United States
Environmental Protectio
^1 Agency
Office of Water
EPA 823-R-24-001
July 2024
Contaminants to Monitor in Fish and
Shellfish Advisory Programs:
Compilation of Peer Review-Related
Information
-------
Table of Contents
Overview 1
Section 1: Overview of Process 2
Section 2: Summary of Peer Reviewers' Suggestions 6
Section 3: Changes Made in Response to Peer Reviewer Comments 8
Section 4: Final Lists of Contaminants to Monitor 15
Appendix 1: Process for Selecting Contaminants to Monitor in Fish Advisory Programs That Was Sent to
Peer Reviewers A-l
Appendix 2: External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory
Programs A-28
i
-------
Overview
State, Tribal, and territorial fish and shellfish advisory programs monitor and analyze fish and shellfish in
waterbodies within their jurisdictions for contaminants. When contaminants occur in high enough
concentrations to potentially affect the health of people eating fish and shellfish from those waters, the
EPA recommends that those programs issue advisories regarding consumption to protect the
consumers. To help state, Tribal, and territorial fish and shellfish advisory programs, the EPA
recommends a set of contaminants to monitor in its Guidance for Assessing Chemical Contaminant Data
for Use in Fish Advisories.
The EPA is updating its fish advisory guidance for states, which was last revised in 2000. As part of that
update, the EPA is adding to its list of contaminants found to accumulate in fish at levels that could be
problematic for human health. The EPA did not investigate whether any contaminants on the existing
list need to be removed. The process the EPA followed included the following activities:
• Performed an extensive literature search for published journal articles using a set of specified
search terms.
• Compiled concentrations in fish and shellfish from articles and toxicity information from U.S.
government sources.
• Calculated if the concentrations in fish and shellfish would exceed thresholds for safely eating 8
ounces per week for people who could become pregnant and the general population or eating 5
ounces per day for frequent consumers of seafood.
• Compiled two lists of contaminants that have been found in fish and shellfish at concentrations
that may be of concern for human health - one list has toxicity information, allowing fish
advisory consumption rates to be calculated, and the other list contains contaminants that do
not have maximum allowable exposure information but the levels found in fish and shellfish are
high enough to warrant monitoring.
The EPA then submitted the process and results (included in Appendix 1) to subject matter experts in
toxicology and human health risk assessment for an independent, external peer review. The peer
reviewers responded to the charge questions and had some suggestions, which a contractor compiled
into a report (included in Appendix 2). The EPA considered these suggestions and made some changes to
the process and the resulting list of contaminants.
This document is arranged in the following manner.
• Section 1 contains a summary of the process that was provided to the peer reviewers.
• Section 2 contains the main areas of suggestions from the peer review report.
• Section 3 groups the comments into areas and described the changes that were made to the
process and list of contaminants in response to the peer reviewers' comments.
• Section 4 contains the final lists of contaminants that the EPA is recommending fish and shellfish
advisory programs monitor.
• Appendix 1 contains the document provided to the peer reviewers.
• Appendix 2 contains the entire peer review report.
July 2024
1
-------
Section 1: Summary of Process
The EPA initiated a systematic screening process to identify any additional relevant compounds that are
not currently included in the current version of its Guidance for Assessing Chemical Contaminant Data
for Use in Fish Advisories (2000 Guidance). The EPA developed a protocol to compile peer-reviewed
articles that provide a basis for choosing any new analytes. This section contains a summary of the
process; the more detailed version that was sent to the peer reviewers is in Appendix 1.
Literature Review
Several environmental science, health, and toxicology databases were searched for relevant peer-
reviewed publications using keywords and inclusion and exclusion criteria. The screening criteria for
selecting publications identified before beginning the literature review included criteria in areas such as
publication status, publication date, language, seafood species, and compounds analyzed. The EPA
included only articles that were peer-reviewed, published in 2000 or later (to capture information
published after the 2000 Guidance), and written in English.
The EPA searched the PubMed, PubChem, Web of Science, Environmental Science and Pollution
Management, Science Direct, and Toxline databases for relevant articles. The EPA also utilized an
internal literature search that had been conducted for harmful algal blooms (HABs) and HAB toxins.
Specific keywords were used to search the literature; some examples are included here, listed by
category.
• Target analytes: PCBs OR contaminant OR constituent OR contamination OR emerging
contaminant OR aquatic contaminant OR chemicals OR pollutants OR metals OR pesticides.
• Aquatic animal types: Finfish OR fish OR freshwater turtles OR shellfish OR bivalves OR
crustaceans OR mollusks AND [edible tissue OR muscle tissue].
These efforts resulted in a compilation of more than 600 articles for further review. The EPA screened
the articles collected during the literature search, and examined articles that had information in at least
two of these areas: contaminant concentration levels in fish or shellfish, BCF or BAF data, oral toxicity
data, and species found in the U.S.
From these articles, the EPA developed a preliminary list of 242 potential contaminants in the following
classes:
• Antibacterial, antibiotic, and
antimicrobial compounds.
Brominated compounds.
Chlorinated compounds.
Cyanotoxins and neurotoxins.
Flame retardants.
Hormones.
Industrial byproducts.
Inorganics.
Metabolites.
Metals.
Organophosphorus esters.
Nanoparticles.
Personal care products.
Pesticides.
Per- and polyfluoroalkyl substances
(PFAS).
Pharmaceuticals.
Phthalates.
Polychlorinated biphenyls (PCBs).
Polychlorinated naphthalenes.
Polycyclic aromatic hydrocarbons.
Sulfonamides.
Other.
July 2024
2
-------
The EPA then extracted concentration data from the articles, sorted them based on what parts were
analyzed (e.g., fillet, whole body, organs), and removed any that were from species not found in U.S.
waters or that were not measuring ambient conditions (e.g., lab dosing studies).1
The EPA refined the list of potential new contaminants by removing compounds without concentration
data, compounds already on the monitoring list in the 2000 Guidance, and mixtures (e.g., BDE-119 +
BDE-120, sum of PFAS), and this resulted in lists of 49 potential contaminants with fillet data, 55 with
whole fish data, and 14 with shellfish data.
Consistent with its approach to developing water quality criteria, the EPA searched for toxicology
information (e.g., reference dose, minimal risk level, cancer slope factor) for each of the contaminants
on the list in the following eight peer-reviewed, publicly available sources:
1. EPA's Integrated Risk Information System (IRIS) program.
2. EPA's Office of Pesticide Programs Pesticide Chemical Search.
3. EPA's Office of Pollution Prevention and Toxics Existing Chemicals.
4. EPA's Office of Water Water Topics.
5. EPA's Office of Solid Waste and Emergency Response Provisional Peer Reviewed Toxicity Values
for Superfund (PPRTV).
6. U.S. Department of Health and Human Services. Agency for Toxic Substances and Disease
Registry (ATSDR) Toxic Substances Portal.
7. Health Canada.
8. California Environmental Protection Agency's Office of Environmental Health Hazard
Assessment - All Public Health Goals.
For PFAS compounds, the EPA used the reference doses and minimal risk levels for perfluorooctane
sulfonic acid (PFOS), perfluorooctanoic acid (PFOA), perfluorononanoic acid (PFNA), and
perfluorohexane sulfonic acid (PFHxS) and the PPRTV for perfluorobutane sulfonic acid (PFBS) that were
used in the proposed national primary drinking water regulation released on March 14, 2023, and the
reference doses in IRIS for perfluorobutanoic acid (PFBA), perfluorodecanoic acid (PFDA), and
perfluorohexanoic acid (PFHxA).
Analyses
For each contaminant with a non-cancer toxicity value, the EPA calculated a non-cancer screening value
using this equation from the 2000 Guidance:
... reference dose x consumer body weight
Non-cancer screening level =
consumption rate
where:
Body weight of adult in general population and of frequent fish consumer = 80 kg
Body weight of pregnant person = 75 kg
1 Fish and shellfish advisory programs are concerned with real-world conditions that reflect what is in the
environment, whereas lab studies can be designed to determine what is possible to occur (e.g., amount of
bioaccumulation), so the EPA used only articles that demonstrated contaminant accumulation in fish and shellfish
in ambient waters.
July 2024
3
-------
Consumption rate of adult in general population and pregnant person = 8 oz/week * 28.35 g/oz * 1
week/7d = 32.4 g/d2
Consumption rate of frequent fish consumer = 142 g/d3
The non-cancer screening levels were compared to the concentration data extracted from the scientific
literature.
EPA analyzed whether the maximum or average concentrations extracted from articles exceeded the
non-cancer screening level for:
o An adult in the general population,
o A pregnant person,
o A frequent fish consumer.
EPA also calculated whether the maximum concentration and average concentration were within 75
percent of the non-cancer screening level for an adult in the general population to see if there were
compounds that could be problematic but not currently accumulating to problematic levels. The EPA did
not find additional compounds to include as a result of those analyses.
A contaminant's presence in fish does not necessarily indicate a human health risk exists. For the
contaminants without non-cancer toxicity values, before the peer review, the EPA calculated a generic
screening level to capture contaminants with fish tissue concentrations high enough to potentially be a
human health concern after reference doses are developed. In its screening level calculations, the EPA
used the lowest final toxicity value (that is, the most stringent toxicity value that was not draft or being
developed) available among the contaminants found in fish. The lowest toxicity value for compounds
that were considered for inclusion on the monitoring list in this evaluation was 3 x 10 s mg/kg-d, which
was the minimal risk level for PFNA. The calculated generic screening level was 7.41 x 10"3 ng/g, which
was compared to the maximum and average concentration data for each compound without toxicity
values. After the peer review, a revised generic screening level was calculated for each class of
contaminants.
For each contaminant with a cancer slope factor, the EPA calculated a cancer screening value using this
equation and constants:
_ ... cancer risk level x consumer body weight
Cancer screening level =
cancer slope factor x consumption rate
where:
Cancer risk level = 10 s
Body weight of adult in general population and of frequent fish consumer = 80 kg
Consumption rate of general population = 32.4 g/d
Consumption rate of frequent fish consumer = 142 g/d
[Note: The 2000 Guidance presented fish meal calculations based on a cancer risk level of 10"5. The
EPA is considering updating that factor to 10 s to be consistent with methods for developing water
2 This consumption rate is based on the recommendation in the U.S. Department of Agriculture and Department of
Health and Human Services' Dietary Guidelines for Americans. 2020-2025 to eat 8-10 ounces per week of seafood.
3 From EPA's Methodology for Deriving Ambient Water Quality Criteria for the Protection of Human Health (2000)
that recommends a default of 142.4 grams/day for subsistence fishers.
July 2024
4
-------
quality criteria. To ensure it captured problematic compounds, the EPA used the 10 s cancer risk
level in these screening level calculations.]
The EPA analyzed whether the maximum or average concentrations extracted from articles exceeded
the cancer screening level for an adult in the general population or for a frequent fish consumer. Some
studies reported only averages for a contaminant; some reported only maximums. The EPA considered a
contaminant for consideration if the screening level was exceeded by the average tissue concentration,
maximum concentration, or both.
Results of Process
EPA developed two draft lists of compounds based on available information:
1. Contaminants to Monitor for Advisories: The compounds on this list are found to occur in edible
tissue of fish or shellfish at levels of concern to human health and have toxicity information in
the form of an EPA oral reference dose, ATSDR minimal risk level, or EPA cancer slope factor.
These are new compounds that the EPA is considering adding to its existing list (in 2000
Guidance) of recommended compounds to monitor for fish advisories. The version sent through
the peer review process included five PFAS compounds, one cyanotoxin, two flame retardants,
and two metals.
2. Contaminants to Monitor to Watch: The compounds on this list are those that may need
advisories in the future. They are on the list because they have documented concentrations in
fish and/or shellfish that could be of concern for human health, based on the generic screening
level used in these analyses, but the federal government has not yet developed toxicity values
such as reference doses or cancer slope factors for them. The version sent through the peer
review process included two cyanotoxins, five flame retardants, seven PFAS compounds, four
pharmaceuticals, and three divisions of chlorinated paraffins.
July 2024
5
-------
Section 2: Summary of Peer Reviewers' Suggestions
The EPA submitted the process it used to create a list of contaminants to add to the monitoring list in its
2000 Guidance to a group of three of subject matter experts in toxicology and human health risk
assessment for an independent, external peer review. The peer reviewers responded to the charge
questions and had some suggestions, which the peer review contractor compiled into a report. This
section summarizes the reviewer comments by charge question. Appendix 2 contains the complete peer
review report.
Charge Question 1: Is the process EPA followed to identify compounds for which fish and shellfish
advisories might be needed reasonable?
All three reviewers agreed that the process EPA followed was reasonable, but also indicated that the
process would benefit from some revision. One reviewer suggested incorporating toxicity values from
databases less focused on North America. Two reviewers noted that using a cancer slope factor as a
screening level for lead is highly unusual; both suggested that EPA's Integrated Exposure Uptake
Biokinetic Model for Lead in Children (IEUBK) model or EPA's Adult Lead Model should be used to
develop a screening level. One reviewer recommended that the draft IRIS reference dose values for PFHxS
(released after the document that was sent to the peer reviewers was written) and PFNA (not yet
released) be used for calculating the screening levels for those compounds.
For calculating a generic screening level for contaminants without established toxicity information, all
three reviewers questioned the use of the chronic reference value based upon the PFNA minimal risk
level (MRL) produced by ATSDR. One reviewer stated that EPA should use a well-established toxicity
value with a high degree of scientific consensus regarding the validity of the value and how it was derived.
This reviewer commented this was not the case for PFNA, because IRIS is developing a newer draft
assessment for it. Another reviewer stated that the generic screening level based on the ATSDR MRL for
PFNA is highly uncertain and that ATSDR MRLs are based on animal data, not human data. The third
reviewer expressed that extrapolating the same reference dose across chemical classes seems
unnecessary when it is possible to choose the lowest reference dose within each chemical class.
One reviewer suggested updating the screening level equation to better reflect current state practices
for implementing fish advisories. This reviewer also noted that, for occurrence data, sample maximums
are very unreliable statistics and recommended using a high percentile (e.g., 95th or 99th percentiles) to
represent high data values more appropriately. The same reviewer also recommended against using
lipid-normalized concentrations and suggested that these values be converted to wet weight
concentrations.
Charge Question 2: Is the list of contaminants advisory programs should consider monitoring for
reasonable (e.g., reflects the current range of contaminants detected in fish with potential human
health impacts)?
All three reviewers agreed that the list of contaminants is reasonable and comprehensive. One reviewer
noted that there are other compounds of concern that could be included in the list, but that these
contaminants are not well known or well researched; this reviewer indicated that the list is reasonable
even without the additional contaminants. Another reviewer noted that some states have already
developed fish consumption advisories for many contaminants on the list.
July 2024
6
-------
Charge Question 3: Are there additional contaminants that should be included in the "monitor for
advisories" list or "monitor to watch" list? If so, what are they, and why should they be included?
Two reviewers suggested including additional contaminants. One reviewer recommended adding 6:2 di-
and mono-PAPs and fluorotelomer sulfonates to the lists, as they all have shown high bioconcentration
factor (BCF) values in recent studies. Another reviewer suggested adding additional cyanotoxins, such as
cylindrospermopsin, anatoxin-a, and saxitoxin to the lists, because harmful algal blooms (HABs) are of
current concern. This reviewer noted that cyanotoxin advisories should take into account that HAB
exposures are often short-term, not chronic.
July 2024
7
-------
Section 3: Changes Made in Response to Peer Reviewer Comments
The EPA has grouped the comments provided by the peer reviewers into areas based on their similarity
and described any changes that were made to the process and list of contaminants in response to the
peer reviewers' comments.
Area 1: Toxicity Values
Reviewers commented that the EPA should consider expanding the universe of toxicity values used in
the screening level calculations beyond those published by U.S. agencies, including EFSA, WHO, and
state agencies. They also suggested using an exposure model for lead instead of a cancer slope factor.
In the information considered by the peer reviewers, the EPA had used toxicity information from a list of
specific sources, in alignment with the EPA's methodology for developing water quality criteria to
protect human health (2010) and corresponding criteria update (2015). That methodology lists eight
sources; of those eight the EPA found and used reference doses, cancer slope factors, minimal risk
levels, and similar toxicity information from the following sources:
• EPA's Integrated Risk Information System (IRIS) program.
• EPA's Office of Water programs.
• EPA's Office of Land and Emergency Management Provisional Peer Reviewed Toxicity Values
(PPRTVs) for the Superfund program.
• U.S. Department of Health and Human Services' Agency for Toxic Substances and Disease
Registry (ATSDR).
• California Environmental Protection Agency's Office of Environmental Health Hazard
Assessment.
The EPA has elected not to deviate from the process it uses for water quality criteria. In general, states,
territories and tribes have the flexibility to monitor for any contaminant, and they can use additional
toxicity values than the EPA considered when they calculate and issue fish advisories. The EPA's list of
contaminants for monitoring are simply recommendations to the states, Tribes and territories.
To be consistent in its evaluation process, the EPA followed the same process for all contaminants but
recognizes that lead is often treated differently. Reviewers suggested using a lead model, but lead
exposure models primarily focus on soil contamination as the primary data input, and other exposures
like drinking water and fish consumption are used to refine the risk to the receptor from exposures to
soil. The relative contribution from specific exposure pathways (e.g., water, diet, soil, ambient air) to
blood lead concentrations is situation specific. According to the EPA's Integrated Science Assessment for
Lead (External Review Draft, 2023, EPA/600/R-23/061), 30 causality determinations were made for
human health outcomes from exposure to lead, like cognitive function decrements in children and
cardiovascular effects. With each EPA assessment for lead over time (e.g., 2006, 2013), the
epidemiologic and toxicological evidence demonstrated that progressively lower blood lead levels or
lead exposures are associated with cognitive deficits in children. The EPA decided to keep lead on the
monitoring list for fish advisories because there is no known safe level of exposure to lead.
Area 2: Generic screening level
Regarding the EPA's calculation of a generic screening level for contaminants without established
toxicity information, all three reviewers questioned the use of the chronic reference value based upon
the minimal risk level for PFNA produced by the ATSDR. One reviewer stated that the EPA should use a
well-established toxicity value with a high degree of scientific consensus regarding the validity of the
July 2024
8
-------
value and how it was derived. This reviewer commented this was not the case for PFNA, because IRIS is
developing a draft assessment for it. Another reviewer stated that the generic screening level based on
the ATSDR MRL for PFNA is highly uncertain and that ATSDR MRLs are based on animal data, not human
data. The third reviewer expressed that extrapolating the same reference dose across chemical classes
seems unnecessary when it is possible to choose the lowest reference dose within each chemical class.
The EPA investigated using the lowest reference dose within each class of contaminants. Every
contaminant class that contained a substance needing a generic screening level had at least one
compound with a reference dose from the water quality criteria's list of sources except paraffins and
pharmaceuticals. The EPA used oral human exposure doses for paraffins, and screening doses for
pharmaceuticals in place of reference doses. The EPA used the lowest toxicity value (i.e., reference dose,
human exposure dose, or screening dose) in each contaminant class and calculated screening levels for
each class of contaminants, as described in each of the following subsections.
Cyanotoxins
For cyanotoxins, the EPA used the lowest available reference dose, which was 0.05 ng/kg-d for
microcystins from the EPA's Health Effects Support Document for the Cyanobacterial Toxin Microcystins
(2015). Concentrations of BMAA ((B-methylamino-L-alanine) and DABA (2,4-diaminobutyric acid
dihydrochloride) in fish were higher than the screening level calculated using the microcystins reference
dose, so the EPA retained them on the final list of contaminants.
Flame retardants
For flame retardants, the EPA used the lowest available reference dose for oral exposure, which was 1 x
10"4 mg/kg-d for both BDE-47 and BDE-99 from the EPA's Integrated Risk Information System.
Concentrations in fish of all the flame retardants without toxicity information were lower than the
screening level calculated using the reference dose for BDE-47 and BDE-99, so the EPA removed BDE-49,
BDE-100, Dechlorane 602, Dechlorane 604, and decabromodiphenyl ethane (DBDPE) from the final list
of contaminants.
PFAS
The PFAS group was handled slightly differently. Based on input from the EPA's Office of Research and
Development and as published in the Framework for Estimating Noncancer Health Risks Associated with
Mixtures of Per- and Polyfluoroalkyl Substances (PFAS) (External Review Draft, 2023), the PFAS
compounds were separated into four groups based on whether they were long- or short-chain
carboxylic or sulfonic acids, as shown in Table 1. PFOSA was grouped with the long-chain sulfonic acids.
Table 1: Characterization System of Short-Chain and Long-Chain Perfluoroalkyl Acids3
Total number of carbons
3
4
5
6
7
8
9
10
Number of fluorinated carbons
2
3
4
5
6
7
8
9
Perfluorocarboxylic acids (PFCAs)
Short-chain PFCAs
Long-chain PFCAs
PFPrA
PFBA
PFPeA
PFHxA
PFHpA
PFOA
PFNA
PFDA
Number of fluorinated carbons
3
4
5
6
7
8
9
10
Perfluorosulfonic acids (PFSAs)
PFPS
PFBS
PFPeS
PFHxS
PFHpS
PFOS
PFNS
PFDS
Short-chain PFSAs
Long-chain PFSAs
Notes: PFPrA = perfluoropropanoic acid; PFBA = perfluorobutanoic acid; PFPeA = perfluoropentanoic acid; PFHxA =
perfluorohexanoic acid; PFHpA = perfluoroheptanoic acid; PFOA = perfluorooctanoic acid; PFNA =
perfluorononanoic acid; PFDA = perfluorodecanoic acid; PFPS = perfluoropropanesulfonic acid; PFBS =
July 2024
9
-------
perfluorobutanesulfonic acid; PFPeS = perfluoropentanesulfonic acid; PFHxS = perfluorohexanesulfonic acid;
PFHpS = perfluoroheptanesulfonic acid; PFOS = perfluorooctanesulfonic acid; PFNS = perfluorononanesulfonic
acid; PFDS = perfluorodecanesulfonate. For brevity, Table 1 only includes perfluoroalkyl acids of 3-10 carbons;
the long-chain class of PFCAs and PFSAs can be expanded considerably.
a Table 1-3 from EPA's Framework for Estimating Noncancer Health Risks Associated with Mixtures of Per- and
Polyfluoroalkyl Substances (PFAS) (2024); modification of Table 2-2 from ITRC's Per- and Polyfluoroalkyl
Substances Technical and Regulatory Guidance (2022).
The EPA considered the lowest available draft and final toxicity data within each PFAS group, as shown
in Table 2, to calculate the group's generic screening level for those PFAS without a reference dose (RfD)
or minimal risk level.
Table 2: Toxicity Information Used for PFAS Groups
PFAS Group Name
PFAS in Group Found in
Fish and/or Shellfish
Compound with Lowest Human Health Toxicity
Value and its Value (Source)
Short-chain PFCAs
PFBA, PFHxA, PFHpA,
PFPeA
PFHxA: 5 x 10"4 mg/kg-d (final RfD - IRIS)
Long-chain PFCAs
PFDA, PFDoA, PFNA, PFOA,
PFTeDa, PFTrDA, PFUndA
PFDA: 4 x 10 10 mg/kg-d (draft RfD -IRIS)
PFOA: 3 xlO"8 mg/kg-d (final RfD - OW)
Short-chain PFSAs
PFBS
PFBS: 3 x 10"4 mg/kg-d (final human health
toxicity value - PPRTV)
Long-chain PFSAs
PFDS, PFHpS, PFHxS, PFOS,
PFOSA
PFHxS: 4 x 10 10 mg/kg-d (draft RfD - IRIS)
PFOS: 1 x 10"7 mg/kg-d (final RfD - OW)
Using the reference dose for PFHxA for the short-chain PFCAs without a toxicity value resulted in a
screening level higher than the concentrations of PFHpA and PFPeA that were found in fish, so the EPA
did not add them to the final list of recommended contaminants to monitor. PFBA and PFHxA had
screening levels based on published toxicity values; these were also higher than the fish tissue
concentrations, so the EPA did not add PFBA and PFHxA to the final list.
Using either the draft reference dose for PFDA or the final reference dose for PFOA for the long-chain
PFCAs without a toxicity value resulted in a screening level lower than the concentrations of PFDoA,
PFTeDA, PFTrDA and PFUnDA that were found in fish and shellfish, so they remained on the final list of
recommended contaminants to monitor. PFDA, PFNA, and PFOA had screening levels based on
published toxicity values; these were also lower than the fish tissue concentrations, so PFDA, PFNA, and
PFOA remained on the final list.
For short-chain PFSAs, the only one that was found in fish was PFBS, which has a reference dose. PFBS
concentrations in fish did not exceed the calculated screening levels, therefore the EPA did not include it
on the final list.
Using either the draft reference dose for PFHxS or the final reference dose for PFOS for the long-chain
PFSAs without a toxicity value resulted in a screening level lower than the concentrations of PFDS,
PFHpS and PFOSA that were found in fish, so PFDS and PFOSA remained on the final list and PFHpS was
added to the list. PFHxS and PFOS had screening levels based on published toxicity values; these were
also lower than the fish tissue concentrations, so they remained on the final list.
July 2024
10
-------
Paraffins
The contaminant class of paraffins did not have a compound with a reference dose that could be used to
calculate a generic screening level. The EPA's Office of Pollution Prevention and Toxics developed oral
human exposure doses for paraffins in its TSCA New Chemicals Review Program Standard Review Risk
Assessment on Medium-Chain Chlorinated Paraffins (PMN P-12-0282, P-12-0283) and Long-Chain
Chlorinated Paraffins (PMN P-12-0284) (2015). The EPA used those human exposure doses in place of a
reference dose in the screening level equations and determined that paraffins concentrations found in
fish were not high enough to exceed the screening levels.
Pharmaceuticals
The other contaminant class without a reference dose at the time of the peer review was
pharmaceuticals. After the peer review, for pharmaceuticals the EPA switched from using a generic
screening level to using information from the draft Human Health Drinking Water Benchmarks for
Pharmaceuticals produced by the Office of Water (one of the allowable sources in the process). A
screening level for the general population for each pharmaceutical found in fish were calculated using
the screening doses in the draft benchmarks document. Screening doses were calculated by dividing the
lowest therapeutic doses from FDA labels by a composite uncertainty factor of 3,000 to account for
interspecies extrapolation, intraspecies variation, subchronic-to-chronic study extrapolation,
extrapolation from a lowest-observed-adverse effect level (LOAEL) (i.e., lowest therapeutic dose) to a
no-observed-adverse effect level (NOAEL), and database deficiencies. The screening dose was treated as
equivalent to a reference dose in the screening level equations. Norfluoxetine, norverapamil,
sulfadimethoxine, and triclocarban did not have screening doses in the pharmaceutical benchmarks
document. For those four compounds, the EPA used the lowest screening dose for compounds in the
pharmaceutical category that were found in fish, which was 8.3 x 10 s mg/kg-d for amphetamine, to
calculate a generic screening level. The concentrations of those four compounds in fish were lower than
the screening level, so the EPA did not add them to the contaminant list.
Concentrations in fish no longer exceeded the screening levels for these pharmaceuticals:
• Metformin.
• Sertraline.
• Sulfadimethoxine.
The EPA removed those pharmaceuticals from the draft list of contaminants to monitor. Only
amphetamine had concentrations in fish that exceeded the screening levels, so it remained on the list.
Summary of actions resulting from change in process to generic screening level
Using the generic screening levels calculated for each contaminant class and PFAS group and using the
human exposure doses for paraffins and screening doses for pharmaceuticals yielded the following
changes from the draft list to the final list.
Concentrations in fish fillets and whole fish that previously had exceeded the initial generic screening
level based on ATSDR's MRL for PFNA did not exceed the newly calculated screening levels for these
contaminants (five flame retardants, three classes of paraffins, and three pharmaceuticals):
• BDE-49.
• BDE-100.
• Dechlorane 602.
• Dechlorane 604.
July 2024
11
-------
Decabromodiphenyl ethane (DBDPE).
Short-chain chlorinated paraffins.
Medium-chain chlorinated paraffins.
Long-chain chlorinated paraffins.
Metformin.
Sertraline.
Sulfadimethoxine.
Concentrations in fish fillets and whole fish that previously did not exceed the initial generic screening
level exceeded the newly calculated screening levels for this contaminant (one PFAS):
• PFHpS.
As a result of the changes to the generic screening levels, the EPA removed the five flame retardants,
three paraffin classes, and three pharmaceuticals from the draft list. In addition, the EPA added the one
PFAS compound to the list of contaminants to watch.
Area 3: Contaminant concentration data
For each contaminant, the EPA compared the screening levels to the average and maximum
concentrations in fish and shellfish that were extracted from the journal articles reviewed. One reviewer
raised concerns with the use of maximum and lipid-weight concentrations.
Maximum concentrations
The reviewer said a sample maximum is an unstable summary statistic and subject to extreme results
but recognized that the EPA was limited by what is reported in the literature.
Post peer review, the EPA calculated 75th percentiles of concentration data maximums for contaminants
on the draft list and compared those to the screening levels. As a result, BDE-99 no longer exceeded any
screening levels and was removed from the draft list. In addition, there was not enough data to calculate
a 75th percentile concentration for PFProPrA and thallium, so they were removed from the draft list.
Lipid weight concentrations
The EPA did not include lipid weight concentration data in its analyses unless the only data for a
contaminant was reported in lipid weight form. This circumstance was true for only three compounds
that were placed on the draft list of contaminants to monitor - the flame retardants Dechlorane 602,
Dechlorane 604, and Decabromodiphenyl ethane (DBDPE).
The EPA converted the lipid weight concentrations for these three compounds to wet weight
concentrations using total fat content percentages from Table 10-125 (e.g., 6.61% for "trout, mixed
species") and Equation 10-7 in the Environmental Factors Handbook (2011):
"... wet-weight residue levels in fish may be estimated by multiplying the levels based on fat by the
fraction of fat per product as follows:
L
Cww — Clw
100
where:
Cww= wet-weight concentration, Ciw= lipid-weight concentration, and L = percent lipid (fat) content."
July 2024
12
-------
When the EPA calculated revised screening levels for each contaminant class and used the converted
wet weight concentrations, the three flame retardants no longer had concentrations that exceeded the
screening levels and the EPA removed them from the draft list.
Area 4: Additional contaminants
One reviewer recommended adding fluorotelomer sulfonates to the contaminant list. The study the
reviewer referenced was performed in Norway and was on primarily non-U.S. species and/or measured
contaminant levels in parts that the EPA's fish advisory program recommends not eating (e.g., crab
hepatopancreas, fish liver) so it did not meet our search parameters, but the EPA will continue to be on
the lookout for contaminants in fish that should be added to the monitoring list.
Another reviewer suggested adding additional cyanotoxins, such as cylindrospermopsin, anatoxin-a, and
saxitoxin to the lists. In its analyses performed before the peer review, the EPA analyzed the data it had
found for those toxins, but concentrations were not high enough to exceed the screening levels. If there
is an active harmful algal bloom, advisory programs could monitor and analyze fish and shellfish for the
relevant cyanotoxin and issue a short-term consumption advisory, if warranted.
General comments
Comments on specific wording changes for the process will be addressed when the EPA updates the
applicable section of the Guidance for Assessing Chemical Contaminant Data for Use in Fish Advisories.
The comment on updating the screening level equation to include a hazard quotient and relative source
contribution (RSC) to account for additional exposure pathways will be addressed when the EPA shares
its suite of fish advisory equations with a set of peer reviewers, as part of the update to the fish advisory
guidance. The draft reference dose for PFHxS that was released by IRIS after the development of the
document that went to the peer reviewers was incorporated into the calculations; PFHxS remains on the
list of contaminants to monitor.
Revisions to the Contaminant List After the Peer Review
Table 3 summarizes the changes that the EPA made to the draft list of contaminants after reviewing the
peer reviewers' comments and making changes to the process.
Table 3: Changes to Contaminant Status
Contaminant class
Retained
Added
Dropped
Chlorinated paraffins
Short-, medium-, and
long-chain
Cyanotoxins
BMAA
DABA
Microcystins
Flame retardants
BDE-47
BDE-49
BDE-99
BDE-100
Dechlorane 602
Dechlorane 604
DBDPE
Metals
Lead
Thallium
PFAS
PFDA
PFDS
PFHpS
PFPrOPrA
July 2024
13
-------
Contaminant class
Retained
Added
Dropped
PFDoA
PFHxS
PFNA
PFOA
PFOS
PFOSA
PFTeDA
PFTrDA
PFUnDA
Pharmaceuticals
Amphetamine
Metformin
Sertraline
Sulfadimethoxine
July 2024
14
-------
Section 4: Final Additions to Contaminants to Monitor
The EPA has finalized which new contaminants that it will recommend fish and shellfish advisory
programs include in their monitoring programs. These contaminants have been found to occur in the
edible tissue of fish and shellfish at concentrations that may be of concern for human health. The EPA
has separated the lists into two groups based on the availability of toxicity information.
1. The first list, "Contaminants to monitor for advisories" (shown in Table 1), contains new
contaminants for which the EPA or other federal agencies have released measures of oral
toxicity in humans (e.g., reference dose, cancer slope factor). The EPA recommends that
advisory programs use this list for monitoring and issuing advisories with consumption limits.
2. The second list, "Contaminants to monitor to watch" (shown in Table 2), contains contaminants
for which the EPA or other federal agencies have not yet released assessments of the effects on
human health. The EPA recommends that advisory programs monitor for compounds on this list
to determine if they are accumulating in fish in local waters. The advisory programs could
calculate their own or use another agency's scientifically based measures of oral toxicity in
humans to calculate consumption limits, or wait for such values to be released from a federal
agency.
Table 1. New contaminants to monitor for fish and shellfish advisories
Contaminant Group
Contaminant
Cyanotoxins
Microcystins
Flame retardants
BDE-47
Metals
Lead
Pharmaceuticals
Amphetamine
PFAS
Perfluorodecanoic acid (PFDA)
Perfluorohexane sulfonic acid (PFHxS)
Perfluorononanoic acid (PFNA)
Perfluorooctanoic acid (PFOA)
Perfluorooctane sulfonic acid (PFOS)
Table 2. Contaminants to monitor to watch
Contaminant Group
Contaminant
Cyanotoxins
BMAA ((B-methylamino-L-alanine)
DABA (2,4-diaminobutyric acid dihydrochloride)
PFAS
Perfluorodecanesulfonic acid (PFDS)
Perfluorododecanoic acid (PFDoA)
Perfluoroheptanesulfonic acid (PFHpS)
Perfluorooctanesulfonamide (PFOSA)
Perfluorotetradecanoic acid (PFTeDA)
Perfluorotridecanoic acid (PFTrDA)
Perfluoroundecanoic acid (PFUdA, PFUnA, PFUnDA)
July 2024
15
-------
Appendix 1: Process For Selecting Contaminants To Monitor In Fish Advisory
Programs That Was Sent to Peer Reviewers
Table of Contents
Overview A-2
Background A-2
Literature Search A-3
Preliminary Contaminant List Compilation, Data Extraction, and Exclusions A-5
Researched Toxicity Values A-7
Screening Level Calculations and Analyses A-7
Results of Comparing Concentration Data to Screening Levels A-9
Results of Process A-14
References Cited A-15
Appendix A-17
Table of Tables
Table 1. Contaminants exceeding screening levels in fillet data A-10
Table 2. Contaminants exceeding screening levels in whole fish data A-l 1
Table 3. Contaminants exceeding screening levels in shellfish data A-l 1
Table 4. Contaminants exceeding "generic" screening level in fillet data A-12
Table 5. Contaminants exceeding "generic" screening level in whole fish data A-13
Table 6. Contaminants exceeding "generic" screening level in shellfish data A-13
Table 7. New contaminants to monitor for advisories A-14
Table 8. Contaminants to monitor to watch A-15
Table A-l. Analytes Recommended for Monitoring in EPA's 2000 Guidance A-17
Table A-2. Initial List of Potential Analytes Based on Literature Search A-18
Table A-3. List of Potential Analytes with Fillet Data A-24
Table A-4. List of Potential Analytes with Whole Fish Data A-25
Table A-5. List of Potential Analytes with Shellfish Data A-27
September 2023 A-l
-------
Overview
State, tribal, and territorial fish and shellfish advisory programs should monitor and analyze fish
and shellfish in waterbodies within their jurisdictions for contaminants. When contaminants
occur in high enough concentrations to potentially affect the health of people eating fish and
shellfish from those waters, EPA recommends that those programs issue advisories regarding
consumption to protect the consumers. To help state, tribal, and territorial fish and shellfish
advisory programs, EPA recommends a set of contaminants to monitor in its 2000 version of
Volume 1 of Guidance for Assessing Chemical Contaminant Data for Use in Fish Advisories
(2000 Guidance), shown in Appendix Table A-l. In updating this guidance, EPA reviewed
scientific literature and determined that additional contaminants should be monitored. The
process and new contaminants identified as a result of that process are described here.
EPA performed a literature search to identify scientific information on the contaminants that
bioaccumulate in fish and shellfish and their corresponding concentrations. After searching
multiple databases, using a specified set of search terms, EPA extracted references, removed
duplicates, and screened articles to remove any that contained fish not found in U.S. waters or
contained concentration data only from lab studies. After extracting maximum and average
concentrations of contaminants from the papers, EPA used these concentrations in fish advisory
equations to determine if the levels found would exceed thresholds for restrictions in
consumption of fish.
This document describes the steps for screening scientific literature for compounds that
bioaccumulate in aquatic animals and deciding on their potential inclusion on the lists of analytes
that fish advisory programs should monitor.
Background
States, territories, and tribes issue consumption advisories for substances that occur in fish and
shellfish at levels of concern to human health. Consumption advisories issued by states and tribes
include freshwater, estuarine, and coastal marine fish and shellfish species. The selection of
appropriate target analytes in fish and shellfish contaminant monitoring programs is essential to
the issuance of these advisories. For identifying potential analytes, the primary necessary
element is evidence that the substance occurs in the edible tissue of consumed fish species at a
concentration that may cause exposure to be of concern for human health for identified health
endpoints.
Contaminant occurrence in fish tissue generally requires three conditions:
1. The compound must have been released to the environment in sufficient quantity.
2. The compound must be persistent in water and/or air for transport in the environment
once released.
3. The chemical nature of the substance must cause it to bioaccumulate in food webs due to
an affinity for fish tissues, which vary both by chemical and fish species characteristics.
Contaminants in fish must be persistent and bioaccumulative in nature to advance through the
food web and be ingested by humans to an extent that it could impact the health of consumers.
Persistent substances will avoid being dissolved in the aquatic environment and will remain
September 2023
A-2
-------
relatively intact while being transported in water and/or atmospherically. Persistence is required
because in the aquatic environment, chemical compounds are subject to the considerable solvent
power of the polar water molecule and other chemical and physical factors. These include effects
of the impingement of light (UV) and dissolved gasses in the water column, and interaction with
other dissolved constituents in water, especially in marine environments.
Bioaccumulation is a key mechanism required for compounds to occur at concentrations in fish
that can affect human health, and therefore is needed for a compound to be a candidate for an
advisory. In finfish, the mechanism for bioaccumulation is predation and resulting bio-
magnification up through the food chain. Starting with micro-organisms in sediment or the water
column, predation by larger species causes persistent compounds with an affinity for various
tissue types (e.g., lipids, muscle, bloody organs, blood) to increase at successive trophic levels to
ultimate concentrations in the fish (top predators usually contain the highest concentrations).
Because shellfish filter feed, they can take up toxic substances directly from the water column,
which they then store and accumulate. The accumulated compounds can become harmful to
consumers when ingested.
Developing an advisory for a compound generally requires data on its toxic effects on humans.
Reproductive, developmental (including neurodevelopmental), hepatotoxic (liver), and
immunotoxic are among the most common types of human health effects from exposure to
contaminants in fish. These effects can be quantified by a measure of oral toxicity. Two
measures used in assessing toxicity and risk to humans through a fish consumption pathway are
reference dose and cancer slope factor:
• Reference dose (RfD)\ a metric used to denote an amount of a contaminant that can be
consumed over a time period without adverse health effects. RfDs are typically expressed
as the ingestion in milligrams (mg) of a contaminant per kilogram (kg) of body weight
(of the consumer) per day. It specifically indicates an amount of a chemical to which a
person can be exposed on a daily basis over an extended period of time (usually a
lifetime), with a measure of uncertainty, without suffering a deleterious effect.
• Cancer slope factor (CSF): a metric used to describe the increase in cancer risk resulting
from a given rate of exposure to a substance, usually over a lifetime. Cancer slope factors
are typically expressed as typically expressed in units of proportion of a population
affected per milligram of substance per kilograms of body weight per day (expressed in
units of reciprocal dose (mg/kg-day)"1).
Literature Search
EPA initiated a systematic screening process (not a full-blown systematic review) to identify any
additional relevant compounds that are not currently included in the 2000 Guidance. EPA
developed a literature review protocol to compile peer-reviewed articles that provide a basis for
choosing any new analytes. Several environmental science, health, and toxicology databases
were searched for relevant peer-reviewed publications using keywords and inclusion and
exclusion criteria. These combined efforts resulted in more than 600 articles being compiled for
further review.
September 2023
A-3
-------
Pre-search Definition of Screening Criteria
Before beginning the literature search, EPA identified criteria in five key areas for selecting
publications:
1. Publication status
2. Publication date
3. Publication language
4. Fish species discussed
5. Compounds analyzed in the study.
EPA also defined screening factors for potential contaminants. EPA determined that each
compound should preferably meet all these criteria:
1. Present in fish and/or shellfish
2. Potential to bioaccumulate
3. Prevalent and persistent in the environment
4. Associated with evidence that eating fish and shellfish is a potential exposure pathway
5. Presence of toxicity information, preferably a reference dose or cancer slope factor
generated by the federal government (e.g., EPA, ATSDR)
6. Quantifiable in fish tissue with a validated analytical method capable of determining its
concentration at levels of human health relevance.
Article Inclusion Criteria
EPA included only articles that were:
• Peer-reviewed
• Published in 2000 or later (to capture information published after the 2000 Guidance)
• Written in English.
Sources Searched
EPA searched environmental science, health, and toxicology databases for relevant articles:
• PubMed
• PubChem
• Web of Science
o includes Science Citation Index Expanded, Social Sciences Citation Index, and
Conference Proceedings, Citation Indexes for Science and for Social Science and
Humanities
• Environmental Science and Pollution Management
o includes Aquatic Science and Fisheries Abstracts, Aquatic Pollution and
Environmental Quality, Water Resources Abstracts, TOXLINE, Toxicology
Abstracts, Environmental Abstracts, Pollution Abstracts, and Conference Papers
Index
• Science Direct
• Toxline
EPA also utilized an internal literature search that had been conducted for harmful algal blooms
(HABs) and HAB toxins.
September 2023
A-4
-------
Keywords
The keywords used to search the literature are listed by category in the following bullets.
• Target analytes
o PCBs or contaminant or constituent or contamination or emerging contaminant or
aquatic contaminant or chemicals or pollutants or metals or pesticides
• Aquatic animal types
o Finfish or fish or freshwater turtles or shellfish or bivalves or crustaceans or
mollusks
o AND edible tissue or muscle tissue
• Chemistry
o Persistent, halogenated, chlorinated, brominated, perfluorinated, cyclic,
polycyclic, ortho, meta, para
• Bioaccumulation
o Accumulation or bioconcentration or bioaccumulation or BCF or BAF or
bioaccumulate or bioconcentrate or bioaccumulative
• Toxicity
o Toxic or toxicity or human health benchmark or oral toxicity or RfD or reference
dose or cancer or slope factor
• Location
o United States, Alaska, Hawaii; freshwater, estuarine and marine waters
In addition to searches using the keywords above that were selected to cast a wide net, EPA
consulted publications by federal agencies (e.g., EPA, NIEHS in NIH), states (WA, MN, MI,
ME, NY), the Great Lakes Alliance, and articles listing potential persistent organic pollutants
(e.g., Brown and Wania, 2008; Sun et al, 2016; Muir et al, 2006; Stockholm Convention List) to
identify candidate compounds previously identified by other researchers. Then EPA conducted
literature searches pairing each compound name or its CAS number with these keywords:
o Fish, fish tissue, or concentration
Weight of evidence analysis
EPA screened the articles collected during the literature search, extracted information from the
peer-reviewed publications, and applied weight of evidence (WOE) points. Each publication
received one point based on whether it included information on the contaminant's detection in
fish, BCF or BAF data, oral toxicity data, and species found in the U.S. For example, if an article
had analyte concentration data in a U.S. fish species but no toxicity data nor BCF/BAFs, it
received two points. The points assigned ranged from a minimum of zero to a maximum of four.
Initially only articles with a WOE total of 3 or 4 points were included for further analysis.
However, EPA determined that requiring papers to have concentration data in fish tissue,
BCF/BAF data, and toxicity data artificially restricted the number of potential contaminants.
Articles with a WOE total of 2 or higher were examined for relevance and data.
Preliminary Contaminant List Compilation, Data Extraction, and Exclusions
EPA developed a preliminary list of 242 potential contaminants mentioned in the articles,
compiled during the literature search, with a WOE total of 2 or more (Appendix Table A-2).
September 2023
A-5
-------
Contaminants were found in the following classes:
• Antibacterial, antibiotic, and
antimicrobial compounds
• Brominated compounds
• Chlorinated compounds
• Cyanotoxins and neurotoxins
• Flame retardants
• Hormones
• Industrial byproducts
• Inorganics
• Metabolites
• Metals
Nanoparticles
PCBs
Personal care products
Pesticides
PFAS compounds
Pharmaceuticals
Phthalates
Polychlorinated naphthalenes
Polycyclic aromatic hydrocarbons
Sulfonamides
Other
• Organophosphorus esters
EPA then extracted concentration data from the articles. The articles EPA reviewed for the
compounds' concentration data mining effort contained concentration data in different units,
value formats (e.g., averages, ranges or maxes), and fish tissue types. After extracting the data
from the articles, EPA sorted the data by units. If articles did not provide enough information to
determine concentration units, EPA removed them from the dataset.
Articles reported concentrations in three different units: wet weight, lipid weight, and dry weight.
EPA prioritized wet weight data because it is the most common unit used for benchmark
analyses. To use the concentration data presented as lipid or dry weight, EPA needed additional
% lipid or % moisture data, which was often not available in the articles reviewed. Therefore, if
EPA compiled data for a compound in more than one unit (e.g., there was lipid and wet weight
data), EPA used only the wet weight data in the calculations. Some compounds did not have any
concentrations in wet weight units; these were kept as lipid or dry weight and not modified.
The articles EPA reviewed for compounds contained concentration data not only for fillet tissue,
but also for whole body, organs, blood, etc. Each article EPA reviewed had a different study
objective, and, thus, different fish and shellfish tissues were analyzed. Tissue data were sorted
into several categories: fillet, whole body, organs (including liver, kidney, blood and eggs),
shellfish only, and eel only. Articles which did not specify fish tissue were removed from the
literature review.
There were data from more than 75 species in the concentration data mining effort. Data from
non-native species were removed unless they have been found in a U.S waterbody in the last 10
years. Only two non-native species were retained in this list: Brown trout and Common carp.
These species are invasive and abundant in U.S. waters.
Fish advisory programs managers are most interested in what potentially affects fish in ambient
waters. If the studies were not analyzing ambient conditions (e.g., fish were dosed in a lab
study), then those concentrations were removed.
After sorting the data by tissue type and species, EPA collected maximum and average
concentration data, and converted data, where necessary, to ng/g. Not all articles provided
maximum and average values. If a range of concentrations was reported, where possible, the
maximum value from the concentration range was used as the maximum value. Maximum values
September 2023 A-6
-------
reported as an inequality (e.g., >6) were excluded. Average concentrations reported as a range or
with an inequality symbol (e.g., <4.8) were excluded.
EPA removed compounds without concentration data, compounds already on the monitoring list
in the 2000 Guidance, and mixtures (e.g., BDE-119 + BDE-120, sum of PFAS), and this resulted
in lists of 49 potential contaminants with fillet data, 55 with whole fish data, and 14 with
shellfish data (Appendix Tables A-3, A-4, and A-5).
Researched Toxicity Values
After extracting toxicity information from articles and U.S. government sources such as EPA
IRIS and ATSDR, EPA searched for additional and updated toxicology data.
EPA searched the following eight peer-reviewed, publicly available sources, as described in its
2015 update of human health ambient water quality criteria, to obtain the toxicity values
(reference dose, minimum risk level, and/or cancer slope factor) and used them in screening level
calculations, in this order of preference:
1. EPA's Integrated Risk Information System (IRIS) program
2. EPA's Office of Pesticide Programs Pesticide Chemical Search
3. EPA's Office of Pollution Prevention and Toxics Existing Chemicals
4. EPA's Office of Water Water Topics
5. EPA's Office of Solid Waste and Emergency Response Provisional Peer Reviewed
Toxicity Values for Superfund (PPRTY)
6. U.S. Department of Health and Human Services. Agency for Toxic Substances and
Disease Registry (ATSDR) Toxic Substances Portal
7. Health Canada
8. California Environmental Protection Agency's Office of Environmental Health Hazard
Assessment - All Public Health Goals
For PFAS compounds, EPA used the reference doses for PFOS, PFOA, PFNA, PFBS, and
PFHxS that were used in the proposed national primary drinking water regulation released on
March 14, 2023, and the reference doses in IRIS for PFBA, PFBS, PFDA, and PFHxA. EPA did
not use toxicity values from any sources other than those listed in this section, because of the
variability of methods applied and inconsistency of the existence of adequate quality control
documentation.
Screening Level Calculations and Analyses
For each contaminant with a non-cancer toxicity value, EPA calculated a non-cancer screening
value using this equation from the 2000 Guidance:
reference dose x consumer body weight
Non-cancer screening level = :
consumption rate
where:
Body weight of adult in general population and of frequent fish consumer = 80 kg
Body weight of pregnant person = 75 kg
September 2023
A-7
-------
Consumption rate of adult in general population and pregnant person = 8 oz/week * 28.35
g/oz * 1 week/7d = 32.4 g/d1
Consumption rate of frequent fish consumer = 142 g/d2
The reference doses were provided in mg/kg-day. EPA multiplied them by 1000 to convert them
into micrograms/kg-day in order to calculate non-cancer screening levels in micrograms per
gram (ppm). Concentration data were all converted to micrograms per gram, and the non-cancer
screening levels were compared to the concentration data found in the scientific literature.
EPA analyzed whether:
• The maximum concentration, which is the highest concentration found in reported
maximum concentrations extracted from articles, exceeds the non-cancer screening level
for an adult in the general population
• The maximum concentration exceeds the non-cancer screening level for a pregnant
person
• The maximum concentration exceeds the non-cancer screening level for a frequent fish
consumer
• The average concentration, which is the average of the reported average concentrations
extracted from articles, exceeds the non-cancer screening level for an adult in the general
population
• The average concentration exceeds the non-cancer screening level for a pregnant person
• The average concentration exceeds the non-cancer screening level for a frequent fish
consumer
EPA also calculated whether the maximum concentration and average concentration were within
75 percent of the non-cancer screening level for an adult in the general population to see if there
were compounds that could be problematic but not currently accumulating to problematic levels.
EPA did not find additional compounds to include as a result of those analyses.
A contaminant's presence in fish does not necessarily indicate a human health risk exists. For the
contaminants without non-cancer toxicity values, EPA calculated a "generic" screening level to
capture contaminants with fish tissue concentrations high enough to potentially be a human
health concern after reference doses are developed. In its screening level calculation, EPA used
the lowest final toxicity value (that is, the most stringent toxicity value that was not draft or
being developed) available among the contaminants found in fish. The lowest reference dose for
compounds that were considered for inclusion on the monitoring list in this evaluation is 3 x 10"6
mg/kg-d (PFNA), which was then multiplied by 1000 to convert to micrograms/kg-day. The
calculated "generic" screening level was 7.41 x 10"3 |ig/g. Concentration data found in the
1 This consumption rate is based on the recommendation in the U.S. Department of Agriculture and Department of
Health and Human Services' Dietary Recommendations for Americans. 2020-2025 to eat 8-10 ounces per week of
seafood.
2 From EPA's Methodology for Deriving Ambient Water Quality Criteria for the Protection of Human Health (2000)
that recommends a default of 142.4 grams/day for subsistence fishers.
September 2023
A-8
-------
scientific literature were all converted to micrograms per gram, and the screening levels were
compared to the concentration data. EPA analyzed whether:
• The maximum concentration exceeds the "generic" non-cancer screening level
• The average concentration exceeds the "generic" non-cancer screening level
For each contaminant with a cancer slope factor, EPA calculated a cancer screening value using
this equation and constants:
cancer risk level x consumer body weight
Cancer screening level = : :
cancer slope factor x consumption rate
where:
Cancer risk level = 10"6
Body weight of adult in general population and of frequent fish consumer = 80 kg
Consumption rate of general population = 32.4 g/d
Consumption rate of frequent fish consumer = 142 g/d
[Note: EPA's 2000 Guidance presented fish meal calculations based on a cancer risk level of
10"5. EPA is considering updating that factor to 10"6 to be consistent with methods for
developing water quality criteria. To be conservative and ensure it captured problematic
compounds, EPA used the 10"6 cancer risk level in these screening level calculations.]
To calculate cancer screening levels in micrograms per gram (ppm), EPA multiplied the cancer
slope factors by 1000, converting them from mg/kg-day into micrograms/kg-day. After
converting concentration data to micrograms per gram, EPA compared the cancer screening
levels to the concentration data.
EPA analyzed whether the:
• Maximum concentration exceeds the cancer screening level for an adult in the general
population
• Maximum concentration exceeds the cancer screening level for a frequent fish consumer
• Average concentration exceeds the cancer screening level for an adult in the general
population
• Average concentration exceeds the cancer screening level for a frequent fish consumer
Results of Comparing Concentration Data to Screening Levels
The following subsections describe which contaminants exceeded the specific and generic non-
cancer and cancer screening levels in fillet and whole fish data.
Contaminants exceeding screening levels in fillet data
The 10 analytes in Table 1 had concentrations in fillet tissue that exceeded one or more of the
screening levels that were calculated for each compound. The table shows which screening levels
were exceeded (cancer and/or non-cancer for adult in general population, pregnant person, and/or
September 2023
A-9
-------
frequent fish consumer) and which data type (average concentration, maximum concentration, or
both) exceeded the screening level.
Table 1. Contaminants exceeding screening levels in fillet data
Contaminant
Reason on list
BDE-47
Maximum concentration exceeds non-cancer screening levels for adult in
general population, pregnant person & frequent fish consumer
BDE-99
Maximum concentration exceeds non-cancer screening level for frequent
fish consumer
Lead
Maximum concentration exceeds cancer screening levels for adult in
general population and frequent fish consumer
Microcystins
Maximum concentration and average concentration exceed non-cancer
screening levels for adult in general population, pregnant person, and
frequent fish consumer
PFDA
Maximum concentration and average concentration exceed non-cancer
screening levels for adult in general population, pregnant person, and
frequent fish consumer
PFHxS
Maximum concentration exceeds non-cancer screening level for frequent
fish consumer
PFNA
Maximum concentration exceeds non-cancer screening levels for adult in
general population, pregnant person, and frequent fish consumer
PFOS
Maximum and average concentration exceed non-cancer screening levels
for adult in general population, pregnant person, and frequent fish
consumer and cancer screening levels for adult in general population and
frequent fish consumer
PFOA
Maximum concentration and average concentration exceed non-cancer
screening levels for adult in general population, pregnant person, and
frequent fish consumer
Thallium
Maximum concentration and average concentration exceed non-cancer
screening levels for adult in general population, pregnant person, and
frequent fish consumer
(Note: BDE-47, BDE-99, lead, PFHxS, and PFNA did not have any average concentration data
to extract from literature.)
Contaminants exceeding screening levels in whole fish data
The eight analytes in Table 2 had concentrations in whole fish tissue that exceeded one or more
of the screening levels that were calculated for each compound. The table shows which screening
levels were exceeded (cancer and/or non-cancer for adult in general population, pregnant person,
and/or frequent fish consumer) and which data type (average concentration, maximum
concentration, or both) had a value that exceeded the screening level.
September 2023
A-10
-------
Table 2. Contaminants exceeding screening levels in whole fish data
Contaminant
Reason on list
BDE-47
Maximum concentration exceeds non-cancer screening levels for adult in
general population, pregnant person, and frequent fish consumer
BDE-99
Maximum concentration exceeds non-cancer screening levels for adult in
general population, pregnant person, and frequent fish consumer
Lead
Maximum concentration exceeds cancer screening levels for adult in general
population and frequent fish consumer
PFDA
Maximum and average concentration exceed non-cancer screening levels for
adult in general population, pregnant person, and frequent fish consumer
PFHxS
Maximum concentration exceeds non-cancer screening levels for adult in
general population, pregnant person, and frequent fish consumer
PFNA
Maximum concentration exceeds non-cancer screening level for frequent fish
consumer
PFOA
Maximum concentration exceeds non-cancer screening levels for adult in
general population, pregnant person and frequent fish consumer
PFOS
Maximum and average concentration exceed non-cancer screening levels for
adult in general population, pregnant person and frequent fish consumer
(Note: lead and
3FOA did not have average concentration data to extract from literature.)
Contaminants exceeding screening levels in shellfish data
The five analytes in Table 3 had concentrations in shellfish that exceeded one or more of the
screening levels calculated for each compound. The table shows which screening levels were
exceeded (cancer and/or non-cancer for adult in general population, pregnant person, and/or
frequent fish consumer) and which data type (average concentration, maximum concentration, or
both) had a value that exceeded the screening level.
Table 3. Contaminants exceeding screening levels in shellfish data
Contaminant
Reason on list
Microcystins
Maximum concentration exceeds non-cancer screening levels for adult in
general population, pregnant person, and frequent fish consumer and average
concentration exceeds non-cancer screening level for frequent fish consumer
PFDA
Maximum and average concentrations exceed non-cancer screening levels for
adult in general population, pregnant person, and frequent fish consumer
PFNA
Maximum concentration exceeds non-cancer screening levels for adult in
general population, pregnant person and frequent fish consumer and average
concentration exceeds non-cancer screening level for frequent fish consumer
PFOA
Maximum and average concentrations exceed non-cancer screening levels for
adult in general population, pregnant person and frequent fish consumer
PFOS
Maximum and average concentrations exceed non-cancer and cancer
screening levels for adult in general population, pregnant person and frequent
fish consumer
September 2023
A-ll
-------
Contaminants exceeding "generic" screening level in fillet data
For contaminants without an EPA RfD, EPA calculated a generic screening level using the most
stringent RfD of the compounds found in fish tissue: 3 x 10"6 mg/kg-d. This resulted in a
screening level of 0.00741 |ig/g. The 12 analytes in Table 4 had concentrations in fillet tissue
that exceeded the generic screening level. The table also shows which data type (average
concentration, maximum concentration, or both) had a value that exceeded the screening level.
The analytes in Table 4 all had maximum concentrations in fillet tissue that exceeded the generic
screening level, except the HAB toxin DAB A, for which there was only average concentration
data and that exceeded the screening level. In addition to the exceedances of screening levels for
maximum concentrations, BMAA also had average concentration data that exceeded the
screening level. The average concentration data for PFDoA, PFOSA, and PFUnA did not exceed
the screening level; the rest did not have average concentration data.
Table 4. Contaminants exceeding generic screening level in fillet data
Analyte
Concentration type that exceeds
the generic screening level
BMAA (P-methylamino-L-alanine)
Maximum concentration
Average concentration
DABA (2,4-diaminobutyric acid
dihydrochloride)
Average concentration2
Dechlorane 6023
Maximum concentration1
Medium chain chlorinated paraffins
Maximum concentration1
Perfluorodecanesulfonic acid (PFDS)
Maximum concentration1
Perfluorododecanoate (PFDoA)
Maximum concentration
Perfluorooctanesulfonamide (PFOSA)
Maximum concentration
Perfluorotetradecanoic acid (PFTeDA)
Maximum concentration1
Perfluorotridecanoate (PFTrDA)
Maximum concentration1
Perfluoroundecanoate (PFUdA, PFUnA,
PFUnDA)
Maximum concentration
Perfluoro-2-propoxypropanoic acid
(PFPrOPrA)
Maximum concentration1
Short chain chlorinated paraffins
Maximum concentration1
1 No average data was extracted for this compound.
2 No maximum data was extracted for this compound.
3 Results for this compound are based on lipid data.
Contaminants exceeding "generic" screening level in whole fish data
For contaminants without an EPA RfD, EPA calculated a generic screening level using the most
stringent RfD of the compounds found in fish tissue: 3 x 10"6 mg/kg-d. This resulted in a
screening level was 0.00741 |ig/g. The 12 analytes in Table 5 had concentrations in whole body
tissue that exceeded the generic screening level. The table also shows which data type (average
or maximum concentration) had a value that exceeded the screening level.
September 2023
A-12
-------
Table 5. Contaminants exceeding generic screening level in whole fish data
Analyte
Concentration type that exceeds
the generic screening level
Amphetamine
Average concentration1
BDE-49
Maximum concentration3
BDE-100
Maximum concentration3
DBDPE (Decabromodiphenyl ethane)1
Average concentration1
Dechlorane 602 2
Average concentration1
Dechlorane 604 2
Average concentration1
Metformin
Average concentration1
Long-chain chlorinated paraffins
Maximum concentration3
PFDoA
Maximum concentration
PFUnA
Maximum concentration
Sertraline
Average concentration1
Sulfadimethoxine
Average concentration1
1 No maximum data was extracted for this compound.
2 Results for this compound are based on lipid data.
3 No average data was extracted for this compound.
Contaminants exceeding "generic" screening level in shellfish data
For contaminants without an EPA RfD, EPA calculated a generic screening level using the most
stringent RfD of the compounds found in fish and shellfish tissue: 3 x 10"6 mg/kg-d. This
resulted in a screening level of 0.00741 |ig/g. The analyte in Table 6 had concentrations in
shellfish tissue that exceeded the generic screening level. The table also shows which data type
(average concentration, maximum concentration, or both) had a value that exceeded the
screening level.
Table 6. Contaminants exceeding generic screening level in shellfish data
Analyte
Concentration type that exceeds
the generic screening level
Perfluoroundecanoate (PFUdA, PFUnA,
PFUnDA)
Maximum concentration
Average concentration
September 2023
A-13
-------
Results of Process
EPA is proposing to develop two lists of compounds based on available information:
1. Contaminants to Monitor for Advisories: The compounds on this list are found to occur in
fish at levels of concern to human health and have toxicity information in the form of an
EPA oral reference dose, ATSDR minimum risk level, or EPA cancer slope factor. Table
7 shows the new compounds EPA is proposing to add to its existing list (in 2000
Guidance) of recommended compounds to monitor for fish advisories based on the
process described in this document.
2. Contaminants to Monitor to Watch. The compounds on this list are those that may need
advisories in the future. They are on the list because they have documented
concentrations in fish and/or shellfish that could be of concern for human health, based
on the generic screening level used in these analyses, but the federal government has not
yet developed toxicity values such as reference doses or cancer slope factors for them.
Table 8 shows the compounds EPA has identified based on the process described in this
document.
Contaminants to Monitor for Advisories
EPA proposes to add ten contaminants (five PFAS compounds, one cyanotoxin, two flame
retardants, and two metals) to the Contaminants to Monitor for Advisories list in the 2000
Guidance; these are shown in Table 7. These compounds met the criteria of being documented in
studies as occurring in edible tissue of consumed fish or shellfish species at a concentration that
exceeds the screening level associated with the reference dose (for non-cancer effects) or cancer
slope factor and cancer risk level (for cancer effects) and therefore is of concern for human
health.
Table 7. New contaminants to monitor for advisories
Class
Analyte
PFAS
Perfluorodecanoic acid (PFDA)
Perfluorohexane sulfonic acid (PFHxS)
Perfluorononanoic acid (PFNA)
Perfluorooctanoic acid (PFOA)
Perfluorooctane sulfonic acid (PFOS)
Cyanotoxins
Microcystins
Flame retardants
BDE-47
BDE-99
Metals
Lead
Thallium
Contaminants to Monitor to Watch
EPA proposes to add twenty-one contaminants (two cyanotoxins, five flame retardants, seven
PFAS compounds, four pharmaceuticals, and three divisions of chlorinated paraffins) to a newly
created Contaminants to Monitor to Watch list; these are shown in Table 8. These compounds
September 2023
A-14
-------
met the criteria of being documented in studies as occurring in edible tissue of consumed fish or
shellfish species at a concentration that could be of concern for human health, based on the
generic screening level used in these analyses. These compounds do not currently have
government-issued reference doses or cancer slope factors. After the relevant toxicity values are
developed, these compounds should be re-evaluated to determine if they warrant inclusion on
consumption advisories.
Table 8. Contaminants to monitor to watch
Class
Analyte
Cyanotoxins
BMAA (|3-methylamino-L-alanine)
DABA (2,4-diaminobutyric acid dihydrochloride)
Flame retardants
BDE-49
BDE-100
Decabromodiphenyl ethane (DBDPE)
Dechlorane 602
Dechlorane 604
PFAS
Perfluorodecanesulfonic acid (PFDS)
Perfluorododecanoate (PFDoA)
Perfluorooctanesulfonamide (PFOSA)
Perfluorotetradecanoic acid (PFTeDA)
Perfluorotridecanoate (PFTrDA)
Perfluoroundecanoate (PFUdA, PFUnA, PFUnDA)
Perfluoro-2-propoxypropanoic acid (PFPrOPrA)
Pharmaceuticals
Amphetamine
Metformin
Sertraline
Sulfadimethoxine
Other
Long-chain chlorinated paraffins
Medium-chain chlorinated paraffins
Short-chain chlorinated paraffins
References Cited
Brown, Trevor and Frank Wania, 2008. Screening Chemicals for the Potential to be Persistent
Organic Pollutants: A Case Study of Arctic Contaminants. Environ. Sci. Technol. 2008, 42, 14,
5202-5209. Publication Date: June 11, 2008. https://doi.org/10.1021/es8004514
Muir, Derek and Philip Howard, 2006. Are There Other Persistent Organic Pollutants? A
Challenge for Environmental Chemists. Environ. Sci. Technol. 2006, 40, 23, 7157-7166.
Publication Date: November 2, 2006. https://doi.org/10.1021/es061677a
Stockholm Convention on Persistent Organic Pollutants (POPs): Text and Annexes, Revised in
2019. Secretariat of the Stockholm Convention. September 2000.
Sun, Mei & Arevalo, Elisa & Strynar, Mark & Lindstrom, Andrew & Richardson, Michael &
Kearns, Ben & Smith, Chris & Pickett, Adam & Knappe, Detlef, 2016. Legacy and Emerging
Perfluoroalkyl Substances Are Important Drinking Water Contaminants in the Cape Fear River
Watershed of North Carolina. Environmental Science & Technology Letters. 3 (12), pp 415-419.
September 2023
A-15
-------
DOI: 10.1021/acs.estlett.6b00398.
USEPA. 2015. 2015 update of human health ambient water quality criteria. Found at
https://www.epa.gOv/wqc/human-health-water-quality-criteria-and-methods-toxics#2015.
USEPA. 2016. Chemical-specific Inputs for EPA's 2015 Final Updated Human Health Ambient
Water Quality Criteria. March 16, 2016 revision. Found at
https://www.epa. gov/ sites/production/files/2016-
03/documents/summary of inputs final revised 3.24.16.pdf
USEPA. 2000. Guidance for Assessing Chemical Contaminant Data for Use in Fish Advisories.
Found at https://www.epa.gOv/fish-tech/epa-guidance-developing-fish-advisories#national.
September 2023
A-16
-------
Appendix
Table A-l shows the set of contaminants that EPA recommends state, tribal, and territorial fish advisory
programs to monitor in its 2000 version of Volume 1 of Guidance for Assessing Chemical Contaminant
Data for Use in Fish Advisories.
Table A-l. Analytes Recommended for Monitoring in EPA's 2000 Guidance
Metals
Arsenic (inorganic)
Cadmium
Mercury (methylmercury)
Selenium
Tributyltin
Organochlorine Pesticides
Chlordane, total (cis- and trans-chlordane, cis- and trans-nonachlor,
oxychlordane)
DDT, total (2,4'-DDD, 4,4'-DDD, 2,4'-DDE, 4,4'-DDE, 2,4'-DDT, 4,4'-DDT)
Dicofol
Dieldrin
Endosulfan (1 and II)
Endrin
Heptachlor epoxide
Hexachlorobenzene
Lindane (y-hexachlorocyclohexane; y-HCH)
Mi rex
Toxaphene
Organophosphate Pesticides
Chlorpyrifos
Diazinon
Disulfoton
Ethion
Terbufos
Chlorophenoxy Herbicides
Oxyfluorfen
Polycyclic aromatic
hydrocarbons (PAHs)
Polychlorinated biphenyls
Total PCBs (sum of PCB congeners or Aroclor equivalents)
(PCBs)
Dioxins/furans
DDT = p,p'-dichlorodiphenyl trichloroethane
DDE = p,p'-dichlorodiphenyl dichloroethylene
DDD = dichlorodiphenyldichloro ethane
September 2023
A-17
-------
Table A-2 shows the contaminants that EPA compiled during the literature search with a weight of
evidence score of 2 or more.
Table A-2. Initial List of Potential Analytes Based on Literature Search
Analyte
Analyte Classification
Azithromycin
Antibacterial, antibiotic, and antimicrobial
Cefotiam
compounds
Cefoxitin
Ciprofloxacin
Enrofloxacin
Erythromycin
ICON (active ingredient fipronil)
Lincomycin
Linezolid
Norfloxacin
Norfluoxetine
Ofloxacin
Roxithromycin
Sulfadimethoxine
Sulfamethoxazole
Sulfaquinoxaline
Triclocarban
Triclosan
Polybrominated biphenyls (PBBs)
Brominated compounds
1,3,6,8-Tetrabromocarbazole
1,3,6,8-Tetrachlorocarbazole
1,3,6-Tribromocarbazole
l,8-Dibromo-3,6-dichlorocarbazole
l-Bromo-3,6-dichlorocarbaozle
2,3,6,7-Tetrachlorocarbazole
2,7-Dibromocarbazole
3,6-Dibromocarbazole
3,6-Dichlorocarbzole
Chlorinated compounds
3-Bromocarbazole
3-Chlorocarbazole
Chlorobenzenes
Chloronaphthalene, 1-
Chloronaphthalene, 2-
Medium chain chlorinated paraffins
Polychlorinated biphenyls (PCBs)
Polychlorinated naphthalenes (PCNs)
(B-methylamino-L-alanine (BMAA)
Cyanotoxins and neurotoxins
Cylindrospermopsin
September 2023
A-18
-------
Analyte
Analyte Classification
Diaminobutyric acid dihydrochloride (DABA)
Domoic acid
Microcystins
Saxitoxin
Cyanotoxins and neurotoxins
l,2-dibromo-4-(l,2-dibromoethyl) cyclohexane (TBECH)
BDE, deca-
BDE, octa-
BDE, penta-
BDE-100
BDE-119
BDE-119+BDE-120
BDE-126
BDE-153 (Hexabromodiphenyl ether, 2,2',4,4',5,5')
BDE-154
BDE-155
BDE-17+BDE-25
BDE-180
BDE-183
BDE-197
BDE-198
BDE-203
BDE-204
BDE-206
Flame retardants
BDE-207
BDE-208
BDE-209 (Decabromodiphenyl ether)
BDE-28
BDE-28+BDE-33
BDE-47 (Tetrabromodiphenyl ether, 2,2',4,4'-)
BDE-49
BDE-51
BDE-66
BDE-75
BDE-77
BDE-99 (Pentabromodiphenyl ether, 2,2',4,4',5-)
Debrominated diphenyl ethers (De-BDEs)
Dechlorane 602
Dechlorane 603
Dechlorane 604
Dechlorane Plus
Hexabromocyclododecane (HBCDs)
Hexabromocyclododecane, alpha- (a-HBCDD)
September 2023
A-19
-------
Analyte
Analyte Classification
Hexabromocyclododecane, beta- ((B-HBCDD)
Flame retardants
Hexabromocyclododecane, gamma- (y-HBCDD)
Hexabromocyclododecane, Sum (HBCDD)
Methoxylated brominated diphenyl ethers (MeO-BDE)
Monohydroxylated Polybrominated Diphenyl Ethers (OH-
PBDEs)
Polybrominated diphenyl ethers (PBDEs)
Tetrabromobisphenol A
Androstenedione
Hormones
Estradiol, 17(B-
Estrone
Norethindrone
Testosterone
Chlorinated paraffins
Industrial by-products
Naphthenic acids
Octachlorostyrene
Produced water
Ammonia (NH3)
Inorganics
Bromine
Calcium
Chlorine
Iodine
Nitrate
Nitrite
Orthophosphate
Rare earth elements
Sulphate (S02)
Zinc chloride
Dimethylarsinate (DMA)
Metabolites
Arsenobetaine
Benzoylecgonine
Desmethyldiltiazem
Hydroxypyrene, 1-
Aluminum
Metals
Antimony
Barium
Beryllium
Boron
Cadmium
Cesium
Chromium (III)
Chromium (VI)
September 2023
A-20
-------
Analyte
Analyte Classification
Cobalt
Copper
Iron
Lead
Lithium
Magnesium
Manganese
Molybdenum
Nickel
Potassium
Radiocesium
Rubidium
Silver
Metals
Sodium
Strontium
Thallium
Tin
Titanium
Tungsten
Uranium
Vanadium
Zinc
Copper oxide nanoparticles (CuO NPs)
Nanoparticles
Nanoparticles
Zinc oxide nanoparticles (ZnO NPs)
Radionuclides
Semivolatile organic compounds (SVOCs)
Other
SSCP
Trinitrotoluene, 2,4,6- (TNT)
Aroclor 1016
PCBs
Aroclor 1254
Cyclic volatile methyl siloxanes (cVMS)
Hexamethyldisiloxane (HMDS)
Personal care products
Linear alkylbenzenes (LABs)
Octamethylcyclotetrasiloxane (D4)
Aldrin
Atrazine
Benzene hexachloride, alpha-
Benzene hexachloride, beta-
Pesticides
Copper pyrithione (CuPT)
Hexachlorobenzene (HCB)
Isodrin
September 2023
A-21
-------
Analyte
Analyte Classification
Methoxychlor
Methylarsonate (MA)
Monomethylarsonic acid (MMA)
Monosodium methanearsonate
N,N-Diethyl-meta-toluamide (DEET)
Pesticides
Rhothane (TDE)
Telodrin
Thiabendazole
Triphenyltin
Perfluoro-2-propoxypropanoic acid (PFPrOPrA)
Pentafluorobenzoic acid (PFBA)
Perfluorobutane sulfonate (PFBS)
Perfluorodecane sulfonate (PFDS)
Perfluorodecanoate (PFDA)
Perfluorododecanoate (PFDoA)
Perfluoroheptanesulfonic acid (PFHpS)
Perfluoroheptanoate (PFHpA)
Perfluorohexane sulfonate (PFHxS)
Perfluorohexanoate (PFHxA)
PFAS
Perfluorononanoate (PFNA)
Perfluorooctane sulfonate (PFOS)
Perfluorooctanesulfonamide (PFOSA)
Perfluorooctanoate (PFOA)
Perfluoropentanoate (PFPA)
Perfluoropentanoic acid (PFPeA)
Perfluorotetradecanoate (PFTA)
Perfluorotetradecanoic acid (PFTeDA)
Perfluorotridecanoate (PFTrDA)
Perfluoroundecanoate (PFUdA, PFUnA, PFUnDA)
Albuterol
Alprazolam
Amitriptyline
Amlodipine
Amphetamine
Atenolol
Atorvastatin
Pharmaceuticals
Benztropine
Caffeine
Carbamazepine
Cimetidine
Clarithromycin
Cocaine
September 2023
A-22
-------
Analyte
Analyte Classification
Codeine
Cotinine
Diazepam
Diclofenac
Diltiazem
Dimethylxanthine
Diphenhydramine, 1,7-
Enalapril
Fluoxetine
Furosemide
Gemfibrozil
Glyburide
Hydrochlorothiazide
Hydrocodone
Ibuprofen
Meprobamate
Metformin
Metoprolol
Pharmaceuticals
Miconazole
Naproxen
Nifedipine (Dehydro)
Norverapamil
Oxycodone
Paroxetine
Promethazine
Propoxyphene
Propranolol
Ranitidine
Sertraline
Simvastatin
Triamterene
Trimethoprim
Valsartan
Verapamil
September 2023
A-23
-------
Table A-3 shows the potential contaminants with a weight of evidence score or 2 or more and fillet data
that EPA could extract from the literature.
Table A-3. List of Potential Analytes with Fillet Data
Classification
Analyte
Antibiotics
Sulfaquinoxaline
BDE-100
BDE-153 (Hexabromodiphenyl ether,
2,2',4,4',5,5')
BDE-154
BDE-155
BDE-209 (Decabromodiphenyl ether)
BDE-47 (Tetrabromodiphenyl ether, 2,2',4,4'-)
BDE-49
Flame retardants
BDE-66
BDE-99 (Pentabromodiphenyl ether, 2,2',4,4',5-)
Dechlorane 602
Dechlorane 603
Dechlorane 604
HBCDD, alpha- (Hexabromocyclododecane)
HBCDD, beta-
HBCDD, gamma-
Tetrabromobisphenol - A (TBBP-A)
BMAA ((B-methylamino-L-alanine)
Cyanotoxins and
Cylindrospermopsin
neurotoxins
DABA (2,4-diaminobutyric acid dihydrochloride)
Microcystins
Lead
Metals
Nickel
Thallium
Pentafluorobenzoic acid (PFBA)
Perfluorobutane sulfonate (PFBS)
Perfluorodecanesulfonic acid (PFDS)
Perfluorodecanoic acid (PFDA)
Perfluorododecanoate (PFDoA)
Perfluoroheptanesulfonic acid (PFHpS)
PFAS
Perfluoroheptanoate (PFHpA)
Perfluorohexane sulfonate (PFHxS)
Perfluorohexanoate (PFHxA)
Perfluorononanoate (PFNA)
Perfluorooctane sulfonate (PFOS)
Perfluorooctanesulfonamide (PFOSA)
Perfluorooctanoic acid (PFOA)
September 2023
A-24
-------
Classification
Analyte
Perfluoropentanoic acid (PFPeA)
Perfluorotetradecanoic acid (PFTeDA)
Perfluorotridecanoate (PFTrDA)
Perfluoroundecanoate (PFUdA, PFUnA, PFUnDA)
Perfluoro-2-propoxypropanoic acid (PFPrOPrA)
Other
Medium chain chlorinated paraffins
Radiocesium
Short chain chlorinated paraffins
Table A-4 shows the potential contaminants with a weight of evidence score or 2 or more and whole
body data that EPA could extract from the literature.
Table A-4. List of Potential Analytes with Whole Fish Data
Classification
Analyte
Azithromycin
Antibacterial,
Erythromycin
antibiotic, and
Norfluoxetine
antimicrobial
Sulfadimethoxine
compounds
Triclocarban
Triclosan
BDE-100
BDE-153 (Hexabromodiphenyl ether,
2,2',4,4',5,5')
BDE-154
BDE-155
BDE-47 (Tetrabromodiphenyl ether, 2,2',4,4'-)
BDE-49
BDE-66
BDE-75
Flame retardants
BDE-99 (Pentabromodiphenyl ether, 2,2',4,4',5-)
Chlordene Plus
DBDPE (Decabromodiphenyl ethane)
Dechlorane 602
Dechlorane 604
HBB (Hexabromobiphenyl)
HBCDD, beta-
HBCDD, gamma-
PBEB (pentabromoethylbenzene)
Syn-Dechlorane plus
Metals
Lead
PFAS
Perfluorobutane sulfonate (PFBS)
September 2023
A-25
-------
Classification
Analyte
Perfluorodecanoic acid (PFDA)
Perfluorododecanoate (PFDoA)
Perfluoroheptanoate (PFHpA)
Perfluorohexane sulfonate (PFHxS)
Perfluorohexanoate (PFHxA)
PFAS
Perfluorononanoate (PFNA)
Perfluorooctane sulfonate (PFOS)
Perfluorooctanesulfonamide (PFOSA)
Perfluorooctanoic acid (PFOA)
Perfluorotetradecanoic acid (PFTeDA)
Perfluorotridecanoate (PFTrDA)
Perfluoroundecanoate (PFUdA, PFUnA, PFUnDA)
Alprazolam
Amitriptyline
Amlodipine
Amphetamine
Diazepam
Diltiazem
Diphenhydramine, 1,7-
Pharmaceuticals
Fluoxetine
Gemfibrozil
Metformin
Miconazole
Norverapamil
Ranitidine
Sertraline
Verapamil
Other
Long-chain chlorinated paraffins
September 2023
A-26
-------
Table A-5 shows the potential contaminants with a weight of evidence score or 2 or more and shellfish
data that EPA could extract from the literature.
Table A-5. List of Potential Analytes with Shellfish Data
Classification
Analyte
Cyanotoxins and
Microcystins
neurotoxins
Perfluorobutane sulfonate (PFBS)
Perfluorodecanesulfonic acid (PFDS)
Perfluorodecanoic acid (PFDA)
Perfluorododecanoate (PFDoA)
Perfluoroheptanoic acid (PFHpA)
Perfluorohexane sulfonate (PFHxS)
PFAS
Perfluorohexanoate (PFHxA)
Perfluorononanoate (PFNA)
Perfluorooctane sulfonate (PFOS)
Perfluorooctanoic acid (PFOA)
Perfluorotetradecanoic acid (PFTeDA)
Perfluorotridecanoate (PFTrDA)
Perfluoroundecanoate (PFUdA, PFUnA, PFUnDA)
September 2023
A-27
-------
Call Order 68HERH23F0356 under
Contract GS-00F-079CA
BPA #68HERH23A0019
External Peer Review of the Process for
Selecting Contaminants to Monitor in
Fish Advisory Programs
FINAL PEER REVIEW SUMMARY REPORT
October 2023
Submitted to:
U.S. Environmental Protection Agency
Office of Water, Office of Science and Technology
Standards and Health Protection Division
1200 Pennsylvania Avenue, NW
Washington, DC 20460
Submitted by:
ERG
561 Virginia Road
Building 4 - Suite 300
Concord, MA 01742
%ERG
www.erg.com
A-28
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
CONTENTS
1.0 INTRODUCTION A-30
1.1 Background A-30
1.2 Peer Reviewers A-30
2.0 SUMMARY OF REVIEWER COMMENTS ORGANIZED BY CHARGE QUESTION A-31
2.1 Is the process EPA followed to identify compounds for which fish and shellfish advisories
might be needed reasonable? A-31
2.2 Is the list of contaminants advisory programs should consider monitoring for reasonable
(e.g., reflects the current range of contaminants detected in fish with potential human
health impacts)? A-42
2.3 Are there additional contaminants that should be included in the "monitor for advisories"
list or "monitor to watch" list? If so, what are they, and why should they be included? A-42
APPENDIX A CHARGE TO REVIEWERS A-44
APPENDIX B INDIVIDUAL REVIEWER COMMENTS AND CLARIFICATIONS A-46
REVIEWER 1 A-47
REVIEWER 2 A-50
REVIEWER 3 A-59
A-29
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
1.0 INTRODUCTION
This report documents the results of an independent external peer review of the U.S. Environmental
Protection Agency's (EPA) draft Process for Selecting Contaminants to Monitor in Fish Advisory Programs.
ERG, Inc. (a contractor to EPA) organized this review and developed this report. The report provides
background on the development of the draft document (Section 1.1), describes ERG's peer reviewer selection
process (Section 1.2), and provides reviewers' comments organized by charge question (Section 2.0) along with
a summary of the comments by charge question. Appendix A provides the charge to reviewers and Appendix B
presents the reviewer comments organized by reviewer.
1.1 Background
EPA developed the Guidance for Assessing Chemical Contaminant Data for Use in Fish Advisories (EPA, 2000)
to help state, local, regional, and tribal environmental and public health officials who are responsible for
developing and managing fish consumption advisories. This guidance, which consists of four volumes, is
intended to be used together, since no single volume addresses all the topics involved in the development of
fish consumption advisories. This four-volume guidance set includes:
• Volume 1: Fish Sampling and Analysis
• Volume 2: Risk Assessment and Fish Consumption Limits
• Volume 3: Overview of Risk Management
• Volume 4: Risk Communication
EPA is revising Volumes 1 and 2 to include changes that have occurred since these documents were
published. These changes include, but are not limited to, contaminants of concern in fish, sampling
design approaches, and default values for developing fish consumption limits. Descriptions of each
volume can be found at: https://www.epa.gov/fish-tech/epa-guidance-developing-fish-advisories.
The purpose of this peer review was to review the proposed changes to the target analyte list
found in Volume 1, Fish and Sampling Analysis. Reviewers were asked to evaluate the approach
and process used for updating and selecting target analytes for use in screening studies by state,
territorial and tribal fish advisory programs.
1.2 Peer Reviewers
For this review, ERG identified, screened, and selected reviewers who had no conflict of interest in performing
the review, who are nationally recognized technical experts, and who had experience in one or more of the
following disciplines:
• Toxicology
• Risk Assessment
• Analytical Chemistry
ERG initiated a search process, asking interested candidates to describe their qualifications and respond to a
series of "Conflict of Interest" (COI) analysis questions. ERG carefully screened submissions to identify a pool of
qualified, COI-free candidates. From the set of candidates who met the criteria, ERG proposed a pool of five
candidates to EPA on September 12, 2023. From this pool, ERG selected three experts who collectively best
met the selection criteria. ERG contracted with two and committed the following three experts to perform the
review:
A-30
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
1. Philip Goodrum, Ph.D., DABT; Principal Toxicologist, GSI Environmental
2. Gloria Post, Ph.D., DABT; Research Scientist, New Jersey Department of Environmental Protection
3. Penelope Rice, Ph.D., DABT; Toxicologist and Subject Matter Expert, US Food and Drug Administration
(no fee)
ERG provided reviewers with instructions, the draft document entitled Process for Selecting Contaminants to
Monitor in Fish Advisory Programs, and the charge to reviewers (Appendix A of this report) prepared by EPA.
Reviewers worked individually to develop written comments in response to the charge questions. After
receiving reviewer comments, ERG compiled responses and prepared a summary of comments by charge
question (see Section 2.0) and included the responses and requested clarifications from EPA organized by
reviewer (see Appendix B).
2.0 SUMMARY OF REVIEWER COMMENTS ORGANIZED BY CHARGE QUESTION
This section summarizes reviewer comments by charge question. Each summary is followed by a table
presenting individual reviewer responses to that charge question (see Appendix B for the complete set of
reviewer comments).
2.1 Is the process EPA followed to identify compounds for which fish and shellfish advisories might be
needed reasonable?
All three reviewers agreed that the process EPA followed was reasonable, but also indicated that the process
would benefit from some revision. One reviewer suggested incorporating toxicity values from databases less
focused on North America. Two reviewers noted that using a cancer slope factor as a screening level for lead is
highly unusual; both suggested that EPA's Integrated Exposure Uptake Biokinetic Model for Lead in Children
(IEUBK) model or EPA's Adult Lead Model should be used to develop a generic screening level. One reviewer
recommended that the draft IRIS RfD values for PFHxS (released after the peer review document was written)
and PFNA (when available) be used for calculating the screening levels for those compounds.
For calculating a generic screening level for contaminants without established toxicity information, all three
reviewers questioned the use of the chronic reference value based upon the PFNA minimal risk level (MRL)
produced by the Agency for Toxic Substances and Disease Registry (ATSDR). One reviewer stated that EPA
should use a well-established toxicity value with a high degree of scientific consensus regarding the validity of the
value and how it was derived. This reviewer commented this was not the case for PFNA, because IRIS is
developing a draft assessment for it. Another reviewer stated that the generic screening level based on the
ATSDR MRL for PFNA is highly uncertain and that ATSDR MRLs are based on animal data, not human data. This
reviewer also asked if the generic screening level would be changed after the final IRIS assessment for PFDA is
released since IRIS' RfD for PFDA would presumably be lower than ATSDR's value for PFNA. The third reviewer
expressed that extrapolating the same reference dose across chemical classes seems unnecessary when it is
possible to choose the lowest RfD within each chemical class.
One reviewer suggested updating the screening level equation to better reflect current state practices for
implementing fish advisories. This reviewer also noted that, for occurrence data, sample maximums are very
unreliable statistics and recommended using a high percentile (e.g., 95th or 99th percentiles) to represent high
data values more appropriately. The same reviewer also recommended against using lipid-normalized
concentrations and suggested that these values be converted to wet weight concentrations.
A-31
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
Reviewer
Comments
Reviewer 1
The EPA's process for compiling the preliminary contaminant list appears reasonable, although
it is unclear why nanoparticles are included in the list. The databases which were mined for
toxicity values for their contaminant list appear to be heavily focused on North America. The
European Food Safety Authority (https://www.efsa.europa.eu/en/data-report/chemical-
hazards-database-openfoodtox) and the World Health Organization (https://apps.who.int/food-
additives-contaminants-iecfa-database/) have also published toxicity values for various
contaminants found in food. This reviewer suggests that these databases also be included in
EPA's review process, as this may broaden the chemical space covered by established toxicity
values.
The derivation of screening levels for contaminants with toxicity values (both non-cancer and
cancer values) appears appropriate. However, this reviewer does not consider the EPA's
'generic screening level' appropriate. IRIS' assessment, and the associated lifetime/subchronic
oral RfD value for PFNA exposure which forms the basis for the generic screening level, is still
in draft. In commenting on the draft IRIS assessment, this reviewer pointed out inconsistencies
in the selection of critical endpoint and derivation of the RfD for PFNA. A generic screening
value for use in the risk assessment of a broad chemical space should be based on a well-
established toxicity value with a high degree of scientific consensus regarding the validity of
the value and how it was derived. This reviewer suggests that the EPA select another toxicity
value on which to base their generic screening level.
Reviewer 2
As noted in the review document, the EPA (2000) Guidance for Assessing Chemical
Contaminant Data for Use in Fish Advisories does not include all contaminants found in fish
and shellfish that are currently of concern. For this reason, a process is needed to identify
additional contaminants that should be monitored in aquatic organisms and considered for
consumption advisory development, and EPA presents this process in the review document.
The process followed by EPA likely identified most contaminants that have been detected In
U.S. fish and/or shellfish at levels that might warrant fish consumption advisories. However,
some aspects of the this process should be clarified, as noted in my comments below:
• p. 2, first paragraph. Suggest clarifying whether all states, tribes, and territories must
have and/or actually have fish and shellfish advisory programs, or whether this is
optional and/or only some states have such programs.
• p. 2, Background section, numbered points. Points 2 and 3 are not meaningful as
written because "persistent" and "bioaccumulative" are not quantitatively defined.
Importantly, the degree of persistence and bioaccumulation needed for a contaminant
to reach a concentration of human health concern in aquatic organisms is dependent on
the dose at which toxicity occurs. Contaminants with very low non-cancer or cancer
screening levels can accumulate to levels of concern even if they are not highly
persistent or highly bioaccumulative. It is suggested that this issue be addressed by
adding the word "sufficiently" to points 2 and 3, as shown in bold below. Relevant to
this suggestion, point 1 already includes the word "sufficient" regarding the quantity
released to the environment.
A-32
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
Reviewer
Comments
Suggested revisions:
1. The compound must be sufficiently persistent in water and/or air for transport
in the environment once released.
2. The chemical nature of the substance must cause it to bioaccumulate
sufficiently in food webs due to an affinity for fish tissues, which vary both by
chemical and fish species characteristics.
• p. 3, first full paragraph on bioaccumulation. Related to the comment on
bioaccumulation above, this paragraph should state that the degree of bioaccumulation
needed for a contaminant to post a risk to consumers depends on the concentration of
the contaminant found in the aquatic environment (e.g., water, sediment) and the
concentration of the contaminant in aquatic organisms that results in a human health
risk from consumption of the organism. In other words, a contaminant that is not highly
bioaccumulative can be present in aquatic organisms at levels of concern for
consumers' consumption if it is present in the aquatic environment (water, sediments)
at a high enough concentration and/or if toxicity can occur from very low doses.
• p. 3, second full paragraph, second sentence. "Reproductive, developmental (including
neurodevelopmental), hepatotoxic (liver), and immunotoxic are among the most
common types of human health effects from exposure to contaminants in fish." The
sentence should be revised either to say that these are among the most common types
of non-cancer human health effects or to add carcinogenicity to the list of health
effects.
• p. 4, second set of numbered points in Pre-Search Definition of Screening Criteria
section. Regarding points 2 and 3, it is unclear how "potential to bioaccumulate,"
"prevalent...in the environment," and "persistent in the environment" are defined for
use as criteria. Relevant to comments above, the magnitude of prevalence, persistence,
and bioaccumulation necessary for a contaminant in fish or shellfish to pose a human
health risk from consumption is dependent on the dose at which toxicity can occur.
Additionally, regarding point 5, please note that ATSDR develops minimal risk levels
(MRLs), not reference doses, for non-cancer effects, and ATSDR does not develop
cancer slope factors or other toxicity factors for carcinogenic effects.
• p. 4, Article Inclusion Criteria section. "Published in 2000 or later (to capture information
published after the 2000 Guidance)." A minor comment is that the literature search for
the 2000 Guidance likely ended prior to 2000, since it took time for development and
review of a document prior to the date when it was finalized. Does the 2000 Guidance
include the date when the literature search that was used was performed?
• p. 5, Keywords. The literature search strategy (e.g., how AND, OR, etc. were used with
the keywords listed) should be provided. Additionally, does "states" mean "state
environmental agencies," and how were the states listed selected?
Were publications by the Delaware River Basin Commission (DRBC) and other similar
interstate authorities (e.g., Ohio River Valley Water Sanitation Commission [ORSANCO])
A-33
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
Reviewer
Comments
included? DRBC has conducted multiple studies of emerging contaminants in fish,
including pharmaceuticals, flame retardants, PFAS, and others. See
https://www.ni.gov/drbc/programs/qualitv/cecs.html.
• p. 5, Weight of evidence analysis. The weight of evidence approach, based on the
information presented here, does not appear to be completely logical and supportable.
Specifically, one of the four criteria used to assign points is "species found in the U.S."
However, the Overview section (p. 1) says that articles that contained "fish not found in
U.S. waters" were removed. Based on this statement, it appears that all articles that
were included contained information on species offish found in the U.S. However, this
is inconsistent with the weight of evidence discussion (p. 5), in which it appears that a
paper would be included even if it was not assigned 1 point for "species found in the
U.S." if it was assigned 2 points for meeting two other criteria.
Additionally, it is unclear how a study could meet the criterion for including "BCF or BAF
data" without also meeting the criterion for including "information on the
contaminant's detection in fish." Furthermore, even if a study provided a BCF or BAF for
a contaminant without contaminant concentration data, the study would have been
removed, since it is stated in first full paragraph of p. 7 that "EPA removed compounds
without concentration data...".
Based on the above, it appears that all included studies would have met the following two
criteria: including "information on the contaminant's detection," and including
information on "species found in the U.S."
Also, the criterion for "oral toxicity data" is unclear. Does this mean data from studies of
oral toxicity in mammalian species, or does this mean an oral toxicity factor (e.g.,
reference dose, cancer slope factor)?
• p. 7, Researched Toxicity Values section. The review document does not mention that the
process used for selection of the toxicity values in the review document differs from the
process used by in the EPA (2015) that is cited, and clarification of this issue needs to be
added. Specifically, while the list of eight sources of toxicity values in the review
document is the same as the list used by EPA (2015), the process for selection of the
toxicity factor in the review document is not the same as in EPA (2015). In the review
document, toxicity values were selected from the eight sources listed based on the order
in which the source is listed (i.e., when toxicity values were available from multiple
sources, the toxicity value from the source highest on the list was used). In contrast, EPA
(2015) used a different process to select from among the available toxicity factors. The
description of this process is included in each of the contaminant-specific " Update of
Human Health Ambient Water Quality Criteria" documents (linked from the table of
human health criteria at https://www.epa.gov/wqc/national-recommended-water-
qualitv-criteria-human-health-criteria-table; for example, see
https://www.epa.gov/sites/default/files/2015-10/documents/final-l-l-l-
trichloroethane.pdf) and is copied below:
A-34
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
Reviewer
Comments
"After identifying and documenting all available toxicity values, EPA followed a
systematic process to select the toxicity values used to derive the AWQC for
noncarcinogenic and carcinogenic effects. EPA selected IRIS toxicity values to derive the
updated AWQC if any of the following conditions were met:
1. EPA's IRIS toxicological assessment was the only available source of a toxicity
value.
2. EPA's IRIS toxicological assessment was the most current source of a toxicity
value.
3. EPA's IRIS program was reassessing the chemical in question and had published
the draft Toxicological Review for public review and comment, discussion at a
public meeting, and subsequent expert peer review.
4. The toxicity value from a more current toxicological assessment from a source
other than EPA IRIS was based on the same principal study and was numerically
the same as an older EPA IRIS toxicity value.
5. A more current toxicological assessment from a source other than EPA IRIS was
available, but it did not include the relevant toxicity value (chronic-duration oral
RfD or CSF).
6. A more current toxicological assessment from a source other than EPA IRIS was
available, but it did not introduce new science (e.g., the toxicity value was not
based on a newer principal study) or use a more current modeling approach
compared to an older EPA IRIS toxicological assessment.
EPA selected the toxicity value from a peer-reviewed, publicly available source other
than EPA IRIS to derive the updated AWQC if any of the following conditions were met:
1. The chemical is currently used as a pesticide, and EPA Office of Pesticide
Programs had a toxicity value that was used in pesticide registration decision-
making.
2. A toxicological assessment from a source other than EPA IRIS was the only
available source of a toxicity value.
3. A more current toxicological assessment from a source other than EPA IRIS
introduced new science (e.g., the toxicity value was based on a newer principal
study) or used a more current modeling approach compared to an older EPA
IRIS toxicological assessment."
• p. 7, Researched Toxicity Values section. The hotlinks in this section of the review
document do not work. It is assumed that this will be fixed in the final version.
• p. 7, Researched Toxicity Values section, toxicity values for PFAS. The toxicity values for
PFNA and PFHxS referred to as "reference doses" from the proposed EPA (2023)
National Primary Drinking Water Regulation (NPDWR) are called "chronic reference
values" not "reference doses." The proposed rule states that, for PFNA and PFHxS, "a
A-35
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
Reviewer
Comments
chronic reference value based on an Agency For Toxic Substances And Disease Registry
(ATSDR) intermediate-duration oral Minimal Risk Level" was developed.
A draft IRIS reference dose for PFHxS is now available, and a draft IRIS reference dose
for PFNA will be released soon. Since the draft IRIS PFDA value is used in the review
document, it is suggested that the draft IRIS PFHxS value (and the draft IRIS PFNA
values, when available), which are more recent than the ATSDR/NPDWR values, also be
used. Relevant to this point, a key difference between the recent reference doses for
long-chain PFAS developed by EPA (Office of Water - PFOA, PFOS; IRIS - PFDA, PFHxS)
and the ATSDR Minimal Risk Levels for long-chain PFAS (PFOA, PFOS, PFNA, PFHxS) is
that the EPA toxicity values are based on human data, and they are much more
stringent than the ATSDR values based on animal data.
• p. 7, Screening Level Calculations and Analysis, general comment. For contaminants
with a very low Reference Dose or a high cancer potency (slope factor), general dietary
exposure in the general population may exceed the exposure from consumption of a
weekly fish meal at the screening concentration. In such cases, it is not beneficial from a
public health viewpoint to issue a fish consumption advisory based on the screening
concentration because other foods that do not have the health benefits associated with
fish will be consumed instead of fish, while exposure to the contaminant will still be
above the toxicity value. This situation is much more likely to occur if fish consumption
advisories for carcinogens are based on the 10 s risk level instead of the 10"5 risk level, as
was done in the screening level calculations in the review document (p. 9).
In such cases, alternative approaches for development offish consumption advisories
may be considered. For example, in New Jersey's development offish consumption
advisories for dioxins and related compounds, the lifetime cancer risk resulting from
background dietary exposures to dioxin-like compounds was estimated to be about 10"3.
For this reason, an advisory based on 10"5 or 10 s risk would not result in a reduction of
risk from dioxins and related compounds. Therefore, the advisories were developed
using an alternative approach based on comparison with background dietary exposures.
For the general population, it was recommended that the fish consumption advisory
be based on an intake of dioxin and related compounds equal to the daily background
exposure in the total diet, such that consumption of fish at the advisory level would
result in a doubling of the background exposure. The advisory for the high-risk
population (pregnant and nursing mothers, women of childbearing age, and young
children) considered the fact that consumption of fish is beneficial as part of a healthy
diet. For this population, it was recommended that daily dioxin exposure from
consumption of fish should not exceed twice the exposure of an average meal, and it
was concluded that this exposure was likely to fall within the range of normal dietary
variation.
• p. 7, Screening Level Calculations and Analysis. It should be clarified in the text (not just
in the equation and the footnote) that the screening values for the general population
(adult and pregnant individual) developed in the review document are based on weekly
consumption of one 8-ounce fish meal. This is important because the EPA (2000)
A-36
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
Reviewer
Comments
guidance that is cited (Volume 2, Section 3) provides information for developing
screening values based on several different consumption frequencies.
• p. 8, last paragraph, "generic" screening level. It is recognized that a "generic" toxicity
value for screening of contaminants for which no toxicity value is available is needed.
Based on the bioaccumulative potential and low-dose toxicity of long-chain PFAS, it is
likely that a toxicity value based on a long-chain PFAS such as PFNA will be protective
for most other contaminants. That being said, the "generic" screening level based on a
toxicity value of 3 x 10 s mg/kg/day (based on the ATSDR MRL) for PFNA is highly
uncertain.
Additionally, it is stated that the "lowest final toxicity value (that is, the most stringent
toxicity value that was not draft or being developed)" was used for the generic
screening value. However, as discussed above, IRIS is currently developing Reference
Doses for several long-chain PFAS based on human data, and these IRIS Reference
Doses are lower than the ATSDR MRLs based on animal data. It is likely that the draft
IRIS toxicity assessment for PFDA, which includes a much lower Reference Dose than
the ATSDR PFNA value used here, will be finalized soon. When this occurs, will the
"generic" toxicity value be revised?
For contaminants that do not have a toxicity value in the eight sources listed in the
review document, chemical-specific toxicity values from other sources (e.g., values
developed by state environmental or health agencies other than California EPA) could
be reviewed and considered. It is stated in the section on Researched Toxicity Values
that other sources were not used "because of the variability of methods applied and
inconsistency of the existence of adequate quality control documentation." However, it
is unlikely that chemical-specific values developed by states (e.g., New Jersey,
Minnesota, Massachusetts) using EPA risk assessment guidance are more uncertain
than a "generic" value based on the toxicity value for a different chemical. As one
example, New Jersey has developed a Reference Dose of 1.3 x 10 s mg/kg/day (1.3
ng/kg/day) for perfluoroundecanoic acid specifically for use in fish consumption
advisories. See https://dep.ni.gov/wp-content/uploads/dsr/pfunda-fish-consumption-
trigger.pdf.
• Comments on screening levels in Excel spreadsheet:
o In these spreadsheets, the concentration data in columns G, H, and I are shown in
units of ng/g (which is ppb, although not stated) but the Screening Levels in the
columns to the right are shown in units of ng/g (ppm). This inconsistency in units is
confusing and may easily be overlooked by the reader, and consistent units should be
used.
o The Screening Level for lead is based on cancer risk using the CalEPA (2011) cancer
slope factor because no Reference Dose is available for lead. The reason that there is
no Reference Dose for lead is because there is no known threshold for the
neurodevelopmental effects of lead in children, and these neurodevelopmental
effects are generally the focus of concern regarding risks of lead exposure. If possible,
development of a Screening Level and fish consumption advisory for lead that is
A-37
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
Reviewer
Comments
protective for neurodevelopmental effects of lead in children, using the EPA
Integrated Exposure Uptake Biokinetic Model for Lead in Children (IEUBK) model,
could be considered. New Jersey has used such an approach for its fish consumption
advisories for lead.
o The cancer slope factor for PFOA of 0.0293 mg/kg/day shown in the "fillet-analysis w
tox info" and "shellfish-analysis w tox info" spreadsheets is incorrect. The cancer
slope factor from the cited EPA (2023) reference is 0.0293 ng/kg/day, which is 29,300
mg/kg/day.
o The cancer slope factors for PFOA and PFOS are missing from the "WholeBody-
analysis w tox info" spreadsheet.
Reviewer 3
Yes, overall, the process EPA followed to identify priority compounds is reasonable. However,
EPA might consider revising the documentation and analyte selection process in these areas:
1. Update the equations used to calculate screening levels (SLs) to more closely align
with current fish advisory practices. The current equation cites to the 2000
guidance, which is a special case of a more general equation.
2. Provide more analysis and documentation of the fish tissue concentrations
summarized from the literature, particular for analytes that are selected because
the sample maximum concentrations exceeds the SL.
3. Consider providing different weighting factors to these two conditions:
A. sample maximum > SL and sample mean < SL
B. sample maximum > SL and sample mean > SL
4. Consider refining the decision process for selecting an RfD to serve as a protective
surrogate value when the RfD is missing for a chemical.
5. Derive a SL for lead (Pb) using EPA's lead risk models, rather than the cancer slope
factor.
6. Either exclude the lipid-normalized concentrations, or apply a default assumption
for lipid content to convert the values to wet weight units.
The basis for each recommendation is provided below.
Screening Level (SL) Equations
Separate equations for calculating a fish tissue screening level (SL) are provided for noncancer
and cancer endpoints. The equations are consistent with the 2000 Guidance, but could be
updated to more clearly show the underlying assumptions and to reflect how states currently
implement fish advisories. Applying abbreviations for convenience, the equation presented to
calculate a screening level for noncancer effects (SLnc) on p.7, including the unit conversion
factor (CF) for mass discussed on p. 8, is:
rj RfD X BW r i (mg COPC/kg BW-day) x (kg BW)
nc CR X CF ^ ^ (g ww/day) x 0.001 kg/g
A-38
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
Reviewer
Comments
where,
SLnc =
fish tissue concentration (mg/kg ww)
RfD
chronic oral refence dose (mg/kg-day)
BW
body weight (kg)
CR
average daily fish consumption rate (g ww/day)
CF
conversion factor (0.001 kg per g)
What is implied, but not stated directly, is that the SL is the concentration that, when included
in the calculation of average daily dose (ADD), equals the RfD. In other words, the ratio of the
ADD/RfD is 1, or equivalently, the target hazard quotient (THQ) is 1. Also, in practice, most
state agencies consider fish consumption rate to be the product of the meal size and meal
frequency, which is how different meal frequencies are ultimately determined. Finally, some
agencies also apply a relative source contribution (RSC) to account for additional exposure
pathways that may contribute to a total average daily dose. Considering all of these concepts,
a more general expression for SL is:
THQxRfD xRSCxBW
SLnc ~ x MF) X CF
where,
SLnc =
fish tissue concentration (mg/kg ww)
THQ =
target hazard quotient
RfD
chronic oral refence dose (mg/kg-day)
RSC
relative source contribution
BW
body weight (kg)
MS
meal size (g ww/meal)
MF
average daily meal frequency (meals/day)
CF
conversion factor (0.001 kg per g)
Then, it can be stated that two assumptions used in the SL are: 1) THQ =1 (which would open
the door for some discussion on the science policy decision, and standard conventions used by
USEPA in selecting a target level); and 2) RSC = 1 (which would also open the door for some
discussion on why this is used in the SL derivation, but might be revisited in site-specific
applications).
The product of (MS x MF) is CR, and USEPA can continue to present the CR estimates for typical
and high-end consumers, and briefly discuss what meal frequency these correspond to when
expressed over a period of one month or one year.
A similar general equation can be presented for the SL for cancer endpoints.
A-39
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
Reviewer
Comments
Summary of Occurrence Data on Concentrations in Fish Tissue
The guidance document discuss the literature review methods and data usability criterion. The
occurrence data generated from this process are provided in the Excel file (Screen Level
Calculations.xlsx), grouped into separate worksheets for: 1) fillet data; 2) whole body data; 3)
shellfish data. The occurrence data are distilled down to two summary statistics - "Maximum"
and "Average".
The sample maximum is a very unstable summary statistic, and subject to extreme results that
do not actually represent the conditions found in most water bodies in the United States. The
chances of observing an extreme value actually increases with increasing sample sizes. It is
clear that one of the reasons for selecting the maximum is that the choice of statistics is
limited to a large extent by the information presented in table summaries in the literature - it
is unreasonable to expect to obtain the underlying raw data from most published studies.
However, a preferred (more stable) statistic, that achieves the goal of representing a high-end
value, would simply be an upper percentile (e.g., 95th percentile, or even 99th percentile). A
recommended hierarchy of summary statistics for representing a high-end value is:
• Reported upper percentile (90th, 95th, or 99th)
• Estimate of upper percentile based on an assumed distribution (e.g., mean and
standard deviation are reported, so assume a lognormal distribution to estimate
the corresponding 95th percentile)
• Sample maximum
The following extreme cases of sample maximums are noted by comparing the ratio of the
sample maximum to the arithmetic mean:
Worksheet
Chemical
Maximum
(ng/g)
Average (ng/g)
Ratio of
Max/Average
Fillet
PFDoA
859,000
4.2
204,135
Fillet
PFOS
2,840,000
53.1
53,525
Whole Body
BDE-99
650
0.24
2,708
Given the unreliability of the sample maximum as an indicator of conditions on a national
scale, the rather large set of analytes for which only a maximum is provided (there are no
estimates of the mean) should be carefully considered, at least in terms of the weighting
scores used to rank each analyte. The following counts of analytes for which no "average" is
available are noted, by chemical class:
Worksheet
Chemical Class
Number of Analytes
Missing an Average
Fillet
Flame Retardants
16
A-40
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
Reviewer
Comments
PFAS
12
Metals
1
Chlorinated
1
Cyanotoxin
1
Other (paraffins)
2
Whole Body
Flame Retardants
8
PFAS
3
Metals
1
Other (paraffins)
1
The one metal listed in the table above is for lead. Lead is included in this guidance based on
the cancer slope factor, which is an extremely unusual choice. From my experience as a
toxicologist and frequent participant on EPA's science advisory panels involving lead, lead is
not regulated based on the cancer slope factor at any site, for any medium. USEPA and state
agencies rely instead on the screening levels developed from regulatory models that predict
blood lead concentrations (e.g., IEUBK or Adult Lead Model) from average daily intake. The
USEPA Regional Screening Level tool1 and guidance notes, "EPA has no consensus RfD or SFO
for inorganic lead, so it is not possible to calculate SLs as we have done for other chemicals".
EPA should develop a generic fish tissue level using one of EPA's lead models. For example,
alternative dietary inputs can easily be included in the IEUBK model for children to develop a
protective SL for lead in fish tissue.
Consider also including the number of studies and the number of study values that were
curated from the literature and used to derive the "Maximum" and "Average".
Do not include the tissue concentrations that are lipid normalized, directly in the comparison
to the toxicity values. The units matter in this case. A preferred approach would be to apply a
general assumption for % lipid content to convert the lipid-normalized values to wet weight
concentrations. Or, alternatively, exclude the study results that are expressed only as lipid
normalized values.
Surrogates RfD for Missing Values
EPA elected to the RfD for PFNA (3E-06 mg/kg-day) as the proxy value for analytes without an
RfD because, "it is the lowest final RfD for all contaminants being considered for inclusion in
the monitoring list". In the Excel file, these are listed as "generic SLs" and include chemicals
from a wide range of categories: antibacterials and antibiotics, cyanotoxins, flame retardants,
and pharmaceuticals. This extrapolation across chemical classes seems unnecessary when it is
possible to select from the lowest RfD with the same chemical class.
1 https://epa-prgs.ornl.gov/cgi-bin/chemicals/csl_search
A-41
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
2.2 Is the list of contaminants advisory programs should consider monitoring for reasonable (e.g.,
reflects the current range of contaminants detected in fish with potential human health impacts)?
All three reviewers agreed that the list of contaminants is reasonable and comprehensive. One reviewer noted
that there are other compounds of concern that could be included in the list, but that these contaminants are
not well known or well researched; this reviewer indicated that the list is reasonable even without the
additional contaminants. Another reviewer noted that some states have already developed fish consumption
advisories for many contaminants on the list.
Reviewer
Comments
Reviewer 1
The EPA's list of contaminants reflects those chemicals which have been measured in fish in
published studies. This list is only a subset of the actual contaminants that may be found in
fish, as studies are generally only done on compounds which are widely known to be present in
fish and/or are easy to analyze in fish tissue. It is likely that the EPA's list will fail to capture
compounds, like metabolites of fluorotelomer sulfonates, which are not commonly the subject
of scientific studies in the broader research community. However, there is little the EPA can do
to remedy this issue, short of itself conducting a nontargeted analytical assessment of
contaminants in a range of species from different geographic areas, which would be very time-
consuming and expensive. Given the time- and resource limitations, the EPA's list is
reasonable.
Reviewer 2
The list of contaminants to monitor for advisories in Table 7 appears reasonable. It should be
noted that New Jersey and other states already have developed fish consumption advisories
for many of the contaminants in Table 7. Of the chemicals included on this list, New Jersey has
developed fish consumption advisory triggers and/or waterbody-specific fish consumption
advisories for PFOA, PFOS, PNA, microcystins, and lead. Several other states have also
developed fish consumption advisories for PFAS. California has also developed consumption
trigger for microcystins, and other states may also have developed advisories for contaminants
on this list.
The list of contaminants to monitor to watch in Table 8 also appears to be reasonable.
Reviewer 3
Yes, the range of chemical classes makes sense and appears to be comprehensive. See above
for recommendations on revisiting the approach used to derive SLs for some of these analytes.
2.3 Are there additional contaminants that should be included in the "monitor for advisories" list or
"monitor to watch" list? If so, what are they, and why should they be included?
Two reviewers suggested including additional contaminants. One reviewer recommended adding 6:2 di- and
mono-PAPs and fluorotelomer sulfonates to the lists, as they all have shown high bioconcentration factor (BCF)
values in recent studies. Another reviewer suggested adding additional cyanotoxins, such as
cylindrospermopsin, anatoxin-a, and saxitoxin to the lists, because harmful algal blooms (HABs) are of current
concern. This reviewer noted that cyanotoxin advisories should take into account that HAB exposures are
often short-term, not chronic.
A-42
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
Reviewer
Comments
Reviewer 1
This reviewer suggests including 6:2 di- and monoPAPs and fluorotelomer sulfonates in the
candidate list, as fluorotelomer sulfonates have been previously shown to accumulate in
marine invertebrates (https://pubs.acs.org/doi/10.1021/acs.est.9b00927) and both have had
reportedly high BCF values in published studies
(https://pubs.acs.org/doi/epdf/10.1021/acs.est.2c03734).
Reviewer 2
The list of additional contaminants in the "monitor for advisories" and "monitor to watch" lists
include the contaminants identified through the process described in the review document.
Inclusion of additional cyanotoxins (e.g., cylindrospermopsin, anatoxin-a, and/or saxitoxin)
could be considered since potential risks from fish from waterbodies with harmful algal blooms
(HABs) are of current concern. New Jersey and California have developed fish consumption
triggers for cylindrospermopsin and anatoxin-a, and other states have developed qualitative
advice for consumption of fish where HABs have occurred. Advisories for cyanotoxins should
consider the fact that exposure to cyanotoxins in fish is likely to be short-term or subchronic,
rather than chronic, due to the relatively short timeframe that a HAB persists in a waterbody.
Reviewer 3
1 am not aware of any additional contaminants that would be reasonable candidates to include
in the monitoring lists.
A-43
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
APPENDIX A
CHARGE TO REVIEWERS
A-44
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
Technical Charge to External Peer Reviewers
Contract GSA GS-00F-079CA; BPA #68HERH23A0019
Call Order 68HERH23F0356 (ERG Call Order 002)
September 2023
External Peer Review of the Process for Selecting Contaminants
to Monitor in Fish Advisory Programs
BACKGROUND
EPA developed the Guidance for Assessing Chemical Contaminant Data for Use in Fish Advisories
(EPA, 2000) to help state, local, regional and tribal environmental and public health officials
(intended audience) who are responsible for developing and managing fish consumption
advisories. This guidance which consists of four volumes is intended to be used together, since no
single volume addresses all the topics involved in the development of fish consumption advisories.
This four-volume guidance set includes:
• Volume 1: Fish Sampling and Analysis
• Volume 2: Risk Assessment and Fish Consumption Limits
• Volume 3: Overview of Risk Management
• Volume 4: Risk Communication
EPA is revising Volumes 1 and 2 to include changes that have occurred since these documents
were published. These changes include but are not limited to contaminants of concern in fish,
sampling design approaches, and default values for developing fish consumption limits.
Descriptions of each volume can be found at: https://www.epa.gov/fish-tech/epa-guidance-
developing-fish-advisories.
The objective of this peer review is to evaluate the approach and process used for updating and
selecting contaminants that state, territorial and tribal fish advisory programs should monitor.
CHARGE QUESTIONS
1. Is the process EPA followed to identify compounds for which fish and shellfish advisories might
be needed reasonable?
2. Is the list of contaminants advisory programs should consider monitoring for reasonable (e.g.,
reflects the current range of contaminants detected in fish with potential human health
impacts)?
3. Are there additional contaminants that should be included in the "monitor for advisories" list or
"monitor to watch" list? If so, what are they, and why should they be included?
A-45
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
APPENDIX B
INDIVIDUAL REVIEWER COMMENTS
AND CLARIFICATIONS
A-46
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
COMMENTS SUBMITTED BY
REVIEWER 1
A-47
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
External Peer Review of the Process for Selecting Contaminants
to Monitor in Fish Advisory Programs
Charge Questions
1. Is the process EPA followed to identify compounds for which fish and shellfish advisories might
be needed reasonable?
The EPA's process for compiling the preliminary contaminant list appears reasonable, although it is
unclear why nanoparticles are included in the list. The databases which were mined for toxicity
values for their contaminant list appear to be heavily focused on North America. The European
Food Safety Authority (https://www.efsa.europa.eu/en/data-report/chemical-hazards-database-
openfoodtox) and the World Health Organization (https://apps.who.int/food-additives-
contaminants-jecfa-database/) have also published toxicity values for various contaminants found
in food. This reviewer suggests that these databases also be included in EPA's review process, as
this may broaden the chemical space covered by established toxicity values. The derivation of
screening levels for contaminants with toxicity values (both non-cancer and cancer values) appears
appropriate. However, this reviewer does not consider the EPA's 'generic screening level'
appropriate. IRIS' assessment, and the associated lifetime/subchronic oral RfD value for PFNA
exposure which forms the basis for the generic screening level, is still in draft. In commenting on
the draft IRIS assessment, this reviewer pointed out inconsistencies in the selection of critical
endpoint and derivation of the RfD for PFNA. A generic screening value for use in the risk
assessment of a broad chemical space should be based on a well-established toxicity value with a
high degree of scientific consensus regarding the validity of the value and how it was derived. This
reviewer suggests that the EPA select another toxicity value on which to base their generic
screening level.
2. Is the list of contaminants advisory programs should consider monitoring for reasonable (e.g.,
reflects the current range of contaminants detected in fish with potential human health
impacts)?
The EPA's list of contaminants reflects those chemicals which have been measured in fish in
published studies. This list is only a subset of the actual contaminants that may be found in fish,
as studies are generally only done on compounds which are widely known to be present in fish
and/or are easy to analyze in fish tissue. It is likely that the EPA's list will fail to capture
compounds, like metabolites of fluorotelomer sulfonates, which are not commonly the subject
of scientific studies in the broader research community. However, there is little the EPA can do
to remedy this issue, short of itself conducting a nontargeted analytical assessment of
contaminants in a range of species from different geographic areas, which would be very time-
consuming and expensive. Given the time- and resource limitations, the EPA's list is reasonable.
3. Are there additional contaminants that should be included in the "monitor for advisories" list
or "monitor to watch" list? If so, what are they, and why should they be included?
This reviewer suggests including 6:2 di- and monoPAPs and fluorotelomer sulfonates in the candidate
list, as fluorotelomer sulfonates have been previously shown to accumulate in marine invertebrates
A-48
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
(https://pubs.acs.org/doi/10.1021/acs.est.9b00927) and both have had reportedly high BCF values in
published studies (https://pubs.acs.org/doi/epdf/10.1021/acs.est.2c03734
Requested Clarification from EPA
Question for Reviewer 1:
You suggested that the EPA select another toxicity value on which to base the generic screening level. EPA
did not use the draft IRIS PFNA reference dose in its generic screening level calculation, because IRIS has not
released the draft assessment yet and because draft values can change. The calculation used ATSDR's final
oral MRLfor PFNA from 2021. Using the lowest developed RfD (4 x 10-10 mg/kg-d, IRIS' draft RfD for PFDA)
results in such a low screening level that detection at any concentration would result in a contaminant being
added to the list. What toxicity value would you suggest that EPA use when calculating the generic screening
level?
Reviewer 1 Response:
I don't think it is appropriate to use any given compound's specific value as a generic screening value for any
contaminant without data in the absence of consideration of the structural relatedness of the data-poor
contaminant to the reference chemical. Instead, I would advise that EPA to take a case-by-case approach to
the development of screening levels for data-poor compounds. This assessment should assess the structural
similarity of a data-poor contaminant with contaminants that have specific screening values. If the data-poor
contaminant is structurally-similar to a compound with a data-based screening value, the structural analog's
screening value may be applied to the assessment of exposure to the data-poor compound. Alternatively,
high throughput NAMs may be used to identify toxicologically-similar index chemicals whose screening value
would be appropriate for use in the risk assessment of exposure to a data-poor chemical.
A-49
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
COMMENTS SUBMITTED BY
REVIEWER 2
A-50
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
External Peer Review of the Process for Selecting Contaminants
to Monitor in Fish Advisory Programs
Charge Questions
1. Is the process EPA followed to identify compounds for which fish and shellfish advisories might
be needed reasonable?
Response: As noted in the review document, the EPA (2000) Guidance for Assessing Chemical
Contaminant Data for Use in Fish Advisories does not include all contaminants found in fish and
shellfish that are currently of concern. For this reason, a process is needed to identify additional
contaminants that should be monitored in aquatic organisms and considered for consumption
advisory development, and EPA presents this process in the review document.
The process followed by EPA likely identified most contaminants that have been detected in U.S.
fish and/or shellfish at levels that might warrant fish consumption advisories. However, some
aspects of the this process should be clarified, as noted in my comments below:
• p. 2, first paragraph. Suggest clarifying whether all states, tribes, and territories must have
and/or actually have fish and shellfish advisory programs, or whether this is optional
and/or only some states have such programs.
• p. 2, Background section, numbered points. Points 2 and 3 are not meaningful as written
because "persistent" and "bioaccumulative" are not quantitatively defined. Importantly,
the degree of persistence and bioaccumulation needed for a contaminant to reach a
concentration of human health concern in aquatic organisms is dependent on the dose at
which toxicity occurs. Contaminants with very low non-cancer or cancer screening levels
can accumulate to levels of concern even if they are not highly persistent or highly
bioaccumulative. It is suggested that this issue be addressed by adding the word
"sufficiently" to points 2 and 3, as shown in bold below. Relevant to this suggestion, point
1 already includes the word "sufficient" regarding the quantity released to the
environment.
Suggested revisions:
1. The compound must be sufficiently persistent in water and/or air for transport in
the environment once released.
2. The chemical nature of the substance must cause it to bioaccumulate sufficiently
in food webs due to an affinity for fish tissues, which vary both by chemical and
fish species characteristics.
• p. 3, first full paragraph on bioaccumulation. Related to the comment on bioaccumulation
above, this paragraph should state that the degree of bioaccumulation needed for a
contaminant to post a risk to consumers depends on the concentration of the
contaminant found in the aquatic environment (e.g., water, sediment) and the
concentration of the contaminant in aquatic organisms that results in a human health risk
from consumption of the organism. In other words, a contaminant that is not highly
A-51
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
bioaccumulative can be present in aquatic organisms at levels of concern for consumers'
consumption if it is present in the aquatic environment (water, sediments) at a high
enough concentration and/or if toxicity can occur from very low doses.
• p. 3, second full paragraph, second sentence. "Reproductive, developmental (including
neurodevelopmental), hepatotoxic (liver), and immunotoxic are among the most common
types of human health effects from exposure to contaminants in fish." The sentence
should be revised either to say that these are among the most common types of non-
cancer human health effects or to add carcinogenicity to the list of health effects.
• p. 4, second set of numbered points in Pre-Search Definition of Screening Criteria section.
Regarding points 2 and 3, it is unclear how "potential to bioaccumulate," "prevalent...in
the environment," and "persistent in the environment" are defined for use as criteria.
Relevant to comments above, the magnitude of prevalence, persistence, and
bioaccumulation necessary for a contaminant in fish or shellfish to pose a human health
risk from consumption is dependent on the dose at which toxicity can occur.
Additionally, regarding point 5, please note that ATSDR develops minimal risk levels
(MRLs), not reference doses, for non-cancer effects, and ATSDR does not develop cancer
slope factors or other toxicity factors for carcinogenic effects.
• p. 4, Article Inclusion Criteria section. "Published in 2000 or later (to capture information
published after the 2000 Guidance)." A minor comment is that the literature search for
the 2000 Guidance likely ended prior to 2000, since it took time for development and
review of a document prior to the date when it was finalized. Does the 2000 Guidance
include the date when the literature search that was used was performed?
• p. 5, Keywords. The literature search strategy (e.g., how AND, OR, etc. were used with the
keywords listed) should be provided. Additionally, does "states" mean "state
environmental agencies," and how were the states listed selected?
Were publications by the Delaware River Basin Commission (DRBC) and other similar
interstate authorities (e.g., Ohio River Valley Water Sanitation Commission [ORSANCO])
included? DRBC has conducted multiple studies of emerging contaminants in fish,
including pharmaceuticals, flame retardants, PFAS, and others. See
https://www.ni.gov/drbc/programs/qualitv/cecs.html
• p. 5, Weight of evidence analysis. The weight of evidence approach, based on the
information presented here, does not appear to be completely logical and supportable.
Specifically, one of the four criteria used to assign points is "species found in the U.S."
However, the Overview section (p. 1) says that articles that contained "fish not found in
U.S. waters" were removed. Based on this statement, it appears that all articles that were
included contained information on species of fish found in the U.S. However, this is
inconsistent with the weight of evidence discussion (p. 5), in which it appears that a paper
would be included even if it was not assigned 1 point for "species found in the U.S." if it
was assigned 2 points for meeting two other criteria.
A-52
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
Additionally, it is unclear how a study could meet the criterion for including "BCF or BAF
data" without also meeting the criterion for including "information on the contaminant's
detection in fish." Furthermore, even if a study provided a BCF or BAF for a contaminant
without contaminant concentration data, the study would have been removed, since it is
stated in first full paragraph of p. 7 that "EPA removed compounds without concentration
data...".
Based on the above, it appears that all included studies would have met the following two
criteria: including "information on the contaminant's detection," and including
information on "species found in the U.S."
Also, the criterion for "oral toxicity data" is unclear. Does this mean data from studies of
oral toxicity in mammalian species, or does this mean an oral toxicity factor (e.g.,
reference dose, cancer slope factor)?
• p. 7, Researched Toxicity Values section. The review document does not mention that the
process used for selection of the toxicity values in the review document differs from the
process used by in the EPA (2015) that is cited, and clarification of this issue needs to be
added. Specifically, while the list of eight sources of toxicity values in the review document
is the same as the list used by EPA (2015), the process for selection of the toxicity factor in
the review document is not the same as in EPA (2015). In the review document, toxicity
values were selected from the eight sources listed based on the order in which the source
is listed (i.e., when toxicity values were available from multiple sources, the toxicity value
from the source highest on the list was used). In contrast, EPA (2015) used a different
process to select from among the available toxicity factors. The description of this process
is included in each of the contaminant-specific " Update of Human Health Ambient Water
Quality Criteria" documents (linked from the table of human health criteria at
https://www.epa.gov/wqc/national-recommended-water-qualitv-criteria-human-health-
criteria-table: for example, see https://www.epa.gov/sites/default/files/2015-
10/documents/final-l-l-l-trichloroethane.pdf) and is copied below:
"After identifying and documenting all available toxicity values, EPA followed a systematic
process to select the toxicity values used to derive the AWQC for noncarcinogenic and
carcinogenic effects. EPA selected IRIS toxicity values to derive the updated AWQC if any
of the following conditions were met:
1. EPA's IRIS toxicological assessment was the only available source of a toxicity
value.
2. EPA's IRIS toxicological assessment was the most current source of a toxicity
value.
3. EPA's IRIS program was reassessing the chemical in question and had published
the draft Toxicological Review for public review and comment, discussion at a
public meeting, and subsequent expert peer review.
4. The toxicity value from a more current toxicological assessment from a source
A-53
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
other than EPA IRIS was based on the same principal study and was numerically
the same as an older EPA IRIS toxicity value.
5. A more current toxicological assessment from a source other than EPA IRIS was
available, but it did not include the relevant toxicity value (chronic-duration oral
RfD or CSF).
6. A more current toxicological assessment from a source other than EPA IRIS was
available, but it did not introduce new science (e.g., the toxicity value was not
based on a newer principal study) or use a more current modeling approach
compared to an older EPA IRIS toxicological assessment.
EPA selected the toxicity value from a peer-reviewed, publicly available source other than
EPA IRIS to derive the updated AWQC if any of the following conditions were met:
1. The chemical is currently used as a pesticide, and EPA Office of Pesticide Programs
had a toxicity value that was used in pesticide registration decision-making.
2. A toxicological assessment from a source other than EPA IRIS was the only
available source of a toxicity value.
3. A more current toxicological assessment from a source other than EPA IRIS
introduced new science (e.g., the toxicity value was based on a newer principal
study) or used a more current modeling approach compared to an older EPA IRIS
toxicological assessment."
• p. 7, Researched Toxicity Values section. The hotlinks in this section of the review
document do not work. It is assumed that this will be fixed in the final version.
• p. 7, Researched Toxicity Values section, toxicity values for PFAS. The toxicity values for
PFNA and PFHxS referred to as "reference doses" from the proposed EPA (2023) National
Primary Drinking Water Regulation (NPDWR) are called "chronic reference values" not
"reference doses." The proposed rule states that, for PFNA and PFHxS, "a chronic
reference value based on an Agency For Toxic Substances And Disease Registry (ATSDR)
intermediate-duration oral Minimal Risk Level" was developed.
A draft IRIS reference dose for PFHxS is now available, and a draft IRIS reference dose for
PFNA will be released soon. Since the draft IRIS PFDA value is used in the review
document, it is suggested that the draft IRIS PFHxS value (and the draft IRIS PFNA values,
when available), which are more recent than the ATSDR/NPDWR values, also be used.
Relevant to this point, a key difference between the recent reference doses for long-chain
PFAS developed by EPA (Office of Water - PFOA, PFOS; IRIS - PFDA, PFHxS) and the ATSDR
Minimal Risk Levels for long-chain PFAS (PFOA, PFOS, PFNA, PFHxS) is that the EPA toxicity
values are based on human data, and they are much more stringent than the ATSDR
values based on animal data.
• p. 7, Screening Level Calculations and Analysis, general comment. For contaminants with a
very low Reference Dose or a high cancer potency (slope factor), general dietary exposure
A-54
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
in the general population may exceed the exposure from consumption of a weekly fish
meal at the screening concentration. In such cases, it is not beneficial from a public health
viewpoint to issue a fish consumption advisory based on the screening concentration
because other foods that do not have the health benefits associated with fish will be
consumed instead of fish, while exposure to the contaminant will still be above the
toxicity value. This situation is much more likely to occur if fish consumption advisories for
carcinogens are based on the 10 s risk level instead of the 10"5 risk level, as was done in the
screening level calculations in the review document (p. 9).
In such cases, alternative approaches for development offish consumption advisories may
be considered. For example, in New Jersey's development of fish consumption advisories
for dioxins and related compounds, the lifetime cancer risk resulting from background
dietary exposures to dioxin-like compounds was estimated to be about 10"3. For this
reason, an advisory based on 10"5 or 10 s risk would not result in a reduction of risk from
dioxins and related compounds. Therefore, the advisories were developed using an
alternative approach based on comparison with background dietary exposures.
For the general population, it was recommended that the fish consumption advisory be
based on an intake of dioxin and related compounds equal to the daily background
exposure in the total diet, such that consumption offish at the advisory level would result
in a doubling of the background exposure. The advisory for the high-risk population
(pregnant and nursing mothers, women of childbearing age, and young children)
considered the fact that consumption of fish is beneficial as part of a healthy diet. For this
population, it was recommended that daily dioxin exposure from consumption of fish
should not exceed twice the exposure of an average meal, and it was concluded that this
exposure was likely to fall within the range of normal dietary variation.
• p. 7, Screening Level Calculations and Analysis. It should be clarified in the text (not just in
the equation and the footnote) that the screening values for the general population (adult
and pregnant individual) developed in the review document are based on weekly
consumption of one 8-ounce fish meal. This is important because the EPA (2000) guidance
that is cited (Volume 2, Section 3) provides information for developing screening values
based on several different consumption frequencies.
• p. 8, last paragraph, "generic" screening level. It is recognized that a "generic" toxicity
value for screening of contaminants for which no toxicity value is available is needed.
Based on the bioaccumulative potential and low-dose toxicity of long-chain PFAS, it is
likely that a toxicity value based on a long-chain PFAS such as PFNA will be protective for
most other contaminants. That being said, the "generic" screening level based on a
toxicity value of 3 x 10 s mg/kg/day (based on the ATSDR MRL) for PFNA is highly
uncertain.
Additionally, it is stated that the "lowest final toxicity value (that is, the most stringent
toxicity value that was not draft or being developed)" was used for the generic screening
value. However, as discussed above, IRIS is currently developing Reference Doses for
A-55
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
several long-chain PFAS based on human data, and these IRIS Reference Doses are lower
than the ATSDR MRLs based on animal data. It is likely that the draft IRIS toxicity
assessment for PFDA, which includes a much lower Reference Dose than the ATSDR PFNA
value used here, will be finalized soon. When this occurs, will the "generic" toxicity value
be revised?
For contaminants that do not have a toxicity value in the eight sources listed in the review
document, chemical-specific toxicity values from other sources (e.g., values developed by
state environmental or health agencies other than California EPA) could be reviewed and
considered. It is stated in the section on Researched Toxicity Values that other sources
were not used "because of the variability of methods applied and inconsistency of the
existence of adequate quality control documentation." However, it is unlikely that
chemical-specific values developed by states (e.g., New Jersey, Minnesota, Massachusetts)
using EPA risk assessment guidance are more uncertain than a "generic" value based on
the toxicity value for a different chemical. As one example, New Jersey has developed a
Reference Dose of 1.3 x 10"6mg/kg/day (1.3 ng/kg/day) for perfluoroundecanoic acid
specifically for use in fish consumption advisories. See https://dep.nj.gov/wp-
content/uploads/dsr/pfunda-fish-consumption-trigger.pdf.
• Comments on screening levels in Excel spreadsheet:
o In these spreadsheets, the concentration data in columns G, H, and I are shown in
units of ng/g (which is ppb, although not stated) but the Screening Levels in the
columns to the right are shown in units of ng/g (ppm). This inconsistency in units
is confusing and may easily be overlooked by the reader, and consistent units
should be used.
o The Screening Level for lead is based on cancer risk using the CalEPA (2011)
cancer slope factor because no Reference Dose is available for lead. The reason
that there is no Reference Dose for lead is because there is no known threshold
for the neurodevelopmental effects of lead in children, and these
neurodevelopmental effects are generally the focus of concern regarding risks of
lead exposure. If possible, development of a Screening Level and fish consumption
advisory for lead that is protective for neurodevelopmental effects of lead in
children, using the EPA Integrated Exposure Uptake Biokinetic Model for Lead in
Children (IEUBK) model, could be considered. New Jersey has used such an
approach for its fish consumption advisories for lead.
o The cancer slope factor for PFOA of 0.0293 mg/kg/day shown in the "fillet-analysis
w tox info" and "shellfish-analysis w tox info" spreadsheets is incorrect. The
cancer slope factor from the cited EPA (2023) reference is 0.0293 ng/kg/day,
which is 29,300 mg/kg/day.
o The cancer slope factors for PFOA and PFOS are missing from the "WholeBody-
analysis w tox info" spreadsheet.
A-56
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
2. Is the list of contaminants advisory programs should consider monitoring for reasonable (e.g.,
reflects the current range of contaminants detected in fish with potential human health
impacts)?
Response: The list of contaminants to monitor for advisories in Table 7 appears reasonable. It
should be noted that New Jersey and other states already have developed fish consumption
advisories for many of the contaminants in Table 7. Of the chemicals included on this list, New
Jersey has developed fish consumption advisory triggers and/or waterbody-specific fish
consumption advisories for PFOA, PFOS, PNA, microcystes, and lead. Several other states have
also developed fish consumption advisories for PFAS. California has also developed
consumption trigger for microcystins, and other states may also have developed advisories for
contaminants on this list.
The list of contaminants to monitor to watch in Table 8 also appears to be reasonable.
3. Are there additional contaminants that should be included in the "monitor for advisories" list
or "monitor to watch" list? If so, what are they, and why should they be included?
Response: The list of additional contaminants in the "monitor for advisories" and "monitor to
watch" lists include the contaminants identified through the process described in the review
document.
Inclusion of additional cyanotoxins (e.g., cylindrospermopsin, anatoxin-a, and/or saxitoxin)
could be considered since potential risks from fish from waterbodies with harmful algal blooms
(HABs) are of current concern. New Jersey and California have developed fish consumption
triggers for cylindrospermopsin and anatoxin-a, and other states have developed qualitative
advice for consumption of fish where HABs have occurred. Advisories for cyanotoxins should
consider the fact that exposure to cyanotoxins in fish is likely to be short-term or subchronic,
rather than chronic, due to the relatively short timeframe that a HAB persists in a waterbody.
Requested Clarification from EPA
Questions for Reviewer 2:
1. When calculating a generic screening level for contaminants that do not have a toxicity value in the
eight sources listed in the review document or developed by state environmental or health
agencies, what do you recommend EPA do?
2. Are you in favor or opposed to the idea of a generic screening level in those cases?
3. If in favor, what reference dose do you recommend EPA use?
Reviewer 2 Responses:
1. I addressed this question in the last paragraph of my comments on the generic screening level
(copied below), and EPA should review that part of my comments. In summary, for contaminants
that do not have a toxicity value in the eight sources listed in the review document, I recommend
that chemical-specific toxicity values from other sources (e.g., values developed by state
environmental or health agencies other than California EPA) be reviewed and considered.
A-57
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
2. If no chemical-specific toxicity value that is considered to be acceptable is located, then I agree that
use of a generic value is necessary.
3. By its nature, the generic screening level is highly uncertain. The generic value should be selected
with the expectation that it will be protective for most or all chemicals without toxicity values. EPA's
proposed generic screening value of 3 x 10-6 mg/kg/day, based on the ATSDR MRL for PFNA, is
acceptable from a numerical viewpoint. However, the EPA document states that the generic value
selected is "the lowest final toxicity value (that is, the most stringent toxicity value that was not
draft or being developed) available among the contaminants found in fish". The final ATSDR MRL for
PFOS (2 x 10-6 mg/kg/day) is lower than the ATSDR MRL for PFNA (3 x 10-6 mg/kg/day) which was
selected as the generic screening value. As such, it is unclear why the lower PFOS MRL was not
selected as the generic screening value.
My earlier comments about the generic screening level are on p. 6 of my response to the charge
questions [included again below],
FROM p. 6 OF MY EARLIER RESPONSES TO CHARGE QUESTIONS:
p. 8, last paragraph, "generic" screening level. Itis recognized that a "generic" toxicity value for screening of
contaminants for which no toxicity value is available is needed. Based on the bioaccumulative potential and
low-dose toxicity of long-chain PFAS, it is likely that a toxicity value based on a long-chain PFAS such as PFNA
will be protective for most other contaminants. That being said, the "generic" screening level based on a
toxicity value of 3 x 10-6mg/kg/day (based on the ATSDR MRL) for PFNA is highly uncertain.
Additionally, it is stated that the "lowest final toxicity value (that is, the most stringent toxicity value that
was not draft or being developed)" was used for the generic screening value. However, as discussed above,
IRIS is currently developing Reference Doses for several long-chain PFAS based on human data, and these
IRIS Reference Doses are lower than the ATSDR MRLs based on animal data. It is likely that the draft IRIS
toxicity assessment for PFDA, which includes a much lower Reference Dose than the ATSDR PFNA value used
here, will be finalized soon. When this occurs, will the "generic" toxicity value be revised?
For contaminants that do not have a toxicity value in the eight sources listed in the review document,
chemical-specific toxicity values from other sources (e.g., values developed by state environmental or health
agencies other than California EPA) could be reviewed and considered. It is stated in the section on
Researched Toxicity Values that other sources were not used "because of the variability of methods applied
and inconsistency of the existence of adequate quality control documentation." However, it is unlikely that
chemical-specific values developed by states (e.g., New Jersey, Minnesota, Massachusetts) using EPA risk
assessment guidance are more uncertain than a "generic" value based on the toxicity value for a different
chemical. As one example, New Jersey has developed a Reference Dose of 1.3 x 10-6 mg/kg/day (1.3
ng/kg/day) for perfluoroundecanoic acid specifically for use in fish consumption advisories. See
https://dep.ni.gov/wp-content/uploads/dsr/pfunda-fish-consumption-trigger.pdf.
A-58
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
COMMENTS SUBMITTED BY
REVIEWER 3
A-59
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
External Peer Review of the Process for Selecting Contaminants
to Monitor in Fish Advisory Programs
Charge Questions
1. Is the process EPA followed to identify compounds for which fish and shellfish advisories might
be needed reasonable?
Yes, overall, the process EPA followed to identify priority compounds is reasonable. However, EPA
might consider revising the documentation and analyte selection process in these areas:
1. Update the equations used to calculate screening levels (SLs) to more closely align with
current fish advisory practices. The current equation cites to the 2000 guidance, which is a
special case of a more general equation.
2. Provide more analysis and documentation of the fish tissue concentrations summarized
from the literature, particular for analytes that are selected because the sample maximum
concentrations exceeds the SL.
3. Consider providing different weighting factors to these two conditions:
A. sample maximum > SL and sample mean < SL
B. sample maximum > SL and sample mean > SL
4. Consider refining the decision process for selecting an RfD to serve as a protective
surrogate value when the RfD is missing for a chemical.
5. Derive a SL for lead (Pb) using EPA's lead risk models, rather than the cancer slope factor.
6. Either exclude the lipid-normalized concentrations, or apply a default assumption for lipid
content to convert the values to wet weight units.
The basis for each recommendation is provided below.
Screening Level (SL) Equations
Separate equations for calculating a fish tissue screening level (SL) are provided for noncancer and
cancer endpoints. The equations are consistent with the 2000 Guidance, but could be updated to
more clearly show the underlying assumptions and to reflect how states currently implement fish
advisories. Applying abbreviations for convenience, the equation presented to calculate a
screening level for noncancer effects (SLnc) on p.7, including the unit conversion factor (CF) for
mass discussed on p. 8, is:
RfD XBW
^ ^ (g ww/day) x 0.001 kg/g
CR X CF
where
SLnc
BW
RfD
fish tissue concentration (mg/kg ww)
chronic oral refence dose (mg/kg-day)
body weight (kg)
A-60
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
CR = average daily fish consumption rate (g ww/day)
CF = conversion factor (0.001 kg per g)
What is implied, but not stated directly, is that the SL is the concentration that, when included in
the calculation of average daily dose (ADD), equals the RfD. In other words, the ratio of the
ADD/RfD is 1, or equivalently, the target hazard quotient (THQ) is 1. Also, in practice, most state
agencies consider fish consumption rate to be the product of the meal size and meal frequency,
which is how different meal frequencies are ultimately determined. Finally, some agencies also
apply a relative source contribution (RSC) to account for additional exposure pathways that may
contribute to a total average daily dose. Considering all of these concepts, a more general
expression for SL is:
THQ X RfD X RSC X BW
SLnc ~ (MSl
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
chances of observing an extreme value actually increases with increasing sample sizes. It is clear
that one of the reasons for selecting the maximum is that the choice of statistics is limited to a
large extent by the information presented in table summaries in the literature - it is unreasonable
to expect to obtain the underlying raw data from most published studies. However, a preferred
(more stable) statistic, that achieves the goal of representing a high-end value, would simply be an
upper percentile (e.g., 95th percentile, or even 99th percentile). A recommended hierarchy of
summary statistics for representing a high-end value is:
• Reported upper percentile (90th, 95th, or 99th)
• Estimate of upper percentile based on an assumed distribution (e.g., mean and standard
deviation are reported, so assume a lognormal distribution to estimate the corresponding
95th percentile)
• Sample maximum
The following extreme cases of sample maximums are noted by comparing the ratio of the sample
maximum to the arithmetic mean:
Worksheet
Chemical
Maximum
(ng/g)
Average (ng/g)
Ratio of
Max/Average
Fillet
PFDoA
859,000
4.2
204,135
Fillet
PFOS
2,840,000
53.1
53,525
Whole Body
BDE-99
650
0.24
2,708
Given the unreliability of the sample maximum as an indicator of conditions on a national scale,
the rather large set of analytes for which only a maximum is provided (there are no estimates of
the mean) should be carefully considered, at least in terms of the weighting scores used to rank
each analyte. The following counts of analytes for which no "average" is available are noted, by
chemical class:
Worksheet
Chemical Class
Number of Analytes
Missing an Average
Fillet
Flame Retardants
16
PFAS
12
Metals
1
Chlorinated
1
Cyanotoxin
1
Other (paraffins)
2
Whole Body
Flame Retardants
8
PFAS
3
A-62
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
Metals
1
Other (paraffins)
1
The one metal listed in the table above is for lead. Lead is included in this guidance based on the
cancer slope factor, which is an extremely unusual choice. From my experience as a toxicologist
and frequent participant on EPA's science advisory panels involving lead, lead is not regulated
based on the cancer slope factor at any site, for any medium. USEPA and state agencies rely
instead on the screening levels developed from regulatory models that predict blood lead
concentrations (e.g., IEUBK or Adult Lead Model) from average daily intake. The USEPA Regional
Screening Level tool1 and guidance notes, "EPA has no consensus RfD or SFO for inorganic lead, so
it is not possible to calculate SLs as we have done for other chemicals". EPA should develop a
generic fish tissue level using one of EPA's lead models. For example, alternative dietary inputs can
easily be included in the IEUBK model for children to develop a protective SLfor lead in fish tissue.
Consider also including the number of studies and the number of study values that were curated
from the literature and used to derive the "Maximum" and "Average".
Do not include the tissue concentrations that are lipid normalized, directly in the comparison to
the toxicity values. The units matter in this case. A preferred approach would be to apply a general
assumption for % lipid content to convert the lipid-normalized values to wet weight
concentrations. Or, alternatively, exclude the study results that are expressed only as lipid
normalized values.
Surrogates RfD for Missing Values
EPA elected to the RfD for PFNA (3E-06 mg/kg-day) as the proxy value for analytes without an RfD
because, "it is the lowest final RfD for all contaminants being considered for inclusion in the
monitoring list". In the Excel file, these are listed as "generic SLs" and include chemicals from a
wide range of categories: antibacterials and antibiotics, cyanotoxins, flame retardants, and
pharmaceuticals. This extrapolation across chemical classes seems unnecessary when it is possible
to select from the lowest RfD with the same chemical class.
2. Is the list of contaminants advisory programs should consider monitoring for reasonable (e.g.,
reflects the current range of contaminants detected in fish with potential human health
impacts)?
Yes, the range of chemical classes makes sense and appears to be comprehensive. See above for
recommendations on revisiting the approach used to derive SLs for some of these analytes.
3. Are there additional contaminants that should be included in the "monitor for advisories" list
or "monitor to watch" list? If so, what are they, and why should they be included?
I am not aware of any additional contaminants that would be reasonable candidates to include in the
monitoring lists.
1 https://epa-prgs.ornl.gov/cgi-bin/chemicals/csl_search
A-63
-------
External Peer Review of the Process for Selecting Contaminants to Monitor in Fish Advisory Programs
Requested Clarification from EPA
Question for Reviewer 3:
When calculating a generic screening level, you recommended selecting the lowest RfD from the same
chemical class. If a chemical class does not have a contaminant with a final reference dose (e.g., paraffins),
what would you recommend that EPA do?
Reviewer 3 Response:
For chemicals that are part of a chemical class that does not have a final toxicity value (reference dose or
oral cancer slope factor), there are 3 options that can be pursued:
1. Apply toxicity values developed by another program office of USEPA
2. Use established computational toxicity models to identify a suitable surrogate chemical/chemical
class from which to estimate the toxicity value. There is an extensive effort within EPA to develop
tools and frameworks for just this purpose.
3. Do not develop a fish advisory at this time. The rationale would be that Approaches 1 and 2
introduce too much uncertainty to develop a risk-based fish tissue concentration.
In the case of paraffins (polychlorinated n-alkanes), EPA has published several reviews of the literature on
animal toxicity and human epidemiological data. The most recent was prepared by EPA/OPPT for the TSCA
Section 5 New Chemicals Program. The December 22, 2015 report is available online here:
https://www.epa.gov/sites/default/files/2015-12/documents/dover -
standard review risk assessment p-12-0282-0284 docket O.pdf
• Section 4 (Human Health Hazard Overview) summarizes the literature on medium-chain chlorinated
paraffins (MCCP) and long-chain chlorinated paraffins (LCCP).
• Tables 9 and 10 present dose estimates from drinking water and fish consumption based on
modeled exposure concentrations.
• Sections 6.1.2 and 6.2.2 present the exposure and risk evaluations relevant to oral exposure via fish
consumption, and how toxicity values were derived from selected points of departure (PODs)
derived from key animal toxicity studies.
While the use of these toxicity values, with attribution to EPA/OPPT, introduces some uncertainty in the fish
advisory calculations, there is less uncertainty applying these values than using PFNA as a surrogate for all
chemicals.
A-64
------- |