for risk assessment.
These studies were not identified or made public.
EPA should make excluded studies available or, at a
minimum, provide a list of excluded studies with an
explanation of how EPA applies its hierarchy of
preferences
Different lines of evidence are routinely used in TSCA chemical
assessments because of data availability, sources, underlying
documentation, and quality varies. EPA preferentially relies on a
variety of test and analog data. In the absence of suitable test
data, predictive modeling tools may be used.
EPA clarified under Figure 1-6 that lower quality data from 39
sources were not integrated based on EPA's integration
approach (i.e., higher quality data from other sources were used;
in these cases, the hierarchy of preferences was not a factor in
the decision). EPA also added that the data integration approach
for releases and occupational exposure data is discussed in
Appendix C of the document titled Risk Evaluation for N-
Methylpyrrolidone (2-Pyrrolidinone, 1 Methyl-) (NMP),
Supplemental Information on Occupational Exposure
Assessment ( b). EPA will seek peer review of
its Systematic Review protocol, including the hierarchy of
approaches to exposure estimation.
53
PUBLIC COMMENTS:
There is concern that EPA may continue to
inadvertently exclude useful information from review
in future risk evaluations. This problem is compounded
by unclear review criteria that change due to the
iterative nature of data collection and screening.
The timeframe for development of the TSCA Scope documents
was very compressed and the first ten chemical substances were
not subject to prioritization, the process through which EPA
expects to collect and screen much of the relevant information
about chemical substances. As a result, EPA had limited ability
to develop a protocol upfront. For these reasons, the protocol
development was staged in phases while conducting the
Page 189 of 205
-------
While EPA has provided general inclusion criteria in
Appendix G of the NMP Problem Formulation, EPA
has not provided information on how it applies these
criteria to exclude relevant studies. EPA also states that
that application of review criteria is subject to change
with each risk evaluation.
assessment work (see Section 3.1 of the Application of
Systematic Review in TSCA Risk Evaluations for more
discussion of this step).
EPA is in the process of revising the data quality criteria for
future assessments based on feedback from the NASEM TSCA
Committee.
EPA published the literature search strategy and title/abstract
inclusion/exclusion criteria for NMP in the Strategy for
Conducting Literature Searches for NMP: Supplemental
ument to the TSCA Scope Document.
EPA will publish a protocol document for the next TSCA risk
evaluations. Furthermore, EPA has received feedback from the
NASEM TSCA Committee on its systematic review process and
will carefully review their recommendations for future
chemicals.
34, 48
PUBLIC COMMENTS:
The TSCA systematic review method is not evidence-
based, lacks transparency, is not peer reviewed, and is
likely to have resulted in a biased evidence base for the
risk evaluation. Inadequate methods were used to
assess risk of bias, including financial conflicts of
interest. The method relies on numerical scores that
falsely imply a relationship between scores and effect
or association.
EPA/OPPT's quality evaluation method was developed
following identification and review of various published
qualitative and quantitative scoring systems to inform our own
fit-for-purpose tool. The development process involved
reviewing various evaluation tools/frameworks (e.g., OHAT
Risk of Bias tool, CRED, etc.; see Appendix A of the
Application of Systematic Review in TSCA Risk Evaluations
document and references therein), as well as soliciting input
from scientists based on their expert knowledge about
evaluating various data/information sources specifically for risk
assessment.
EPA will publish a protocol document for the next TSCA risk
evaluations. Furthermore, EPA has received feedback from the
NASEM TSCA Committee on its systematic review process and
will carefully review their recommendations for future
chemicals.
48
PUBLIC COMMENTS:
EPA has revised its searching and screening procedures to
include all studies in the systematic review process (screening,
Page 190 of 205
-------
EPA states that it "leveraged information presented in
previous assessments when identifying relevant key
and supporting data." The supplemental documents do
not contain the phrasing "key and supporting
information." There has been, and continues to be, a
lack of clarity on how EPA chose and evaluated the
key sources. EPA should define what "key and
supporting" information is.
data evaluation) for the next set of TSCA chemical risk
evaluations. In other words, no key and supporting studies will
bypass any step in the systematic review process.
EPA defines key and supporting data in a footnote in Section
1.5.1, "Key and supporting data and information are those that
support key analyses, arguments, and/or conclusions in the risk
evaluation."
40, 34,
48,51,
53
PUBLIC COMMENTS:
Suggestions for improving the systematic review
process include:
o Detailed documentation and transparency of how
information was identified and evaluated,
o Follow best practices in the field to simplify the
data quality criteria and to synthesize and integrate
each evidence stream,
o Do not be overly stringent and exclude studies
based on a single criterion,
o Submit the process for review to the NAS.
o Develop a protocol prior to commencing the
systematic review,
o Consider using an existing peer-reviewed method
such as the that used by the National Toxicology
Program's Center for the Evaluation of Risks to
Human Reproduction, the Institute of Medicine, or
EPA's IRIS program.
EPA will publish a protocol document for the next TSCA risk
evaluations. To refine that protocol, EPA is reviewing existing
peer-reviewed systematic review approaches, consulting with
EPA's IRIS program, and considering feedback from the
NASEM TSCA Committee.
Systematic review - Evidence integration
48
PUBLIC COMMENTS:
It is recommended that EPA conduct separate evidence
synthesis and determinations about the certainty of the
evidence for each stream of evidence and describe how
different streams of evidence are integrated to draw
conclusions.
When synthesizing and integrating evidence for each human
health hazard endpoint, EPA considered quality, consistency,
relevancy, coherence and biological plausibility as specified in
Application of Systematic Review in TSCA Risk Evaluations.
For NMP, EPA considered each of these factors qualitatively in
characterizing the weight of the scientific evidence and overall
Page 191 of 205
-------
confidence in selected PODs. The weight of the scientific
evidence for hazard endpoints is described Section 3.1.3. The
rationale for selection of specific PODs is described in Section
3.2.5.6 and the strengths and weakness and overall confidence
in each POD are described qualitatively in Section 3.2.6.
Transparency of citations - General handling of CBI data
59, 54
PUBLIC COMMENTS:
Under TSCA Section 14(b)(2)(A), the law's
restrictions on disclosure of CBI do not apply to "any
health and safety study which is submitted under this
Act" for a chemical substance that "has been offered
for commercial distribution." Therefore, any
information reported to or obtained by EPA from a
health and safety study must also be disclosed.
The regulations are explicit that tests to determine the
chemical and physical properties and fate and transport
behavior of a substance fall within the definition, along
with studies of a chemical's human health effects and
ecotoxicity and assessments of human and
environmental exposure, 40 CFR ง 720.3(k).
EPA should exercise its discretion to protect CBI in
Health and Safety studies authorized under TSCA by
appropriately balancing the competing interests of
transparency and protection of compensability.
Information that would qualify for protection includes
the submitter's identity, the identities of employees
who worked on the study, and confidential commercial
information in the study (e.g., financial statistics,
product codes, information that discloses processes
used in the manufacture or processing of the chemical,
or the portion of a chemical in a mixture); information
needed for regulatory acceptance where
The key and supporting studies EPA relied on as the basis for
quantitative analysis in the final risk evaluation are publicly
available. At the time of the draft risk evaluation, several studies
were not available to EPA. These studies were therefore cited
but not used as the basis for quantitative analysis. EPA has since
received two two-generation reproduction studies from NMP
Producers Group. EPA posted these studies to the public docket,
evaluated study quality using the systematic data quality
criteria, and incorporated relevant information from the studies
into the weight of the scientific evidence considered in hazard
characterization and dose-response analysis.
Where possible, EPA revised citations to refer to primary
sources that are publicly available rather than relying on
secondary sources. EPA's conclusions are based on data
reasonably available from primary sources. In some places,
EPA notes the existence of additional studies cited and
summarized in dossiers, but these are only included to provide a
complete picture of reasonably available information, not to
serve as the basis for EPA decisions.
Page 192 of 205
-------
compensability (commercial value) of the study may
be an issue.
54
PUBLIC COMMENTS:
Section 26 clearly contemplates that not all information
relevant to a Section 6 risk evaluation will be publicly
available. "Subject to Section 14" means that Congress
contemplated that CBI contained in information relied
upon to support a scientific decision by the agency may
not be subject to disclosure. Section 26(j)(4) requires
that EPA identify the "list of studies considered" in a
risk evaluation along with their results, but not that the
entire studies themselves would necessarily be made
public. The practical effect of EPA's decision not to
use the two NMP producers' more recent studies was
to forego its obligations under Section 26 to use the
best available science and apply the weight of the
scientific evidence.
EPA thanks the commenter for their input. EPA has added
additional studies to docket # EPA-HO-OPPT-2019-0236
available at www.reeulations.eov. Releasing these studies
ensures EPA's risk evaluation process is transparent, robust, and
uses the best available science. EPA received these studies after
publishing the draft NMP risk evaluation.
59
PUBLIC COMMENTS:
EPA appears to have appropriately rejected relying
on the missing sources submitted by the NMP
Producer's Group because the agency correctly
determined that the health and safety studies could
not be claimed CBI, as requested by the submitter,
and indicated that it could not rely on these studies
if it did not make them public.
For 10 sources from the NMP Producer's Group,
EPA has not described any claim(s) of
confidentiality that it believes justifies withholding
them. These information sources are all "health and
safety studies" that cannot receive CBI protection
under TSCA.
EPA's obligation to disclose these references
cannot be satisfied merely by releasing "robust
EPA appreciates the comment. EPA remains committed to a
transparent and reproducible systematic review process to ensure
that the information the Agency relies on in its risk evaluations
meets the scientific requirements in TSCA Section 26. EPA has
emphasized that in order to evaluate the quality of a study, the
Agency needs access to the complete study methodology and a
complete set of data tables and summary statistics for all
endpoints. Without this information, EPA does not have a basis
to judge the quality of a study through our TSCA systematic
review process or to assess the conclusions by applying a weight
of the scientific evidence approach. Since the release of the draft
risk evaluation, EPA received the NMP Producers' Group
studies and has added the studies to docket # EPA-HO-OPPT-
2019-0236 available at httos://www.reeulations.eov/. Releasing
these studies ensures EPA's risk evaluation process is
transparent, robust, and uses the best available science. EPA
Page 193 of 205
-------
summaries" but requires public access to the full
studies.
received these studies after publishing the draft NMP risk
evaluation in November 2019. These studies from the NMP
Producers Group provide the agency with additional information
on developmental and reproductive toxicity.
54
PUBLIC COMMENTS:
Although EPA claims the NMP Producers Group
proposed options that would make the CBI publicly
available in a restricted manner (e.g., in a reading room
not subject to mechanical reproduction), EPA rejected
them for reasons that are not persuasive.
EPA states that deviating from the practice of making
information available in the dockets "may create a
vulnerability for the Agency now, and in the future, in
its implementation of TSCA section 6."
EPA's protestations are unpersuasive in light of the
well-established importance CBI has for U.S.
companies and the U.S. economy. Precedence exists
for EPA's use of restricted-access reading rooms in
other comparable circumstances.
For example, amendments to the RMP statutory
authority in CAA ง 112(r) allow public access, through
reading rooms, to paper copies of offsite consequence
analysis information.
Possible solutions for handling CBI: If EPA posts
unredacted studies on its website, study owners may
lose the compensation value for use of those studies in
other jurisdictions to which they would otherwise be
entitled. This can result in substantial harm to their
competitive positions. The potential loss of
compensability also leads to reluctance to submit
studies voluntarily, such as in this instance with the
NMP draft risk evaluation. Options that EPA can
employ to address this issue include:
EPA seeks to establish a transparent and reproducible systematic
review process to ensure that the information the Agency relies
on in its risk evaluations meets the scientific requirement in
TSCA Section 26. In order to evaluate a study's quality, EPA
needs access to the complete study methodology and a complete
set of data tables and summary statistics for all endpoints.
Without this information, EPA does not have a basis to judge the
quality of a study through our TSCA systematic review process
or to assess the conclusions by applying a weight of the
scientific evidence approach. Since the release of the draft risk
evaluation, EPA received the NMP Producers' Group studies
and has added the studies to docket #EPA-HO-OPPT-2019-0236
available at www.reeulations.eov.
Page 194 of 205
-------
o Confidentiality agreements: Any stakeholder
willing to sign a confidentiality agreement would
receive a full copy of the study, including any CBI
redacted from the public version of the study,
o Public reading rooms: Unredacted paper copies of
studies could be made available in EPA public
reading rooms, as long as they could not be
mechanically duplicated,
o Voluntary submissions: EPA has a strong interest
in voluntary submission of studies on Section 6
chemicals, even if there is not complete public
disclosure of the studies. Whether or not companies
that own studies choose to make them available to
EPA depends in part on whether or not EPA will
make the studies public in a manner that results in
loss of compensability.
Where EPA lacks statutory authority to require
submission of studies on Section 6 chemicals, as with
European companies preparing REACH dossiers, EPA
has no alternative but to accept studies with redactions
designed to preserve compensability. Even where EPA
has statutory authority to require submission of studies
on Section 6 chemicals, EPA should prefer to receive
the studies without having to exercise that statutory
authority.
Transparency of citations -NMP Producers Group 1999 studies
33, 54
PUBLIC COMMENTS:
The NMP Producers Group sponsored two
reproductive toxicity studies to help clarify the findings
from Exxon (1991), the key study for derivation of
PODs for reproductive effects.
The NMP Producers Group was agreeable to providing
full study reports to EPA but requested that they not be
At the time the draft risk evaluation was released, EPA did not
have access to unredacted versions of either of the NMP
Producers Group 1999 studies. NMP Producers Group was
initially hesitant to share the studies if they were to be made
public. NMP Producers Group has since provided unredacted
versions of both studies to the agency. EPA has reviewed the
studies using the systematic review data quality criteria,
Page 195 of 205
-------
publicly posted and that they be protected as CBI. EPA
did not respond to the Group's request and rejected this
proposal as insufficient for reasons that were not
clearly explained.
EPA did not explain in the draft risk evaluation why it
does not have complete access to the full reports. Is it
possible that EPA made an unnecessarily restrictive
interpretation of TSCA Section 14?
integrated results into hazard identification, weight of the
scientific evidence, and dose-response analysis, and posted the
studies to the public docket. EPA generally expects to make the
information it uses for decision-making publicly available,
consistent with and subject to the requirements of TSCA
Section 14.
General feedback on TSCA risk evaluations for existing chemicals
54
PUBLIC COMMENTS:
EPA should convene a broader discussion with other
EPA program offices about how OPPT should
coordinate with the other EPA program offices on how
OPPT should address substances already specifically
regulated by these other offices, as well as substances
that EPA addresses through more general regulatory
requirements.
EPA should articulate the principles and approaches
that will form the foundation of EPA's intra-agency
coordination efforts and provide the public the
opportunity to comment.
EPA communicated with other program offices within the
agency throughout the risk evaluation process, including at
scoping, problem formulation, and both draft and final risk
evaluation. These discussions included regulatory requirements
and processes of the various environmental statutes. EPA will
continue to have these conversations with other offices at the
Agency for the next round of chemicals to be evaluated under
TSCA Section 6. See Section 1.4.2 of the risk evaluation
regarding EPA's approach to exposure pathways and risks
addressed by other EPA-administered statutes.
In the 2017 Procedures for Chemical Risk Evaluation Under the
Amended Toxic Substances Control Act (82 FR 33726, July 20,
2017), EPA committed to, by codifying, interagency
collaboration to give the public confidence that EPA will work
with other agencies to gain appropriate information on chemical
substances. This is an ongoing deliberative process and EPA is
not obligated to provide descriptions of predecisional and
deliberative discussions or consultations with other federal
agencies. In the interest of continuing to have open and candid
discussions with our interagency partners, EPA is not intending
to include the content of those discussions in the risk evaluation.
34, 53,
31, 32
PUBLIC COMMENTS:
SACC meetings should not be scheduled before the
close of the comment period on the draft risk
EPA appreciates the comment and will consider whether a
longer comment period is warranted for future draft risk
evaluations. EPA did extend the public comment period for the
Page 196 of 205
-------
evaluations. This limits the information available to the
SACC, depriving the Committee of the benefit of
comments that could have enabled a more focused and
informed review.
The current tight public comment deadlines
compromise stakeholder's ability to comment. A
thorough review of the draft risk evaluation was not
possible.
The final deadline for submission of public comments
should be extended in the future by an additional 30
days - to 90 days - at least through the next round of
20 high priority chemicals.
draft risk evaluation of NMP by two additional weeks to give
stakeholders more time to review and comment on the draft
document.
SACC
SACC COMMENTS:
It was suggested that representatives from OSHA and
NIOSH attend these evaluation reviews to help answer
questions that the Committee continues to have on
issues related to occupational safety and health.
OSHA and NIOSH were able to comment on this document
during interagency review. EPA will consider adding
representatives from OSHA or NIOSH attend future peer review
meetings.
SACC
SACC COMMENTS:
Recommendation: Consider including GHS
classification information on the subject chemical.
The GHS classification is the primary mechanism for
communicating hazards of a chemical in an industrial
setting through SDSs and labeling. The GHS provides
a way to compare relative hazards across substances
and could provide useful context to readers.
EPA appreciates this suggestion and will consider this approach
for future risk evaluations.
38
PUBLIC COMMENTS:
An important Expert Consensus Statement published in
Nature Reviews Endocrinology earlier this year
identifies 10 "key characteristics of endocrine-
disrupting chemicals as a basis for hazard
identification." The EPA TSCA program could gain
much benefit from incorporating these current
Thank you for recommending this resource.
Page 197 of 205
-------
scientific approaches into its chemical risk evaluations,
including NMP.
38
PUBLIC COMMENTS:
EPA should consider utilizing mechanistic data when
evaluating chemicals and should incorporate these
current scientific approaches into its chemical risk
evaluations.
In response to this comment, EPA has further explored
mechanistic evidence for NMP. While there is very limited
mechanistic data available for NMP, EPA has identified some
evidence that NMP is a bromodomain inhibitor. This suggests a
plausible mechanism for male reproductive effects. Additional
discussion of this mechanistic evidence has been added to
Section 3.2.4.2. The mechanistic studies supporting this
discussion were evaluated using the systematic review data
quality criteria. EPA is in the process of modifying the
systematic review approach to incorporate mechanistic data
earlier in the risk evaluation process. Future risk evaluations
may be supported by available mechanistic data identified
through systematic review.
39
PUBLIC COMMENTS:
There are some fundamental flaws in the Simon et al.
(2016) implementation of Bayesian/probabilistic
methods.
If adopting a Bayesian approach, it is recommended
that EPA adopt the WHO/IPCS (2017b) framework in
its probabilistic analysis. Numerous tools are available
for implementing it, including an Excel spreadsheet
tool APROBA available on the WHO website
(WHO/IPCS, 2017a), an RShiny web app
APROBAweb (Chiu, 2018), and as part of the
Bayesian Benchmark Dose online web system,
benchmarkdose.org (Shao and Shapiro, 2018).
EPA must incorporate all reasonably available information,
defined in 40 CFR 702.33 as "information that EPA possesses
or can reasonably generate, obtain, and synthesize for use in risk
evaluations, considering the deadlines specified in TSCA
Section 6(b)(4)(G) for completing such evaluation." Due to
time and resource constraints associated with the deadline for
completing the NMP risk evaluation, EPA cannot implement a
Bayesian framework comprehensively for this risk evaluation as
the information is not reasonably available; however, EPA will
consider incorporating more probabilistic modeling into future
risk evaluations under TSCA.
54
PUBLIC COMMENTS:
There are several risk assessment models that EPA
might use to communicate its TSCA risk
characterizations, such as HESI's Risk21 Project and
Web Tool. EPA should consider this approach, or a
EPA will investigate the methods and principles behind the
HESI Risk 21 application and consider using its visualizations
in future risk evaluations.
Page 198 of 205
-------
similar approach that provides an effective visual
representation of the potential range of risks, to aid in
communication of the risk characterization for the
conditions of use.
SACC
SACC COMMENTS:
One Committee member suggested providing a section
to the document that clearly explains the public health
implications (not benefits) of the overall findings. This
could be a standard part of the executive summary or a
stand-alone statement. For example, the summary for
NMP could highlight that dermal exposure from direct
skin contact and from vapor-to-skin contributes the
largest fraction of dose compared to exposure via
inhalation.
In addition, the summary could review the extent of
exposures (geographically and via population numbers)
and provide estimated numbers of occupationally
exposed workers in some of the most affected
occupations.
EPA appreciates the comment and will consider including
additional information where feasible in future risk evaluations.
It should be noted that Table 2-4 "Estimated Number of
Workers in the Assessed Industry Uses of NMP" provides
estimates by occupational exposure scenario of the number of
workers potentially exposed in manufacturing, chemical
processing and other occupational scenarios.
53
PUBLIC COMMENTS:
At the SACC meeting, the SACC and EPA considered
stopping SACC's review of each draft evaluation so it
only convenes occasionally to discuss issues of
emerging science that could affect how EPA conducts
evaluations.
The SACC should continue to conduct review of each
draft risk evaluation at least through the next group of
20 TSCA risk evaluation chemicals.
EPA may need to further adjust its approach to risk
evaluations to consider aggregated effects and legacy
uses.
EPA appreciates these comments and will consider this input for
future risk evaluations. Regarding aggregate exposures, 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. EPA defines aggregate
exposures 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). 40 CFR 702.33. EPA considers the
reasonably available information and used the best available
science to determine whether to consider aggregate or sentinel
exposures for a particular chemical.
Page 199 of 205
-------
The SACC's continued input on these and other
matters, through review of each draft risk evaluation,
would assist EPA in establishing a consistent approach.
Page 200 of 205
-------
References
Akesson. M t .irnerup. \ I \, Jons son. BA. (2004). Evaluation of exposure biomarkers from
percutaneous absorption of N-methyl-2-pyrrolidone. Scand J Work Environ Health 30:
306-312.
Anundi. H; Langworth. S; Joh an son. G; Lind. ML; Akessoi lis. L; Itkes. N: Soderman. E;
Jonsson. BA; Edling. C. (2000). Air and biological monitoring of solvent exposure during
graffiti removal. Int Arch Occup Environ Health 73: 561-569.
http://dx.doi.ors .004200000157,
Babich. MA; Chen. SB; Greene. MA; Kiss. CT; Porter. WK; Smith. TP; Wind. ML; Zamula.
WW. (2004). Risk assessment of oral exposure to diisononyl phthalate from children's
products. Regul Toxicol Pharmacol 40: 151-167.
http://dx.doi.off .vrtph.2004.06.005.
Bader. M; Van Thriel. C. (2006). Human volunteer study on biomarkers of N-Methyl-2-
Pyrrolidone (NMP) after inhalation exposure. Hannover and Dortmund, Germany:
Hannover Medical School and University of Dortmund, Institute of Occupational
Physiology.
Bader. M; Wrbitzks aszkewicz. M; Schaper. M; van Thriel. C. (2008). Human volunteer
study on the inhalational and dermal absorption of N-methyl-2-pyrrolidone (NMP) from
the vapour phase. Arch Toxicol 82: 13-20. http://dx.doi.ors 0204-007-0230-5.
Bartsch. W; Sponer stmann. K; Fuchs. G (1976). Acute toxicity of various solvents in
mouse and rat - LD50 of ethanol, diethylacetamide, dimethylformamide,
dimethylsulfoxide, glycerine, N-methylpyrrolidone, poly ethylene-glycol 400, 1,2-
propanediol and Tween 20. Arzneimittelforschung 26: 1581-1583.
BASF. (19831 Unpublished data, study No. 83/1 12, 31 Aug 1983. BASF AG.
BASF. (1986). Department of Toxicology, study no. 85/289, 05 Feb 1986 (unpublished).
BASF. (1989). Department of Ecology, unpublished data, project No. 1035/88.
https://hpvchemicals.oecd.org/ui/SponsoredChemicals.aspx.
BASF. (2001). Department of Experimental Toxicology and Ecology, unpublished data, project
No. 00/0969/51/1. https://hpvchemicals.oecd.org/ui/SponsoredChemicals.aspx.
Beyer. BK; Chemoff. N; Danielsson. BR; Davis-Bruno. K; Harrouk. W; Hood. RD; Janet u,
Liminga. UW; Kim. JH; Rocca. M; Rogers. J; Scialli. AR. (2011). ILSI/HESI maternal
toxicity workshop summary: maternal toxicity and its impact on study design and data
interpretation. Birth Defects Res B Dev Reprod Toxicol 92: 36-51.
http://dx.doi.ors 32/bdrb .20281.
Brown. RP; Del idstedt. SL; Rhomberg. LR; Beliles. RP. (1997). Physiological
parameter values for physiologically based pharmacokinetic models. Toxicol Ind Health
13: 407-484. http://dx.doi.. 74823 3 797013 00401.
Di Toro. DM. (1984). Probability model of stream quality due to runoff. J Environ Eng 1 10: 607-
628. http://dx.doi.Org/10.1061/fASCE10733-93 72(19841110:3(6071
DTI (2004). Survey of chemical substance in consumer products. (42).
E. 1. Dupont De Nemours & Co. (1990). Initial submission: reproductive and developmental
toxicity of l-methyl-2-pyrrolidinone in the rat with cover letter dated 10/01/92. (OTS:
Page 201 of 205
-------
OTS0555618; 8EHQNum: 8EHQ-1092-11957; DCN: 88-920010214; TSCATS RefID:
440618; CIS: NA).
EaglePicher Technologies. (2020a). Comments of EaglePicher Technologies, LLC on the draft
TSCA risk evaluation of n-methylpyrrolidone (NMP): EPAHQOPPT-2019-0236.
(EPA-HQ-OPPT-2019-0236-0049). Washington, DC: U.S. Environmental Protection
Agency. https://beta.reeiilations.eov/dociiment/EPA-HQ-OPPT-2019-0236-0049.
EaglePicher Technoloeies. (2020b). Conference call with EaglePicher Technologies on n-
methylpyrrolidone. (EPA-HQ-OPPT-2016-0743-0113). Washington, DC: U.S.
Environmental Protection Agency, https://beta.reeulations.gov/document/EPA-HQ-
QPPT-2016-0743-0113.
Exxon Biomedical. (1991). Multigeneration Rat Reproduction Study with N-Methylpyrrolidone,
Project Number 236535 [TSCA Submission], (OTS#: 0532510; New Doc ID: 40-
91107125; Old Doc ID: 42114F1-2). Wayne, USA: GAF Corp.
Flick Isness. CE; Jackh. R; Buesen. R; Klue. S. (2009). Embryotoxic potential of N-
methyl-pyrrolidone (NMP) and three of its metabolites using the rat whole embryo
culture system. Toxicol Appl Pharmacol 237: 154-167.
http://dx.doi.ore/10.1016/j .taap.2009.02.024.
Fox. JR; Hoean. KA; Davis. A. (2016). Dose-response modeling with summary data from
developmental toxicity studies. Risk Anal 37: 905-917.
http://dx.doi.on isa. 12667.
Fuiifilm Holdings America Corporation. (2020). FUJIFILM comments for docket ID # EPA-
HQ-OPPT-2019-0236 for CASRN 872-50-4, n-methylpyrrolidone (NMP). (EPA-HQ-
OPPT-2019-0236-0050). Washington, DC: U.S. Environmental Protection Agency.
https://beta.reeiilations.eov/dociiment/EPA-HQ-OPPT-2019-0236-005Q.
Gentry. PR; Covington. TR; Andersen. ME; Clewel (2002). Application of a
physiologically based pharmacokinetic model for isopropanol in the derivation of a
reference dose and reference concentration [Review], Regul Toxicol Pharmacol 36: 51-
68. http://dx.doi.ore/10.1006/rtph.2002.1540.
Gerike. P; Fischer. WK. (1979). A correlation study of biodegradability determinations with
various chemicals in various tests. Ecotoxicol Environ Saf 3: 159-173.
http: //dx. doi. or e/http s: //dot. or y 10 101 Oil I m '190009-5.
Hass. II; Jakobsen. BM; Lund. SP. (1995). Developmental toxicity of inhaled N-
methylpyrrolidone in the rat. Pharmacol Toxicol 76: 406-409.
Hass. nd. SP; Eisner. J. (1994). Effects of prenatal exposure to N-methylpyrrolidone on
postnatal development and behavior in rats. Neurotoxicol Teratol 16: 241-249.
http://dx.doi.ore/10 J 016/0892-0362(94)90045-0.
Haufroid \ . .Kieeer. VK; Jeggli. S; Eiseneeeer. R; Berrunl \ U < >son. D; Hotz. P.
(2014). Biological monitoring and health effects of low-level exposure to N-methyl-2-
pyrrolidone: A cross-sectional study. Int Arch Occup Environ Health 87: 663-674.
http://dx.doi.on 10 100 s00420-013-09^
Health Safety Laboratory. (2007). Protective glove selection for workers using NMP containing
products -Graffiti removal. (HSL/2007/41). United Kingdom: Health and Safety
Laboratory, http://www.hse.eov.uk/research/hsl pdf/2007/hsl( if.
Hines. RN. (2007). Ontogeny of human hepatic cytochromes P450 [Review], J Biochem Mol
Toxicol 21: 169-175. http://dx.doi.ori 02/ibt.20179.
Page 202 of 205
-------
Huntingdon Life Sciences. (1998). [14c]-N-Methylpyrrolidone: Topical application: Dermal
absorption study in the rat. (982974). Huntingdon Life Sciences, Ltd.
Intel Corporation. (2019). Intel comments to: Science Advisory Committee on Chemicals
(SACC) on the draft Toxic Substances Control Act (TSCA) risk evalulation for N-
Methylpyrrolidone (NMP). (EPA-HQ-OPPT-2019-0236-0037). U.S. Environmental
Protection Agency. https://beta.regulations.gov/document/EPA-HQ-OPPT-2019-Q236-
0037.
Intel Corporation. (2020). Comments of Intel to the United States Environmental Protection
Agency on the draft Toxic Substances Control Act (TSCA) risk evaluation for n-
methylpyrrolidone (NMP). (EPA-HQ-OPPT-2019-0236-0064). U.S. Environmental
Protection Agency. https://beta.regulations.gov/document/EPA-HQ-OPPT-2019-0236-
0064.
John said. EK; Koukourital > 1 J! 1 h akaran. i rM unengraber. LL; Mines. RN: McCarvci 1 H i
(2003). Human hepatic CYP2E1 expression during development. J Pharmacol Exp Ther
307: 402-407. http://dx.doi.on 10 M : S jpet iO: 0\'< I: I.
Kftzek. K; Ruzicka. J; Julinova. M; Husarova. L; Houser. J; Dvorackova. M; Jancova. P. (2015).
N-methyl-2-pyrrolidone-degrading bacteria from activated sludge. Water Sci Technol 71:
776-782. http://dx.doi.Org/l0.2166/wst.2
Lee. KP; Chromev. NC: Culik. R; Barnes. J.R; Schneider. PW. (1987). Toxicity of N-methyl-2-
pyrrolidone (NMP): Teratogenic, subchronic, and two-year inhalation studies. Fundam
Appl Toxicol 9: 222-235. http://dx.doi.oi '0272-0590(87190045-5.
LICM. (2020a). Comment on docket no. EPA-HQ-OPPT-2019-0236, Toxic Substances Control
Act (TSCA) draft risk evaluation for n-methylpyrrolidone (NMP). (EPA-HQ-OPPT-
2019-0236-0056). Washington, DC: U.S. Environmental Protection Agency.
https://beta.regiilations.gov/dociiment/EPA-HQ-QPPT-2019-0236-0056.
LICM. (2020b). Conference call with Lithium Ion Cell Manufacturers" Coalition (LICM) on n-
methylpyrrolidone. (EPA-HQ-OPPT-2016-0743-0114). Washington, DC: U.S.
Environmental Protection Agency, https://beta.regulations.gov/document/EPA-HQ-
QPPT-2016-0743-0114.
LICM. (2020c). Supplemental NMP information to document PPE protection for submission to
docket from Lithium Ion Cell Manufacturers. (EPA-HQ-OPPT-2016-0743-0116).
Washington, DC: U.S. Environmental Protection Agency.
https://beta.regiilations.gov/dociiment/EPA-HQ-QP
Li got Li 1 * 1 r on. D; Haufroid. V. (2003). Contribution of CYP2E1 to N-methyl-2-pyrrolidone
metabolism. Arch Toxicol 77: 261-266. http://dx.doi.org/10.1007/s00204-003-044Q-4.
Mallev. LA; Kennedy. GL; Elliott. GS; Slone. TW; Mellert. W; Deckardt. K; Kuttler. K;
Hildebrg iton. MI; Parod. RJ; Griffiths. JC. (2001). Chronic toxicity and
oncogenicity of N-methylpyrrolidone (NMP) in rats and mice by dietary administration.
Drug Chem Toxicol 24: 315-338. http://dx.doi.org/10.1081/D( >262.
NIQSH. (1993). Health Hazard Evaluation report no. HETA-93-0844-241 1, Rosebud Company,
Atlanta, Georgia (pp. 93-0844). (HETA-93-0844-2411). Washington, DC: U.S.
Department of Health and Human Services.
Nishimura. S: Yasui. H; Miyauchi. H; Kikuchi. Y; Kon ebavashi. T: Tanaka. S:
Mikoshiba. Y: Omae. K; Nomiyama. T. (2009). A cross-sectional observation of effect of
exposure to N-methyl-2-pyrrolidone (NMP) on workers' health. Ind Health 47: 355-362.
Page 203 of 205
-------
K cers Group. (1999a). Two generation reproduction toxicity study with n-
methylpyrrolidone (NMP) in Sprague Dawley rats: Administration in the diet. (Project
No. 97-4106). Millestone, NJ: Huntingdon Life Science.
K cers Group. (1999b). Two Generation Reproduction Toxicity Study with N-
Mythylpyrrolidone (NMP) in Wistar Rats - Administration in the Diet. (Project No.
70R0056/97008). Ludwigshafen, Germany: Department of Toxicology of BASF
Aktiengesellschaft.
OECD. (2007). S1DS initial assessment report on 1 -methyl-2-pyrrolidone. Organization for
Economic Cooperation and Development.
http://webnet.oecd. org/hpv/ui/handler.axd?id=84daa4ac-feb7-4b5a-9839-20(: 2.
OSHA. (2017). Chemical Exposure Health Data (CEHD) provided by OSHA to EPA. U.S.
Occupational Safety and Health Administration.
Pavan. IP; Boudry. I; Bevdon. >ry. JP; Gran del aude. MC: Ferrari. E; Andre. JC. (2003).
Toxicokinetics and metabolism of N-[(14)C]N-methyl-2-pyrrolidone in male Sprague-
Dawley rats: in vivo and in vitro percutaneous absorption. Drug Metab Dispos 31: 659-
669. http://dx.doi.o id.31.5.659.
Poet. TS: Kirnu ier. M: van Thriel. C: Gargas. ML; Hinderliter. PM. (2010).
Quantitative risk analysis for N-methyl pyrrolidone using physiologically based
pharmacokinetic and benchmark dose modeling. Toxicol Sci 113: 468-482.
http://dx.doi.oo 93/toxsci/kfp264.
Prestige. (2010). RaLex oven cleaner material safety data sheet. http://ralex-
uk.com/?product=oven-cleaner.
RIVM. (2013). Annex XV Restriction Report: Proposal for a Restriction. (Version 2). The
Netherlands: National Institute for Public Health and the Environment (RIVM).
http s: //ech a. europa. eu/docum ents, 'nmp annex xv report en.pdf.
Saillenfait. AM; Gallissot. F: Langonne. I: Sabal (2002). Developmental toxicity of N-
methyl-2-pyrrolidone administered orally to rats. Food Chem Toxicol 40: 1705-1712.
http://dx.doi.on 30278-6915(02100115-1.
Saillenfaป \ \ I illissot. F; Morel G. (2003). Developmental toxicity of N-methyl-2-
pyrrolidone in rats following inhalation exposure. Food Chem Toxicol 41: 583-588.
http://dx.doi.org/ i 0. i 0 i 6/S0278-6915(02)00300-9.
Schmidt. R. (1976). Tierexperimentelle Untersuchungen zur embryotoxischen und teratogenen
Wirkung von N-MethylPyrrolidon (NMP). Biol Rundsch 14: 38-41.
Semiconductor Industry Association. (2019a). Comments of the Semiconductor Industry
Association (SIA) to the Science Advisory Committee on Chemicals (SACC) on the draft
Toxic Substances Control Act (TSCA) risk evaluation for n-methylpyrrolidone (NMP).
(EPA-HQ-OPPT-2019-0236-0031). Washington, DC: U.S. Environmental Protection
Agency, https://www.regutations.gov/document?D=EPA-I PT-2016-0743-0019.
Semiconductor Industry Association. (2019b). NMP Supplemental Data: Container Handling.
Washington, DC: Semiconductor Industry Association (SIA).
Semiconductor Industry Association. (2019c). SIA N-Methylpyrrolidone Risk Management
Measures and Worker Exposure Monitoring Results. Washington, DC: Semiconductor
Industry Association (SIA).
Semiconductor Industry Association. (2020). Comments of the Semiconductor Industry
Association (SIA) on the draft Toxic Substances Control Act (TSCA) risk evaluation for
n-methylpyrrolidone (NMP). (EPA-HQ-OPPT-2019-0236-0052). Washington, DC: U.S.
Page 204 of 205
-------
Environmental Protection Agency, https://beta.reeulations.gov/documemt/EPA- HQ-
QPPT-2019-0236-0052.
Shaver. TN. (1984). FATE OF ETHEPHON AND N-METHYL-2-PYRROL1 DONE IN SOIL
AND COTTON PLANTS. Arch Environ Contam Toxicol 13: 335-340.
http://dx.doi.on !4.
Sitarek. K; Stetkiewicz. J. (2008). Assessment of reproductive toxicity and gonadotoxic potential
of N-methyl-2-pyrrolidone in male rats. Int J Occup Med Environ Health 21: 73-80.
http://dx.doi.ore/10.2478/vl0001-008-0006-z.
Sitarek. K; Stetkiewicz. J; Wasowicz. W. (2012). Evaluation of reproductive disorders in female
rats exposed to N-methyl-2-pyrrolidone. Birth Defects Res B Dev Reprod Toxicol 95:
195-201. http://dx.doij irb.21001.
Solomon. GM; Morse. EP; Garbo. Ml; Milton. DK. (1996). Stillbirth after occupational exposure
to N-methyl-2-pyrrolidone: A case report and review of the literature [Review], J Occup
Environ Med 38: 705-713. http://dx.doi.cnv 10 i0^ 000 I '< I r<\ 0 000 0001 I.
Solomon. HM; Burses; nedv. GL. Jr.; Staples. RE. (1995). l-methyl-2-pvrrolidone
(NMP): Reproductive and developmental toxicity study by inhalation in the rat. Drug
Chem Toxicol 18: 271-293. http://dx.doi.org/10.3109/0148Q5495t
(1991). Guidelines for developmental toxicity risk assessment. Fed Reg 56: 63798-
63826.
(1994). Guidelines for Statistical Analysis of Occupational Exposure Data: Final.
United States Environmental Protection Agency :: U.S. EPA.
(2012). Estimation Programs Interface Suite for Microsoftฎ Windows, v 4.11.
Washington, DC. https://www.epa.gov/tsca-screening-tools/epi-suitetm-estimation-
program-interface.
(2018). Problem formulation of the risk evaluation for n-methylpyrrolidone (2-
pyrrolidinone, 1-methyl-). (EPA-740-R1-7015). Washington, DC: Office of Chemical
Safety and Pollution Prevention, United States Environmental Protection Agency.
https://www.epa.gov/sites/prodiiction/files/2018-06/dociiments/nmp -
(2019a). Draft Risk Evaluation for N-Methylpyrrolidone, Supplemental Information
on Human Health Benchmark Dose Modeling. (Docket EPA-HQ-OPPT-2019-0236).
(2019b). Risk evaluation for n-methylpyrrolidone (2-pyrrolidinone, 1 methyl-)
(NMP), supplemental information on occupational exposure assessment (Draft).
van Raaii. MTM; Janse iersma. AH. (2003). The relevance of developmental toxicity
endpoints for acute limit setting. (601900004). Bilthoven: RIVM.
Wells. DA; Digenis. GA. (1988). Disposition and metabolism of double-labeled [3H and 14C]
N-methyl-2-pyrrolidinone in the rat. Drug Metab Dispos 16: 243-249.
Page 205 of 205
------- |