g% pf^j| United States
Environmental Protection
mm Agency
Office of Water
4304T
EPA-820R25002
January 2025
EPA Response to External Peer Review Comments on the
Draft Sewage Sludge Risk Assessment for Perfluorooctanoic
Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic
Acid (PFOS) CASRN 1763-23-1
January 2025
U.S. Environmental Protection Agency
Office of Water, Office of Science and Technology
Health and Ecological Criteria Division 1200 Pennsylvania Avenue, NW Washington, DC 20460
-------
DISCLAIMER
This document has been reviewed in accordance with U.S. Environmental Protection Agency
policy and approved for publication. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
TABLE OF CONTENTS
I. INTRODUCTION 1
External peer review process 1
Charge to peer reviewers 2
II. EPA RESPONSE TO GENERAL IMPRESSIONS 4
III. EPA RESPONSE TO CHARGE QUESTION COMMENTS 13
QUESTION 1: Programmatic, statutory, and regulatory context 14
QUESTION 2: Problem Formulation 20
QUESTION 3: Ecological Risk Assessment 26
QUESTION 4: Pathway evaluation, model selection, model parameterization, risk
estimation and risk discussion 32
Specific editorial and technical comments 62
l
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
I. INTRODUCTION
Versar Global Solutions, an independent contractor for the United States Environmental
Protection Agency (EPA), coordinated an external letter peer review of the Draft Sewage Sludge
Risk Assessment for Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 AND Perfluorooctane
Sulfonic Acid (PFOS) CASRN 1763-23-1 report. The peer review was conducted for the EPA's
Office of Water, Office of Science and Technology, Health and Ecological Criteria Division in
August of 2024.
Assessing the potential risk of pollutants found in biosolids is a priority of the EPA's Biosolids
Program. The EPA identifies pollutants found in sewage sludge through open literature reviews
and sewage sludge surveys to assess their potential risk to public health and the environment.
Sewage sludge that has been treated in accordance with 40 CFR part 503 for land application is
often called biosolids by the EPA (although others treat the terms sewage sludge and biosolids as
synonyms).
The EPA's PFAS Strategic Roadmap includes conducting a biosolids risk assessment for two
PFAS compounds, perfluorooctanoic acid (PFOA) and perfluorooctane sulfonic acid (PFOS).
The current version of this assessment is a draft. After the risk assessment process is complete,
the EPA will consider risk management options for PFOA and PFOS in biosolids, if appropriate.
External peer review process
Versar conducted an independent search for scientific experts with one or more expertise in
evaluating: fate, transport, exposure and risk from PFOA/PFOS in terrestrial environments (e.g.,
agricultural sites or superfund sites); modeling groundwater transport and aquifer contamination
resulting from soil contamination; uptake of chemicals from soil into plants; uptake of chemicals
by livestock.
As a result of this search, Versar identified and contacted 39 experts. Of these experts, Versar
received seven positive responses expressing interest and availability to participate. The
remaining 32 experts either had a conflict of interest, were not available during the peer review
timeframe, or did not respond to the invitation. For each interested and available peer reviewer,
Versar evaluated their qualifications and conducted conflict of interest (COI) screening to ensure
that the experts had no COI. Versar selected the following five scientific experts to serve as peer
reviewers:
Xindi Hu, ScD
Mathematica Policy Research
Ramon Lavado, PhD
Baylor University
Charles Newell, PhD, PE, BCEE
GSI Environmental
1
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
P. Barry Ryan, PhD
Emory University
Marc-Andre Verner, PhD
Montreal University
Charge to peer reviewers
The peer reviewers were asked to evaluate the scientific and technical merit of the draft
document and provide their responses to the following charge questions:
1. Programmatic, statutory, and regulatory context. The Clean Water Act1 and its
implementing regulations2 directs EPA to assess and manage risks associated with the
disposal or use of sewage sludge to protect public health and the environment. Please
comment on the extent to which the description of EPA's Biosolids Program, including
its statutory and regulatory authorities, provides the context and basis for the risk
assessment (see Section 1).
2. Problem formulation. Please comment on the characterizations of the environmental fate,
exposure, and toxicity of PFOA/PFOS in the problem formulation section of the risk
assessment (see Section 2, Appendix A).
3. Ecological risk assessment. As described in Section 2.6.3, available data indicate that risk
to aquatic and terrestrial life from land application of biosolids contaminated with PFOA
and/or PFOS is expected to be lower than the potential risk to human health from
same/similar biosolids applications. As a result, the quantitative risk assessment (risk
estimates, risk characterization) is scoped to focus solely on human health. Please
comment on the conceptual and technical basis for limiting the assessment of ecological
risk and focusing on human health for PFOA/PFOS.
4. Pathway evaluation, model selection, model parameterization, risk estimation and risk
discussion. The draft risk assessment shows that human health risks associated with
drinking water and diet are expected to be the greatest among modeled scenarios and
pathways. Please comment on the technical basis and clarity of the following elements in
supporting this interpretation. Be as detailed as possible about deficiencies and suggested
improvements:
a. Models and parameters selected for these scenarios and pathways, including
comment on the inclusion of relevant scenarios (e.g., pasture farm) and pathways
(e.g., groundwater, fish consumption) - see Sections 2.9.2, 2.9.3, Appendix B and
Appendix C.
b. Modeling of groundwater behavior for PFOA and PFOS. (see Section 2.9.2 and
Appendix C focused on groundwater modeling).
c. Risk estimation and discussion, including clarity of results (see Sections 3 and 4)
and description of variability, uncertainty, and sensitivity (see Section 5,
Appendix D).
1 33 U.S. Code § 1345 - Disposal or use of sewage sludge
2 40 CFR Part 503 - Standards for the use or disposal of sewage sludge
2
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
d. Comparison of modeled results to biosolids investigations conducted in Michigan
and Alabama (see Section 5.3).
3
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
II. EPA RESPONSE TO GENERAL IMPRESSIONS
4
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
Reviewer 1
COMMENT: This document describes the risk assessment for PFOA and PFOS in sewage
sludge. Overall, the document is clearly written, and the assumptions and methodological
choices are reasonable. The rationale for selecting humans as the most sensitive receptor is
spelled out, and the four use and disposal scenarios for biosolids cover a wide range of exposure
pathways. Central tendency models used a 1 ppb concentration in biosolids, a concentration that
is much lower than those reported in the literature, to estimate risks in a "low-dose" setting. At
this concentration, EPA found that risk estimates exceeded both cancer and non-cancer
acceptable risk levels. I believe the conclusions of this report are sound, but some moderate
revisions could potentially make the document clearer. One thing that stood out is the language
used to characterize the human health risks. Exceedances of acceptable/tolerable exposure levels
were described as "significant human risks". I believe they should be described as levels
exceeding what is considered as acceptable/tolerable, or risk levels that can't be considered as
negligible.
RESPONSE: Thank you for the comment that, overall, the document is clearly written with
sound methodological choices. The EPA agrees that "significant human health risks" is not well-
defined and has revised this text to "exceeds the acceptable risk level." The EPA also included
text to define the "acceptable risk level" in this context.
COMMENT: Also, it may be difficult for readers to interpret the results in the context of
widespread PFAS contamination. For example, consuming drinking water at the maximum
contaminant level (MCL) of 4 ppt for PFOA alone would lead to an excess cancer risk in the
order of 2 E-03, which is similar to the highest risk estimates for sewer sludge disposal. It would
be valuable to describe whether we expect risk in farm families to exceed risk in the general
population, and by approximately how much (an example using some of the reported biosolids
concentrations would be useful). A recent paper (https://pubmed.ncbi.nlm.nih.gov/38941944/)
could be used to illustrate how much higher exposures can be in farm families compared to the
general population.
RESPONSE: The EPA's goal is to use this draft risk assessment, once finalized, to determine if
the concentrations of PFOA and PFOS present in sewage sludge may adversely affect public
health and the environment when the sewage sludge is disposed of or reused. Future risk
mitigation actions (regulatory or otherwise) would consider the scale of risk reduction that is
possible through various risk reduction activities. The commentor's suggestion to contextualize
risks sourced from biosolids by comparing them to other sources of exposure to the general
population is currently outside the scope of this draft risk assessment. Additionally, though the
EPA is planning a National Sewage Sludge Survey3 that would include PFAS analysis of sewage
sludge across the U.S., these national-scale data are currently not available.
COMMENT: Another issue is that while modeled PFOA/PFOS concentrations in media and risk
estimates are presented, the calculated ADDs and LADDs that are used for risk calculations are
not in the report. I suggest including tables presenting the dose estimates.
3 https://www.federalregister.gov/documents/2024/10/10/2Q24-23474/agencY-information-collection-activities-
submission-to-the-office-of-management-and-budget-for
5
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
RESPONSE: The EPA agrees that ADDs and LADDs are useful to many readers and has added
tables with these values to Section 3.
COMMENT: In terms of format, there are some issues with figures and tables. All figures and
tables should be checked for numbering (most tables do not have a number), and abbreviations
should be spelled out under tables/figures.
RESPONSE: Thank you for pointing out this error. The EPA has made these corrections.
Reviewer 2
COMMENT: I found the document to be very well written. While not an easy read as it is a
technical document, it was organized in a logical manner with sufficient information presented in
each section such that the selection could be read alone by a reader interested in that specific
subject matter. At times, this resulted in redundancy, but such redundancy was more than offset
by the utility of the approach and clear organizational structure. The Table of Contents gives a
clear understanding of the structure of the document affording the reader the opportunity to
peruse different sections as desired or warranted by the need to understand different aspects of
the risk assessment.
The document was very well-referenced with the primary focus on previous EPA review
documents, where available, and primary literature when EPA reviews were not available. I
noted a few locations where references were not yet available. As this is still a Draft document,
these were noted in the text by highlighting with an indication that such would be filled in at an
appropriate time, e.g., when a publication had been accepted in the literature. However, it is not
clear how this would affect the final version if publication of the specific reference was delayed
or the submitted manuscript withdrawn.
Considering the science behind the written document, I found the approach to determining risk
associated with sewer sludge to be presented as a well-though out design in general and
specifically for the PFOA/PFOS pair of PFAS. The Executive Summary, with one or two
exceptions, puts forward the process used to perform the risk assessment, the results, and their
meaning in an abbreviated form. The body of the document is dense, comprehensive, and
thorough. It is a difficult read, but the complex nature of the problem at hand merits a complex
analysis that, in turn, requires a detailed presentation.
RESPONSE: Thank you for your comment. With regard to references to EPA documents
pending publication, these documents have since been published and the public comment draft
risk assessment includes the citations to the public versions.
COMMENT: The choice of scenarios deserves discussion here. I believe the scenarios tested are
appropriate for assessing risk from general deployment of sewer sludge containing PFOA and
PFOS. Further, as noted in the text, these scenarios serve to model similar scenarios with the
same level of contamination for PFAS in general and, perhaps, even other classes of
6
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
environmental contaminants. The roadmap set forth here could be used in other types of risk
assessments that are of interest to EPA, of course, with data appropriate for the contaminants of
interest and processes that affect their movement through the environment.
Of particular interest is the data gaps noted in the text of this document. Data are lacking for
many of the parameters needed to understand risk more fully. These data gaps lead to
uncertainties in the results or, in some sense equivalently, broad spread in the risk assessment
results. The selected scenarios were chosen to span a wide range of potential risks as well as a
substantial range of likely values affecting the outcome. However, real-world data for some
(many?) parameters are sparse or even non-existent, tempering the utility of such an analysis.
The treatment of such uncertainties is, perhaps, the weakest component of the report. In my
specific comments below, I have made some suggestions on how this might be improved and
concerns about biased results for exposure and concomitant risk addressed. It is my sense that
there is more to be included in the discussion of these uncertainties and I urge the EPA authors to
consider this in their final Report.
RESPONSE: Thank you for your comments on the discussion of uncertainties in the draft risk
assessment. To clarify, empirical data are available for nearly all the parameters used for
modeling in this risk assessment, with the exception of specific uptake factors for each fruit or
vegetable that may be grown on a farm. These comments are discussed alongside the more
detailed comments from this reviewer in Question 4.
COMMENT: In their final assessment of their analysis, the authors state that the deterministic
approach that they have taken coupled with a small number of scenarios chosen for evaluation is
sufficient for their purposes. Indeed, their deterministic approaches indicate that there is
significant risk in the vast majority of deterministic analyses they made and for all scenarios. A
Monte Carlo approach, which would have been my initial thought for this analysis, would have
required substantially more work, might have afforded assessment of a probability distribution
for risk, and been, possibly, more defensible. However, their argument regarding the utility of
the information obtained from the deterministic analysis precludes the need for a more
sophisticated approach. When a simple approach using the best quality data available and
appropriate models of the fate and transport of contaminants through the environment tells you
that essentially any scenario results in significant risk, even one with low level of contaminants
in the environment, the more sophisticated approach is not going to tell regulators anything more
than the simpler approach. It will just take more time and cost more.
The authors are to be commended. This Report is a Tour de Force- comprehensive, well-
developed, properly analyzed, and well-done overall.
RESPONSE: Thank you for your comment. The EPA agrees that in this case the deterministic
approach is sufficient to potentially inform risk managment actions. The EPA appreciates that
the reviewer found the written rational for that decision (that additional probabalistic modeling
will only result in higher risk findings because it aims to protect at the 95th percentile rather than
the median percentile) clear and convincing.
Reviewer 3
7
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
COMMENT: The risk assessment presented in this document is very well done, providing a
clear and easy-to-follow evaluation of the potential human health risks associated with the land
application and disposal of sewage sludge containing PFOA and PFOS. The assessment
effectively highlights the widespread presence of these chemicals in sewage sludge, their
persistence in the environment, and their significant toxicity to humans, including their potential
to cause cancer and other serious health effects. By employing a conservative screening approach
using the BioSolids Tool and central tendency deterministic modeling, the assessment identifies
substantial risks even at low concentrations of PFOA and PFOS, underscoring the need for
careful consideration of sludge management practices.
One notable strength of the assessment is its comprehensive analysis of various exposure
pathways across different scenarios, including agricultural and non-agricultural settings.
However, the primary focus of the risk assessment is on human health, with less emphasis on
ecological impacts. This human-centric approach, while critical, represents a limitation in that it
does not fully address the broader environmental implications of PFOA and PFOS
contamination. Despite this, the assessment provides a robust foundation for understanding the
risks posed to humans and offers valuable insights for mitigating these risks through informed
decision-making.
RESPONSE: Thank you for this comment. The EPA will respond to the specific items on this
topic discussed in Question 3.
Reviewer 4
COMMENT: This document is well-written, technically thorough, and establishes a solid
foundation for achieving the objectives of the Sewage Risk Assessment for PFOS and PFOA. I
commend the development team for their hard work on this complex draft assessment and have
tried to provide constructive, useful input for their consideration. I have summarized my main
comments below:
As I understand it, the overall goal of this risk assessment is to estimate the magnitude of risks
under different biosolids use and disposal scenarios on a central tendency (median) risk basis.
However in several parts of the calculation a "conservative" approach (which is designed to
overestimate risk) are used. I recommend removing all "conservative" calculation steps, input
data, and risk thresholds to and replace them with ones based on median / central tendencies.
The current groundwater exposure model has one unrealistic calculation step, the assumption
that the exposure concentration is equal to the highest concentration found in the top 2 meters of
the aquifer. This assumption runs counter the hydraulics and behavior of any drilled domestic
water well anywhere in the country. A more accurate approach should be evaluated:
o Use average concentrations over a realistic well screen length;
o Recognize that most or almost all domestic wells are either: a) screened at the bottom of
the shallowest unconfined aquifer unit, or b) screened in deeper aquifer units.
RESPONSE: Thank you for this comment; the EPA will respond to the commentor's specific
8
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
concerns on this topic under question 4.
COMMENT: A key issue in the groundwater modeling work is if a PFAS-specific model that
accounts for unusual properties of PFOA/PFOS should be applied, or to use EPACMPT model
which has been used before by USEPA for risk assessments. While the situation is complex and
somewhat unsatisfying, in the end I agree with the current approach to use EPACMPT is
sufficient as a compromise method. I agree that the USEPA should continue evaluating the
availability of groundwater and vadose zone models as this assessment is finalized. If there are
significant improvements to these PF AS-specific screening models, then a switch to these more
PFAS-centric models may be merited.
RESPONSE: Thank you for your comment of support that the use of EPACMTP is the best
available option. The EPA will continue to monitor the developments of PFOA and PFOS-
specific modeling tools as this assessment is finalized.
COMMENT: The groundwater exposure concentrations should be compared to the Maximum
Concentration Limits (MCLs) for PFOS and PFOA. Having a risk assessment that evaluates the
drinking water pathway without mention of MCLs seems to be missing a critical piece of the
USEPA risked-based regulatory framework for managing PFOS and PFOA.
RESPONSE: The commenter appears to have confusion regarding the definition of the MCL,
which is not a health-based value. Maximum Contaminant Levels (MCLs) are the highest level
of a contaminant that is allowed in drinking water. MCLs are set as close to the health-based
Maximum Contaminant Level Goals (MCLGs) as feasible using the best available treatment
technology and taking cost into consideration. The biosolids risk assessment compares modeled
exposure concentrations to health-based values, resulting in risk estimates for a suite of potential
exposure pathways. Any future risk mitigation/management actions would consider detection
thresholds for PFOA and PFOS in sewage sludge (based on the data provided in Table 9 in EPA
Method 1633, the limit of quantification for sewage sludge could be expected to range from 1.6-
4 ppb and the method detection limit is estimated to be 0.7 ppb, however, the exact values will
be determined by each laboratory) and other relevant factors.
COMMENT: I recommend that this risk assessment address the uncertainty in the current
USEPA toxicology calculations by briefly describing other regulatory and scientific viewpoints
for the toxicity parameters shown in Tables 4 and 5.
RESPONSE: It is unclear what the commenter is referring to by "uncertainty in the current
USEPA toxicology calculations." The EPA is relying on the conclusions of the agency's Final
Human Health Toxicity Assessment for Perfluorooctanoic Acid (PFOA) and Related Salts (US
EPA, 2024a) and Final Human Health Toxicity Assessment for Perfluorooctane Sulfonic Acid
(PFOS) and Related Salts (US EPA, 2024b). These documents represent the best available
science on PFOA and PFOS human health toxicity and were written in accordance with
longstanding EPA policies and guidance on toxicity assessment. These final toxicity assessments
for PFOA and PFOS reflect revisions made to respond to consensus recommendations from the
9
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
Science Advisory Board peer reviewers and thousands of public comments.
The EPA's final human health toxicity assessment for PFOA (US EPA, 2024a) considered all
publicly available human epidemiological, animal toxicological, mechanistic and toxicokinetic
evidence relevant to studies that evaluated health effects after oral PFOA exposure. Overall, the
available evidence indicates that PFOA exposure is likely to cause hepatic, immunological,
cardiovascular, and developmental effects in humans, given sufficient exposure conditions (e.g.,
at levels in humans as low as 1.1 to 5.2 ng/mL and doses in animals as low as 0.3 to 1.0
mg/kg/day). These judgments are based on data from epidemiological studies of infants,
children, adolescents, pregnant individuals, and non-pregnant adults, as well as short-term (28-
day), subchronic (90-day), developmental (gestational), and chronic (2-year) oral-exposure
studies in rodents. The EPA derived and considered multiple candidate reference doses (RfDs)
from both epidemiological and animal toxicological studies across the four non-cancer health
outcomes that the EPA determined had the strongest weight of evidence (i.e., immune,
cardiovascular, hepatic, and developmental). Decreased serum anti-tetanus and anti-diphtheria
antibody concentrations in children (Budtz-Jorgensen and Grandjean, 2018), decreased infant
birth weight (Wikstrom et al., 2020), and increased total cholesterol in adults (Dong et al., 2019)
were selected as the co-critical effects for the overall oral RfD of 3 x 10"8 mg/kg/day (US EPA,
2024a). With respect to uncertainty, this RfD was derived by applying a total uncertainty factor
(UF) of 10 to account for intraspecies variability.
Consistent with EPA's Guidelines for Carcinogen Risk Assessment (US EPA, 2005a), the EPA
reviewed the weight of the evidence across epidemiological, animal toxicological, and
mechanistic studies and concluded that PFOA is Likely to Be Carcinogenic to Humans via the
oral route of exposure. Epidemiological studies provided evidence of kidney and testicular
cancer in humans and some evidence of breast cancer in susceptible subpopulations. Chronic oral
animal toxicological studies in Sprague-Dawley rats reported Leydig cell tumors (LCT),
pancreatic acinar cell tumors (PACT), and hepatocellular tumors. PFOA exposure is associated
with multiple key characteristics of carcinogenicity. Available mechanistic data suggest that
multiple modes of action (MOAs) could be involved in the renal, testicular, pancreatic, and
hepatic tumorigenesis associated with PFOA exposure in humans and animal models. The EPA
derived and considered multiple candidate cancer slope factors (CSFs) for PFOA from both
epidemiological and animal toxicological studies across multiple tissue types and organ systems
(i.e., kidney, liver, pancreas, testes). The oral slope factor of 0.0293 (ng/kg/day)"1 (29,300
(mg/kg/day)"1) for renal cell carcinoma (RCC) in human males from Shearer et al. (2021) was
selected as the basis of the overall CSF for PFOA (US EPA, 2024a).
The EPA's final human health toxicity assessment for PFOS (US EPA, 2024b) considered all
publicly available human epidemiological, animal toxicological, mechanistic and toxicokinetic
evidence relevant to studies that evaluated health effects after oral PFOS exposure. Overall, the
available evidence indicates that PFOS exposure is likely to cause hepatic, immunological,
cardiovascular, and developmental effects in humans, given sufficient exposure conditions (e.g.,
at levels in humans as low as 0.57 to 5.0 ng/mL and doses in animals as low as 0.0017 to 0.4
mg/kg/day). These judgments are based on data from epidemiological studies of infants,
children, adolescents, pregnant individuals, and non-pregnant adults, as well as short-term (28-
day), subchronic (90-day), developmental (gestational), and chronic (2-year) oral-exposure
studies in rodents. The EPA derived and considered multiple candidate RfDs from both
epidemiological and animal toxicological studies across the four non-cancer health outcomes that
the EPA determined had the strongest weight of evidence (i.e., immune, cardiovascular, hepatic,
10
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
and developmental). Decreased infant birth weight (Wikstrom et al., 2019) and increased total
cholesterol in adults (Dong et al., 2019) were selected as the co-critical effects for the overall
oral RfD of 1 x 10-7 mg/kg/day (US EPA, 2024b). With respect to uncertainty, this RfD was
derived by applying a total UF of 10 to account for intraspecies variability.
Consistent with EPA's Guidelines for Carcinogen Risk Assessment (US EPA, 2005a), the EPA
reviewed the weight of the evidence across epidemiological, animal toxicological, and
mechanistic studies and concluded that PFOS is Likely to Be Carcinogenic to Humans via the
oral route of exposure. Epidemiological studies provided evidence of bladder, prostate, liver,
kidney, and breast cancers in humans, although evidence was limited or mixed for some cancer
types. Animal toxicological studies supported findings from human studies. Bioassays conducted
in Sprague-Dawley rats reported hepatocellular tumors, pancreatic islet cell tumors, and thyroid
follicular cell tumors after chronic oral exposure. Some studies observed multisite tumorigenesis
(liver and pancreas) in male and female rats. PFOS exposure is associated with multiple key
characteristics of carcinogenicity. Available mechanistic data suggest that multiple MO As play a
role in pancreatic and hepatic tumorigenesis associated with PFOS exposure in animal models.
The EPA derived and considered multiple candidate CSFs from animal toxicological studies
across multiple tissue types or organ systems (i.e., liver and pancreas). The oral slope factor of
39.5 (mg/kg/day)"1 for combined hepatocellular adenomas and carcinomas in female rats from
Butenhoff et al. (2012) and Thomford (2002) was selected as the basis of the overall CSF for
PFOS.
For more information, please see the Final Human Health Toxicity Assessment documents (US
EPA 2024a;b), the EPA response to the related SAB review (US EPA 2023b), the Final PFAS
National Primary Drinking Water Regulation Rulemaking (US EPA, 2023a), and the Responses
to Public Comments on Per- and Polyfluoroalkyl Substances (PFAS) National Primary Drinking
Water Regulation Rulemaking (US EPA, 2024c).
COMMENT: Dioxin is a more potent carcinogen than either PFOS or PFOA. In 2003, the
USEPA decided against regulating dioxin in sewage sludge after conducting a risk assessment. I
recommend USEPA perform a quantitative comparison of the dioxin and PFOS/PFOA risk
assessments to: 1) highlight similarities and differences between these chemical classes; 2)
provide context for current regulatory decisions; and 3) ensure consistency in the risk assessment
approaches. This comparison should be placed in or next to Section 5.3.
RESPONSE: While the EPA disagrees that information on Dioxins and PCBs should be added to
the PFOA and PFOS Draft Risk Assessment, the EPA will discuss these three items in response
to specific comments by Reviewer 4 in the Charge Question portion of this document.
Reviewer 5
COMMENT: The document provides a risk assessment of the potential human health and
environmental risks associated with the land application and disposal of sewage sludge
containing PFOA and PFOS. The accuracy of the information is supported by recent and well-
documented data showing elevated concentrations of these contaminants in US sewage sludge,
including comparing modeled concentrations and observed concentrations in Alabama and
Michigan. The report authors built upon EPA or other agency's previous work when a relevant
11
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
report or assessment exists and conducted a literature search of peer-reviewed and grey literature
when needed. For the fate and transport models, the inclusion of regional location-based
parameters to model a wide range of climate conditions (dry, moderate, and wet) increased the
applicability of this risk assessment to diverse settings across the nation.
The clarity of the presentation is strong, with a clear structure outlining the assessment process.
The report effectively distinguishes between high-end conservative risk assessments and central
tendency deterministic modeling. It details the scenarios modeled, including agricultural reuse
and disposal practices, and explains the implications of these scenarios in terms of human health
risks. The conclusions drawn are sound, based on conservative and median modeling results that
indicate significant health risks even at low concentrations of PFOA and PFOS. The decision not
to pursue further probabilistic modeling at this time, in favor of focusing on sharing current
results and mitigation strategies, is consistent with the mission of protecting human health and
the environment.
RESPONSE: Thank you for your comment. The EPA agrees that in this case the deterministic
approach is sufficient to potential inform risk managment actions. We appreciate the feedback
that the rationale for this approach, as written in the draft risk assessment, is clear and
convincing.
12
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
III. EPA RESPONSE TO CHARGE QUESTION COMMENTS
13
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
QUESTION 1: Programmatic, statutory, and regulatory context
14
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
Prosrammatic. statutory. and regulatory context. The Clean Water Act and its implementing
regulations directs EPA to assess and manage risks associated with the disposal or use of
sewage sludge to protect public health and the environment. Please comment on the extent to
which the description of EPA '.s Biosolids Program, including its statutory and regulatory
authorities, provides the context and basis for the risk assessment (see Section 1).
Reviewer 1
COMMENT: The document clearly sets the stage for the risk assessment. Information is
provided on the disposal of sewage sludge, land applications, and potential pathways of
exposure. The background on previous rounds of regulations, namely regarding organic
compounds like dioxins and PCBs helps put the current assessment in context.
RESPONSE: Thank you for your comment.
Reviewer 2
COMMENT: This Section provides a history of regulation for metals and chemical compounds
in biosolids and sewer sludge. The text indicates the type of modeling structures that have been
used and their results. In particular, the focus on specific metals, PCBs and related compounds
were modeled using multiple scenarios (14 in all covering pathways encompassing 17 different
scenarios). The authors then suggest in the final sentence that a similar assessment should be
carried out for PFOS and PFOA in this assessment.
There appears to be sufficient regulatory support to consider PFOA and PFOS in this assessment
and the section provides the context clearly under the Clean Water Act and its Amendments.
I note no specific comments on material in this section. However, I do expect follow-up on the
use of deterministic modeling strategies only as compared with the Monte Carlo strategies
implemented in the dioxins assessment.
RESPONSE: Thank you for your comment. The dioxins assessment used Monte Carlo modeling
approaches to estimate the 50th, 75th, 90th, 95th, and 99th percentile exposure levels. As described
in section 2.9.1 of the draft risk assessment (modeling plan), the assessment of PFOA and PFOS
started with a high-end screening risk assessment, assuming a reasonable maximum exposure
scenario (see Appendix E in the draft risk assessment). Due to the high-risk estimates in the
screening analysis, the EPA next conducted a deterministic modeling approach targeting "central
tendency" exposures (the 50th percentile) to better understand the potential scope and magnitude
of potential risks under different use and disposal scenarios. As described in section 4.9, based
on the results of these 50th percentile modeling exercises, the EPA determined that further Monte
Carlo modeling for PFOA and PFOS would not change the risk conclusions. The different
modeling approaches for PFOA/PFOS and dioxins are appropriate due to the different risk
findings in each case. Please see the EPA's responses to Reviewer 4 in this section for more
information about the differences between the PFOA/PFOS and dioxin risk assessments.
Reviewer 3
15
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
COMMENT: The EPA's Biosolids Program is governed by the Clean Water Act (CWA) and its
amendments, specifically Section 405(d), which mandates the establishment of regulations to
manage sewage sludge (biosolids) and protect public health and the environment from potential
adverse effects. This statutory framework requires the EPA to set numeric limits and
management practices that mitigate risks associated with toxic pollutants in biosolids. The
Program's regulations under 40 CFR part 503 outline the acceptable practices for land
application, surface disposal, and incineration of biosolids, including criteria for pollutant
concentrations and site management practices. This regulatory backdrop is essential for
understanding the scope and objectives of the risk assessment for PFOA and PFOS.
The risk assessment process described in the document is designed to evaluate the potential
health and ecological risks associated with the land application and disposal of sewage sludge
containing PFOA and PFOS. The extent to which the description of the EPA's Biosolids Program
informs this assessment is evident in the structured approach taken to evaluate different scenarios
of biosolids use. These scenarios include land application on agricultural lands and reclamation
sites, and disposal in surface disposal sites. The assessment considers various exposure pathways
and populations, such as farm families, community-supported agriculture participants, and
individuals consuming freshwater fish, reflecting the Program's comprehensive approach to
managing potential risks associated with biosolids.
Moreover, the historical context provided by previous risk assessments under the Biosolids
Program illustrates the evolution of the EPA's approach to managing chemical risks in biosolids.
Past regulations and assessments focused on metals and other contaminants, with subsequent risk
evaluations addressing additional pollutants such as dioxins and PCBs. The framework
developed from these earlier assessments serves as a model for the current evaluation of PFOA
and PFOS, incorporating lessons learned and methodological advancements. This continuity
ensures that the risk assessment for PFOA and PFOS builds upon a solid foundation of
regulatory and scientific understanding, adapting to the specific challenges posed by these
persistent and bioaccumulative substances.
In conclusion, the description of the EPA's Biosolids Program, including its statutory and
regulatory authorities, provides a crucial context for the risk assessment of PFOA and PFOS in
biosolids. It outlines the regulatory framework within which the assessment operates, the
historical evolution of risk management practices, and the structured approach to evaluating
potential risks. This comprehensive context ensures that the risk assessment is both relevant and
robust, addressing the specific challenges posed by these emerging contaminants while building
on established regulatory practices and scientific methodologies.
RESPONSE: Thank you for your comment.
Reviewer 4
COMMENT: Recommend adding context about how many people are likely to impacted by the
groundwater pathway modeling scenarios. This maybe only possible on an order of magnitude
basis. For the specific current scenario, this analysis might start with the population of the
16
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
farming/ranching community in the U.S. and estimating how many farms: 1) applied biosolids,
and 2) drink water from a domestic water well, and 3) have a well that immediately
downgradient of the biosolids application area, and 4) have a well with a well screen that only
draws water from the top two meters of the aquifer (i.e., not screened in deeper aquifers or at the
bottom of their aquifer). The goal is to give the public a rough estimate for how common this
scenario occurs.
RESPONSE: Unfortunately, the number of people potentially impacted by groundwater
contamination sourced from sewage sludge reuse and disposal is larger than the universe of
potentially impacted people suggested by this commentor. The number of sewage sludge use and
disposal sites includes not only farms with biosolids land-application, but also unlined or clay
lined lagoons, surface disposal sites, land reclaimed using biosolids, forestry sites, golf courses,
playgrounds, and other sites where biosolids are land applied. The number, size, and location of
these sites is currently unknown. Therefore, the number of homes with potentially impacted
drinking water sources (whether these drinking water sources are groundwater or surface water)
has not been quantified as part of this draft risk assessment.
The goal of this draft risk assessment is to determine if PFOA and PFOS may be present at
concentrations in sewage sludge that may adversely affect public health or the environment (see
the Clean Water Act section 405(d)). If the EPA decides it would be appropriate to develop risk
mitigation measures, it would consider the scope and scale of the potential beneficial outcomes
of each risk mitigation activity using the best available information.
COMMENT: Recommend USEPA provide additional context on the comparison of the dioxin
sewage sludge vs. PFOS/PFOA in sewage sludge. This analysis could be presented in or near
Section 5.3. There appear to be large differences in the overall risk in sewage sludge these
chemicals. Was this due to the starting concentrations, toxicology, fate and transport, or
exposure factors?
RESPONSE: Comments related to the biosolids risk assessment for dioxin are out of scope;
nevertheless, the EPA is providing some information comparing the dioxin assessment to the
current PFOA/PFOS assessment. There are many differences between PFOA and PFOS and
dioxins that have resulted in different findings in their respective risk assessments. First, the
concentrations of dioxin (expressed as toxic equivalent concentration or TEQ) was far lower in
the 2001 National Sewage Sludge Survey (NSSS) than the 1 ppb value used here for PFOA and
PFOS. Most of the samples in the 2001 NSSS had TEQ values between 7 and 55 ng/kg (part per
trillion or ppt); these data were used to generate a distribution of concentrations for the 2003
dioxin risk assessment. There are currently no national survey data on PFOA and PFOS
concentrations in U.S. sewage sludge; however, data from states and peer-reviewed journal
articles indicate that many PFOS sludge concentrations are near 10 parts per billion (ppb) with
an upper range value being above 100 ppb (see draft risk assessment, appendix A). In short,
PFOS concentrations in sewage sludge appear to be about 1000 times higher than dioxin
concentrations were at the time the EPA developed the dioxins risk assessment.
Second, PFOA and PFOS are more mobile than dioxins, resulting in greater plant uptake and
transport to groundwater and surface water. This greater mobility of PFOA and PFOS resulted in
17
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
groundwater being included as a pathway for draft risk assessment whereas it was scoped out of
the dioxin assessment entirely. Similarly, in the dioxins assessment, the uptake of dioxins to
above grounds plants was limited to dioxins becoming airborne from the soil and then being
absorbed or landing on the plant as particulates. For PFOA and PFOS, studies demonstrate that
these chemicals are transported to plant vegetative tissue via the plant vascular system.
Third, the reviewer has focused on the toxicity of one congener (2,3,7,8-TCDD); however, many
of the congeners found in sewage sludge have toxic equivalence factors much less than one, i.e.,
a fraction of the toxicity of 2,3,7,8-TCDD. Specifically, TEQs range from 1 to 0.00001, see
Table 2-2 of the 2003 dioxins assessment (US EPA, 2003). This makes it difficult to directly
compare the hazard of PFOA or PFOS in sewage sludge to dioxins in sewage sludge. In the case
of PFOA and PFOS, the EPA is relying on toxicity values (reference doses, cancer slope factors)
that are specific to these two chemicals.
In summary, dioxins and PFOA/PFOS differ in their occurrence in sewage sludge, their fate and
transport behaviors in natural systems, and their human health toxicity. These differences have
resulted in differences between the risk conclusions for dioxins in biosolids and the preliminary
risk conclusions for PFOA or PFAS in biosolids.
COMMENT: Recommend this risk assessment compare the Section 3 media groundwater (and
surface water?) concentrations to the Maximum Concentration Limits (MCLs) for PFOS and
PFOA for all drinking water ingestion scenarios. This comparison should be done in the tables
and the text.
RESPONSE: The EPA disagrees that is it appropriate to compare surface water and groundwater
concentrations to MCLs in the context of this risk assessment. As described previously (see page
9), the MCLs for PFOA and PFOS finalized under the PFAS Drinking Water Rule take into
account best available treatment technology in drinking water and cost to drinking water
systems.
Reviewer 5
COMMENT: The description of EPA's Biosolids Program, including its statutory and regulatory
authorities, provides a helpful context and basis for the risk assessment. First, the text
highlighted Section 405 of the Clean Water Act (CWA) as the primary legal authority under
which the EPA is required to establish and review regulations to protect public health and the
environment from the adverse effects of pollutants in sewage sludge. Then, it introduced the
Standards for the Use or Disposal of Sewage Sludge (40 CFR Part 503).
The historical regulatory context is particularly important, as it shows the evolution of the EPA's
approach to managing risks associated with sewage sludge. The document traces the
development of the EPA's sewage sludge regulations from the first rule in 1993, which
established pollutant limits and management practices for ten metals, to subsequent risk
assessment around dioxins and dioxin-like PCBs. This historical context establishes a foundation
for understanding the risk assessment of PFOA and PFOS by connecting it to previous regulatory
actions and frameworks used by the EPA. Introducing prior work before discussing the risk
18
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
assessment of PFOA and PFOS in sewage sludge is helpful because it provides relevant
background knowledge for understanding the assessment results, and helps the reader to
appreciate the wide range of outcomes that can result from the risk assessment, including setting
numeric standards if considerable risk is found, and not setting a numerical standards if there is
considerable sufficient safety margin.
RESPONSE: Thank you for your comment.
19
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
QUESTION 2 : Problem Formulation
20
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
Problem formulation. Please comment on the characterizations of the environmental fate,
exposure, and toxicity of PFOA/PFOS in the problem formulation section of the risk
assessment (see Section 2, Appendix A).
Reviewer 1
COMMENT: The state of the science regarding factors influencing PFOA/PFOS partitioning in
the environment is clearly described. Historical measurements of PFOA and PFOS in biosolids
are presented, and the potential for precursors to contribute to PFOA/PFOS concentrations in
biosolids is highlighted.
RESPONSE: Thank you for your comment.
Reviewer 2
COMMENT: As a screening tool the models used appear well formulated if somewhat
simplified. The Box equations on pages 42 and 43 illustrate this. While including essential
features of concentrations, exposures, and risk, certain nuances are left out. However, the effect
of these details may be expected to be insignificant given the approaches, e.g., high-end and
median estimates of exposure and risk discussed.
RESPONSE: Thank you for your comment. It is unclear what "certain nuances" the commenter
is referring to.
Reviewer 3
COMMENT: The description of the processing of influent and sewage sludge at wastewater
treatment plants (WWTPs) highlights an intriguing aspect of environmental toxicology. The
breakdown of fluorinated precursors into perfluorooctanoic acid (PFOA) and perfluorosulfonic
acid (PFOS) during wastewater treatment and sludge processing underscores a critical area of
concern. This process is particularly relevant given that these chemicals are persistent and
hazardous. Precursors such as perfluorooctane sulfonamidoethanol-based phosphate diesters,
fluorotelomer alcohols, and polyfluorinated iodides are commonly transformed into PFOA and
PFOS, which then accumulate in biosolids used in land applications.
The research into these transformations is of high priority in environmental toxicology due to the
significant environmental and health implications. Laboratory studies and real-world
observations indicate that biosolids' treatment and land application are crucial pathways through
which these hazardous chemicals are introduced into soils and the broader environment. The
presence of these precursors in consumer products and their subsequent release into wastewater
emphasizes the need for a focused assessment on PFOA and PFOS.
Given the identified data gaps in understanding these precursors' occurrence, environmental fate,
and degradation pathways, further research is necessary. The current risk assessment prioritizes
PFOA and PFOS due to their established risks, but future assessments should consider the
21
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
broader scope of PFAS precursors and their impacts. The findings could lead to refined policy
decisions regarding managing these substances, enhancing our ability to mitigate environmental
and health risks.
RESPONSE: Thank you for your comment. The EPA agrees that research to better understand
the environmental fate and effects of precursors to PFOA and PFOS is important.
Reviewer 4
COMMENT: Section 2.6.1.1 of the Risk Assessment provides a detailed, well-written, and
thorough explanation of the USEPA's Human Health Toxicity results. Recommend providing
the readers of the document with a brief mention of regulatory examples and human toxicity
studies that result in alternative viewpoints of PFOS/PFOA safe doses. For example:
• Australia's PFOA drinking water guideline is 560 ng/L vs. a USEPA interim health
advisory level of 0.004 ng/L and a USEPA MCL of 4 ng/L.
• Burgoon et al., 2023 summarizes the work of 24 scientists from 8 countries who
concluded the PFOA safe dose is between 10-70 ng/kg/day vs. USEPA's 0.03 ng/kg/day
RfD (Table 4). Burgoon et al. (2023) states: "However, this range is well above the
single values of both EFSA (2020) and EPA (2023). The principal reasons for the larger
disparity between this provisional range with these latter two single values is the
unanimous judgment of the international collaboration that the existing human cancer
and noncancer data are not sufficiently credible as a basis of the PFOA safe dose in the
absence of mechanistic data that are relevant to humans at serum concentrations seen in
the general population. In this regard, Health Canada, the WHO and Food Standards of
Australia and New Zealand are in agreement with the Collaboration—the use of human
data is not sufficiently credible as the basis for the PFOA safe dose. "
Finally while this is likely be untenable for the purpose of this USEPA risk assessment, is it
possible that the Australia /Burgoon et al. toxicity values above are closer to the goal of reporting
a "Central Tendency" for a biosolids risk assessment than using the current USEPA toxicology
factors in Tables 4 and 5? Were the USEPA toxicological values developed in part using data
from unusual populations and therefore may not to represent a toxicological "central tendency"
of the response to PFOS/PFOA exposure to the population in the U.S.?
RESPONSE: As described previously, the EPA is relying on the conclusions of the agency's
Final Human Health Toxicity Assessments for PFOA and PFOS (US EPA, 2024a;b) because
these documents represent the best available science on PFOA and PFOS human health toxicity
and were written in accordance with longstanding EPA policies and guidance on toxicity
assessment. See the EPA's prior response on this topic on page 9 of this document.
As to the differences between the EPA's toxicity assessments and similar documents from other
sources, one explanation for differing conclusions between health agencies are the differing
methods and guidance used to develop the assessments. The EPA uses established systematic
review practices to identify, evaluate, synthesize, integrate, and quantify evidence in a chemical
database (US EPA 2022). Other health agencies, including the WHO, do not follow these same
practices and, as a result, may arrive at different conclusions. Notably, the WHO has recently
22
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
withdrawn their proposed drinking water guidelines for PFOA and PFOS. Additionally, the EPA
followed agency guidance, such as the Guidelines for Carcinogen Risk Assessment to determine
the cancer classifications for PFOA and PFOS (US EPA, 2005). The classification systems used
by other agencies (e.g., IARC, UK COT, CalEPA) differ from those used by the EPA; the
application of different systems may result in different conclusions by other agencies. However,
CalEPA's final public health goals are also generally supportive of the EPA's cancer
classifications for PFOA and PFOS (CalEPA, 2024). As a final example, some agencies, such as
the WHO, have published guidance values that are not solely health based (i.e., they consider
feasibility, analytical methods, etc.) and, therefore, cannot be directly compared to the EPA's
MCLGs, which are based solely on health effects information.
Further, the EPA disagrees with comments stating that the epidemiological database for PFOA is
too uncertain to support a classification of Likely to Be Carcinogenic to Humans. As described
similarly in both the draft and final toxicity assessments for PFOA as well as the Maximum
Contaminant Level Goals for Perfluorooctanoic Acid (PFOA) and Perfluorooctane Sulfonic Acid
(PFOS) document (US EPA, 2024d), the available epidemiological data support an increased risk
of both kidney and testicular cancers associated with PFOA exposure. There is also evidence that
PFOA exposure may be associated with an increased breast cancer risk, based on studies in
populations with specific polymorphisms and for specific types of breast tumors. Taken together,
these results provide consistent and plausible evidence of PFOA carcinogenicity in humans.
Additionally, while genotoxicity is one potential MOA for carcinogenicity, there is no
requirement that a chemical be genotoxic for the EPA to classify it as either Carcinogenic to
Humans, Likely to Be Carcinogenic to Humans, or Suggestive Evidence of Carcinogenic
Potential according to the Guidelines for Carcinogen Risk Assessment. Importantly, the SAB
PFAS Review Panel supported the rationale for the Likely to Be Carcinogenic to Humans
designation for PFOA in its final report.
Reviewer 5
COMMENT: Section 2 provided a comprehensive characterization of the environmental fate,
exposure, and toxicity of PFOA/PFOS, which offered helpful background information for the
risk assessment.
Section 2.1 is about the general literature search strategy; it is good to see that the authors of the
report also referenced "grey literature" in their search.
Section 2.2 is about the chemical and physical properties of PFOS and PFOA, and their
environmental behavior including fate and transport. The discussions around the transformation
and degradation of precursors are particularly important, as this process is responsible for a
considerable proportion of PFOA and PFOS load in the sewage sludge.
Section 2.3 is about the sources to wastewater treatment plants and biosolids, in which the
authors cited relevant literature to discuss the source of PFOS and PFOA despite the phase-out of
domestic manufacturing of these compounds. Here I wish the authors expanded more on any
quantitative information they can find regarding the contribution of different sources of
PFOS/PFOA to WWTPs in various geographies across the country. This review activity can be
23
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
nicely combined with the literature summary in section 2.4, where the authors synthesized recent
studies over the past 15 years on PFAS occurrence data in biosolids. An easy improvement is to
add a column to Tables A and B in Appendix A, and list major source of PFAS to the WWTP
included in the study, if the paper reported any.
RESPONSE: Thank you for the helpful comment. There is a varying level of information
available in the studies on the sources of PFOA and PFOS to the monitored biosolids. The EPA
included information in the text and Appendix A where studies highlighted the sources (e.g.,
Michigan and Vermont). Since most studies do not pinpoint the PFOA and PFOS source, the
EPA is planning the upcoming influent study to gather nationwide PFAS data on industrial and
domestic sources and concentrations in POTW influent, effluent, and sewage sludge.
COMMENT: Section 2.5 is about PFOA and PFOS accumulation in animal and plants, and it is
relevant for a few different purposes of the report: the discussion of PFOA/PFOS accumulation
in human is relevant for understanding reference dose and potential health effects, as well as
internal dose calculation; the discussion of other animals is relevant for the exposure pathway of
fish consumption; the discussion of plants uptake is relevant for the exposure pathway of
consuming contaminated produce. I find the discussions about phytoremediation interesting but
not particularly relevant to the focus of this report.
Section 2.6 is about the effects on human and aquatic and terrestrial biota. For human toxicity,
the authors discussed the health effects of different exposure pathways and included reference
dose (RfD) as well as cancer slope factor (CSF) for PFOA/PFOS from the recent EPA toxicity
assessment. For ecological effects, the authors included a helpful table comparing the freshwater
aquatic life water quality criteria in Table 3. For terrestrial organisms, since the amount of data
available is less, listing the reference doses equivalent in the paragraph is helpful to set the stage
for the discussion in section 2.6.3.
Section 2.7 is about exposure pathways for humans and aquatic and terrestrial biota. Section
2.7.1 can be more strengthened with adding quantitative information about the contributions of
various exposure pathways. This information likely differs for different geographical regions, so
the authors can discuss a few scenarios such as communities with contaminated drinking water,
general population, occupation exposure, and children.
RESPONSE: The EPA agrees that information about relative contributions of each exposure
pathway in each sewage sludge use or disposal scenario is useful in the context of exposures
from biosolids use and disposal; this information is currently available in Section 4 (Risk
Estimation). This assessment does not include discussion of exposure scenarios where there may
be sources of PFOA and PFOS exposure other than biosolids (i.e., non-biosolids occupational
exposures, communities with drinking water contamination from another source, consumer
products) because this risk assessment is scoped to consider only risks from sewage sludge use
and disposal.
24
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
COMMENT: Section 2.8 introduces the four modeling scenarios: crop farm scenario, pasture
farm scenario, surface disposal scenario, land reclamation scenario, and other land application
scenarios. For each scenario, the authors developed a conceptual model to describe the pathways
from source to environmental media, to exposure pathways, and to receptors. These are helpful
to set up the framework before getting into the model parameters in the subsequent sections. The
last scenario has considerable uncertainties around the key parameters and processes, so they
were assessed qualitatively.
Section 2.9 is the analysis plan, which walked through the progression of risk assessment models
from high-end deterministic models, to central tendency deterministic models, and last to
probabilistic models. This is a logical progression as it went from most conservative to more
realistic scenarios. This section also includes the various models EPA selected for assessing the
different steps of PFOS and PFOA fate and transport: including from soil to groundwater, from
groundwater to surface water, from groundwater to drinking water wells, and leaching through
lined or unlined surfaces. This section also has detailed information on how EPA obtained
parameters for these models and their literature review strategy. When there is an existing
assessment or report from EPA or another agency, those conclusions are prioritized over
individual studies. For peer-reviewed studies, field studies are favored over lab studies, and
studies with biosolids applications are favored over other sources of PFAS contamination.
RESPONSE: Thank you for your comment.
25
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
QUESTION 3: Ecological Risk Assessment
26
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
Ecological risk assessment. As described in Section 2.6.3, available data indicate that risk to
aquatic and terrestrial life from land application of biosolids contaminated with PFOA and/or
PFOS is expected to be lower than the potential risk to human health from same/similar
biosolids applications. As a result, the quantitative risk assessment (risk estimates, risk
characterization) is scoped to focus solely on human health. Please comment on the
conceptual and technical basis for limiting the assessment of ecological risk andfocusing on
human health for PFOA/PFOS.
Reviewer 1
COMMENT: The rationale for limiting the assessment to human health is sound considering that
human health-based thresholds are more stringent, which means that guidelines established based
on risk to humans would be protective for the terrestrial and aquatic ecosystems. It could be
noted that animal/plant toxicity experiments are typically conducted at much higher doses, which
may not allow the observation of low-dose effects.
RESPONSE: Thank you for your comment that the rationale for limiting the scope of this draft
risk assessment to human health is sound based on the relative stringency of potential human-
health protective practices compared to the stringency of potential practices protective of aquatic
or terrestrial wildlife.
The EPA disagrees with the statement that there are not experimental data available regarding
the potential for low-dose effects in aquatic animals. The EPA's final Aquatic Life Criteria for
PFOA and PFOS include summaries of many studies conducted over chronic exposure scenarios
that include observations of sublethal and potentially sensitive endpoints (US EPA, 2024e;f).
These studies are available for many species of fish and aquatic invertebrates. The aquatic life
criteria found multiple studies to evaluate surface water concentrations protective of 95% of
aquatic organisms. Even given these data, surface water concentrations with the potential to
impact human health via drinking water or eating fish are much lower than surface water
concentrations potentially harmful to aquatic life. For example, the PFOA and PFOS 2022
interim drinking water health advisories were well below 1 ppt and the 2024 final health-based
MCLGs for PFOA and PFOS were each zero, whereas the final chronic aquatic life criteria are
250 ppt for PFOS and 100,000 ppt for PFOA. Risks associated with fish consumption also are
expected to occur multiple orders of magnitude below the EPA's national recommended aquatic
life criteria (see draft risk assessment section 2.6.3).
The EPA agrees that there are limited PFOA and PFOS toxicity data available for terrestrial
organisms, such as birds and other wildlife. Toxicity data for terrestrial plants and soil
invertebrates are also limited and primarily focus on acute (mortality) effects (see section 2.6.2.2
of the draft risk assessment). The EPA will continue monitor the available ecotoxicological
literature for terrestrial organisms.
Reviewer 2
COMMENT: Data on non-human toxicity of PFOA and PFAS is limited with only a few studies
done examining effects on plants, invertebrates, birds, and livestock/game. These limited data
result in substantial uncertainty in expected health impact and ecological risk in these media.
27
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
They do, however, suggest that the impact on these classifications of living organisms may be
substantially less than those observed in humans. However, these must be strongly tempered by
the relative number of investigations on them compared to human studies.
It is my opinion that the data on non-human life forms is currently too sparse to consider
ecological risk assessments. I do think this data gap compels EPA to fund or carry out in-house
research programs evaluating these effects in order to obtain more reliable estimates of exposure
and effect. Such data would be useful in parameterizing models to look at the effects of control
strategies or likely impacts of current levels on ecological risk assessments. Further, such studies
may suggest whether the need for control strategies focused on ecological risk are even
necessary. The current values thought to be protective of ecological systems exceed- often far
exceed- those values being considered protective of human health. Hence implementing values
protective of human health likely would prove sufficient for protecting ecological health with a
large margin of safety. Additional data collection on PFOA/PFOS in ecological systems is
warranted but is unlikely to result in regulations that would place such systems under greater
regulatory control than regulations based upon human protection.
In summarizing my view of the ecological risk evaluation work presented in this Report, I
believe EPA has made a correct decision to focus on the human health outcomes but should
follow the literature coming out currently focusing on ecological impact with an eye towards
developing additional ecological risk parameters. At this point, I do not feel enough data are
available to make reasonable estimates for modeling.
RESPONSE: Thank you for your comment that available data suggest impacts of PFOA and
PFOS exposure to plants, invertebrates, birds, and livestock appear to be substantially less
sensitive than the impacts observed to these exposures in humans. As stated previously, the EPA
agrees that there are fewer studies available observing effects in non-human terrestrial animals
and terrestrial plants than are available for humans, and that this lack of data warrants discussion
in the draft risk assessment. The EPA has added such discussion. The EPA will continue monitor
the available ecotoxicological literature for terrestrial organisms.
Reviewer 3
COMMENT: The conceptual and technical rationale behind limiting ecological risk assessment
for PFOA and PFOS and focusing primarily on human health appears fundamentally flawed and
potentially shortsighted. This approach underscores a troubling tendency to prioritize human
health risks over broader ecological impacts, which could lead to inadequate ecosystem
protection.
Firstly, the assertion that adverse effects on plants, invertebrates, fish, and birds occur at higher
concentrations than those affecting humans does not necessarily justify minimizing ecological
risk assessments. Ecosystems are complex and interconnected, and focusing narrowly on human
health could ignore subtle yet significant ecological impacts. For instance, the health of plants
and invertebrates can influence broader ecological functions, including nutrient cycling, food
web dynamics, and habitat structure. Adverse effects in these species might reflect direct harm
28
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
and disrupt these essential ecological processes, ultimately affecting biodiversity and ecosystem
resilience.
RESPONSE: As stated above, surface water concentrations with the potential to impact human
health via drinking water or eating fish are much lower than surface water concentrations
potentially harmful to aquatic life. For example, the PFOA and PFOS 2022 interim drinking
water health advisories were well below 1 ppt and the 2024 final health-based MCLGs for PFOA
and PFOS were each zero, whereas the final chronic aquatic life criteria are 250 ppt for PFOS
and 100,000 ppt for PFOA. Risks associated with fish consumption also are expected to occur
multiple orders of magnitude below the EPA's national recommended aquatic life criteria (see
section 2.6.3).
The EPA agrees that there are limited PFOA and PFOS toxicity data available for terrestrial
organisms, such as birds and wildlife. Toxicity data for terrestrial plants and soil invertebrates
are also limited and primarily focus on acute (mortality) effects (see section 2.6.2.2 of the draft
risk assessment). The EPA will continue monitor the available ecotoxicological literature for
terrestrial organisms.
COMMENT: Moreover, there needs to be more adverse effects reported in livestock to equate to
a comprehensive understanding of environmental impacts. Livestock health is only one aspect of
the broader ecological system, and its absence from studies does not negate the potential for
other wildlife or plant species to suffer from exposure to PFOA and PFOS. This narrow focus
could overlook cumulative and indirect effects that may become apparent only after prolonged
exposure or under specific environmental conditions.
RESPONSE: The EPA is confused by the suggestion that this assessment narrowly focuses on
adverse effects to livestock. The EPA agrees that adverse effects in livestock would be a distinct
concern from adverse impacts to terrestrial wildlife or ecosystems. Moreover, the EPA does not
indicate in the draft risk assessment that the presence or absence of adverse effects in livestock
gives significant indication of the potential for adverse effects terrestrial wildlife like insects,
other mammals, or disruptions to ecosystems. As written, the assessment summarizes available
data on terrestrial wildlife toxicity separately from the available data on livestock toxicity in
section 2.6.2.2.
COMMENT: From a technical standpoint, the decision to base risk thresholds primarily on
human health rather than ecological criteria overlooks significant concerns. The observed effect
levels for soil and aquatic environments being in the range of 10's to 100's mg/kg for PFOA and
PFOS, compared to much lower thresholds for protecting human health, suggests a substantial
disparity in protective measures. This discrepancy indicates that current aquatic life criteria and
soil thresholds may be insufficiently stringent to prevent ecological harm, potentially allowing
harmful concentrations of these substances to persist in the environment.
Additionally, the notion that human health-based thresholds will inherently protect ecological
systems is problematic. Ecosystems only sometimes respond straightforwardly to contaminants,
and different species and environmental compartments can have varying sensitivities to
pollutants. The assumption that setting stringent human health standards will automatically
29
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
ensure ecological protection is an oversimplification that needs to account for the diverse ways
pollutants can affect the environment. For example, fish might accumulate PFOA and PFOS at
levels that are not directly harmful to human health but could still have detrimental effects on
fish populations and aquatic food webs.
Furthermore, the argument that more studies on wildlife effects could bridge the gap between
ecological and human health thresholds is insufficient. The current reliance on human health
criteria as a proxy for ecological protection risks inadequate safeguards for ecosystems.
Immediate and rigorous ecological risk assessments are crucial to address potential gaps and to
ensure that both human health and environmental health are comprehensively protected.
In summary, the approach of prioritizing human health over ecological risk for PFOA and PFOS,
while understandable given the potency of these chemicals in human health contexts, needs to be
revised. It fails to account for the complexities of ecological systems and the potential for
indirect and cumulative environmental impacts. A more balanced approach that integrates both
human and ecological health considerations is essential for effective and comprehensive risk
management.
RESPONSE: The EPA disagrees with this comment. As described previously, the EPA's
recently finalized Aquatic Life Criteria for PFOA and PFOS include chronic, low-dose studies
with observations of potentially sensitive non-lethal effects (US EPA, 2024e;f). It is possible for
chemicals to have larger effects (increased toxicity) on human health than on other organisms, or
vice versa, which would result in a "disparity" in the observed effect levels and the resulting
protective thresholds for humans as compared to other organisms. Given the currently available
data, it appears that PFOA and PFOS have substantially longer half-lives in human beings and
are more toxic to humans than to other terrestrial or aquatic organisms. That said, the EPA
acknowledges that there are far fewer studies available on toxicity to terrestrial wildlife than are
available on toxicity to humans. The EPA will continue to evaluate the available data on
ecological effects.
Specifically with regards to fish, the draft biosolids risk assessment finds that fish accumulate
PFOA and PFOS levels into the edible portions of their body that are potentially harmful to
human health at very low concentrations of PFOA and PFOS in surface water (far lower than
currently available detection limits; see sections 3 and 4 of the draft risk assessment). Even very
low levels of PFOA or PFOS in fish tissue result in risks above the EPA's acceptable threshold
for human health (1-in-1-million cancer risk level). These same levels of PFOA and PFOS in fish
tissue do not exceed the EPA's final chronic tissue-based aquatic life criteria for PFOA and
PFOS.
While the EPA agrees that there could be complicated ecosystem responses to chemicals, the
reviewer has not provided any citations to quantitative data of these effects for use in this
assessment of PFOA or PFOS. As described previously, the EPA finds that currently available
data indicate that human health assessment will result in lower (more protective) media
concentration thresholds than the terrestrial or aquatic ecological assessments. The EPA will
continue to monitor the literature for ecotoxicity studies, particularly for plants and wildlife.
30
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
Reviewer 4
COMMENT: No comments.
Reviewer 5
COMMENT: The decision to limit the assessment of ecological risk and focus primarily on
human health in the risk assessment of PFOA and PFOS is based on the observation that the
toxicity thresholds for these chemicals are significantly lower in humans than in aquatic and
terrestrial life. Specifically, the concentrations of PFOA and PFOS that pose risks to human
health, particularly through ingestion, are much lower than the levels that cause adverse effects
in plants, invertebrates, fish, and birds. For instance, soil and water concentrations protective of
human health are orders of magnitude more stringent than those required to protect ecological
systems. Given the more potent nature of PFOA and PFOS toxicity in humans, the EPA has
chosen to prioritize human health endpoints in the biosolids assessment. This approach is
conceptually and technically justified because protecting human health at such low thresholds is
likely to also provide adequate protection for ecological receptors, even though the exact risk
levels for wildlife may still need further investigation. This strategy ensures that human health is
not compromised while still offering a degree of protection to the broader environment.
RESPOSE: Thank you for your comment that the EPA's approach is conceptually and
technically justified, even with the understanding that risks to wildlife may require further
investigation and research.
31
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
QUESTION 4: Pathway evaluation, model selection, model parameterization, risk estimation
and risk discussion
32
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
Pathway evaluation. model selection. model parameterization. risk estimation and risk
discussion. The draft risk assessment shows that human health risks associated with drinking
water and diet are expected to be the greatest among modeled scenarios and pathways. Please
comment on the technical basis and clarity of the following elements in supporting this
interpretation. Be as detailed as possible about deficiencies and suggested improvements:
a. Models and parameters selectedfor these scenarios and pathways, including comment
on the inclusion of relevant scenarios (e.g., pasture farm) and pathways (e.g.,
groundwater, fish consumption) - see Sections 2.9.2, 2.9.3, Appendix B and Appendix
Reviewer 1
COMMENT: The exposure scenarios are reasonable given they pertain to the individuals most
likely to be exposed to PFOS/PFOA present in biosolids, i.e., individuals living on farms where
biosolids have been applied.
When reading the section on transport models, I was confused as to what approach was selected.
On page 39, the text states that "no currently available transport models reliably predict the
timing of PFOA and PFOS impacts to groundwater after surface application" and that
"Consistent with previous sewage sludge risk assessments, this assessment will consider the peak
groundwater concentrations when calculating risks". However, the following sections describe
models (e.g., EPACMTP) that were used to calculate transport. I think the overall process could
be clearer.
RESPONSE: Thank you for your comment. The EPA has revised the text for improved clarity
(see Section 2.9.2). There are several aspects of subsurface fate and transport models that the
EPA evaluated when deciding which model to use, including how well the model captures peak
concentration in the aquifer and how well the model captures the timing of the chemicals
arriving to groundwater. Based on this assessment of available models, the EPA concluded that
the EPACMTP model was the best available for this portion of the draft risk assessment.
COMMENT: Although the data on parameters like BCFs and BTFs are relatively sparse, the
values selected for the models seem reasonable. However, I'm unsure the 10-year exposure
period is conservative enough for farm families who could potentially be exposed for longer.
That being said, estimated risks would be more elevated if a longer exposure period is used,
which wouldn't change the conclusions.
RESPONSE: The EPA disagrees that bioconcentration factor (BCF) and biotransfer factor (BTF)
data are sparse for PFOA and PFOS. While additional scientific data could be useful to improve
the understanding of plant and livestock uptake, the quantity and quality of studies measuring
PFOA and PFOS uptake into plants and livestock are relatively robust. Empirical data were
available for all of the uptake and accumulation parameters (BAFs, BTFs, BCFs) used for
calculations in this assessment, which distinguishes this assessment from screening assessments
that often included modeled or estimated parameters. The number of fish BAF studies is also
large compared to other chemicals.
33
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
The EPA agrees that many farm families reside on their property for more than ten years;
however, as noted by the commentor, a modeling scenario with a longer duration of exposure on
the farm would likely increase the risk findings of this assessment and therefore not change the
conclusions. The EPA selected a residency time of ten years consistent with the goal of targeting
a "central tendency" exposure scenario.
COMMENT: Calculations were made using a 1 ppb PFOS/PFOA. Measured concentrations in
biosolids often far exceed this value. Is the relationship between starting concentrations in
biosolids and risk estimates linear? In other words, would risk estimates be 5 times higher if the
starting PFOA concentration was 5 ppb? Adding this information would help readers estimate
the risks at higher starting biosolids concentrations (e.g., average concentrations in Maine).
RESPONSE: Yes, the models currently have a linear relationship between the starting
concentration of PFOA and PFOS in sewage sludge and the calculated risk levels for each
pathway. As you describe, this means that a sewage sludge with a starting concentration of 10
ppb PFOA or PFOS would have a final risk level (hazard quotient or cancer risk level) 10 times
the reported value in this draft risk assessment. The EPA has added a description of this linear
relationship to the draft risk assessment (see Section 3.1 in the draft risk assessment).
COMMENT: Lactational exposure was not included in the assessment. Breastfed infants in farm
families could potentially be receiving a much higher daily dose than their parents. I understand
including breastfeeding as a route of exposure would add a layer of modeling (pharmacokinetic)
in this assessment, and results would be difficult to interpret in this context, but I was surprised
to see that only a small section (page 110) is dedicated to this. If the EPA decides not to include
this route of exposure, I suggest adding a strong rationale for not doing so in one the previous
sections.
RESPONSE: As the commentor mentions, this topic is currently described in Section 4.8 of the
draft risk assessment. Because the draft risk assessment already finds that there are unacceptable
risk levels using the lower drinking water intake value for non-pregnant or lactating adults, it was
not necessary to add additional risk tables with risk results calculated specifically for pregnant
and lactating women. Instead, as described in the draft risk assessment, risks to this population
would be 14-71% higher than the risks presented for the drinking water pathways in each
scenario. The EPA agrees that this population is important to consider when weighing risk
reduction options and communicating with the public.
Reviewer 2
COMMENT: Pasture and crop farms are well visualized with appropriate modifications of
scenarios. Land reclamation is modeled similarly to Pasture Farm but at a higher level of
application.
The High-end Deterministic: The High-end Deterministic Risk Screening tool is very, very
conservative using a series of 95th percentile estimates and exposure. As noted on Page 37 Line
8- "could result in excess risk". One may argue that there is more risk estimates are possible if
one were to perform Monte Carlo models, but such seems unlikely given the compounding of
low-probability, i.e., 95th percentile, values for several parameters. I disagree with "reasonable
34
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
maximum exposure" argument, however. I would think that compounding a 95th percentile
concentration with a 2.4L/day intake rate, also a 95th percentile, puts one at something on the
order of a 99th to 99.9th or even higher percentile level.
RESPONSE: The commenter appears to misunderstand the Biosolids Tool (BST) and how the
results are interpreted for the purpose of screening. The purpose of the BST is to prioritize
compounds for further risk assessment. Choosing conservative values for parameters like the
biosolids concentration and the exposure factors is appropriate for this goal.
COMMENT: Central Tendency Deterministic: The Central Tendency Deterministic modeling
approach offers a more realistic assessment of the exposure and risk than does the High-End
Deterministic approach. However, it gives no real assessment of the risk in context of a
population. Data showing that such an approach does lead to an estimate of, say, the median
exposure and risk to some (or the general) population would be more compelling. The argument,
however, is made heuristically, which is, of course, a weaker argument. Additional thought and
justification for these assumptions is warranted.
RESPONSE: By parameterizing models with median values for each scenario, it is expected that
the model will output a median risk level for the target population in each scenario. Section
2.9.3.8 provides more information about the targeted population. Importantly, this biosolids risk
assessment is primarily focused on a potential risks to a farm family because that population is
likely to have the highest exposure to PFOA and PFOS from land applied biosolids. This
approach is consistent with past agency practice when conducting biosolids risk assessments
under CWA section 405 (US EPA, 1992; US EPA, 2003).
Reviewer 3
COMMENT: The selection of independent models for the PFOA and PFOS assessment is highly
appropriate for addressing the complex fate and transport of these contaminants. By employing a
series of specialized models tailored to different environmental scenarios—such as crop farms,
pasture farms, reclamation sites, and surface disposal sites—the EPA ensures a comprehensive
evaluation of how PFOA and PFOS move through various environmental media. The initial
models focus on the sorption and movement of these chemicals through soil and groundwater,
which is crucial for understanding their potential pathways and concentrations in different
settings. Subsequent models that track runoff, erosion, and leaching, and estimate concentrations
in surface water and groundwater, provide detailed insights into the dispersion and potential
impact on drinking water sources. The final step, using uptake factors to calculate concentrations
in food products and subsequently assessing risk to human health, ensures that all relevant
exposure pathways are considered. This multi-tiered approach allows for a thorough assessment
of PFOA and PFOS across multiple environmental compartments and human exposure scenarios,
providing a robust framework for evaluating risks and informing regulatory decisions.
The approach described is highly appropriate and adequate for the risk assessment of PFAS.
Prioritizing exposure factors specific to home-produced foods for agricultural site models, and
using the most current data from the EPA's Exposure Factors Handbook, ensures that the
35
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
assessment accurately reflects real-world conditions and exposure scenarios for farm families.
Additionally, employing regionally representative parameters and default values from peer-
reviewed EPA models enhances the reliability of the environmental fate and transport
predictions. This thorough and up-to-date methodology provides a robust framework for
accurately assessing PFAS risks, accounting for both specific and general environmental and
exposure factors.
The selection and parameterization of the models for assessing PFOA and PFOS risks are
executed with exceptional precision. The incorporation of specific fate and transport
considerations, along with well-chosen models for surface water, groundwater, air dispersion,
and plant and animal uptake, demonstrates a comprehensive and detailed approach. The model
parameters, including toxicity values, sewage sludge concentrations, physical and chemical
properties, and various uptake factors, are thoroughly vetted and aligned with the latest research
and data. Additionally, the careful selection of studies to inform these models ensures robustness
and reliability. This meticulous approach leaves no room for further comment on this section, as
it is exemplary in its execution.
RESPONSE: Thank you for your comment.
Reviewer 4
COMMENT: No comments.
Reviewer 5
COMMENT: The selected modeling approach involves using several independent models to
estimate the concentrations of PFOA and PFOS across different environmental media, such as
soil, surface water, and groundwater, within various scenarios (e.g., crop farm, pasture farm,
reclamation site, and surface disposal site). The scenario about sewage sludge incinerators is
qualitatively discussed due to the uncertainties around PFOS and PFOA destruction in
incinerators. The first two scenarios are discussed in more detail because they represent the
higher exposure dosage from a human health perspective. This rationale is justified. Section
2.9.2 then walked through different PFOA and PFOS fate and transport models, including
surface soil, surface water, groundwater, air dispersion, and plant and animal uptake. The choice
of using existing EPA models is appropriate, such as using EPA's 3MRA model to assess soil
surface processes and using EPA's VVWM model to assess surface water processes. The plant
and animal update processes are described with several equations multiplying together soil
concentration and bioconcentration factor. I don't have any critique for the equations used for
plants and fish. However, for livestock, are we assuming the ingested PFOS/PFOA will be
absorbed at 100%? Usually, there is a bioavailability factor for the amount of PFAS livestock
can absorb from feed as well.
RESPONSE: As stated in the draft risk assessment in Section 2.9.3.6, "There are no data
available on PFOA and PFOS bioavailability to livestock specifically from feed, water, or soil;
this assessment assumes 100% is available when orally ingested." Though there are no studies
available that specifically study the bioavailability of PFOA and PFOS in feed, this assumption
is supported by closely examining the studies used understand accumulation in livestock. The
studies used to derive BTFs for livestock include a variety of exposure scenarios for the
36
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
experimental animals. In some cases, the animals are exposed through water only, in other cases
the animals are exposed feed only, and in other cases the animals are sampled from a pasture
farm where they have exposure from feed, water, and soil. When comparing the PFOA BTFs
dervied for chicken eggs from the Wilson et al. 2020 study (animals exposed only through water)
and the Kowalzak et al., 2020 (animals exposed only through contaminated feed), the calculated
BTFs are nearly identical. This indicates that if there is a reduced bioavailability of PFOA in
chicken feed, that effect is likely negligible. In the case of dairy cows, the BTFs selected for this
study (from Vestegren et al., 2013) were derived by calculating the exposure from feed and
water combined. If there were a reduced bioavailability of PFOA or PFOS in feed in dairy cows,
this would already be factored into the BTF calculation. For beef cattle, the BTFs were also
derived using data from pasture-fed cows (Vestegren et al., 2013 and Drew et al., 2021), so these
factors also inherently consider differences in bioavailability between feed and water in the
calculated values. Note that part of the reason previous assessments included assumptions about
reduced bioavailability in feed compared to water is because the BTFs in these assessments were
modeled, not measured. By using BTFs derived from empirical experiments with multiple
sources of livestock exposure, the uncertainty regarding bioavailability across livestock exposure
pathways is reduced or eliminated.
COMMENT: In Section 2.9.3, the process to parameterize the models is explained. I agree with
prioritizing existing assessments or reports from the EPA or another agency over individual peer-
reviewed studies. The choice of using RfD from the most recent final human health toxicity
assessments for PFOA and PFOS is appropriate. It is also very encouraging to see that EPA was
able to obtain all the parameters from the first two higher tiers of the data hierarchy focusing on
biosolids applications, this helps with strengthening the relevance of the parameters to this risk
assessment.
RESPONSE: Thank you for your comment.
b. Modeling of groundwater behavior for PFOA and PFOS. (see Section 2.9.2 and
Appendix C focused on groundwater modeling).
Reviewer 1
COMMENT: The modeling presented in this section is beyond my expertise. Nevertheless, I
have found the text to be clear and understandable for a reader like me.
RESPONSE: Thank you for your comment.
Reviewer 2
COMMENT: My expertise in groundwater modeling is limited. For example, I am not familiar
with the EPACMTP modeling system as the work I have done has used PRZM for vadose zone
transport and MODFLOW for groundwater. That being said, my brief review of EPACMTP
suggests a more simplified model when compared to MODFLOW with fewer parameters that
37
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
may be appropriate for this screening level survey. It is my experience that more parameters
introduce more uncertainty due to the uncertainty in each parameter compounding the whole.
RESPONSE: Thank you for your comment. The EPA agrees that MODFLOW is a more
complicated flow model than the EPACMTP. MODFLOW has many more parameters and given
that this risk assessment is modeling hypothetical farms in various regions of the U.S., rather
than actual sites with a large quantity of measured site-specific field data, EPACMTP is a more
appropriate model for this work. Additionally, EPACMTP is more appropriate than PRZM for
vadose zone modeling because EPACMTP is more flexible to the conceptual models used in the
biosolids risk assessments. For example, EPACMTP allows for the drinking water well to be
placed away from the field where the application occurs, while PRZM requires that the well be
located on the field.
Reviewer 3
COMMENT: The approach outlined for assessing PFOA and PFOS risk through the use of the
3MRA and EPACMTP models is methodologically sound and reflects a thorough understanding
of the contaminants' behavior. The 3MRA source modules effectively capture the maximum
mass flux of PFOA and PFOS from the top layer of soil or surface disposal units during the
application period, providing a comprehensive estimate over a 150-year timeframe. This long
modeling period ensures that even delayed or residual leaching is accounted for.
The subsequent use of EPACMTP to model the transport of PFOA and PFOS through the vadose
zone and into groundwater is also appropriate. By considering the variability in vadose zone
depth and acknowledging the unique challenges posed by PFOA and PFOS—such as their
surfactant properties and interaction with soil minerals—this approach demonstrates a nuanced
understanding of these substances. However, the fact that EPACMTP has not traditionally been
parametrized for air-water interface effects suggests a potential limitation. Given PFOA and
PFOS's distinct behavior at this interface, it might be prudent to evaluate whether additional
model adjustments or supplementary methods are needed to fully capture these effects. Overall,
while the approach is robust, attention to the unique properties of PFOA and PFOS could further
refine the risk assessment.
RESPONSE: Thank you for your comment. The EPA includes a discussion of air-water
interphase effects in Appendix C, but new research continues to be published on this topic. The
EPA will continue to evaluate sub-surface transport models to describe PFOA and PFOS
contamination of aquifers as this risk assessment is finalized and acknowledges that EPACMTP
was not designed to capture air-water interface effects. The EPA will also investigate which
modeled environments are most impacted by air-water interface effects.
COMMENT: The procedure described in the model implementation is generally a good
approach for protecting groundwater resources, particularly in the context of a national risk
assessment. The assumptions made, such as placing drinking water wells at the center of the
buffer and focusing on the highest concentration areas of the groundwater plume, are
conservative. This ensures that even in worst-case scenarios, human health and the environment
are protected. By assuming that exposures occur during the years with the highest media
concentrations, the model effectively captures a maximum risk scenario, which is appropriate for
38
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
broad, national assessments. Although this approach may overestimate risks for individuals with
wells located on the fringe of the plume or deeper below the water table, the overarching goal is
to safeguard public health. Overestimation of risk, in this case, is justified as it ensures that
potential dangers are not overlooked, thereby protecting groundwater as a crucial source of
drinking water. However, while these assumptions are suitable for national assessments, they
might not accurately reflect the risks at specific sites where conditions differ. Despite this, the
precautionary nature of the procedure aligns with the fundamental aim of risk assessment, which
is to ensure safety and protect environmental and public health resources.
RESPONSE: The EPA appreciates the support for the strategy used in the draft risk assessment
to assess the potential for risks via the groundwater to drinking water pathway. As described
previously, the goal of this draft risk assessment is to determine if PFOA and PFOS may exist in
concentrations in sewage sludge which may result in adverse effects in humans or the
environment. Because drinking water wells may intersect with the contaminated plume and may
be present near the land-application site, the EPA finds that it is appropriate to assess levels and
risks in groundwater in this scenario.
The EPA agrees that the current risks to those living near biosolids use or disposal sites who use
groundwater as a source of drinking water are difficult to quantify at this time. Currently, the
EPA does not have data on the number of biosolids land application sites, the PFOA and PFOS
concentrations of sewage sludges that were land applied, the size of any resulting groundwater
plumes, or the number of people who are currently, or may in the future, use the contaminated
groundwater as a source of drinking water. Quantifying the size of this potentially impacted
population is outside the scope of the risk assessment.
Reviewer 4
COMMENT: Page 41: "The hypothetical drinking water well in EPACMTP is represented by
four observation locations placed at 0.5, 1.0, 1.5, and 2.0 meters below the water table to ensure
the maximum groundwater concentration is observed. "
Page 32 and 56 of Appendix C: "The well depths were limited to the top 2.0 m below the water
table (1) to be consistent with a residential well scenario (these wells are generally shallow
because of the higher cost of drilling a deeper well) and (2) to produce a conservative estimate
of risk (because the infiltration rate is generally lower than the groundwater seepage velocity,
groundwater plumes tend to be relatively shallow).
Page 45: "While these assumptions may overestimate risk to a specific person at a specific site,
they are reasonable for the purpose of risk assessment since they serve to protect human health
and the environment."
This reviewer appreciates the complexity of the groundwater calculation and the difficulties in
formulating and applying a representative modeling approach, particularly with regards to
concepts of being conservative in certain calculations with the intention of protecting human
health and the environment. However, the goal of this groundwater pathway calculation is to
identify the potential scope and magnitude of risks under different biosolids use and disposal
scenarios on a central tendency (median) risk basis. Applying an overly cautious approach even
39
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
in one step of the calculation may greatly overestimate risk and jeopardize the accuracy and
applicability of the central tendency risk assessment. Overall it is important to strive for a
balanced and scientifically accurate assessment of potential exposure and not apply multiple
"conservative" (overprediction) calculation steps.
RESPONSE: Groundwater is an important resource that requires protection. It is appropriate to
report modeled concentrations of groundwater that are within the plume originating from the
biosolids application site, even though it is possible that a well is located above or below the
region of most impacted groundwater. Of course, it is also possible that a homeowner has a well
located upgradient from the land application site or below an aquitard; building in such
assumptions that represent lower or no groundwater impacts to the risk assessment would not
advance the EPA's goals of identifying potential human health impacts. Further, the commentor
appears to assume that the concentration of PFOA or PFOS rapidly drops off below the water
table, such that the median and maximum groundwater concentration are significantly different.
The EPA has added modeled vertical profiles of groundwater concentration to the modeling
discussion that illustrate PFOA and PFOS concentrations are relatively consistent with depth,
whether the well is located 10 meters or 5 meters from the land application site. In this way, the
median and maximum concentrations of groundwater would be similar over a wide range of well
depths.
COMMENT: The method that is currently used in Section 2.9.2 and Appendix C to translate
modeled groundwater concentration in the underlying aquifer to an exposure concentration is one
such example of a highly conservative (overprediction) calculation step., e.g., the text explicitly
states this step is a "conservative estimate." This step selects the maximum modeled
concentration value of PFOS in groundwater at one offour modeled depths between the water
table and 2 meters depth in the aquifer. The same approach is used for PFOA. Unfortunately,
this conservative estimate is hydraulically incorrect and will significantly overpredict the actual
risk for four reasons described below.
First, groundwater wells screened in a permeable geologic formation (aquifer) average the
concentrations of any constituents of all the water that enters the well. It is physically impossible
for well in an aquifer to selectively remove the highest concentration water in the aquifer and
leave other lower concentration groundwater next to the well behind. Therefore, a simple
adaptation of the existing approach is to average all of the modeled concentrations across the
entire well screen and vertical capture zone. This is a simple, more accurate, and appropriate
way to predict the impacts of groundwater plume on a water supply well and mirrors the well-
established mass flux approach that is accepted both in the scientific literature (e.g., Einarson and
McKay, 2001) and by environmental regulators (e.g., the Interstate Technology and Regulatory
Council, an environmental coalition led by state-regulators, see their mass flux guidance
document, ITRC, 2010).
To implement this screen-averaged approach, an estimated screen length for a domestic water
well is needed. Fortunately, there are commonly accepted general guidelines and empirical data
that can be relied upon to provide this screen length. The best example of guidelines is the
Groundwater & Wells 2nd edition (Driscoll, 1985), perhaps the quintessential reference for well
construction, design, and groundwater hydrology. In the "Design of Domestic Wells" section,
40
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
Groundwater & Wells specifies that "..for farm wells, the screens should be 10 to 15 ft long,
depending on the hydraulic characteristics of the aquifer and the yield requirement. " Other
guidance documents may provide other general lengths for domestic wells, but none indicate that
very short screens (a few inches) are ever used.
While conditions vary significantly across the country, Groundwater & Wells' assumed screen
length is supported by several empirical studies of domestic water well construction. Pope et al.
(2007) studied 2,846 domestic water wells in the Virginia Coastal Plain and determined the most
common (mode) screened interval was 10 feet long and the median screened interval was 15 feet.
A study of 3500 domestic well logs in the Central Valley, California showed that a typical
domestic well screen interval was 8 meters long (26 feet) (Bremer and Harter, 2012).
Therefore overall for a central tendency calculation, a 10-15 foot screen length would likely be
appropriate for a domestic farm well.
RESPONSE: The reviewer appears to have assumed that there is a sharp peak to the
concentrations of the chemicals in the 2.0 meters of groundwater below the water table, such that
a well would dilute contaminated water with non-contaminated water. The EPA has added a
discussion to Appendix C of the draft risk assessment that shows that the modeled contamination
through the first 6-8 m below the water table are fairly constant.
41
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
Vertical profiles of relative PFOA ground water concentrations are shown below for modeled
locations representing the wet climate (Charleston, SC), and the moderate climate (Chicago, IL)
for observations distances 5 and 10 meters down gradient from the edge of the field, and for low
Koc (left column) and high Koc (note that the saturated thickness at Charleston, SC, is 7.6 m).
These profiles show that contamination is roughly constant over the top 6-8 m of the aquifer and
the effect that the reviewer is concerned with is likely small.
o-i -i
e 2
! 4
10
0,0 0.2 0.4 O.e 0,8
RelConc
w
rr
f&
H
n
zr
m
o
Ft"
fB
n
=r
n
0.0 0,2 0.4 0.6 0.8
RelConc
Well distance from
edge of field fm)
J 5
• 10
Figure 1. Modeled Relative Concentrations vs. Depth for CROP, PFOA at Charleston and Chicago at wells located 5 m and 10 m from the edge
of the field.
42
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
The figure below presents vertical profiles of modeled relative concentrations as point
observations and as concentrations averaged over 3 m in the first 10 m below the water table
(again, the saturated thickness at Charleston, SC, is 7.6m). The dots in this figure represent
individual concentrations at a specific depth in the groundwater profile while the x's represent
the vertical averaged concentrations over ten feet. A comparison of point and averaged
concentrations only diverge once the bottom of the plume is encountered. The additional
modeling performed by the EPA indicates that averaging over a 10-foot length (< 3m) would
have little impact on the concentrations of PFOA and PFOS extracted from the groundwater well
that draws water from within the plume.
On
i—t-
iD
n
D-
QJ
fD
r-r
O
• •
m •
• •
¦
¦ ¦
¦
if)
•
¦ ¦
r~t
fD
¦ w
II
¦ :»
¦ «
n
m.
m.
M.
n'
£U
0
M
Chem
X PFOA
PFOS
0.2 0.4 0.6
Concentration
0.2 0.4 0.6
Concentration
0.8
Figure 2 Well depth below water table (m) vs. Relative PFOA and PFOS Concentrations for point observations (solid circles) and 3m (10-feet)
well screen average (cross symbol) for CROP, low Koc (left panels) and high Koc (right panels) at a well located 5 meters away from edge of
field.
The EPA again notes that the Agency has made several assumptions and used modeling
parameters that can be expected to lead to underestimation of risks. These include:
• Not summing any exposure pathways (i.e., no aggregate exposure assessment) for
the farm family and no use of a relative source contribution factor (RSC) to
43
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRJN 1763-23-1
account for other potential expsoures to PFOA and PFOS (e.g., dust, consumer
products, dietary sources off the farm, etc.);
• No consideration of PFOS or PFOA precursors in sewage sludge;
• No consideration of dose additivity of PFOA and PFOS or PFAS of other chain
lenghts (i.e., no cumulative assessment); and
• Assuming a concentration of 1 ppb for PFOA and PFOS in sewage sludge when a
central tendency value in the U.S. is likely higher.
COMMENT: Second, the existing approach assumes a well with 2-meter long screen that starts
right at the water table. Flowever, almost no domestic water wells that are constructed with short
screens directly across the water table because:
• Potential water table fluctuations forces water well drillers to place well screens well
below the water table (see left panel in the figure below from Groundwater & Wells, the
accompanying text states " The drilling contractor must insure that enough potential
drawdown is available to meet present and future yield requirements "). Configuration C
on the left panel below is described as being "constructed properly." Groundwater &
Wells suggested positioning of the screens for domestic wells, shown on the right panel
below, all have the top of the well screen well below the water table.
• Another classic groundwater publication from the USGS concurs, stating "Because
withdrawals from unconfined aquifers result in dewatering of the aquifers, wells in these
aquifers are normally screened only in the lower part in order to obtain the maximum
available drawdown. " (Heath, 1984). Water well construction guidance from USEPA
(1975) states "If the formation being screened is homogeneous and the ground water is
unconfined (water table conditions) screen the lower one third of the formation. "
md
k":
A
i
8
~
c
I
J.. 1
Normal
water table ;
h—-
...
j-
E
r
Lowered
water table
i
• hi
LJ
Figure 13.17. Adequate long-term yields ire ob-
tained by Instilling a w i ll screen of adequate length
at the proper depth. Enough potential drawdown
lust be available to meet future yield demand*.
• Drillers are also cognizant of geochemical problems that occur in wells that are screened
44
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRJN 1763-23-1
across the water table. For example Alberta's "Design and Construction of Water Wells"
states:
Ensure that the pumping water level in
the well never goes below the top of the
slot openings or perforations. This will
prevent oxygen exposure to the aquifer
which would enhance bacterial growth
and reduce well yield.
RESPONSE: The figures presented above show that there is little difference between point or
averaged concentrations from the first 2 m below the water table and the next 4 to 6 m until the
bottom of the modeled contaminant plume is encountered. While it is possible that a deeper well
would result in a lower exposure, that does not alleviate the impacts to the groundwater resource
demonstrated by the modeled contamination. Deeper wells farther down gradient from the field
may also draw from groundwater that is also contaminated (see schematic diagram figure C1-1
in appendix C).
COMMENT: Third, many domestic wells are do not extract water from the shallowest aquifer
underlying the property, a key assumption in the current groundwater exposure scenario. For
example, the USGS empirical database of domestic wells drilled in the Virginia Coastal Plain,
state states that "Contrary to widely held assumptions, only 22 percent of domestic wells in the
Virginia Coastal Plain are completed in the shallow, unconfined surficial aquifer to which the
water is returned directly by home septic systems. Fifty-three percent of the wells are completed
in six deeper confined aquifers, and the remaining 25 percent are completed in the Potomac
aquifer and confining zone, the deepest units in the confined system " (Pope et al., 2007). This
percentage will be different in different regions, but in general there are many domestic wells
that are screened in deeper aquifers and do not extract groundwater from shallow surficial
aquifers. Note that Bremer and Harter's (2012) paper does indicates significant risk of domestic
wells pumping septic tank leachate, but it focuses on overlapping sources and wells on a regional
scale, a scenario which does not seems to be the focus of this biosolids risk assessment. This
paper might be useful if more detailed, regional risk assessments are performed.
RESPONSE: While there is empirical evidence that many domestic water wells in a region may
be deeper than the those modeled here, the economics of well drilling tend to favor depths that
provide sufficient supply in productive surficial aquifers for a residence without having to reach
lower, confined units. For example, some states only require domestic wells to be greater than 10
feet below ground surface with limitations on maximum well depth based on casing size and
materials used (North Carolina Administrative Code Title 15A Subchapter 2C, Section .0107
paragraph (b) (3) and (d)). Exposure to contamination in a surficial aquifer may not occur in all
cases, however, focusing only on deeper sources of ground water ignores the potential impacts to
the ground water resource that many in the U.S. rely on.
45
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRJN 1763-23-1
COMMENT: Fourth, any partially penetrating pumping well, even one with a very short screen,
will draw water from a thicker part of the aquifer (e.g., see converging flowlines from USEPA,
2008 figure below). Therefore the water entering any well is not confined to streamlines directly
horizontal to the well screen but draws water from a larger vertical interval in the aquifer than
the well screen length as one moves upgradient.
RESPONSE: EPA agrees that an active well pulls ground water from above and below the well
screen from a capture zone. In the schematic diagram the commentor provides, where the
direction of groundwater flow is from right to left, the capture zone for a well located just
downgradient of a field with biosolids land application would include the shallow aquifer
directly below the field. This groundwater is likely to have relatively high concentrations of
PFOA and PFOS. For this reason, including consideration of the entire capture zone in our
drinking water modeling would likely not result in a lower groundwater risk finding, despite
presenting a much more complicated modeling framework. At the location of the well, the
vertical profile modeling presenting in the preceding figures indicate that there is not much
differentiation in model concentrations throughout the first 6-8 m below the water table.
COMMENT: Overall, these four factors lead to the following recommendation. For the unusual
cases where a near-water table, shallow, and downgradient well is used at a farm, the calculation
should assume that it draws water from at least 20-30 feet thickness (or the entire assumed
saturated thickness of the aquifer), accounting for 1) a typical screen length at a farm is likely to
be around 10-15 feet long; 2) the typical well is likely to screened at a minimum of at least 5 to
10 feet below the water table, and 3) a partially penetrating well that captures flow from both
above and below the screened interval. Therefore an average modeled PFOA and PFAS
concentration in a 20-30 feet thick interval (to a maximum of the aquifer thickness minus about 5
feet) in the lower part of the aquifer should be used as the exposure concentration for drinking
groundwater, not a single-depth maximum concentration in the top two meters of the aquifer.
RESPONSE: The EPA modeled saturated thickness ranging from 7.6 m at the wet climate site to
21.3 m at the dry climate site. The previous figures demonstrate that contamination of the first 6-
8 meters of ground water below the water table whether averaged or not, are very constant until
the bottom of the contaminant plume (or the bottom of the aquifer in the case of wet climate site)
are reached. Additionally, it appears that the figure included by the reviewer demonstrates that
the flow from deep in the aquifer is not entering the well, which appears to contradict the
46
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
reviewer's statement that the EPA should average over the entire "saturated thickness of the
aquifer." Ignoring the shallower contamination below the water table would ignore potential
exposures as well as impacts to the groundwater resource.
COMMENT: Page 42, Line 10: "While EPACMTP estimates arrival times of aquifer
contamination at the water table that are, in some cases, much longer than those that have been
observed at biosolids application sites in Maine and Michigan, but closer to those observed
breakthrough times than models that incorporate air-water interface effects and nonlinear
adsorption. For this reason, EPACMTP was selected as being more appropriate for modeling
vertical transport through the soil column. "
Appendix C, Page 29: "Overall, we observe that the vadose zone module in EPACMTP would
produce higher (i.e., risk-conservative) PFAS concentrations at the water table because the
model does not have the ability to address PFAS-specific retention behavior at the AWI. "
The selection of the vadose zone model (EPACMTP vs. HYDRUS vs. ANALYTICAL) is a
difficult modeling issue. On one hand one would want to include all of the PFOS/PFOA
retention processes, including air/water partitioning, to provide the most accurate modeling
results. One the other hands even EPACMTP appears to provide travel-time-to-groundwater
results in some cases that do not much observed travel times at certain field sites. The difference
between the AWI models (HYDRUS and ANALYTICAL) appear to be greatest for PFOS and
the 10-m deep water table (Figure Cl-13), but only after centuries of PFOS migration. While the
EPACMPT solution is not technically pleasing, it appears to be adequate for this limited, specific
application in the biosolids risk assessment. As indicated on page 92,1 agree USEPA should
continue to evaluate the availability of groundwater and vadose zone models as this assessment
is finalized.
RESPONSE: Thank you for your comment that the analysis presented in Appendix C is
sufficient to justify the EPA's choices for groundwater model and that the EPACMPT model is
appropriate for the context of this national draft risk assessment.
Reviewer 5
COMMENT: The technical basis of using EPACMTP to estimate the subsurface transport of
PFOA and PFOS is sound. EPACMTP has been used within EPA for a long time and has been
used for modeling the vertical transport of other contaminants through the vadose zone to
groundwater. The challenge of using EPACMTP is traditionally this model has not been
parameterized to estimate air-water interface effects, which are important for PFOA and PFOS
due to their surfactant properties. In Appendix C, EPA evaluated three models for their relevance
to PFOA and PFOS vertical transport: EPACMTP, HYDRUS ID with HD1 Pro Module, and
Guo et al (2022) model. The strengths and limitations of EPACMTP have been discussed above.
For Guo et al (2022) model, it is more specialized for PFAS and incorporates factors like air-
water interface effects and nonlinear adsorption. These factors tend to result in longer delays in
the transport of PFOA and PFOS to groundwater and lower peak groundwater concentrations.
However, it is noted that this model may overestimate the time required for PFOA and PFOS to
reach groundwater compared to real-world observations. HYDRUS model performed similar to
the Guo et al model but in one of the tests the numerical solution became unstable (10m soil
47
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
column, dry environment). The choice of EPACMTP is defensible because it provides
reasonable estimates of flow and transport under eight scenarios that reflect a broad range of
hydrogeological conditions. In addition, EPACMTP's ability to provide estimates that are closer
to observed breakthrough times at biosolid application sites in Maine and Michigan compared to
other models justified its selection.
RESPONSE: Thank you for your comment in support of EPA's decision to use EPACMTP in
this context.
c. Risk estimation and discussion, including clarity of results (see Sections 3 and 4) and
description of variability, uncertainty, and sensitivity (see Section 5, Appendix D).
Reviewer 1
COMMENT: The risk estimations are relatively clear, but I felt like the document was missing
some description/discussion. The risk calculations rely on cancer slopes and RfDs that are only
briefly described. I would suggest adding a paragraph for each of these values, briefly detailing
the critical study, modeling, uncertainty factors, limitations, etc.
RESPONSE: Section 2.6.1 includes numerous paragraphs describing the critical studies, health
effects, and relevant target populations for PFOA and PFOS. The EPA finds that this information
is the most relevant for contextualizing the results of the draft biosolids risk assessment. If the
reader seeks additional information on these toxicity assessments, they are referred to the EPA's
2024 Final Toxicity Assessments.
COMMENT: It would also be important to put results into context. For example, exceedances
were estimated for PFOA in groundwater at concentrations at or below 4 ng/L (the most recent
MCL), which is a concentration that is likely to be observed in many municipal drinking water
systems and private wells. Without estimates of aggregated exposures (and pharmacokinetic
modeling to estimate serum concentrations), it is difficult to evaluate the extent of overexposure
in farm family members in the scenario assuming 1 ppb PFOS/PFOA in biosolids.
RESPONSE: This draft risk assessment uses the RfDs and CSFs presented in the EPA's Final
Toxicity Assessments for PFOA and PFOS as a comparator to exposures from each potential
pathway of exposure. This assessment does not attempt to compare exposures to PFOA and
PFOS from biosolids to exposures to PFOA and PFOS from all other sources. Converting the
exposure values from this assessment to estimated serum levels that could be compared against
measured serum levels in the general population would be complex and is outside the scope of
this assessment.
Reviewer 2
COMMENT: In the Table found on Pages 77-78, one notes little difference between 1- and 10-
year averages (Page 78). Further, Climate partitioned in Dry and Weet, is a stronger influence,
especially Dry Climate/High Koc Groundwater (8 orders of magnitude). Can this be correct?
About one order of magnitude for Moderate and Wet Climates. Larger disparities for PFOS (29
orders of magnitude for PFOS Dry Climate) and five for moderate while only 2 for Wet.
48
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
RESPONSE: It is expected that climate would have a significant impact of the groundwater
results in each scenario. As described in Section 2.9.3.12 and Appendix B, there are many
correlated parameters in the LAU model and EPACMTP that are related to climate/modeling
location. Further, the sensitivity analysis presented in Appendix D indicates that depth from
ground surface to water table (one of the parameters that is dependent on the climate/location) is
highly sensitive in groundwater outcomes. The fate and transport of PFOA and PFOS applied to
land varies significantly across the diverse geography of the U.S.
COMMENT: Crop Farm Results:_The Media concentrations, developed in this model, which
lead directly to exposures and risk, vary over 2-3 orders of magnitude. I expect these are skewed
to the low end in real measurements, but most of the risk is in the group with substantial
exposures possibly in more than one medium. How are these combined in the Deterministic
models? It is here that Monte Carlo approaches may give more useful information.
RESPONSE: As described in sections 2.7.1 and 5.2.1, this draft risk assessment does not attempt
to aggregate risks across multiple pathways. The risks reported in tables presented in section 4
represent only the risks from the designated pathway {i.e., groundwater to drinking water). The
EPA interprets these risks to approximate median risks to the exposed population from the stated
pathway in each modeled scenario, assuming a starting concentration of 1 ppb PFOA and PFOS
in sewage sludge.
COMMENT: I am concerned about the substantial variability in the model results varying over
many orders of magnitude from essentially zero concentration and exposure resulting essentially
no risk to much more substantial values. The dependency of Koc levels and Climate chosen calls
into question the tole such modeling might have in regulation. If the model gives results that vary
across a range from essentially zero to something near or above what may be concern, one might
question their utility. Ground-truthing with more data is necessary, especially in light of the large
differences noted for Koc and the lack of good data on conditions affecting this value in site-
specific cases. Some expansion in Discussion would be warranted here.
RESPONSE: The EPA respectfully disagrees with the assertion that the modeling presented in
this assessment has limited utility because it includes results from scenarios that represent wide
variations in risks across various pathways. This modeling underpins the key conclusions of this
assessment: across a wide range of hydrogeological settings and use or disposal settings, multiple
pathways of exposure may result in exceedances of acceptable risk levels. For example, in
hydrogeological settings where groundwater risks are low (such as areas with deep groundwater
aquifers and little rain), risks through other exposure pathways (soil, surface water, fish) may be
elevated. This is important information to understand the scope and scale of potential impacts in
various regions of the US.
COMMENT: Pasture Farm Results: Similar criticisms of the Pasture Farm results as noted for
Crop Farm results apply. The authors draw parallels between other scenarios not investigated and
Pasture Farm results suggesting importance of identifying potential solutions to this problem. As
49
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
was the case for Crop Farm results, Pasture Farm results are strongly affected by assumed Koc.
A series of papers by Shin, et al., modeled fate and transport of PFOA from an industrial site in
West Virginia. In their complex model, they elected to use foe, the organic content of the soil,
which changes depending on local conditions, as a free parameter to adjust the results to
measured data. This may be generally applicable but suffers from a lack of site-specific data or
sufficient general data to use this idea effectively and may seem arbitrary. However, EPA in this
Report uses scenarios with substantial variance in Koc, which depends critically on foe. Thus the
model used here reflects the same concerns as Shin, et al., attempted to account for in their work.
EPA may wish to include discussion of ways of "fixing" some of these parameters or performing
sensitivity analyses on their effects beyond the use of scenarios.
RESPONSE: The EPA has included a sensitivity analysis, see Appendix D. To clarify, Koc does
not depend on foe. Kd is calculated using Koc and foe, as described in Appendix B and C.
COMMENT: Page 90 and 91 figures show that the model behavior displaying an roughly
exponential increase in concentrations followed by a similar exponential decay after source
removal. This is first-order Differential Equation behavior with Koc acting as a source
dampener. The detailed structure of the model changes the results some to show the minor
oscillatory behavior on the overlaying exponential, but a simple model gives results that are
qualitatively similar to the more detailed model. Exposure over any time period can be inferred
by integration of the differential equation solution to get "lifetime exposure" or exposure over an
extended period, which may be associated with risk.
The groundwater concentrations reflect the delay associated with binding in the vadose zone.
Essentially the PFAS move with different "speeds" through the vadose zone- a type of
"chromatography" again- with PFOS progressing more slowly through vadose zone. Again, this
can effectively be modeled as a first-order differential equation as might be done for "retention
time" in chromatography.
The argument is made that, because both the median and 95th percentile deterministic models
indicate unacceptable risks in many/most scenarios, it is unnecessary to perform Monte Carlo
type analysis. I think more discussion is needed in this section. However, I do agree with their
assessment that it is sufficient to perform these conservative, deterministic approaches in light of
the results suggesting concern for nearly all scenarios.
RESPONSE: Thank you for the comment that the justification the EPA provided for proceeding
with the central tendency modeling approach is clear and understandable. Comments regarding
the potential to model the fate and transport of PFOA and PFOS in soil, surface water, and
groundwater through simple mathematical equations is discussed when this issue is raised again
under "specific editorial and technical comments."
COMMENT: There is clearly an introduction of bias in the estimates of risk due to zero
concentration assumptions. Concentrations, exposures, and risk cannot be less than zero at other
50
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
locations. Thus, assumption of zero risk elsewhere can, at best, be accurate, but only in the
unlikely event that any individual encounters no PFOA/PFOS or precursors in locations other
than those associated with sewer sludge exposures. There is essentially no probability of such an
event for any individual given the ubiquitous environmental distribution of these "forever
chemicals." Additionally, lack of assessment of either aggregate exposures across multiple
sources of PFPA/PFOS or cumulative exposure from simultaneous PFOA/PFOS/Pre-cursor
exposure bias risk as well. While the authors address these issues in passing, I think the reader of
the Report would be better served if these uncertainties were given more attention, perhaps is a
separate section on such effects.
RESPONSE: As mentioned by the commentor, these items are identified as "systematic
uncertainties resulting in underestimation of risks" in Section 5.2.1. These factors are also
included in the executive summary as further support of the report's key finding that PFOA and
PFOS exposures from biosolids use and disposal actions may result in unacceptable risks. The
EPA finds that the existing discussion is sufficient for the purpose and scope of the draft risk
assessment.
COMMENT: Discussion of overestimate of risk Page 112 Line 18-24 seems to be a stretch and
quite speculative.
RESPONSE: The commentor is referring to the following text in the draft risk assessment:
"The current modeling scenario assumes that a farm will receive yearly applications of
biosolids for 40 consecutive years. This may be an overestimate of the loading for a farm,
but the EPA does not have data to indicate the frequency of application at the same site
across the country. The current biosolids regulations allow land application to happen
yearly if the amount of biosolids land applied is consistent with the nutrient needs of the
crops grown at the farm and this assessment attempts to reflect that part of the regulation.
The regulatory framework also considers that a farm may receive biosolids for up to one
hundred years."
Unfortunately, the EPA does not have data available on the typical duration of biosolids land
application at a given site. Given this lack of data, the EPA will continue modeling the duration
of biosolids land application used in prior risk assessments (US EPA 1993; 2003).
COMMENT: While the statement given in the last sentence of the first paragraph of Section
5.2.3 may, indeed, be valid, the justification for it given earlier in the paragraph is somewhat
opaque and gives this reviewer no strong sense that it might be the case beyond the assertion
made by the EPA authors. However, the EPA authors are indeed experts in this field and have
spent substantial amounts of time thinking about this specific problem. Thus, I am reluctant to
attempt to override their statements. Nevertheless, some discussion to justify this conclusion is
warranted.
RESPONSE: The commentor is referring to the following text in the draft risk assessment:
51
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
"Most of the random uncertainties included in this report stem from modeling parameters
where there are data limitations, resulting in an over or underestimation of the "true"
conditions. For example, exposure factors used in this assessment (drinking water intake,
fish intake, intake of various types of foods) are based on surveys conducted at various
times in the US. These surveys vary in sample size and methodology and may be
imperfect measurements of "true" consumption behavior. These surveys also do not
capture all potentially relevant consumption behavior, like the consumption of animal
livers, which are known to have higher levels of PFOA and PFOS than muscle tissues. As
a result, the mean or median of the survey may be over or underestimating reality.
Despite these uncertainties, the EPA believes this assessment relies on the best available
datasets for exposure factors."
The EPA considers the most recent edition of the Exposure's Factor Handbook to represent the
best available information for exposure assessment.
COMMENT: The discussion Random Uncertainties is important, but limited (Page 112 Lines
26-41). While quite brief, the authors touch upon some of the essential details. The discussion of
parameters uncertainty need not be limited to the Koc values as many of the parameters are
uncertain. For example, the acid dissociation constant for PFOA is not really calculable due to
surfactant effects, as noted early in the report. Often there is a chain of assumptions- Kow
Koc that, for there surfactants is not an easy ask. The binding to soil is likely to span orders
of magnitude depending upon environmental conditions.
The discussion on Page 110 Line 23ff is about bias rather than a strict uncertainty as exposures
cannot be less than zero and, thus, exposure s underestimated as long as there is any
PFOA/PFOS present in the individual's exposome that is not accounted for by Sludge/biosolid
exposure. One may consider it a model uncertainty, but also a parameter uncertainty. This bias is
acknowledged implicitly in PI 11/L24ff It is not clear that there is sufficient data for many
parameters to identify these uncertainties as systematic, random, or model specification. Such is
often the case in highly parameterized, complex models.
RESPONSE: Thank you for your comment. To clarify, the acid dissociation constant is the pKa,
not the Kow. Kow was not used to determine the Koc for PFOA or PFOS in this assessment.
Reviewer 3
COMMENT: This section discusses the results of modeling the concentration and exposure of
PFOA and PFOS through three individual exposure pathways within various biosolids use or
disposal scenarios. The concentrations in different media (such as milk, soil, water, and beef) are
modeled assuming an initial concentration of 1 ppb in sewage sludge. The sensitivity of these
results is influenced by climate conditions (dry, moderate, wet) and the Koc values, with
exposures presented for both low (10th percentile) and high (90th percentile) Koc scenarios.
Crop: The discussion presented in the paragraph is generally good, as it provides a detailed and
comprehensive analysis of the modeled concentrations of PFOA and PFOS in various
environmental media within the crop farm scenario. It effectively highlights the key findings,
such as the differences in concentrations across groundwater, surface water, soil, fish tissue, and
52
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
crops, while also considering the impact of different variables like climate, Koc values, and plant
uptake factors. The discussion also acknowledges the limitations and uncertainties in the data,
particularly in the uptake factors for fruits and vegetables, which adds transparency and
credibility to the analysis. However, the paragraph is quite dense, and the flow of information
could be improved by organizing the content into more clearly defined sections or bullet points
to enhance readability. Overall, the discussion is informative and well-supported by the modeling
results, but it could benefit from better structure and a more concise summary of the key points.
RESPONSE: Thank you for this comment. The EPA has revised the text in this section to
enhance readability.
COMMENT: Pasture farm: The discussion in this paragraph is well-structured and thorough,
providing a clear comparison between the pasture farm scenario and the crop farm model,
particularly in terms of PFOA and PFOS concentrations across various media, such as soil,
groundwater, surface water, and animal products. It effectively highlights the impact of the
absence of soil tilling in the pasture model, which leads to higher soil and surface water
concentrations, subsequently increasing fish tissue contamination. The analysis of exposure
pathways for dairy cows and chickens, along with the implications of different Koc settings on
contaminant levels in milk, beef, and eggs, is detailed and informative. However, the paragraph
could be improved by being more concise and focusing on the most critical findings, as it
currently presents a lot of data that might overwhelm the reader.
RESPONSE: Thank you for this comment. The EPA has edited this section for clarity.
COMMENT: Reclamation site: The reclamation scenario models the concentrations of PFOA
and PFOS in various environmental media, including groundwater, surface water, soil, fish, milk,
beef, eggs, and chicken. The scenario, which involves a single application of biosolids at a
higher rate than the pasture farm scenario, generally results in lower concentrations of PFOA and
PFOS across all media compared to the pasture farm model, which assumes annual applications
over 40 years. Groundwater concentrations are particularly low, especially in soils with low
sorption capacity, such as sandy soils or those with high pH. While most media concentrations
are below detectable levels, PFOS in fish and eggs remains consistently detectable due to its high
bioaccumulation potential. Overall, this scenario suggests that a single application of low
concentration biosolids poses minimal risk of significant groundwater contamination but
highlights the persistent presence of PFOS in certain media.
Sewage sludge disposal site: This assessment of the surface disposal scenario is very well done.
It effectively models groundwater concentrations across unlined, clay-lined, and composite-lined
disposal sites, accurately identifying the highest contamination levels in unlined sites, with
PFOA concentrations ranging from 0.024 to 25 ng/L and PFOS up to 2.2 ng/L. The analysis
correctly notes that clay-lined sites result in slightly lower concentrations, while composite-lined
sites show minimal infiltration, with PFOA concentrations up to 0.014 ng/L and negligible PFOS
infiltration. The discussion also thoughtfully considers the impact of different climate conditions
on groundwater concentrations, including factors such as water table depth, infiltration rates, and
rainfall-induced dilution.
53
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
The analysis in this section effectively outlines and categorizes the different types of
uncertainties—systemic and random—present in the assessment, providing a clear and thorough
discussion of their potential impact on risk estimations. The description of systemic uncertainties
is particularly well done, highlighting how certain assumptions, such as low starting
concentrations of PFOA and PFOS or the exclusion of precursor chemicals, could lead to an
underestimation of risks. The analysis also acknowledges how these factors might not fully
capture the complexities of real-world scenarios, such as long-term exposure or background
contamination levels.
Additionally, the discussion of uncertainties that could lead to overestimation of risks, such as
the use of greenhouse study data for plant uptake and assumptions about biosolids application
frequency, is well-articulated. The explanation of random uncertainties, particularly those related
to variability in site conditions and consumption behaviors, further enhances the credibility of the
analysis by acknowledging the limitations of the data and models used. Overall, the discussion is
very well done, and I do not have further comments on this section.
RESPONSE: Thank you for your comment.
Reviewer 4
COMMENT: Recommend that the text compares the drinking water concentration results
groundwater concentrations to the MCLs for PFOS and PFOA and explain from a risk
perspective what it means if groundwater or surface water concentration is above the MCL, and
what it means if the concentration is below the MCL.
RESPONSE: As described previously, MCLs are not health-based values. See prior responses on
this topic on pages 9 and 22.
COMMENT: For Sections 4.3, 4.4, 4.6, and 4.7, recommend that the drinking water pathways
values that exceed the acceptable threshold (yellow-shaded values) also have some type of
indication if the MCL is also exceeded.
RESPONSE: As described above, MCLs are not health-based values and are, therefore, not
appropriate comparators for this risk assessment. See prior responses on this topic on pages 9 and
22.
COMMENT: Recommend re-evaluating if lxlO"6 risk level out of the 10"5 to 10"6 range is
faithful to the goal of performing a central tendency calculation. Again, the key to this type of
calculation is not to have any intentional over-conservative (overestimation) of risk or intentional
underestimation of risk.
RESPONSE: It is the EPA's longstanding scientific judgment across its programs that, unless
data indicate otherwise, human carcinogens exhibit linear "non-threshold" dose-responses, which
means that there is no level without risk. The target cancer risk level used in determining if there
is an unacceptable risk level is a policy decision. Prior EPA risk assessments in the biosolids
program have indicated that when small populations are expected to be exposed to the chemical
54
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
of concern, it is appropriate to use a higher (less protective) target cancer risk level than 1 in 1
million (lxlO"6). For example, the Guide to Biosolids Risk Assessment (US EPA, 1995) states
that a higher target cancer risk level is appropriate if meeting that higher target would result in
"only a fraction of a person to several persons being at risk out of the total U.S. population."
Given that the population of people with exposures to PFOA and PFOS via sewage sludge use or
disposal is potentially large and the cancer slope factor is high, the EPA takes the policy position
that a 1 in 1 million target risk level is appropriate for this draft risk assessment. Using a cancer
risk level of 10"6 is health-protective and consistent with other Agency actions, e.g., national
recommended ambient water quality criteria for the protection of human health (US EPA, 2000).
Protecting public health includes protection of humans with increased cancer risk due to greater
susceptibility. The 10"6 risk level has been generally considered to be public-health protective for
a range of susceptibilities based on interindividual differences among humans, although this risk
level may not completely account for all susceptible individuals, including those with certain
diseases, genetic polymorphisms, co-exposures to other chemicals, and/or exposed during
especially sensitive life stages (US EPA, 2005). The EPA and other Federal agencies often use a
cancer risk level of 10"6 to guide the development of management actions in polices involving
cancer risk (Castorina & Woodruff, 2003). Additionally, this long-standing practice of using a
cancer risk level of 10"6 is consistent with national efforts to eliminate cancer as a leading cause
of death in the U.S. by decreasing cancer cases.1
Reviewer 5
COMMENT: Section 3.1 presents the modeled concentration and exposure results for individual
exposure pathways in each of the four scenarios. In each scenario, a table is used to summarize
the modeled concentrations in environmental media, followed by a discussion of the strength of
the evidence and source of uncertainty. In general, these environmental media concentrations are
consistent with both the mechanistic understanding of PFOS and PFOA's environmental
behavior (such as PFOA is more abundant in water, and PFOS adsorb more strongly to soil and
bioaccumulates more). The authors compared estimated concentrations across different scenarios
and provided mechanistic explanations, such as the one-time application of biosolids in the
reclamation scenario typically resulted in lower concentrations than the pasture farm scenario.
When appropriate, the authors also pointed out the source of uncertainties and areas of future
research to reduce those uncertainties, for example, the data limitation on the uptake factors of
fruits and vegetables are mentioned in the crop farm scenario, and the lack of BTF on PFOA
uptake in cows raised for beef is mentioned in the pasture farm scenario. Besides the four
detailed scenarios, the authors also discussed qualitatively how sewage sludge could impact
environmental concentrations of PFOS and PFOA in other uses of biosolids including home
gardening, low-public contact use cases, and incineration.
Section 3.2 provides information on how modeled concentration varies over time for moderate
climate scenarios. One limitation of this section is that the effect of precursors transforming into
PFOS and PFOA over time was not considered. The reason has been clearly stated and can be
understood. Due to the high sensitivity of modeled concentrations in Koc values, the authors
correctly chose to present the results for high Koc and low Koc separately. In section 3.2.3, when
the breakthrough time (the time it takes for PFOS/PFOA to reach nearby wells after biosolids
application) is estimated by the groundwater model, it is clear that they were extremely
1 See https://www.whitehouse.gov/cancermoonshot/.
55
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
overestimated by the model. I think the reason is likely because precursors are not explicitly
accounted for in the model and they play an important role in groundwater transport of PFAS.
The authors listed a few other reasons for why the model overestimated this time, such as
microbial weathering, leaching due to freeze-thaw cycles, and macropores. However, I think the
role of precursors needs to be added to page 92's discussion as well.
RESPONSE: Thank you for this comment. Though it is possible that precursors to PFOA and
PFOS may have fate and transport characteristics that lead to faster migration to groundwater,
most PFOA and PFOS precursors also contain carbon-fluorine alkyl chains of equal length or
longer than PFOA and PFOS. Because these precursors also contain long chains of fluorinated
carbons, it is unlikely that these precursors will move more quickly through the soil column than
PFOA and PFOS (see Brusseau, 2023 "Influence of chain length on field-measured distributions
of PFAS in soil and soil porewater). The presence of precursors is more likely to increase the
magnitude of the groundwater concentration than it is to increase the time it takes for PFOA and
PFOS to reach the groundwater. More data on the fate and transport characteristics of PFOA and
PFOS precursors in the vadose zone would be useful in better understanding their transport
behaviors.
COMMENT: Section 4 presents quantitative results from the risk characterization for the four
scenarios, and then discusses qualitative considerations for the other scenarios. The lifetime
cancer risk is calculated by multiplying the lifetime average daily dose with the cancer slope
factor. For non-cancer risk, the average daily dose is compared to the reference dose. I
understand the need to separately show each exposure pathway so that their individual
contributions can be made clear, however, I don't think it is prudent to compare the reference
dose for the aggregate exposure to the average daily dose of each exposure pathway, because this
could lead to under-estimation of the risk. Take the table on page 96 as an example, under the
high Koc and dry climate conditions, for adults even though the HQ for individual exposure
pathways does not exceed one, the sum of them has exceeded one, representing considerable
health risks. Section 4.9 listed reasons why EPA is not conducting additional modeling exercises
at this time, which I completely agree with. I think from the central tendency modeling results, it
is clear that considerable health risks from biosolids application exist, and therefore the more
urgent next step is to identify actions for mitigating such risks.
RESPONSE: Thank you for your comment. The EPA agrees that when exposures are added
from multiple pathways, risk increases. The EPA also acknowledges that there are scenarios in
which one person could be exposed through multiple pathways. For example, some farm families
with biosolids land application on their property may be largely self-sufficient, sourcing nearly
all of their produce, animal products, and water from their property. These families would have
biosolids-related exposures from multiple pathways. Other farm families may source some, but
not all, of their food products from potentially contaminated sources {i.e., they may drink milk
from cows on their farm, but not grow any vegetables or fruits impacted by biosolids land
application). Still more individuals may be impacted by a single pathway of biosolids-related
exposures, such as a person who fishes from an impacted waterbody but has no other sources of
biosolids-related exposures, or an individual whose drinking water source is impacted, but
otherwise sources food from non-impacted sources. By presenting risks for each pathway, it is
easier to conceptualize risks to other populations with biosolids-related exposes from one or
56
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
more pathways of exposure. For example, by presenting exposures by pathway, risk levels are
presented that can be used to estimate the size of the populations who may have drinking water
exposures of PFOA or PFOS from biosolids-impacted groundwater, but no other biosolids-
related exposure. This universe of potentially exposed individuals is larger than the population of
farm families living on farms with biosolids land application. Because there are potential
exceedances of acceptable risk levels in individual pathways, it is important to consider the total
population with biosolids-related exposures, acknowledging that individual exposures within this
group will vary.
COMMENT: Section 5 discusses variability, uncertainty, and sensitivity. It also compared
modeled results with observed results in three states. I think organizational-wise, section 5.3
should be its own section as it does not fit in the discussion of model uncertainties.
RESPONSE: The EPA agrees with this suggestion and has reorganized this section of the draft
risk assessment.
COMMENT: The discussion of variability in section 5.1 is very general and does contribute
much information, so I would suggest removing this section.
RESPONSE: The EPA believes that discussion of variability is important to include even though
data are not available to quantify the variability expected that this time.
COMMENT: Section 5.2 organizes model uncertainty into systemic uncertainty (which then
divides into uncertainties that result in underestimating and overestimating risks) and random
uncertainty. The authors had a comprehensive discussion of the uncertainties that may result in
underestimating the risks. As mentioned above, I think the last point about this assessment does
not quantify aggregate exposures is not an uncertainty, but more of a modeler's choice. EPA has
been using aggregate exposures for other chemicals in the past, which is why the relative source
contribution term is coined. I would encourage the authors to think about developing RSCs for
the four quantitative scenarios as the average daily doses from different exposure pathways have
been estimated.
RESPONSE: The EPA has edited this section of the draft risk assessment to better explain the
rationale for presenting individual risks per pathway, which includes added discussion of the
uncertainties and variabilities that we believe exist around knowing how many pathways are
relevant to various potentially-impacted populations. Due to the fact that single pathways of
exposure may result in exceedances of acceptable risk levels and that there are currently
unknowns in the scope and scale of the potentially impacted population, the EPA finds that
presenting pathway-specific risks is this most efficient way of presenting risks at this time, even
though this presentation does not quantify the total risks to people exposed to multiple sources of
biosolids-related exposures.
d. Comparison of modeled results to biosolids investigations conducted in Michigan and
Alabama (see Section 5.3).
Reviewer 1
57
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
COMMENT: The document indicates there is no complete dataset to evaluate modeled
concentrations in different media after biosolids application. In the absence of adequate data to
evaluate the model (e.g., PFAS concentration in biosolids and application rate), EPA used
incomplete input data from Michigan and Alabama and compared modeled to measured
concentrations. These comparisons provide some qualitative indication that the model generates
reasonable outputs (e.g., PFOA being more mobile in water than PFOS, PFOS more likely to be
detected in milk than PFOA). Quantitatively, the comparisons suggest that modeled
concentrations are in the same ballpark as measured concentrations, but uncertainty is
substantial. Overall, these comparisons provided some support for the modeling described in the
document, but it is unclear whether the model is accurate at sites with lower sludge
concentrations (e.g., 1 ppb) for which data is unavailable.
RESPONSE: Thank you for this comment. The EPA has added a discussion of a biosolids
investigation in Ottawa, Canada, which had a land application scenario with PFOA and PFOS
concentrations closer to those modeled in this assessment (see section 6.1 in the draft risk
assessment).
Reviewer 2
COMMENT: Decatur, Alabama values are close to what is modeled and are likely sufficiently so
to in some sense "validate" the screening-level model. The Michigan modeling results, although
perhaps giving modeling results a bit further from these measured values is also of utility.
However, the data are very limited. Only two locations are noted and the monitoring occurs over
only a brief time. The modeling was done to simulate many years- up to 1000 years in some
cases- with assumptions made that parameters do not change. Further, the discussion regarding
these differences would be of interest to the readers of the Final Report.
RESPONSE: Thank you for your comment. The EPA has added an additional case study to this
section and additional discussion on the differences between the case studies and the scenarios
included for modeling in the draft risk assessment.
Reviewer 3
COMMENT: The comparison of modeled results to biosolids investigations conducted in
Michigan and Alabama is very well described and thoroughly executed. From 1990 to 2008, the
Decatur Utilities Dry Creek WWTP in Alabama processed wastewater containing PFAS from
local industries, resulting in significant contamination across approximately 2,000 hectares of
farmland. Studies by 3M and the EPA revealed substantial PFAS contamination in groundwater,
surface water, and soils, with PFOA being more mobile in water and PFOS more strongly bound
to soils. The observed trends, such as PFOS being more likely to be detected in milk than PFOA,
align with the higher uptake factors modeled for PFOS. The assessment's modeled results
suggest that if biosolids applied at these sites had similar PFAS concentrations to those reported
in the 3M study, the expected contamination levels would be within observed ranges for PFOA
but higher for PFOS, highlighting the complexities of accurately modeling PFAS contamination.
58
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
In 2018, a similar investigation in Michigan revealed high concentrations of PFOS (2,150 ppb)
and lower levels of PFOA (1-5 ppb) in biosolids applied to several land application sites.
Subsequent sampling showed significant contamination in soil, surface water, and beef tissue,
with PFOS levels in soil ranging from 2,480 to 96,700 ppt and PFOA up to 1,530 ppt. Although
the observed PFAS concentrations in Michigan were higher than expected, likely due to the
higher concentrations of PFAS in the biosolids, the study's findings align broadly with the
modeled scenarios. This comparison demonstrates that, despite some uncertainties, the models
provide a reasonable approximation of real-world contamination scenarios, effectively capturing
the challenges in assessing the long-term environmental impacts of PFAS in biosolids.
RESPONSE: Thank you for your comment that the comparison of modeled and observed
concentrations was effective and informative.
Reviewer 4
COMMENT: Recommend comparing the PFOS/PFOA biosolids risk assessment results with the
dioxin sewage sludge risk assessment, which concluded the regulation was not warranted. I
believe the cancer slope factor for 2,3,7,8-TCDD is much higher than PFOS or PFOA. What are
the significant differences in risk, and if so, are they due to the starting concentrations,
toxicology, fate and transport, or exposure factors?
RESPONSE: See prior responses to comments regarding the differences between dioxins and
PFOA/PFOS for biosolids risk assessment.
Reviewer 5
COMMENT: Section 5.3 compares modeled results to real-life observations of PFOA and PFOS
in Decatur, Alabama, and Wixom, Michigan. I would recommend dropping section 5.3.3 because
there are no results available from Maine to be discussed.
RESPONSE: Given that there is widespread awareness of agricultural contamination in Maine,
the EPA finds that it is important to explain that the site-specific data from these sites are not
publicly available at this time.
COMMENT: This ground-truthing exercise is helpful to support the range of the model
estimates with observational data. When interpreting the results, it is important to keep in mind
that these two cases represent the high-end contamination scenarios and there is considerable
uncertainty around the concentrations of PFOA and PFOS in the biosolids being applied. Due to
the retrospective nature of the comparison, the authors could not find all the data they needed to
compare model estimates to the various studies of PFOA and PFOS impact around biosolids land
application sites. For the Alabama example, they drew some general observations but, in my
opinion, these trends are so general that you don't need to go through a sophisticated modeling
exercise to know, such as "PFOA is more mobile in water and PFOS is more strongly sorbed to
soil". For the Michigan example, some more quantitative comparison was made between the
observed data and the model estimates, but since there is a considerable difference in the
biosolids PFAS concentration and the application practice, the comparison reads very hand-
wavy. I would encourage the authors to dig a little deeper and think about what other interesting
59
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
comparisons can be made. One idea is to look at the relative ratios of environmental
concentrations in different compartments such as soil, water, milk, etc., and compare those ratios
between the real-life observations and model estimates. This will generate insights into the
relative importance of various exposure pathways.
RESPONSE: The EPA agrees that it is difficult to compare the data presented on existing
biosolids site assessments for PFOA and PFOS due to missing information. Though the
suggestion to compare modeled and measured ratios of environmental concentrations is an
interesting one, this approach would also be plagued by site-specific conditions potentially
dominating the observed conditions. Below are sets of plots showing the relative non-cancer risk
contributions for PFOA for each pathway of exposure, as calculated for scenario, climate, and
Koc setting (low vs high Koc conditions). These plots show that the percent contribution of risk
for each pathway varies significantly by Koc condition, and also varies per scenario and climate
setting. Lacking information on soil composition and sorption potential in each of these existing
sites, it is difficult to use our model to understand relative ratios of concentrations in various
media.
60
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
PFOA. high Koc. relative non-cancer risk contribution
PFOA, low Koc, relative non-cancer risk contribution
Crap
Pasture
• •
• •
• <
Reclamation
Pathway
5
Dry
Beef
Chicken
EflflS
Exposed fruit
Exposed vegetable
Fish
Groundwater
Milk
Prote ctect fruit
Protected vegetable
Root vegetable
Soil
Surface water
Crop
Pasture
)•
Reclamation
A¥|
Pathway
Dry
s
Beef
Chicken
Egss
Exposed fruit
Exposed vegetable
Fish
Groundwater
Milk
Protected fruit
Protected vegetable
Root vegetable
Soil
Surface water
Wet
61
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
Specific editorial and technical comments
62
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
III. SPECIFIC OBSERVATIONS
Reviewer 1
Pg.
Paragraph
Comments or Questions
Response
Sum.
4
The summary states that "that there are
significant human health risks derived from
land-applying sewage sludge that contains 1 ppb
of PFOA and PFOS". The use of "significant"
lacks specificity, and could be confused with
statistical significance. An option would be to
describe risks as "human health risks above what
is considered to be acceptable".
The EPA agrees and has made these edits.
1
4
The usual terminology for PFAS is per- and
polyfluoroalkyl substances
The EPA agrees.
15
2
The long half-lives of PFOA and PFOS are also
due to reabsorption in the kidney. I would
suggest adding a sentence or two describing this
process.
The EPA agrees and has added this information.
15
3
There are some data on serum PFAS
concentrations in individuals living on farms
where biosolids have been used
(httDs://Dubmed.ncbi.nlm.nih.gov/38941944/). I
suggest adding some information in this section.
The EPA has added this information.
18
3
When discussing sufficient exposure conditions
(sentence on lines
13-16), the data for humans should be presented
as "serum concentrations" rather than "doses".
The EPA agrees and has made these edits.
18
3
I would suggest adding accelerated puberty and
altered ossification in the sentence on lines 28-
30 given that these were used for the previous
PFOA risk assessment.
See discussion on these endpoints in the PFOA
Toxicity Assessment, pg 5-12: "For the current
assessment, EPA preferentially selected endpoints for
which there were a greater number of studies supporting
the observed effect. For example, for the 2016 PFOA
HESD, EPA derived a candidate RfD based on the co-
63
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
Pg.
Paragraph
Comments or Questions
Response
critical effect of accelerated male puberty reported by
Lau et al. (Lau et al., 2006). Results of the current
assessment's literature search showed that no high or
medium confidence studies supporting that observed
effect have been published since 2016. As Lau et al.
(Lau et al., 2006) was also the only study identified in
2016 that reported an acceleration of male puberty (a
second study reported a delay in male puberty
(Butenhoff et al., 2004a) and there were several other
developmental endpoints (e.g., reduced offspring weight
and survival, delayed eye opening) that were supported
by multiple studies), EPA did not further consider this
endpoint from Lau et al. (Lau et al., 2006) for POD
derivation in the present assessment. Similarly, upon
further evaluation during the current assessment of the
co-critical effects of reduced forelimb and hindlimb
ossification in pups reported by Lau et al. (Lau et al.,
2006), it was determined that an unexplained non-linear
dose-response trend adds uncertainty to selection of the
LOAEL as the POD. As reduced ossification was only
observed at the highest dose tested (10 mg/kg/day) by
the one other study (Yahia et al., 2010) that tested dose
levels close to the LOAEL from Lau et al. (Lau et al.,
2006) (1 mg/kg/day) and because no studies identified
during literature searches for the current assessment
reported this effect, EPA relied on other endpoints from
Lau et al. (Lau et al., 2006) that were amenable to BMD
modeling, exhibited dose-dependent response trends,
and were supported by at least one other study in the
available literature."
19
1
Same comment as above, i.e., "doses" should be
replaced by "serum concentrations" for the
human data.
The EPA agrees and has made these edits.
64
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
Pg.
Paragraph
Comments or Questions
Response
20
Section
2.6.1.3
There is some experimental data on dermal
absorption in humans
(https://pubmed.ncbi.nlm.nih.gov/36191486/). It
could be briefly summarized, although it does
not necessarily apply to dermal absorption
through soil or water.
Given that this information has not yet been
incorporated into an existing toxicity assessment from
the EPA or AT SDR, we are hesitant to reference this
single study (see the literature hierarchy described in
section 2.9). Given that this study finds low rates of
dermal absorption consistent with existing assessment
conclusions, this study will not be added to the risk
assessment.
28
Figure 2
Looks like there's a missing arrow going from
Pathway 14 to Adult, Child.
The EPA agrees and has made these edits.
31
Figure 3
Same comment as for Figure 2
The EPA agrees and has made these edits.
35
Figure 7
Same comment as for Figure 2
The EPA agrees and has made these edits.
45-46
Tables 4-5
These Tables are presented earlier in the
document.
Given that many readers flip through the assessment
sections, we find that the redundancy of key
information is acceptable in this context.
90
Figure 5 (and
other figures
reporting
water results)
I suggest using ppt (ng/L) for water
concentrations. Readers are going to be most
familiar with these units given recent guidelines.
The EPA agrees and has made these edits.
91
Figure 12
The figure number jumps from 5 to 12.
The EPA has corrected the table and figure numbers.
94
Equation Al-
24
Why are units different for the PFOA oral cancer
slope factor? I suggest using the same unit for
PFOS and PFOA (i.e., (mg/kg-day)-l) to be
consistent with the equation
The EPA agrees and has made these edits.
101
1
On line 5, fish should be included as one of the
highest risk pathways for PFOA.
The EPA agrees and has made these edits.
101
2
The last sentence of the paragraph states that
"there may be significant risk posed by PFOA
levels in milk from farms with biosolids land
application that fall below detectable limits". I
suggest toning down as exceeding an acceptable
threshold does not necessarily mean that there is
The EPA agrees and has made these edits.
65
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
Pg.
Paragraph
Comments or Questions
Response
a "significant risk", only that we don't consider
risk to be acceptable/tolerable.
Reviewer 2
Page
Paragraph
Comments or Questions
Response
Executive
Summary
Page 1
Line 20ff
I think the Executive Summary should contain some
discussion of likely effects on regulations resulting
from this document in
The goal of this document is to focus
on the risk assessment - potential risk
mitigation options will be discussed in
other documents and forums.
2
Line 18ff
Perhaps the highlighted missing discussion would
address the issue noted in the previous comment.
However, I do not know this at this time. If there were
no plans to include discussion of regulatory impact, I
would urge the authors to include one here.
The goal of this document is to focus
on the risk assessment - potential risk
mitigation options will be discussed in
other documents and forums.
13
Line 6ff
The authors discuss dividing 110 available samples
into five composite samples as done by Venkatesan
and Halden. More discussion needed as to how this
was done in order to determine whether the methods
lead to realistic assessments of concentrations in these
media. On the surface, these seems like too few
composites to give a meaningful assessment of central
tendency and variability, especially noting the
variability in these composite samples themselves for
PFOA and PFOS. More discussion may be of interest
to readers of this document. However, this reference
may itself include such discussion that could then be
alluded to.
For the Venkatesan and Halden 2013
study, the authors generated 5
composite samples by randomly
dividing the 110 available samples
from the 2001 National Sewage Sludge
Survey (NSSS). Each composite
sample encompassed 21 to 24 discrete
samples (Venkatesan and Halden,
2013). According to the Venkatesan
and Halden 2013 study, the purpose of
the technique was to create national
baseline levels for the 13 PFAS with
the composite samples, and this
methodology was utilized in previous
studies (McClellan and Halden, 2010;
Chari and Halden, 2012; Venkatesan
and Halden, 2013b). The next NSSS
that the EPA is currently planning in
66
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
Page
Paragraph
Comments or Questions
Response
collaboration with the POTW Influent
PFAS Study will focus on obtaining
current national concentration data on
PFAS in sewage sludge. Collecting
and analyzing PFAS in individual
samples of sewage sludge from 200 to
300 POTWs throughout the United
States will help in better understanding
the variability across the nation.
15
Line 33
Am I to assume that the publication is "in press" or it
would not have been referenced in an EPA
assessment?
Yes, this had been pending release
from the EPA, but is now available.
The draft assessment has been edited
accordingly.
17
Line 13
Why is "evidence indicates" bold? Emphasis or typo?
This formatting was copied from the
original toxicity assessments for PFOA
and PFOS. The formatting has
significance in these documents per the
systematic review process, so the EPA
elected to maintain the formating for
this document.
21
Line 17
Reference to EPA document missing. Highlighted.
The EPA has made these edits.
26
Lines 34-36
Repetitive with earlier text. Is this intentional so that
those reading only this section get the specific
definition of aggregate (and later cumulative)
exposure?
Yes, this information was intentionally
repeated. While there was a goal to
minimize repetition, many readers
jump around the document rather than
reading from beginning to end.
27-34
Figures
Both cartoons and flow diagrams are well visualized.
Thank you for your comment.
49ff
Lines Iff
There is uneven treatment here. This section appears
as an annotated literature review with sections
associated with each paper while other sections
discuss the references in a written-text manner. This
was a bit jarring to me as a reader.
We will edit to ensure literature is
cited in a consistent fashion.
67
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
Page
Paragraph
Comments or Questions
Response
78
Table
There is a large range of media concentrations. This
suggests to me there would be substantial difference
in exposures and thus risks associated with various
potential media -> exposure dose effects
continuum. Is this discussed?
Koc is a big factor that likely depends on various
other hard-to-measure factors. This will contribute to
model uncertainty.
Discussed in response to question 4.
81
Table
Groundwater and Chicken vary substantially by
Climate. Some explanation of this "correlation" is of
interest. Does it come later?
The EPA has added discussion on this
topic.
85
Line 23
There is a high degree of uncertainty in uptake rates of
PFOA and PFOS due to lack of data. This is a
substantial data gap for this modeling exercise. One
could argue that this, and potentially other data gaps,
calls into question the results. Can some sensitivity
analyses be done to assure the reader that this effect is
negligible or important in this analysis? This can
guide further data collection.
It is not clear if the commentor is
referring to the uptake into edible
plants, feed plants, livestock, or fish.
All of these factors vary linearly with
risk. See Appendix D for the
sensitivity analysis.
88
Line lOff
I am concerned about the "known artefact in the
numerical modeling of 3MRA's Land Application
Unit..." It would seem that this artefact needs to be
explained more fully. From what I can gather, it is
caused by the shutoff of land use of sludge at this
time. However, some of the modeling work goes out
as much as 1000 years. The assertion that this artefact
has negligible effect on the risk calculations, but I do
not see a defense of the assertion here, which gives me
pause.
The existing text explains clearly the
cause of the "artifact" and why it does
not influence the final risk
calculations.
88
Line 18
Is this not simply "chromatography" in the soil
column with the end of the "column" being
The reviewer is referring to the
following sentence in a paragraph
68
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
Page
Paragraph
Comments or Questions
Response
groundwater, the media of interest, or the time scale of
interest?
describing the mechanics of the soil
column model used in this assessment:
"The advective component of the
transport equation moves contaminant
mass down to the next layer (and
ultimately, out the bottom of the LAU)
at discrete time intervals equal to the
time it takes for dissolved
contaminants to traverse a layer via
convective transport." This sentence is
discussing the details of how the model
manages advective vs convective
transport of material over time, which
is not analogous to chromatography.
89-90
Figures
I have a thought here: These figures all look like
exponential growth towards a saturation level
followed by exponential decline after source is
"turned off'- a simple time-dependent solution to a
first-order differential equation. From the point of
view of a screening tool, would it not be similar to just
assume such a model? The little wiggles, etc., would
not appear to be major perturbations that would
change the risk estimates in a manner that would
affect the conclusions of little long-term risk, or
higher risk for certain choices of parameters. Occam's
razor may suggest this; the simplest model sufficient
for the purpose is likely the best.
It is not possible to predict the
outcome of the complex set of fate and
transport models simply by observing
that sometimes, the output of these
complex models takes the shape of a
common mathematical form. The
models selected for this draft risk
assessment are the standard peer-
reviewed models for these scenarios
used by the EPA. This draft risk
assessment is not a screening
assessment.
89-93
I have another thought here.: Given the large time
span modeled, The 40-year offset depositing PFOA
and PFOS coupled with the greater Koc value for
PFOS is "lost" in the "mists of time." The curves are
very similar, just offset by about 80 years which has
This draft risk assessment displays risk
conclusions for the maximum
measured groundwater concentration
over time, consistent with the policy
69
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
Page
Paragraph
Comments or Questions
Response
little significance over the 1000-year modeling period.
Further, there is no effect after about 400 years for
groundwater as all of the PFOA/PFOS makes it
through the soil column to the groundwater in that
time span. These "forever chemicals" will be around
"forever" and will impact the groundwater far into the
future. I suppose if one were an individual alive
during the intervening period where PFOA has
propagated, but not PPFOS these offsets would
matter. But, to paraphrase Keynes, in the long run we
someone will get the exposure. See Figure 13 Page 92
compared to Figure 14 Page 93.
goal of protecting groundwater
resources for future use.
94
Table
The CSF Description has units of (mg/kg/dy)"1 but the
Value column lists the units as (ng/kg/dy)"1,
This has been corrected for
consistency.
96
Table
The Row labels are misaligned as are some of the data
values using exponential notation. Perhaps three
separate tables would be more readable.
The EPA agrees and has made these
edits.
106
Table
Observation: No risk at all for PFAS and no risk for
Composite Liner for PFOA. I believe this is discussed
later. There are row label misalignments in these two
tables.
The EPA assumes the commenter is
referring to PFOS, not PFAS. The EPA
has corrected table formatting issues.
109
Lines 24ff
The discussion of why Mone Carlo analysis is not
warranted is hard to follow and seems, at points, to be
self-contradictory. I urge the authors to re-read it and,
perhaps, re-work the description. If they feel it is clear
enough, then perhaps it is my inability to decipher that
is suspect.
The EPA has edited the discussion of
Monte Carlo analysis for clarity.
110
I agree that EPA should consider the effects of
precursors on risk going forward and incorporate into
these models as data becomes available and of
significant number to merit selection of parameters for
decomposition.
Thank you.
70
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
Page
Paragraph
Comments or Questions
Response
114
Line 13
The Washington, et al., 2010 reference is not called
here but rather appears inline in the text.
The EPA agrees and has made these
edits.
116
Line 33
The colloquial term "-2000 times less" is not an
appropriate term as it is ill-defined. Use l/2000th if
that is what is meant.
The EPA agrees and has made these
edits.
Reviewer 3
No specific revisions or changes are required. The risk assessment does not require just the conceptual visualization of farm scenarios.
Reviewer 4
Page
Line
Comments or Questions
Response
ES-1
18
"Both chemicals are amongst the most potent carcinogens
assessed by the EPA to date." Recommend adding that
this toxicity is only one factor in the risk assessment and
give the example that the while one other chemical,
2,3,7,8-TCDD dioxin, is a more potent carcinogen that
PFOS or PFOA, it was not deemed to be a significant
enough risk in biosolids to warrant regulation by USEPA.
The cancer assessment for 2,3,7,8-TCDD
dioxin was not finalized by the EPA, as the
IRIS program determined that the non-cancer
RfD would be protective of cancer effects. As
described to prior comments, there are many
differences between dioxins and PFOA/PFOS
that result in different risk findings or
tentative risk findings in biosolids.
13
13
Recommend showing the reduction in PFOS and PFOA
concentrations in biosolids over time more clearly in a
table or graph. Is it possible these concentrations will
continue to decline over time?
The EPA presents a discussion of current
PFOA and PFOS sources in section 2.3, and
occurrence information is summarized in
section 2.4 and Appendix A. Note that peer-
reviewed studies in the tables in Appendix A
are shown in chronological order. Recent
U.S. studies on industrially impacted
biosolids have shown high levels of PFOS
despite the phase-out. The EPA is planning a
National Sewage Sludge Survey to obtain
PFAS occurrence information in sewage
71
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
Page
Line
Comments or Questions
Response
sludge from 200-300 of the largest POTWs in
the U.S., which will help elucidate potential
industrial and domestic PFAS sources
nationwide.
15
31
Recommend reporting if any elevated serum populations
have reported any significant health impacts and if so,
what they are.
For a detailed review of human epidemiology
studies for PFOA and PFOS please refer to
the EPA's 2024 final toxicity assessments for
PFOA and PFOS.
27
30
17
In the Farm Scenarios "Conceptual Visualization"
recommend showing conceptual groundwater flow
direction, a shallow plume at the top of the aquifer, and
the assumed vertical configuration of the downgradient
well screen (Minor point: check Figure numbers on page
30)
This detail is not included in the conceptual
model diagram, which is meant to be
schematic, but is rather included in Appendix
C. Thank you for the comments on figure
numbers; the EPA has corrected these typos.
30
2
Same comment as for Figure 1.
See above.
38
38
Some additional explanation would be helpful for the
Bumb et al. (1992) reference. Is the point that this soil
moisture curve paper provides an expression of capillary
pressures in the capillary fringe?
Additional description has been added to this
paragraph to explain that AWI retention may
also be relevant to the saturated zone because
there may be air entrained in pore spaces, as
discussed in Bumb et al. 1992.
39
5
"Most of the mass of PFOA and PFOS can remain in the
vadose zone for decades, centuries, or longer." The
authors should report what percent of the modeled PFOA
and PFOS mass that was applied in the biosolids
remained in the biosolids after 150 years; do those results
match this statement above about "most of the mass"?
Thank you for this comment. The text has
been edited to clarify that the specific
percentage of mass retained in the vadose
zone may vary by location. However,
modeling in this draft risk assessment does
find that most of the mass (>50%) is retained
in the vadose zone in the modeled locations
and the EPA has additionally added a citation
to a publication that has the same finding.
39
28
Agree with the observation that some of the simple PFAS
leaching screening models appear to overestimate the
Thank you.
72
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
Page
Line
Comments or Questions
Response
time to reach groundwater for certain scenarios.
39
29
Agree with the text that describes how the simplest,
screening version of these models appear to overestimate
vertical PFAA travel time for some cases. A site-specific
calibrated version of more sophisticated AWI-types
model would not have the overestimation problem but this
is not helpful to this risk assessment.
Thank you.
39
32
"Consistent with previous sewage sludge risk
assessments, this assessment will consider the peak
groundwater concentrations when calculating risks,
regardless of the timing of their occurrence, to avoid
underestimating risks through this pathway. " USEPA
should confirm that no model runs great than 150 years
were used to assess risk results shown in Section C.3..
Note that Figure 13 on page 92 extends to 600 years.
While the land application unit model is run
for 150 years, the groundwater model
(EPACMTP) runs until the peak
concentrations are observed. The EPA has
edited this text for clarity.
41
13
Recommend providing information on the transport of
PFOS/PFOA out of the hypothetical reservoir: does
VVWM account for PFOS/PFOA in the reservoir
outflow?
VVWM includes a first-order dissipation rate
due to flow moving contaminant out of the
water body - See Equation D. 1 in Appendix
D. To maintain a steady depth in the
reservoir, the flow rate out of the reservoir
equals the flow rate in - (e.g., runoff) - that
rate is a function meteorology and field
scenario (see Table D-3).
41
41
This approach of using the maximum value in the top 2
meters may significantly overpredict the actual risk in
some cases. Any users of a drilled groundwater well will
not be drinking water with the highest concentration in
the formation but an average concentration across the
entire screened interval and capture zone.
See the EPA responses to comments in
Question 4.
46
31
To following underlying objective a "central tendency",
recommend using a mid-range Kd for sorption (the
geometric mean of low and high Kd/s) rather than
Given the sensitivity of Kd, reporting high
and low values is more valuable than a single
median value.
73
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
Page
Line
Comments or Questions
Response
breaking out the low/high results. This would greatly
simplify the data presentation in the results tables in
Section 3, and more importantly provide a central
tendency result.
55
9
The dispersivity values in Table B-7 seem appropriate for
this modeling effort.
Thank you.
77
78
81
83
Recommending reporting the MCL for PFOS and PFOA
(4 ppt) in all of the different exposure tables in Section
3.1 for the drinking water pathways.
See responses to prior comments.
77
26
Recommend that the text explains from a risk perspective
what it means if groundwater or surface water is above
the MCL, and what it means if it is below the MCL.
See responses to prior comments.
99
100
102
103
106
For Sections 4.3, 4.4, 4.6, and 4.7, recommend that the
drinking water pathways values that exceed the
acceptable threshold (yellow-shaded values) also have
some type of indication if the MCL is also exceeded.
See responses to prior comments.
87
37
Minor point: check figure numbers in this section. Are
Figures 6-11 are missing?
The EPA has corrected these figures.
94
14
Recommend re-evaluating if lxlO"6 risk level, which was
selected out of the 10"5 to 10"6 range, is faithful to the goal
of performing a central tendency calculation. Again the
key to this type of calculation is not to have any over-
conservative (overestimation of risk) or underestimation
of risk assumptions, calculation steps, input data.
See response to prior comment.
Reviewer 5
Page
Paragraph
Comments or Questions
Response
5
2
The list of 14 exposure pathways is helpful but it will
be easier to understand if this list is turned into a
graphic. It is also not clear to me the difference
These pathways were identified in historic
assessments and are slightly modified for the
purpose of PFOA and PFOS draft risk
74
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
Page
Paragraph
Comments or Questions
Response
between pathway 1 (sludge-soil-plant-human) and
pathway 2 (sludge-soil-plant-home gardener).
Pathway 1 seems to encompass pathway 2.
assessment, as explained in the text. The
graphics are provided for the pathways
relevant to PFOA and PFOS, which are the
focus of this assessment.
39
5
The header "soil surface modeling" should have its
own numbering.
The EPA has revised accordingly.
53
table
This table needs a number and a title. Also, it will be
helpful to sort the table by parameter so that it is easy
to compare between PFOA and PFOS. For example,
BCF for forage and silage have the same value for
these two chemicals, but BCF Veg for PFOA is a lot
higher than that for PFOS.
The EPA re-configured the table to sort the
table by crop types.
67
Bottom table
Can you explain why the BAF for TL4 is lower than
that of TL3?
Yes, unlike other organic compounds that
typically have higher BAFs in higher trophic
level fish, the empirical data for PFOA and
PFOS show that these chemicals do not
follow the same trend.
69
last
Typo in the last sentence, it should say "the adult
unprotected vegetable intake equates to one serving
of unprotected vegetables every day."
Thank you; the EPA corrected this typo.
70
first
Typo in the last sentence, it should say "the adult root
vegetable intake equates to five servings of root
vegetables a week."
Thank you; the EPA corrected this typo.
73
1
Why does the exposure factor for dust ingestion have
different units for adults, children 12-19, children 6-
11, and children 1-5. Shouldn't it be all in the unit of
mg/kg/day?
The data sources for soil and dust ingestion
available in the EPA Exposure Factors
Handbook report these intake rates in units
of mg/day. The bodyweight parameter
(2.9.3.11) is used to convert these ingestion
rates to mg/kg-day when calculating risk.
73
2
Typo in the NHANES survey years, it should say
1999-2006.1 also wonder if a more recent figure
should be used instead?
Thank you for identifying this typo. This
data represents the most up-to-date edition of
75
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
Page
Paragraph
Comments or Questions
Response
the bodyweight chapter of the EPA's
exposure Factors Handbook.
76
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
References
CalEPA. (2024). Public Health Goals for Perfluorooctanoic Acid and Perfluorooctane Sulfonic
Acid in Drinking Water. Available on the Internet at:
https://oehha.ca.gov/media/downloads/air/public-health-goal-
document/pfoapfosphgfinaldraft040524.pdf
Castoria, R. & Woodruff, T.J. (2003). Assessment of Potential Risk Levels Associated with U.S.
Environmental Protection Agency Reference Values. Environmental Health Perspectives,
111, 10, 1318-1325.
US EPA (US Environmental Protection Agency). (1992). Technical Support Document for Land
Application of Sewage Sludge. Office of Water, Office of Science and Technology. EPA
822/R-93-001a.
US EPA. (1995). A Guide to the Biosolids Risk Assessments for the EPA Part 503 Rule.
EPA/832-B-93-005. https://www.epa.gov/sites/default/files/2018-ll/documents/guide-
biosolids-risk-assessments-part503.pdf
US EPA. (2000). Methodology for deriving ambient water quality criteria for the protection of
human health (2000). EPA-822-B-00-004. Office of Water, Washington, DC.
https://www.epa.gov/sites/default/files/2018-10/documents/methodology-wqc-protection-
hh-2000.pdf
US EPA. (2003). Exposure Analysis for Dioxins, Dibenzofurans, and Coplanar Polychlorinated
Biphenyls in Sewage Sludge. Technical Background Document.
https://www.epa.gov/biosolids/exposure-analvsis-dioxins-dibenzofurans-and-coplanar-
polvchlorinated-biphenyls-sewage
USEPA. (2022a). ORD Staff Handbook for Developing IRIS Assessments. EPA 600/R-22/268.
US EPA. (2023a). Per- and Polyfluoroalkyl Substances National Primary Drinking Water
Regulation. https://www.regulations.gov/document/EPA-HQ-OW-2022-0114-0027
US EPA. (2023b). EPA Response to Final Science Advisory Board Recommendations (August
2022) on Four Draft Support Documents for the EPA's Proposed PFAS National Primary
Drinking Water Regulation. https://www.regulations.gov/document/EPA-HO-OW-2Q22-
0114-0043
US EPA. (2024a). Office of Water Final Human Health Toxicity Assessment for
Perfluorooctanoic Acid (PFOA). 815R24006.
US EPA. (2024b). Office of Water Final Human Health Toxicity Assessment for Perfluorooctane
Sulfonic Acid (PFOS). 815R24007.
77
-------
External Letter Peer Review of Draft Sewage Sludge Risk Assessment for
Perfluorooctanoic Acid (PFOA) CASRN 335-67-1 and Perfluorooctane Sulfonic Acid (PFOS) CASRN 1763-23-1
US EPA. (2024c). Responses to Public Comments on Per- and Polyfluoroalkyl Substances
(PFAS) National Primary Drinking Water Regulation Rulemaking. 815R24005.
https://www.epa.gov/svstem/files/documents/2024-04/pfas-comment-response-
document final-508 v2.pdf
USEPA. (2024d). Per- and Polyfluoroalkyl Substances (PFAS) Occurrence and Contaminant
Background Support Document for the Final PFAS National Primary Drinking Water
Regulation. 815R24013.
US EPA. (2024e) Final Freshwater Aquatic Life Ambient Water Quality Criteria and Acute
Saltwater Aquatic Life Benchmark for Perfluorooctanoic acid (PFOA). 842R24002.
US EPA. (2024f) Final Freshwater Aquatic Life Ambient Water Quality Criteria and Acute
Saltwater Aquatic Life Benchmark for Perfluorooctane sulfonate (PFOS). 842R24003
78
------- |