United States
Environmental Protection
Agency
EPA-452.R-96-001 hi/-
June 1996
Air
Mercury Study
Report to Congress
Supplemental Materials
for SAB Review
SAB REVIEW DRAFT
U S Environmental Protection Agency
R;gion 5, Library (PI-12J) ^
77 West Jackson Boulevard, 12tn
Chicago. IL 60604.3590
&EPA
Office of Air Quality Planning & Standards
and
Office of Research and Development
c66011-2-8
-------
10
r-~
i-y
SUPPLEMENTAL MATERIALS FOR SAB REVIEW
it
C->
•so Table of Contents
v;^
I. Administrative Materials
II. Scientific Materials
Dietary Exposure to Methylmercury
1. Guide to Figures includes text and graphs showing estimated exposure to methylmercury from fish
and shellfish.
2. Macintosh D.L., J.D. Spengler, H. Ozkaynak, L.-h. Tsai, and P.B. Ryan (1996) Dietary exposures
to selected metals and pesticides. Environmental Health Perspectives 104(2):202-209.
3. Mahaffey K.R., 10 May 1996. Note on "Support for Mercury Exposure through Fish".
Neurobehavioral/neurodevelopmental Effects of Methylmercury Exposure
1. Watanabe C., and H. Satoh. 1996. Evolution of our understanding of methylmercury as a health
threat. Environmental Health Perspectives 104 (Supplement 2): 367-379.
2. Grandjean P., P. Weihe, P.J. Jorgensen, T. Clarkson, E. Cernichiari, and T. Videro (1992) Impact
of maternal seafood diet on fetal exposure to mercury, selenium, and lead. Archives of
Environmental Health 47:185-195.
3. Papers on Seychelle Islands Mercury Studies published in NeuroToxicology, Volume 16, winter,
1995; provided under separate cover.
4. Two additional recently published papers on mercury in water and on mercury and cardiovascular
disease/any death.
-------
SUMMARY OF THE ADMINISTRATION REVIEW OF THE MERCURY STUDY
REPORT TO CONGRESS
For Submission to the Science Advisory Board
June 1996
Office of Air Quality Planning and Standards
and
Office of Research and Development
U.S. Environmental Protection Agency
U.S. Environmental Protection Agency
Region 5, Library (PL-12J)
77 West Jackson Boulevard, I2th Floor
Chicago, IL 60604-3590
-------
After subsequent revision, the report was again distributed for review on March 22. 1996
in anticipation of an April 15. 1996 release The fourth section of this document represents the
final review comments. At this point, the decision was made to refer the report to the Science
\d\tsory Board (SAB). As a result, only a few of the latest comments were incorporated into the
document The SAB is receiving the Mercury Study in essentially the same form as the March
22. 1996 draft.
-------
Summary of the Administration Review of the Mercury Study Report to Congress
This document represents a compilation of the comments the U.S. Environmental
Protection Agency (EPA) received from other federal agencies and offices regardins the Mercur\
Stud> Report to Congress, and the EPA's responses to such comments. The information
compiled represents the review of 3 versions of the draft report: each successive version
incorporated the EPA's response to suggested changes and comments.
Comments on the draft Mercury Study were provided by the Food and Drug
Administration (FDA), the National Marine Fisheries Service (NMFS). the National Oceanic and
Atmospheric Administration (NOAA). the Office of Management and Budget (OMB), the U.S.
Department of Agriculture (USDA). Council of Economic Advisors (CEA). Council of
Environmental Quality (CEQ), Office of Science and Technology Policy (OSTP). Department of
Energy (DOE). U.S. Fish and Wildlife Service (FWS). the National Institute of Environmental
Health Sciences (NIEHS). and the Centers for Disease Control (CDO.
Background and Chronology of the Review
In September 1995, at the request of Dr. Lynn Goldman. EPA Assistant Administrator for
the Office of Prevention. Pesticides and Toxic Substances, a special panel was convened by the
Department of Health and Human Services Environmental Health Policy Committee (EHPC).
This EHPC panel facilitated dialogue between EPA and FDA on health-related issues contained
in the Mercury Study. The panel reviewed the executive summary and assisted in the revisions to
that volume and preparation of appropriate public health messages.
The charge to the EHPC panel, comments provided by FDA and the panel, and EPA's
responses to these comments are found in the first section of this document. The final
disposition of the EHPC panel is also included in a letter to the Assistant Secretary for Health.
Department of Health and Human Services from the Chair of the Ad Hoc Panel on
Methylmercury. Dr. Richard Jackson.
Subsequent to review by the EHPC panel, the Mercury Study was submitted to the OMB
for Administration clearance. At this point, the study was due to be released on December 15.
1995. The OMB distributed the document to the above agencies for review. The second section
of this document represents the comments received during the December 1995 time frame, and
EPA's responses to them.
In mid-December, the EPA decided not to release the report as planned on account of the
pending publication of data from the Seychelles Islands study. After publication of the partial
results of the Seychelles study, the decision was made to release the report. Consequently, a
second review of the revised study was undertaken by the other federal agencies and
Administration offices. The third section of this document represents the review comments of
the December 14. 1995 draft and EPA's responses to them.
-------
SECTION ONE
REVIEW OF THE MERCURY STUDY BY THE AD HOC METHYLMERCURY
PANEL CONVENED BY THE ENVIRONMENTAL HEALTH POLICY COMMITTEE
This section contains the following documents in order:
1. Charge to the EHPC Ad Hoc Panel
2. FDA comments, October 13. 1995
3. Ad Hoc Panel comments, October 16, 1995
4. EPA response to FDA and EHPC comments (including revised Executive Summary),
dated October 20, 1995
5. FDA comments dated October 24, 1995
6. EPA response to FDA comments of October 24, 1995
7. Disposition memorandum from Ad Hoc EHPC Panel to Assistant Secretary for Health
-------
ISSUES FOR CONSIDERATION BY THE COMMITTEE
Assessment of the extent of risk posed by mgestion of methylmercury require
identification of adverse health effects or hazard, understanding of human dose-response
to methylmercury, and delineation of conditions that result in methylmercury exposures.
Interpretation of the level of risk produced must describe accompanying uncertainties
and variability in the assessment or characterization of risk.
Human Dose-Response to Methylmercury
Ingestion of methylmercury resulting in overt, clinically evident poisoning has
resulted from at least three food sources: fish, mercury-treated grain, and pork. In
addition, epidemiological evaluation of populations who consume fish has provided
additional data aimed at identifying human dose-response to methylmercury at an
exposure range associated with minimal prevalence of the most subtle indicators of
methylmercury toxicity.
With regard to assessment of human dose-response to methylmercury, the
Committee is charged with identifying:
1. What are the most subtle indicators of methylmercury toxicity in maternal-
fetal pairs and children?
EPA will provide a copy of the NIEHS "Report to Congress" on this topic.
2. What is the contribution of various studies of human dose-response to
methylmerciry given the end-points evaluated? What can these contribute
to current understanding of dose-response to methylmercury? What
inference can be drawn from these studies?
EPA will provide: the draft volumes of the EPA Report to Congress on
Human Health Effects, Exposure to Mercury, and Risk Characterization;
copies of primary references used from Minamata and Niigana; references
used in evaluation of mercury toxicity in Iraq; the official report to the
Swedish Government by Kjeltstrom et al. (1989) on methylmercury
exposure and effects in a fish-consuming population in New Zealand; and
publications on methylmercury exposure from fish and indicators of
methylmercury toxicity among Cree Indians.
3. What are the inferences that can be drawn about methylmercury exposures
to the populations described in "2" above?
4. What is the contribution of current studies on dose-conversion equations
-------
Change to Committee
mercury exposure based on these data.
2. What inferences can be drawn about fish consumption by the highest 5%
of the general population based on general dietary survey data, particularly
cross-sectional data?
EPA will provide their analysis of fish consumption data in the "Mercury
Study Report to Congress".
3. What is the role of specific studies on groups recognized to consume
larger than typical quantities offish from specific locations (for example, as
seen in publications on recreational anglers, subsistence fishers, members
of some Native American Tribes)? How does EPA correctly represent the
distribution of people who consume high quantities of fish?
EPA will provide analysis of fish consumption data presented in the
"Mercury Study Report to Congress".
Characterization of Risk from Methylmercury Contamination of Fish
Any risk assessment acknowledges uncertainty and variability in interpretation of
data for hazard, dose-response and exposure assessment. Specifically, the Committee
is charged to:
1. Given results of the above deliberations, how could EPA better represent
the general population of fish consumers with respect to uncertainty and
variability?
2. Given results described above, how could EPA better represent specific
subpopulations at the highest end of the distribution of fish consumers (for
example, anglers, members of some Native American Tribe, persons
consuming large quantities of fish for health-based reasons) with respect
to uncertainties and variability in their consumption patterns?
-------
Change to Committee
used to translate indicators of body-burden of mercury (for example, hair
and/or blood mercury concentrations) to quantities of methylmercury
ingested?. What do currently available dose-conversions equations tell us
about individual variability in half-times of methylmercury among humans?
What do currently available studies tell us about subpopulations of concern
(specifically, women of child-bearing age, pregnant and/or lactating women,
and children)?
EPA will provide copies of publications cited in the EPA "Mercury Study
Report to Congress" and relevant volumes of the EPA "Mercury Study
Report to Congress".
5. What do the studies described in "2" above contribute to understanding of
the magnitude of methylmercury exposures identified in the cited studies?
What does this information infer about human mercury exposures with
respect to "No Observed Adverse Effect Levels" and "Lowest Observed
Adverse Effect Levels"?
Human Exposure to Methylmercury
In the United States methylmercury exposures are caused by consumption offish.
It is noted that the United States Food and Drug Administration is concerned with fish
in interstate commerce, whereas the United States Environmental Protection Agency is
charged with regulation of primary sources of emission of mercury into air and release
of mercury into water that result in contamination of fish. These two agencies both rely
on estimation of methylmercury exposure from fish, but differ in which populations they
evaluate and the types of exposures considered. The United States Food and Drug
Administration focusses on the general population consuming fish purchased
commercially. The United States Environmental Protection Agency focusses on special
subpopulations (including anglers, some Native American Tribe members, subsistence
fishers) who may consume substantial quantities of fish from a local source. These
differences are acknowledged by the two agencies.
The Committee is charged with:
1. What are the relative merits of expressing fish consumption on a "per
capita" or on a "per user" basis based on dietary survey data from the
general United States population?
EPA will provide their analysis of fish consumption data and estimates of
-------
'.M3-95 04:53?M FROM :?A:'3-S CrSAN/OD TO 3!513569;4/5
?002
FDA Draft, Oct. 13,1995
Recommended Revisions to
the Executive Summary of the EPA Mercury Report
Page 2-1, last line,
It would be better to say that how properties of water bodies influence MeHg
concentrations in fish is not veS understood at this i^ nlbst than entirety understood
It must be clearly stated that the models used to ^fi^T die relationship between mercury
emissions and methyimercuiy concentrations in fish, relate entirely to estimates in
freshwater tub and not to marine speciet. The models seem to tefl us little of the
relationship between emissions and levels of MeHg in marine species.
Page 3-2
We would fike to sec the whole section on warnings against fish consumption in the
Unites State* elimmaucd. There is no evidence that the warnings correlate with the
emissions that EPA is trying to control In fact, there does not seem to be any evidence
that mercury levels in fish have been influenced at all by industrial sources of mercury.
Page 3-1, qaestkra 3
Rewrite to: •How do U.S. EPA's estimates of bow much methyimercury is harmful to
humans compare with levels calculated by the U.S. Food and Drug Administration, WHO,
and the states?"
Page 3-1, "What ctTMta~~.", seeoad paragraph
The sentence referring to several epidemics needs to be very specific. There were exactly
hap epidemics involving industrial conttnvptted fidi Both occurred in Japan.
The severity of exposure in these two epidemics needs to be stressed:
Two epidemics of methyimercury poisoning have occurred through fish consumption. In
each '"«**?¥£?, eAiieuis) exposures resulted from the consumption of large amounts offish
that were heavily contaminated with methyhncrcury from local industrial pollution. The
best known of these two epidemics occurred...."
Page 3-2, tecoad paragraph
Line 6: "if the mother ate large amounts offish contaminated....'*
Page 3-2, fourth paragraph
Last
Rewrite tn- "Ti^ qtfaer mrtatvyy wfrprc *Wt* *"«« *^ITI CTTlTamtnat^ "** **gft IfVfll of
methyimercury (eg., grain and pork), severe methyimercury poisoning has resulted with
pathological "
1
-------
.•>:3-95 04:5Crii rROM rCA.rH orSAN/OD i'O 3l5iJ3bsi4i3 P004
the higher exposure level. This is not consistent with what was estimated on the baas of
the Iraqi study.
Page 3-4, last paragraph
Please, add more about the StycheDe Islands Study
Page 3-4, last paragraph
These studies, unlike the Faroe and SeycheUea Islands studies, have only been
underway for a short period of time. Any pronouncements of observed effects are
exceeding preliminary, and what role methytmercury may play in these studies wiD take
some time to sort out Again, in light of the stated policy to not include unpublished
studies, the Executive Summary could cite their existence without details or discussion.
Page 3-5.
The role of Food and Drug Administration has been largely ignored in this section. The
activities of the FDA predate those of EPA, the states, and the World Heahh
Organization. A narrative including some information from our overview would make this
section more comprehensive.
Page 3-5, Heading
Rewrite to: "How do U.S. EPA's girimatag of how much methylmercury is harmful to
humans compare with levels calculated by the U.S. Food and Drug Administration, WHO,
and the Hates?"
Note: EPA does not discuss what the states think is harmful, they just mention the
ofadvUoriw.
Please see FDA Overview for details.
Page 3-5, third paragraph.
It is stated that a cross-comparison of limits on MeHg is iacUhated by the data from the
WHO (reference not cited, presumed to be the 1990IPCS monograph). It is difficult to
see how the information provided in Table 3-1 provides insights into MeHg concentrations
mlrvcr, kidney, feadief^ete. The poo* the author and this
needs to be rewritten so that the reader can get a clear understanding of bow the various
limits which are created using a number of different approaches, are related to the
information in the table.
Paragraph 3 could be followed by:
"Studies on hair mercury levels in the U.S. have been conducted under contract to the
U.S. National Oceanic and Atmospheric Administration. One study from 1983 reports
observations for 1431 women participants in a nationwide fish consumption gurvey. A log-
normal distribution of mercury hair levels was observed.The mean hair mercury level was
0.48 ppm and 99.72% had hair mercury levels less than 3.9 ppm. The maximum level
observed was 6.3 ppm. A second study from 1985 involved 341 pregnant women in the
-------
Page 3-2, third paragraph
Delete "from fish" in line 4. (..how methyimercury 0 produces neurological disease in
humans.)
From line 4 and down the page: Rewrite to:
"In Mlnimata and NUgata, the epidemics resulted from the long term consumption of
large amounts offish contaminated with abnormally high levels of methyunercury. These
poisonings demonstrate that severe exposures to tnethylmeruuy UXMU £bh can produce
human disease. Since the extent of poisoning in these events was severe, the finding of
more subtle effects associated with lower exposures was difficult. Subtle indicators of
poisoning are important for identifying levels of exposure which wifl not cause adverse
effects. The U.S. EPA calculates one measure of such levels, the so-called reference dose
or RfD. This is die amount of material (with an order of magnitude uncertainty) which can
be consumed on a dauy basis over a lifetime without harm."
The Iraqi poisoning was also severe but evidence was obtained on the occurrence of more
subtle effects from individuals •who were less highly exposed. Nauroiogic changes were
observed in Iraqi cnfldren who had been exposed in utero. The endpoints included lateness
in walking, late talking, and scores on tests of nervous system function.
X7.SJEPA has calculated an R£D from the maternal hair mercury levels (measured in all of
the mothers in the study) and the incidence of neurologic changes in the Iraqi children. In
calculating the mercury level in hair which was associated with no adverse effects, the
U.SJEPA. ..... "
tecond paragraph
Second yr'ffcf- should be rewritten;
"Due to variability m ithe way individuals process me&ylniercury w the body and tlie lack
of data on reproductive effects, an uncertainty factor of 10 was used to derive the RfD
from the benchmark dosex"
Page 3-4, third paragraph.
Ftntt line: Then *T9tignifloantuaG*tiirai»t in the KID." The word significant naadi to
be added to provide a qualitative expression of the degree of uncertainty associated with
the RfD. This wffl provide the reader with an appreciation of the level of confidence in
this cumber.
Page 3-4, fourth paragraph.
We would like EPA to consider deleting this for several reasons:
The report of the New Zealand study waa not published in the peer-reviewed literature
and aa has been stated many times by the EPA authors, only peer-reviewed publications
can be included. The study of the Cree Indians does not support EPA's dose-response
estimate. As noted by the authors of the study, there was no dote-response rfiafionafaap in
that the effect on the mala offspring waa only noted at the lower exposure levd and not at
-------
Q4-.50PU FROM FDA:?3-3 CrSAN/OD TO 9151256974,5 ?006
Page 3-15, paragraph 3.
There is no evidence that children are any more sensitive than adults to MeHg, and to
assert that the effects seen in the fetus are applicable to children is not scientifically sound.
The neurogenic events that may be effected by MeHg in the fetus are not the same as
those that occur in children. To argue that because children consume more fish on a body
•weight basis is a reason to apply this RfD is misleading. Expressing the exposure on a
body weight basis accounts for the differential consumption of children and cannot be used
to justify using an RfD calculated from a fetal endpoint. The reader is left with the
impression that there is scientific evidence that children are as sensitive as the fetus and
more so than adults to MeHg when no such evidence exists, either in animals or humans.
Page 3-1S, third paragraph
Unlike environmental monitoring data, information on body burden of mercury (hair
mercury data) in populations of concern is available. Hair mercury data for women of child
bearing age suggest that the mercury exposure modeling estimates may provide unreliable
indications of the extent of mercury exposure among the U.S. population
This conclusion needs to be reflected throughout the report.
Page 3-18, fourth paragraph
This paragraph which is supposed to delineate the limitations of risk characterization does
not seem adequate. It does not mention the significant uncertainties associated with use of
the Iraqi study to calculate an RfD nor does it acknowledge the significant uncertainties
associated with exposure analysis (see comments above). Again we will refer to our
overview, in which we reviewed the potential problems associated with the use of the Iraqi
study.
-Small population (83 mother/infant pairs) and only four subjects with exposure
levels below those associated with adult effects and whose exposure levels
approach exposure levels seen in fish-eating populations, particularly in the United
States.
- Possible differences in tcoticoldnetics between MeHg in grain versus MeHg in
fish,
- Significant nutrient interaction* with M«Hg (c.g., selenium and zinc) that may
occur with fish consumption but not with grain consumption.
. Uncertainty in measurement of nervous system endpoints. Although the date of
the onset of walking was used as an index of effects of MeHg, the specific age of
the children was not known because the date of birth is not noted or observed in
this culture. Tins was also highly dependent on the recall of the mother because the
interview was performed retrospectively (many months to yean later).
- Significant differences in the type and duration of exposure - high level, short
duration in Iraq and low level steady-state exposure among fish consumers.
-------
?DA:?3-3 CFSAN/OD TO 315!3w.T) to be at 5% risk of neurologic disorder in
the infant. This estimate is predicated on the expectation that their hair mercury levels
corresponding to thia consumption level occur in the 1 0-20 ppm range. Observations from
the NOAA sponsored studies of hair mercury levels in women and pregnant women
suggest that very few women in the U.S. are likely to be exposed to mercury at levels
associated with increased risk to the fetua.
Figures 3-1-
As discussed, these graphs will lead to considerable risk of communication problems if
they are included aa is.
-------
SENT SY'NIEHS : 10-16-35 , i:Q5PM : of I nndDA OFFICE-
October 16, 1995
TO: Kate Mahaffey
phone . 513-569-7957
fax 513-569-7475
FROM: Sheila Newton
phone 202-205-8180 (today 301-496-3511)
fax 202-205-9478 (today 301-496-0563)
Subject: Methylmercury Report cements
I am sending you what I received from Drs. Jackson, Sinks, Coyer,
and Levine. I have also tried to insert their suggestions where
appropriate in my copy of Volume I. I am sending you the relevant
pages with markings. ; I started to include the PDA comments, which
you received separately, but decided not to complete them—we may
want to discuss them independently on our phone call.
I will send this with a Fine setting and hope that it is readable.
If the fax copy cannot be deciphered, please let me know ASAP at
the NIH phone number (301-496-3511).
The fax includes 24 pages counting this cover page.
As I mentioned on the phone, we will convene our conference call
tomorrow (October 17) at 3:00 p.m. Eastern time. I will contact
you with the number to phone in for the call as soon as I hear from
CDC.
Sheila
cc: Panel participants
-------
i-uor* . ocuicjuH urnui
1J
can be oonsuaed on a daily basis over a lifetime without harm.
The new DA R« was calculated using a atudy of Iraqi children
•uffering advarae neurologic effects tftar being axposed la utero
to Mthyl Mxcury whan treated teed gr»in was ground into flour
for br««d.
Th« SPA, »nd th« TOA vik^d that tha CtPC rvviav tba poaitlonft ef
the tw agwiQiat and aaXa raconwndations to aaaiit tlM agaaeiaa
in raachiog a aen««n»\w petition.
1) AM study that farm th« baiia of tha XfD baa ••varal
liaitationa (••9.; •tudy lisa, ralativaly high doc* axpocura*;
and petantial o«nf eundara} . ]
a) Two vtudiaa without many of ta*M iivitationc art a«ar
oonplation*
3} FDA will attafeliab an action law«l Cor natttyl marcury in tiah
one* tba naw atodia* arc available.
4) Taa VA RfD ia iadapcndant of oonaidarationc regarding the
anthropogenic relvaee ot aareury into the ataoepher*.
9) TJw report will be misinterpret**. People will he oonfuced by
a neeeage that oooevminq f lah will remit la expoewre to aetayi
ouy above the
1) tf\ hae developed the report in an open proceaa involving
external Pm review as wall as intezagency review. (
2) Tae •»t*bliah»ent of an XfD ie relevant to evaluate the riak
to human health from expoture to aetxrary,
1) Tne BfD ie based on aa eetiaated Ve Obeerved Adverse Jtfeot
Level (MOA1L) thet is a factor ef 10 greeter than the IfD. In
addition the vuwertainty aurreuading the RID is an order of
0A believe* the 9X0 i* an important part of the
The feed supply does net appear to be oomprceiieed at this
_*•. tvoduoere of fish end iittuscrieo responsible for •ere
•missions should be avers of the danger from mercury entering
food obeia end further contaminating fiah. women who consume
re them 100 gram* of fish eeoh day/ espeeielly from local
iroes where meroury lovels in fish ars high* eheuld else be
of the potential problem.
8) The DA has until October 20th to make changes in the report.
-------
Haalth Policy coodttaa (UPC)
public laalth, ftmrvio* (paa> , DBHB
Spaeial Raviav Panal (OP)
•unmary oft
Maveury ftudy Raport to coagraaa
Off iea of Air Quality Planning and standards
tnd
otfioa ot Maaarob and D*valop«ant
o.f. InvironBavital Protaotlon A^aaey
Chairt
ilobart 7. JaokAttB, MO, MVH
Director, national Oaatar for InYiraraiantal Ha«lthr CDC
_
Coyar
Xiohard Larina
Aaaila navtaa
a Ink*, PU>, MaooUta Director for Soiaaea,
Th* Cl«aa Air Vot AaandMiita of lf90 (CAA) Mtabliahad aaotioa
na(n) (i) (B) viaea nqairM tna o§ IPA to atudy taa ixpaeta of
maccury aic pollution. Tba report oontiata of 7 Telama* and om
Voluaa It Pindino> ani H>OQ«aandatiqna
VeltaM III Inrantdry of Anthropoganio xarourr BBiMieni in tba
stataa
volu»a in i An AaaaaCTant af ixpoaura fm Aathropoganie Nureuyy
•adaaiena in tte Onitad Stataa
iv: Haalta iffaet» of Haroury aad Maroury Ccaipounda
Vi An »«alaf:iaal l««*»r»«nt for AntteofOfanio ttaroucy
in tn« O&itad Stataa
voluma VXi ctaxaotaximatloo of Bonn Haalta and vildlifa RlaXa
froa Aatnropofanio Karcury BHaalona in th* Oaitad stmtaa
supplamant Ai •MMTT of tba atutfy and aolaatifio
Durinv tte iatanganey xaviav of th« docunant ta« rood and
AdBinUtrmtion (IDA) iftjomad tba BPA of coooarua ragardlng aov
t&a daounant daalt vit& iaauaa r awarding hnain aapnaiara to nathyl
fro* f iaa nonaunption. Partieularly, PDA
ab«ut taa a0ta»lifiMant af a raf aranoa doa« (KTD) fear nathyl
that
•fD ia tha anonnt of nat&yi naroury
-------
SENT 3Y--NIEHS ; 10-16-95.; 1.-07PM
saoila/ bora aro ay ooaBonta for tho ravrita of veltmo I. i
won't aoatioft opociflo laoguaqa. x will ouggaat vhoro in tho
doeuaaat X think tho » con««n»u» point* ahould bo oatorod Into
voluaa 1.
Doooription of Iraqi tad Mlaiaata opiaovioo ohould atato that toa
lovolo of aothyl aoreary oenaoaod voro aueh highor than tho
lovolo ourrwitly ropertod In too Of rood oupply.
Pago 3-3
Tho firat tiao tho ftfD io montlonod A oloar •tatoaant tnatt
l) it •otiaatao a oaf a lovol, not a haaardouo lovol. a)
iaoorporatoa a aargin of Mfaty a< voll ao vajoortalnty, a) la
boood on limitod Iraqi datar 4) eurrtot itudioo aftould dooroaao
tho u&oortainty ourrouading both tho bonahaark and tho JttD.
Tho vording ro tho uncartiijjty (pg 3-4) oxphuiioo that tho
ro may not ba •antitivo. I think thU aay bo trxao,
aovovor> it iatft tho only iaoua rogarding tho unoortaijity. Tho
iaouoa ro tho uncertainty ineiodo population diotribvtiOBo of
fiah oonouaption and aorcury hooy burdona, variability botvoaa
blood and hair ooroury ooncantrmtioiw, and tho vory ooall numbo*
of oboorvationo in tho xroqi otudy that for* tho baoio of tho
bonehjoark. foao of tho unoortaiaty ounraonding tho KfD ohould bo
addTCOMd with tho publieation of otadioo in progrooo. Ttoo fDA
hao dotorwinod that tbo Mtabliohvant of an aotion lovol aheuld
vait until tho DOW otudiao havo roduood tho lovol of ttnoortainty
nurrouadlng tho RfD.
Z baliovo it vocthvhilo to oxplain that an BfD dooo not oot a
otaadard for moaraury lovol* la fi*h. VoopLo eon eosunaao lov^«
of mthyl Mrowy at thio lovol ovory day without axporionoing
hoalth probloao.
I aloo think that IFA ohoaid bo oloor that dotonining tho
proportion of aothyj. aorcurv ia fiah from anthropocpoaio wvoo
natural r«l«*M of Mroury into tho oavironaont U not pooolblo.
TBO woight of ovidonoo OTapporto an aaaooiation botnoon local
aoreury oaisaiono and inoroaaiag aoronry lovolo in locally eaught
fiah. Tho oTidonoo bocoaaa voakor vith iaeraaaiag diatanoo froa
oaiioioa oourcoo. THO hoalth hatard ia baaod oa too total
ooaooHtratioa of aothyl morcury in fiah boeauao tho ao«t
iapoctaat oacpoouxo aoureo to haaano ia diatary fiah oonaoaption.
Tho diooaooioa on pago 3-« rogaediag tho population at riak would
bo a good plaoo for fM to otrooo that tho, population la
boliovod to bo atfvorooly
,
affootod by aothyl aavoory ia tho Pf
fiah auppiy. Tao approprlata eaationo ahould aloo bo aoBtionad
(o,g.; w naod hottor data on population ?ena«apti«J»ffia* by
typo of fiah and rogionally caught fiah, hoaan body taraona , and
-------
J10-16-95 ; 1-07P* ;
SRP vea r»rm*d an D*M and r«c«ivaa copies of the draft
report on Fridty Ootober 6, 1985.
we set in Washington DC on Tuesday, ootnber 10, 1995. SPA and
7PA presented their viewpoints to the panel. Reeearchers briefed
tha panal on tha original zrvqi atudy and provided pcaiialaary
IMonatian on tte study oanduetad in tna fayofcalla islands. las
panal briafly eorront Uf fish supply.
4) Tha threshold vmlua for mathyl marcury oausing htnatn Haalth
affaots is iapxaois*. Data ragarding fish consuiptiofl tad Mthyl
•areury body hurdans for tha as population «ad sansitiva
Sttbccpulrtions is limitad or
S) Miullas that owild ha^tw aatiHta tha thraahold vaiua for
aathyl Mxoury ara in proojvass. I»prov«d oxpocara Mtiaatas for
tho Ui population art not in prograss.
•) OA has sat an *f& at o.l ug/xg/day that iaoorporatos a ranga
of unoartainty and is subject to ravlsion if batter data
available.
7) An IB* aotion level far aethyl Berevry would be based upon a
oosplax risk assessiaiit/yieh managament/riak baaafit process that
would eonaidar the IFA Xfs.
S) »A and FDA abauld agree an an appropriate risk oaasxaioaftian
•essegt lev Ytlnns X of the report.
•) Slven the neerteinties, farther researoh is needed.
The panel's contribution «t this point, given the li»itad UJM,
ia %• eneewafe v& and FBA develop the ri«k oosaunieation fee
the resort by revising volus* z and working together oa a press
raleaee with o/« and A'a.
-------
SENT 3Y--N1EH5 ;10-16-95 ; 1:08FM ; dtniotM OFFICE-
Axntmd anthropogenic sources, Figures 3-4, 3-5, and 3-«
show cbe quantity of freshwater flab oonmmed by children (3304 14
years or younger 1 and adult voaen of child-bearing ag« (is through
44 years) win ace rrca/nracez /.!«& once ur wrc cU/riay che
dietary
arpoaure to MeettyiMrcury frctt PISH and Shell flab:
n.s. wcna& of anid-Bflariog Age iflia sac riafi once or ffiire ffrery 3
Day*
Figure 3-2
B*tiaat«d Sxpo0ur« to MabbyliMrcury from Finh and Shell £i«hi
U.S. Male children Mho at ?4«fc Once or Jfore Svery 3 B*y*
figure 3-3
••tiaated Sxpo«uc« Co M»thyla*reurf frt» filh and Shell fiah:
U.S. F«a^i« Childna Hbo £kt J*iak Ctac» or Jfere flvery
Figure 1-*
Bxpoaure to MBtbylaereury from ?ra»h*»t«r
u.s. wof«B of Oilld- Bearing Aga Mio sac /reahwccr ntati Ctacv or
Jfcurv Svmry 3 O*ym
Flgur* 3-5
Bsclnacad ixpotura to Macbyiiwroury froa Pre»h*»t«r Fi«ht
c.«. Mala Qxtldren mto ate Ar*cftwttC«r rlah Oac« or Mar* Arvrjr 3
Figure 3-6
Estimated Sxposure to MethyliKreurr Croat PrcMMttCer
n.s. r«nal« Children Hbo Jtee mchMter «sh Ctace or «ore Xrerr 3
aays
: 9S:SI :
-------
SEMT BY'-NIEHS ; 10-16-95 '• 1 = 08PM '• a£iricoUA OFFICE- :*o/iJ
CCODWACB OH MC^g'TY H'iV^T RCP^tt t*7
Ricaara a. Levine, M.O.
Deputy Branch Chiaf
Epidemiology Branch
Division of Epideniology, Statistics,
aad prtreation fteaearcn
National Inotitute of Child Health tan Hunan Development
October 13, 1995
pJ-3 par. 4:
Roword;
0.3. EFR calculated an RfD ucing data on repoj-ead
CBAnytiM In «l rraqi coildrao who had been txpOMd in utaxe; ehair
nethoeg bad oat«s th« oontudaacad bre^d jr vciaotlaca.
Th« •exidi** ac» conductafl in Ch« seycnelle islands &•!•(•
i**e *e«taae« o/ cftl* pUttgrapA since it refer* co uqpubllatwd
•aearlal
pa-4. boctoB pax:
Delete this paxa^nyp^ since it refers to unpublla&ed aateriaa
pj-s top par:
...Often these warninos 4r* iatuad baaed OB loaal condition*, it
abouOd be noted Cfiat In ftm V.M.. rtllZa frMtsatar fiab may ^
greater lefelJ o£ Mttnylmxcuxy, ajqpoaure of perooam co
c&xzoercnry rro§ eating risa ^0 sueb jor» lifceJy co occur
susptlon o/ oceaaio fJ»h. aad not £o«i frwaawBcar fi«h.
pa -7 to par:
w (iAta lira a&OVB in Figuxaa 3-1, 3-2, and 3-3, wfeich -
v«. total quantity o£ Clak aad •halltiab consunad by callori— _„__
14 or younger) or «dult wamn of chill-baarino age (15 through 44
yvaaral vho at* fiah o«c« or aora durlar a Cnraa-day dlataxy
CSVXZ It/91. H—~"T° freatnAcer flah nay bo otf particular
t S6-CI-OI
-------
SENT BY'NIEflS
JlO-16-95 ;
"
OFFICE-
MERCURY STUDY REPORT TO CONGRESS
VOLUME t
HNDINGS AND RECOMMENDED ACTIONS
DRAFT
of Air QMlttr rtaaatet «ri
-------
Data> October 12, 1995
TOi shell* wewton
B-Mall SVEWTOHIOASSK.SSW.KMS.OOV
FAX 202 205-9471
suggeationa Cor tranaeattal to EPA re: Mercury oocuaant
I thought that the nine pointa eat out by Dr Jackaon at the
conference call aerve aa a vary good fraae worX tor •
to IPA.
I •ugg««t that v» •ocourage EPA to placa aor» ••phaala on n««d to
control aoorcaa of nareury pollution bacauao of potential h««lth
a
-------
...acapm ,«.,
2. MERCURY IN THE ENVIRONMENT
Ai a chemical element, mercury cannot be eraatad or destroyed, Tba *un» mouse ha»
odttad OB the piaoat since the urtb wa» formed. Mercury, however, can cyde m th»
enviroamoBi M put of both oatunl and human (uduopofeafe) aeeMites. Both measured data
tad the resulte from (label "*~HtBt have lad to the undemanding tbat tnthropofeaie mercury
(ten mfrfknif attributable to human activities) equal or exceed thoae from aatonl
source* («4* volcanic activity). Human activity has the overt! effect of makas awe metouiy
btolofletlUy available. Fmarhv of meccury from human tetMiy are thought to oantribute from
lMtwMO-40 «d 75 pacceot of the oonaex total annual iaput of mercury 10 the atmotphera.
Meeiund data aad mixfrltni naulo mdleate that the mount of merouy moWfaad tad released
into the bkaphere hai iacreaaed »toce the befttmfag of the tnrtmtrtal
Oocc cnitted to the tdnaaphere, mercury can depoiit to tfe* atrth in, dKbreot wmyt and at
diflerent rates, dependiBf on it> phyaical and chemical Com. The feem of mecwwy «oritted
mfluaQcaa in tanospheric ata tod tnaaport, at do eondttoos tpedfle to in «ke of release (for
exaaple, stack heifbtaBd nearby ternia). Tbe merawy bound to paniculate matter and the
divalent Conn of uexooy caa be depnetted dote to the i minim aouree («*. within 10
ktiometen). Use eiemaatal form of mercory it known to cfeculaxe hi the atmosphere for about
one year before it 1» etddaed ud depented. Tm* relatively kmf reaideoce daw k sufficient far
the emiuiooi to be mted into the ftooal atmoaphere and dapoaised onto land or water bodiaa
remote from the easfiOM •owee, iadadbf oataide of the United Sutea. Tb» rewlt ia tbat
metcury depoanioa fe a tocaJ, refteoaf aad fjbbal mm*.
.^•vw* *
MerpatycaaDedenogieddfcBCtfrtowaBatbommig
run-off aad eater ponda. atreaaa, men, likaa, etc. The water bocSea contain mieroorgaai«B»
that h*»e the metaboBc capacity to cany out ehonicaitaaea^BBwbaAoWmetcaiy tomotayl
fro«T«Hpfoowiigm»*ylniatcury. Bodiaa of »« without thamteootiaBigmi ate appear tobe
cmpabkofmetJbyUtiafmeTtury. The caeaieal properbei of methylmareaiy permit it to poaetrate
aembranea. Meth^Bemny » the fibm of mercury to which human* and wOdlife an
.
generally expoied, ujuafly from eatinf nab which nave aceuaimated nwwry ta thaw mtaele ekwie,
_ The concentre tten factor is
*in multiples lathe hundreds of tiwosaoda. Thai is, aa animal at a higher
<
-------
UST OF TABLES AND FIGURES
2tft
3-1 WHO Dm on Marctuy in Hair 3.3
±2 ttdauted Population to tb* United Sutai that Coareima Hih, Eaductinf AJvfcB
And Hawtil j.14
3.3 WfldBft Ctltttit for Ntetvy ! ."! 3-17
3U Manury SOOTOC* WUfa Sufideat Dti« to EttiauM NirionmJ Fmintnm 3.19
3*5 SOUTM Cttaffpclcs tad Poiution Prewmioa/Muvid Sep*r«k>a Oppartuitiet 3*23
Potntul M0euty BrnMoa lUductioM «04 Cora far 1uUM»il Souroi Cuafork» .. 3-23
3.1 Sroaircd Byororc to Meifayteercmy from Ffcb tad Shcflfiih: US. Women of '
Chfld-BMriof Af* .W ll\Q .Wfcr 341
3-2 BctiauMd &nosure ID Mmhyiimroay from Rah aad »»•**"->*• UA Mils
CUIdra . . IP; vJ? rrt^"................................................ 3^9
3*3 Fttiaitud Bxpoturr to Methytaarcury from Fsb and "rTBr*" ui Feoute
Cbfldrea.. VJlrtO. fcySr. 3-10
3-4 BMiBMad Scpoaw* to MttturtoBrcusy front Fnabwttw Fiak U5. Wovca of
Chfld-Betrtat Afi . Whd .^tk
3-5 Ewteuted Brporor. to M«thytm«cury from fteabH«i«r Fbh: UJ. MaU duldmi 7.r
3-4 ERfaatted BUMMUM to M«tbyta»«reary from Frcalnvtur FUh: U«S. FMH!«
Chfldm. .WV>a .?^fr 3«13
3*7 Total Mcmay W«4-Dcy Doporition (fine CMC AM^M) 3-21
1995 v( DRAFT: Do Not Ote of Quota
-------
SENT SY:N!SiS ..
ocm or-Atrox m»cooi«r
: 1-25PM :
OFF ICE"
FT-
tx**
As io the Japanese poisoning epidemic*, the tips tad symptoms of disease were
predcoffliatiy those of the nervous system: difficulty with petfeberil »<•$•*«
disturbances, tocnnrillttatton, impaina«ttt ofwilfcaafr thirrad tpoodi and maa»aaMt OMUL
duldm wore «ffaaefl, u weQ « tdults. Of treat concern «w the observation tfaat iaftnts, bora
of mothan who bad eeotiuned the aethyfBcrwiiy-contamiiused pita (partcuttriy during the
•eoond trimester of pregnancy) could *how nerou* iy*iem damage erea though toe aother wm»
only tUiatly affiscted beneit -iS -H« teves _. .
inetdenee of several endpouts (todudtof late waDdn^ Ute talkmt s^ «c«es OB tests erf nervou*
system fttoetfon) were mathemaocaly modeled to determine a mercury level in hair (measured in
att the mothers in the study) which was associated with no adverse eflscts. These affect* were
delays a asosor ead.lenfiiasje dgMeiunutem defined by thes ._ _ __ I."
Inability to dt without support by age 12 months;
inability of the iaawt to puQ hinvei&herself » a itandmf position by tSmontba;'
to weft TWO steps without support by 2 yetra of agt; and
to respond to sonpie verbal weamjatoatioa by age 2 years amonf children
whh food hearlBg.
In rslnj'i^tthft tmtfemy igjgi fa> *»g which we* Mseaiafffirt with no advene cffcca, die'
UA EPA rhCT" the benchmark (which is the lower bound for 10 percent risk of aeurotofical
changes) base* on modehnf of ail effects m chfldreo. T*e 10 percent risk levei was 11 ppm hair
3eptembef20.1»S
3-3
DILAFT: Do Not O» or Quote
-------
SEM BY'NIBiS ;lO-16-95 ; l:25PM = BEThESDA OFFICE-
aero i Oi'Anrox i w i«i;g|ji«r ;vui .iv~ii-«g i 4-« nocrv wu-.
emfesio» and aethytaercury concentration! to Of b are a ttmplificatkm of more complex aad
sometimes not wall undentood procejsef that occur in natuzc.
Ar* Tktrt Wantons Acafawt Hak Ceampttosi la the UMtsd Sartta?
Current levels of mercury in freshwater Osh in die United States ere such that advisories
have been issued la 40 states that wara agamst the consumption of certain amount* and specie* of
fish met are ooBunnineted with mercury. Nine stataa (currant as of August 1995) have statewide
advisories (Lsu advisories posted on every freshwater body in that state). Such widespread
cootammadoa » a concern tin several reasons mcittdmf the potential risk to £bh consumers hi the
to f'Hisrttir^ flahen. Native Amerioaii outnmLvabws. are aho *•*''• 5*h^< by a eentamttiated
fishery. Fffforr»«^ «r*et**> *" p*1**"1'^ **>* *to<*i* |J«ithgrT «• M ripiflart ritk from yaanaary
oontamiaation.
Th« eoooopuc mifr;*"^* of the impact of aMreary eontaamanoe on freshwater fisheries
cannot be accurately predicted, but it is mtuinVe thai fbhte| revenues eould be hnpacted should
mercury contamination wonea. la 1991.31 miiioc adub Amaricaas took advantife of fresbwatw
fishinf opportunitiet, ependj&i a total of $15.1 bffltoa OB ue»U water spordkUaf trips aad
equipment. ^.—r ~-~-j~n~ p^. . M^ »k^». t/> .h> ^««^fe viAatty rf *K— -^rf
other reiatcd eatcrprtes. In Adomas atone, mereury fish advisories issued in an 8-eoury area
were cathDated to result in a toss of $5 mfflioo in revenues in 1993.
Septeaabey 20,1995 VI DRAfT: Do Not Ota or Quote
-------
S^,,BYU:.N.IAE2!S« ,.„„»«' -««i;1.0rJ^rSa; ,1:«2.'«r«;. ^u^DA'SiSSi
,r/J
analyzed the Irtqi data and identified a 30 percent risk to tbe innot of abnormal ocurolopal
Sifn^vbec nuteraal hair mercury concentrations were over 70 ntfc. Ueiac-a&.*ddUonai
" " *" ~-WHO/ttC3 ~" "
ifbonatenat hair ceficettntfoa-war raw 20 MCmercwyjmB ofrbair
WKO/IFCS raeatnmeaded that at a preventive measure, in a aubpopulatiao that
large amount* of fifth (for example, oae serving or 100 anaw per day), hair levels for women of
cbttd-beadag age abould ba moaJtorad for metbybnamny.
Tbe WHO/gCS euimeteU (1990) that a dally methytmemay intaica of 0.48/uf mercuryykg
body weight wffl. not cauae any advent aflaeu to adntti and that a oMthylmercur? tataka of 3 to 7
*tflk| body w«(bi/d«y would i»ult in a <5 pareant inenate in kb« iaddaaee of panatbeaia ia
adulta. Rok to tbu axtont would ba anociatcd witb bail aatcxtty eoooantratlon of appreodmatdy
50 to 125 *if |ran hair. By oompargon, tba US.BFA^NAraooideaa, or tba aacnmt of
aacUiylmgiuay any penon (laetuittng cbiMren indpre|Bant*vaaa)caninfe>tcve^daywttboiu.
•ham k ki body vajajat par day. CbjUran have a Ufbar cxpoaure to awtbyiaMzcury than
doadula. Tie UA EPA eatimttea in tbb lUport that intake of toocbyfanercuzy by cbildrra agea
Wrtb tbsougb U yean rtsidinf u> the contmeotal US.« 044 Mavk| body *oi|JttWay for mala and
O81 *»f/kf body watajhMay for f
ia fbe Ertmeted Stoa efthe fopvJatln at Matt
The people in tbe Ui who consume flab in axceei of 100 fnas/day (e.^ on
are oonaidered to be tbe populadoo at risk, in pertnimr, women of chad-bearing age and
children aged 14 and younger ace of concern becaoin of the ad>ma««fleca of medayimero
^f {jg^topint'aervotti syvwm. TVic acrvfcrtiLl U-** • -^»^^wiv • pi • IP • •^^ «^^^— '— • • ^ —™ ™ • - ni' —— -r—™ — » -
popoiacion of women in tbe United State* ware developed from tbe Untod State Departowat of
AfcfcttMnt oonflnuaif Surweya of fadWduai Food Oonavmpcion for tbe period 1989-1991
(CSPIX09/91). Tb»»uifyiy>«c»ii-rtt«hu*»LUJ^kuiotCT« three-day period. Demon
mercury eoaeentratioBt in nab wen avaOable from tbe Natioeal Macine FikbeoBa Service and
from publbhed iaforanlon on meroay ia ftetbwuer fbb. The»e eoureei have been u»ed in tha
Report to atHmaie luecLjUuen'iir) iatabe ftorn fidu
Oeyictahrr 20,1995 54 DRAFT: Do Not Qte or Quote
-------
iA/\«r VOM'cOXel
to
rv>eruJLrvj
biurd***, \JO\i>o^iuA
concentration for aMtytaercury. A doie-coovenion equation war
intake of 14ML**tini^~"—T*T body.waitWday. that when;
Tblbbd
\toesumite. a diity
^aj.tndMdual wvD
eoncenuaaoB. of 11
Dttt on the bebwior of mercwy is the humeo body were ueoi
of nwnuzy iageitad per d«y tt tnii ao advene effoet level Coaddcraiiooi i
way indmduak proeew netfaybnercuiy io the body cad the effect of Itek of i
effects were included la the RfD eekultrion. Tbe-JtfD tor methytmerenr
to be
ktofhaVber .
There are uncertainties in the RfD, which wore evaluated in qualitative
uncertainty intryses. These tnityias indicated that paresthesia (aumboeu «
>faat, and oocasiooafly around the aomh. hi adults is not
because it is wbjeet to the patient's raeofuooa of the <
ft TBCOaTQ^Dff IDfl CeleVeU^jRBav 0VvBVp UaW IMev '
r, that these were nihartmiat denelopaaeuril delays (for <
ehfld batef aubla to walk two steps without support at two yeen of age).
the physioloffe faeton which wave uwd in eatiraathif the infested aercuryj
the RJD for otftbyhsercury is a reasonable estimate aad b very I ' "
health.
proteodra of Ainneo
This dose-response estimate is supported by srtditinnal stodk* io children exposed in
Mere. Thee* hwhicle hweatipiions tmoag Oee Indian* b Caoeda %ad New reaianden
offish, to these studies the heir coocentrstion of mercery h vied to
reoverdme, Condnetaas by the hr*esti|Bon hi their offldal reports cite
rcaryconceBtrat
monitor
Currently a aoabcr of muarch nadies are
bt- -?>
th« ouaatioo of
"" studies
fc'the North
tadtheUasied
DSWSM Reftoy (ATSDR)). United teiute have become tvaihWe bom the SeycteUe Ishuden
botB;V9 soothers who e
One of the
quantitiei of Oteet Lakes
offish Mirtnt
taadks is of iafsnts wboee notbea comusaed tefe
'. IgJjeja^.l»3 preBateary raeoitt indicated a
BKXhan i«jM«l«d ftonaumprioa
at hiih levels of Oteat Lakes fish tamf preanancy. There ai«\ My eha«eefcm flan that have
toWcbikt eaimtopeMBt (•>», *C8»X so qaurioc mutt be amdaed to
i leiluisiloB of the JBcHft^r^ that mercury eaposurcs are causative to these results.
tVvc level of uunCBrtcuWy ^u->-rourvdL r/va Hvx
-------
'• l'-33PM
DtlficiDA OFFICE-
NCbH/UO-
-------
The RfD for neihytoercury • QUaflJ* of/kf body wcigbtWty or
weight/day. Data oa actfaylnercury comumptioa frost flih for women of chad-tearing ige (ages
15 through 44 yetn) • bow tb* 50th perecauife for adult female* to be 0.08 paykf body Wg
with 75th pereentils expmure to be 0.16 pg acrooy/kf body nwlfbsMay. Cnatequeatty
one half of the feoarai U.S. population of women of efcfldUbaari&f age who
oxuuae fifth at least once wtthia a 3-dey period axe predicted to infatt aetfeytaiarcury %t tb«
RID. These data are showa hi Tifurei 3-1, 3-2, and 3-3, which describe the total quantity of fifth
•ad ihellflth eaauaaad by ehSdran (afe» 14 y*an otWgMyjJ^feK^fflMR^f^0^**1^
aee (15 thsoafh 44 yean) who eat fish often eooua)roWlbetteB»a«ftfiA aodlbtaea in the
three-day dfetny ixtrway, CSFH 19/91. Becauae freftttwater flth may be of pertkwlar fageteat
arooad antfarapaaaoic aeunea, 7i|ur« 3-*, 3-5, and 34ahow the quaotity of bathwater fi«h
arooad antfarapaaaoic aeunea, 7i|ur« 3-*, -, and aow te quaotity o atwater «
comuoed by obUdrea (aaad 14 yean oc youaier) aod adult women of ehfld>bearia| aee (J5
through 44yean)wVw>a*«- *V3h af riftk of.
devetopmeatal delay* m infant whose motheo ecejume fto fteejuentiy duni^ pregnancy. Tbe
level of rUt •updated with cqnrtufat conuiapUoa of 100 pant or axve of Bab per day dapead*
oa the acrcurf coaceattaUon to the fifth enea. Beetvae panoat who am mmarmini Sab tt 100
fraoM per day or bifher auy be more tduat oa ttio ipaciBc aoviowi then to the caaeral
populatioa, the toUowtof anxnatet of die sita of the populatioa judaad to be tt hl|heft ri»k of
fiaethytaiereury enoiure are bated oa coanimprinci of fifth is the aaomt of 100 graav pet day or
aiore. <
I
The mnaber of waajea of ckftWsearinf aa* was detenniaed in the 1990 UJL cewua, Tac
ceoiuft etdmtta4 that the total fieaale popalatiaa afOK 15 throufh 44 yeua «e§ 9U32,000 m the
4Scoatifttowtiaie», The CSF3X fflffl reported 30J percent of wooea ftfat 15 throufh 44 yean
consume fitfa at taaK oace iD tbe 3-day period, or 17,731,000 woaatt who am fifth coeriman
ThJi doet not aeea that the o*er »5 peneoi of women avoid fi»h eoatiRaatly; father that fitb
dttootq>pww«»-diataiyhemduia|the3*dtyfur*B)rpariod. gtaeiumef «e«. coatiaMed flar
a longer period, p higher paneat of WMtaalft would Hkily coawme 6tb.
Baead oat the ravened quaatttfo of fifth comuned tod on the obierwttc* that 2 percent
of people turveyed neon &h coowoption two or awre tea* duriof tbo 3-day period, it WM
eittaated that 2 to 5 percent of ttepopolattoa of WOVMO who eoojaaw fifth a«t, oo tvertje, 100
• I»»M fit 4>ak *^» <4a« fltia HMiBMB *^M» >HMU^M1 14S1UO m^A
fm» of fifth par day. Th» meam that becw-o J5S/W) «ad S8MOO women of chfleVbeatiBf aae
anatrlak. Pia^Qc heahh nathdcs i&dicate that M peneat of woajtn in the aae gronp 15
through4^ye«n(«TepreIMntfe*|tr«>*ar. TW aaaaber of prefnant womea coo«aai»| fi^> in
the quantity of 100 frao* per day or man become* appnateately §4,000 oc 34,000 women for
the 5 peroaat and 2 paneot erfmam, laapaetiveiy, wbea thece eattautta tte wwded to Oe
oeareftt 1,000. Tbcae data aad ettteetav are wmaarixed la Table 3-2.
leptccnbot 20,1995 3-7 DRAFT; Do Not Cfte or Quote
-------
i e«»ll-J1'1
:I : 26-91-01:
vniav.io iwac
S IfGS
-------
-------
'T*
-------
O-
5
s!
«m^«^
o<»•<•*»• «PF•
pM 'fuel MR) •!
pwaiBd a jwhfn to )imn>N|iia
-------
SENT RY'.NIEHS ;10-16-9S ' 1 = 42W = *t™ESDA OFFICE-
jC-il ui .iiuiiy ,.,»rnni«r 'MI-H -in_ii-Ui t irnuuu : ufku/n
The State of Hiwid aad Aittia were oot indudfid in the above estimates because CSFn
89/91 did oot include these state*. Because of their lubnaatial coutat area* nsh-cootumpaoa
pattern* ia Alaska ud Hawaii we predicted to dUSw from dtose of the 48 ooatifuous state*. The
aumber of wooes of chfld-beartag afe (IS through 44 yean) in Alaska fc estimated to be
approxiaaaaiy 138,000 tod ia Hawai is estimated to be ipprcDomateiy 284,000 peraoaa. The
percentage of pmfuat women for afet 15 throufh 44 yean ia Aluka was 9.7 peneat and in
Hawaii toia percatttaje v*» atao 9.7 percent.
Quldrea afed 14 yatn aad youager tra aao a potaatiaUy at ruk population for
raathyiaieraay exeeaure fean flih. Tbe popoiadoa of eafldraa aavd 14 yean or youafor ia tne 48
oootifuovs ttate» is 53.463,000 with aa addidoaai 151,000 ia Ataka aad 240000 ia BawaiL At
•hown ia Table 3-2, data froaj CSFD, 9SH91 idaatified 25 pareaat of ehDdrea as *"r»'f<*| fi$h at
least once during the >-day «urvey period. Baaed oa khaie cadBUtoi appnaajaately 13306,000
caiWreo afai 14 aad youafcr capaaxly ooatuaw fiab, aad tbe oumber of children eoanuaiiig fiih
ia the qutatity of 100 frtaa par day or moie beoocaa appronaacely 665,000.
Wbeo awtbyavBccocy iataka wet eaproned oa a per kflofran body «eiaht becia, children'*
•xponre wac approatiaattr? 2-3 tana* that of aduhs. Ia tbe awthytaeteaiy potfoaiDi epideraJoi
ia Jepan «nd Inq ehfldvea were •ffacted as well at adult*. Tbafe cflectt were oot seen only ia
children exposed to BMthybaecaay an uem bat iaduded chfldiea «poMd taroufh iagestiag
taethytioemiry from food. Whether or act ehfldrea differ ten «duk* ia aow much
metnyUuercury exporare they tolerate wfchaat thowiaf oeurotefigal efiects Is not kaowa.
Because the aervous lyffeai is oot ft»fly mature uatfl late ia oafldBood aad be
because neUrylaercury
atfvenely dfaca the deveJoptaf oarvou ryuem, it appeao hnprudrm to if&we the pofsfbifity that
advene tteuretotfcal eftos ooaur 'm eaflttoea at toubjecatodctBTBB^gth»wBBeat»eteofi.ufld biBiiniaga. Becaase auay womao aad
obikina in thaae aoins any obtaai a avbitaaaal pordoe of their ete froai a hahad fsofrapaic
reakn or from oaJy a lew tpeatea of fl»h, the HkeUbood that they wifl exeaed the taencaoiazk dose
for methyaaexeury depends oa the eoaceatratfoB of a^hylawreaiy a» the £afc they cootnaied,
TOBOT parteular frooet of «OBMB aad children «ouM beoeflt from a BMTB spacalc evataatiOB.
Aa enapJe of MCB aweiaadoa, «oukt \» Wotoffcal awattarieff based oa Byefeuvy uaiyaea of hair
or blood.
September 20,1993 MS DRAFT: Do Not ate or Quote
-------
SE>T SY'-NIEHS
; 10-16-95 j 1.40PM : . BE1HESDA OFFICE-
• '22
-------
:'s E:A ECAO ::NC;NNAT:
.; ID
D1?ARTMINT OF HIXLTH 4 HUMAN SERVICES pubfc M-w, t^rio.
Memorandum
Date October 24, 1995
From Dr. Michael Bolg«r
Subject EPA Mercury Report to Cengreu
To Dr. Kate M«ha&y
Dr. Sheila Newton
Additional comment to:
EPA* i response to FDA'i October 13 coanment* on their dnlt reyon.
Pag« 3-5, ptrtgnpb 5, with rmyarA to nwulyiU of th» Niigatt poisoning:
The authon' dedtlon not to indude individuals with leveli bdow 20 ppm was based
oa two v«iy good reasons; Pint, the lade of reliability of the analytical technique
bdow 20 ppm. The detection Itaii is oat a measure of reliability and the authon
conclusion on mis Issue has to be taken seriously (not that of someone thousands of
raflea away and who wu oot involved in tbt study). Second, levels bdow 20 ppm
reflected baciqpouad ta*eU offish eonsuiBptieb The lew eed of th« raa^ u thar^bre
higtfy icpcodeut ga tbtw two iuuci, and fbt JncluioiirfinjivWualiwidilcveJs
below 2O ppm results in a iraileadiBg interpretstion of the threshold value; The data
support the tuthon' dctetmiwtUm of the threshold values of 4Q to 70 ppm.
-------
10-16-35 ; 1-.43PM '- ixTHESDA OFFICE-
A» the above loaiyiii demonstrates, tome source typei do not deposit as ouch mercury
near tba facility. Tbe localized impact! from aay oae type of source are highly dependent OQ tbe
•mount of mercury nnrittnrt, the height ud velocity of tbe emisaioas. and the chemical form of
tbe corned mercury. The tang-range transport modeling taaiytit that was performed for this
Report uicased the cumulative impact of til souitet that eatit mercury. lie neutts, fflunrated m
Figure 3-t, showed that mercury deposition from satbiopogeaic sources • ubJqatotn aaoai tbe
UJ5* with land area* eest of the Rocky Mountains beiaf the aott itr^- daax are cyding between
the atmojpbcre, laad aad water. There is a latfe ttaervott- of aezcury already to the eavbomaeat
that ariiti ia the global atmosphere aod tediaeats. Global modeUa| of tbe oereuiy cycle sbowt
that aBtaropogeaie nniiMiom rival ox exceed thove coal oatozai sources, Aat&rcpogcmc
ftmin are believed to ooatribme about SO to 75 percent of tne current total anmai input of
mercury to the atawapbere. Aa etajaau of tbe total a&aual iaput from •atbtopogeak; taincea
3632 Mf (4000 ton). Tbe US. jouroai aeeooat &r about 6 percent of tba global
aad Ifleewiee, global emtsiiuus attribute to rteumftkin IB tbe US. Use kmff-nmg* tsaatpon
aaalysai predicted that of the 220 Mg (243 ton) of US, tOMBCM that ware modeled, 7$ Mg (86
mm), or 35 percent, were depodTBd io the cnortoffttaH 1LS. wbfla 143 Mg (160 teas) or 63
pegoaat, wire ttauiported ouokie of the cooaaeatal U«S. The modenag tftnulatbtt alto predicted
that 3? Mf (36 COBS) are deposited aaavaDy m the U& flrom the global atmospheric reservoir. t ,
•^-
Tbe «oatrfbutioa of taexeury to tbe environment from global sources, natural sauces and Q ^, /
fitirlnm rnniitlniM rmnpHrstei an asarsimcpt of The piihtfr haattfi awl •nvirmiinmTsI Imptrr nf 5 ? v
CUTTMU U.S. emioioo*. ^. ^ p
la tbe r>eauai GeMaf Worn or
There an oo aattoaal luet-ieiiu monitoring data for either air oc btota cvajlabto to
address tfaj* question. Tbe 1992 3a»e> of Cfumkal Radtut ai fltk eoadowed by tbe V.S. BPA
saoold provide baseaae data for fish Ie«eto aadoaaQy. A number of studies have deoooitratad
that local daposttion fa cettam areas (og, ia Mfrmeanrs) has deeUaed ia recent yean. However,
tba? local or tesjtocal rteclfne baa aot beao desBoaatrated oa a gtobal scale (wbaie mercury
S^HHB«^^fi*-k^BB h ^k« eajaA ^^a^ae^a^^a^Msn
ej ayysjn ivu Us* Q9* QVGIeMeVVII^e
Meroury eM§e, however, ha* dedfaed by eboai 75 peneat ia the UJL tnt* ita peak in tbe
eariy 1960*1. Tbfi dedsae caa be aarfinuad to tbe raductioo of mercury uw la a number of
iadiutiial applicanoos, pafticahriy tbe removal of aweary firom pea*. Current efibrti to reduce
ncrouiy m other preduea, partJauhaty bettehea, lead to the prediction that mercury see In the
maoafkctariag leeton wfl] cominue to deeaae slowly. DepoeWoa of mercury io the U.S. is also
todeeHa«iftb«gwc»yea»iaionlBaiafer aeiiril aad matadpalweateeombvawra are
September 20,1993 J-20 DRAFT: Do Not Ote or Quote
-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
\AT.O\AL CENTER -CR ENVIRONMENTAL ASSESSMENT
3;.NC'NNAT! OH 45258
MEMORANDUM
DATE:
SUBJECT:
FROM:
TO:
OFFICE OF
RESEARCH AMD DEVELOPMENT
October 20, 1995
Revisions to the Mercury Study Report to Congress
Rita Schoeny,
Associate Director
Members of EHPC Pane! on Mercury
Attached are materials in response to the advice of the panel. These are notes and
responses to comments from the panel, responses to comments from FDA and pages of
Volume I which have been revised. To save resources I have included only the revised
pages. I'll be happy to send you a complete revised Volume I after we have received it from
the document preparation contractor.
A copy of the revised graphs is not yet available in a form which would transmit
legibly by fax. The changes include revising the figure titles (as shown on attached p vi) and
bracketing the RFD with lines to indicate the range of uncertainty. On the attached pages,
shading indicates revised text Deleted text has been struck out Original page numbers
are hand written on the bottom of the page.
I hope these revisions bring our business to a successful conclusion. Please contact
Sheila Newton or myself if you require further discussion. Thank you for your careful and
timely advice on this highly visible and significant matter. Your effort is much appreciated by
all of us involved with the Mercury Study Report to Congress.
Attachments
cc: Sheila Newton (2131 Switzer Bldg)
Richard Jackson (MSF29)
Robert Goyer(WC-05)
Richard Levine (Rm 7803)
George lutier (MSA3-02)
Thomas Sinks (F29)
William H. Fariand (8601)
Terry Harvey (MS711)
Martha Keating (MD13)
KateMahaffey(MS117)
Rita Schoeny (MS 117)
JeanetteWiltse(8601)
I. Kaye Wachsmuth (HFS-2)
-------
RESPONSE TO PANEL'S COMMENTS ON
VOLUME 1, MERCURY STUDY REPORT TO CONGRESS
The following describe EPA's response to suggestions offered on the Summary
Volume (Volume 1) by the panel convened by the Environmental Health Policy
Committee on EPA's "Mercury Study Report to Congress". These refer to materials
send by Sheila Newton on October 16, 1995. These responses were prepared by
Kathryn Mahaffey with contributions by Rita Schoeny, Glenn Rice, and Martha Keating.
The comments sent by the Food and Drug Administration have been sent
separately and will receive a separate response.
Response to FAX of October 16. 1995 from Sheila Newton
pg. vi. Suggestion for modification of the title of all figures showing estimated
methylmercury exposures from fish and shell fish, etc. Although the legend to the figure
indicates that consumption data are for persons reporting fish consumption at least once
during a three-day survey period, the panel suggested clarifying this further by modifying
the title of the figures. Consequently EPA will add the following words to all the titles:
". . .Who Reported Consuming Fish and Shellfish at Least Once during a Three-
Day Survey"
The specific words used in the underlined portion of the addition will vary with the
graph and by Freshwater Fish. Fish and Shellfish, etc.
pg. 2-1. last line. We will modify "entirely understood at this time" to read "well
understood at this time."
pg. 2-2. These are FDA's comments that are addressed in a separate section.
pg. 3-3. At the end of the first paragraph, we will add the sentence: "In both the Iraqi
and Japanese epidemics, the levels of methylmercury consumed were much higher than
the levels currently reported in the U.S. food supply."
pg. 3-3, fourth paragraph. We will modify this to read: U.S. EPA calculated an RfD
using data reported on neurological changes in 81 Iraqi children who had been exposed
in utero: their mothers had eaten the contaminated bread during pregnancy. The data
were collected by interviewing the mothers of the children and by clinical examination
by pediatric neurologists approximately 30 months after the poisoning episode. The
incidence of several endpoints (including late walking, late talking, seizures or delayed
mental development, and scores on clinical tests of nervous system function) were
mathematically modelled to determine a mercury level in hair (measured in att the
mothers in the study) which was associated with no adverse effects. These effects were
-------
uncertainty surrounding the RfD."
pg. 3-6. third full paragraph should read: "The general United States population that
obtain their fish by purchasing them commercially is not believed to be adversely
affected by methylmercury if they consume fish in moderation. The people in the U.S.
who routinely consume fish in excess of 100 grams/day (e.g., one serving) are
considered to be a population potentially at risk. In particular, women of child-bearing
age are considered of concern because of the adverse effects of methylmercury on the
developing nervous system. Children ages 14 and younger are of concern because they
have higher exposures to methylmercury from fish on a per kilogram body weight basis
compared with adults."
pg. 3-7, first paragraph, additions to the last line of this paragraph: "through 44 years)
who ate fish once or more during the three-day dietary survey period."
Figures 3-1 through 3-6. Figure titles will be modified to include the words "Who
Reported Consuming Fish and Shellfish at Least Once during the Three-day Survey
Period"
The words underlined will vary with the graph and describe the type of fish consumed;
e.g., freshwater fish, fish and shellfish, etc.
pg. 3-15, fourth paragraph. The last line will read: "or Native Americans) remains an
issue."
pg. 3-20, the fourth paragraph will be modified to read:
"The contribution of mercury to the environment from global sources, natural
sources, and previous emissions complicates an assessment of the public health and
environmental impact of current U.S. emissions. Conclusions from the modelling of fate
and transport of anthropogenic emissions of mercury from industries identified in the
Clean Air Act Amendments indicate that there is a plausible link between increases in
mercury emissions from these sources and increases in mercury concentrations in fish.
Modelling-based efforts to quantitatively assess the magnitude of this increase cannot
be fully evaluated because of limitations on monitoring data providing mercury
concentrations in the aquatic food chain. The weight of evidence supports an
association between local mercury emissions and increasing mercury levels in locally
caught evidence. For fish collected at sites remote from the source, at this time it is not
possible to adequately source-apportion mercury present in these fish."
-------
delays n motor and language development defined by the:
Inability to walk two steps without support by 2 years of age;
Inability to respond to simple verbal communication by age 2 years among
children with good hearing;
Scores on physical examination by a neurologist that assessed cranial
nerve signs, speech, involuntary movements, limb tone, strength, deep
tendon reflexes, plantar responses, coordination, dexterity, primitive
reflexes, sensation, posture, and ability to sit, stand, walk, and run;
Assessment of mental development or the presence of seizures based on
interviews with the child's mother."
pg. 3-3: EPA will add an explanation in lay terms regarding the computation of the
LOAEL, NOAEL, and RfD and what they mean. It will be emphasized that the RfD
represents a dose of methvlmercurv that can be consumed for a prolonged period of
time without recognized adverse effects: this dose is the midpoint in an estimated range
that covers approximately an order of magnitude.
pg. 3-4. At the end of the first full paragraph, EPA will add the sentence, "The RfD is
a safe level, not a hazardous level; it is the midpoint in an estimated range of about an
order of magnitude. This range reflects variability and uncertainty in the estimate."
pg. 3-4, within the second full paragraph, the fifth full line will now read: "are also
difficulties with reliability in recording and classifying events like late walking in children,
especially since the data were collected approximately 30 months after the child's birth.
pg. 3-4, fourth full paragraph: "Currently a number of research studies are underway
that further address the question of what exposures to methylmercury in fish are
associated with neurological disease. These studies include more subjects than did the
Iraqi study, are prospective in design, and utilize endpoints that are anticipated to be
more sensitive than the clinical signs and symptoms of methylmercury poisoning
observed in Iraq. These studies are conducted in the Seychelle Islands in the Indian
Ocean, the Faroe Islands in the North Atlantic Ocean, and in the United States
[sponsored by the Agency for Toxic Substances and Disease Registry (ATSDR)]. Data
from these studies, when available, should be useful in decreasing the uncertainty
surrounding both the benchmark dose and the RfD. The Food and Drug Administration
has determined that revisions in the FDA action level for mercury concentrations of fish
in interstate commerce should wart until the new studies have reduced the level of
-------
in the Swedish Environment. In the document Relative Atmospheric Loadings of Toxic
Contaminants and Nitrogen to the Great waters by Joe! Baker et al., it is asserted that
"Atmospheric deposition dominates the flux of mercury to lacustrine systems and the
open ocean, and it appears that modest increases in atmospheric loading could lead
directly to enhanced levels of mercury in biota." This conclusion is also supported by
others including Sorenson et al., in their 1990 journal article entitled "Airborne Mercury
Deposition and Watershed Characteristics in Relation to Mercury Concentrations in
Water, Sediments, Plankton, and Fish of Eighty Northern Minnesota Lakes", and by
Fitzgerald (1990) in "Mercury as a Global Pollutant." (Natural Science, 1993).
pg. 3-1, question 3. EPA and FDA have worked out the following joint language that
indicates FDA's description of how FDA arrived at their mercury levels for fish.
Overview of FDA's approach to methylmercury in fish.
"In 1969, in response to the poisonings in Minamata Bay and Niigata, Japan, the U.S.
FDA proposed an administrative guideline of 0.5 ppm for mercury in fish and shellfish
moving in interstate commerce. This limit was converted to an action level in 1974
(Federal Register 39, 42738, December 6. 1974) and increased to 1.0 ppm in 1979
(Federal Register 44. 3990, January 19, 1979) in recognition that exposure to mercury
was less than originally considered. In 1984, the 1.0 ppm action level was converted
from a mercury standard to one based on methylmercury (Federal Register 49,
November 19, 1984).
The action level takes into consideration the tolerable daily intake (TDI) for
methylmercury, as well as information on seafood consumption and associated exposure
to methylmercury. The TDI is the amount of methylmercury that can be consumed daily
over a long period of time with a reasonable certainty of no harm. FDA (and WHO)
established a TDI based on a weekly tolerance of 0.3 mg of total mercury per person,
of which no more than 0.2 mg should be present as methylmercury. These amounts are
equivalent to 5 and 3.3 ug, respectively, per kilogram of body weight. Using the values
for methylmercury, this tolerable level would correspond to approximately 230 ug/week
for a 70 kg person or 33 ug/person/day. The TDI was calculated from data developed
in part by Swedish studies of Japanese individuals poisoned in the episode of Niigata
which resulted from the consumption of contaminated fish and shellfish and the
consideration of other studies of fish-eating populations.
Based on observations from the poisoning event later in Iraq, FDA has acknowledged
that the fetus may be more sensitive than adults to the effects of mercury (Federal
Register 44: 3990, January 19, 1979; Cordle and Tollefson, 1984, FDA Consumer,
September, 1994). In recognition of these concerns, FDA has provided advice to
pregnant women and women of child-bearing age to limit their consumption of fish known
-------
RESPONSE TO FDA'S RECOMMENDED REVISIONS TO THE
EXECUTIVE SUMMARY OF THE EPA MERCURY REPORT
The following are EPA's response to FDA's comments sent by FAX on October 13
1995.
pg. 2-1, last line. EPA agrees that the words "well understood" can replace the words
"entirely understood" in this line.
pg. 2-2. EPA's view is that the section describing warnings against fish consumption
through fish advisories simply reflect the fact that 40 states have issued warnings against
fish consumption based on mercury toxicity and concentrations of mercury present in
fish. This section should include language making it very clear that the level of risk is
based on the concentration of mercury in fish combined with the quantity of fish
consumed. These are major part of predicting body stores of mercury in groups of
people. In addition, group-to-group variability in these quantitative relationships are
acknowledged and are part of the variability and uncertainty expressed in developing the
reference values for mercury.
The presence of a plausible link between ambient, anthropogenic emissions of
mercury and concentrations of mercury in fish is a major finding of this Mercury Study
Report to Congress. This conclusion has been extensively peer-review.
The EPA is willing to delete the bold heading that introduces this section,
however, we prefer to keep the information intact. The purpose of this information is to
convey the value of the fisheries industry and the potential impacts should mercury
levels in fish increase as a result of continuing deposition of mercury. It is not untrue to
state that 40 states have fish advisories for mercury. Many of these states use the FDA
action level of 1 ppm as a trigger for their own advisories and stili find their freshwater
fish with mercury levels sufficient to trigger an advisory. For example, a recent study of
16 Adirondack lakes found that 26 of the yellow perch sampled (3 to 5 year-old fish) had
mercury levels which exceeded 0.5 ug/g and 7 percent exceeded 1 ug/g. At least one
perch exceeding 0.5 ug/g was found in 9 of the 16 lakes. AS this one example
illustrates, the mercury levels the states are measuring are not insignificant.
As to the second point about the lack of evidence that mercury emissions
contribute to elevated fish mercury levels, there are numerous confirming studies that
support the major conclusions of the "Mercury Study Report to Congress" that
anthropogenic emissions of mercury can indeed be linked to elevated levels in fish. The
importance of atmospheric mercury deposition and its impact on biota has been
demonstrated by two other major mercury investigations: 1) The Mercury in Temperate
Lakes Program in Wisconsin, and 2) a broad-based investigation in Sweden, Mercury
-------
methylmercury intake had been reduced from these high levels (See M. Harada, Critical
Reviews in Toxicology 25:1-24, 1995). It is unclear whether or not the occurrence of
more subtle cases having less severe neurological damage were a chronic type of
Minamata disease or reflect more subtle changes associated with lower exposures to
methylmercury.
Consequently we have reservations about using the terms "high" concentrations
of methylmercury and "high" intakes of fish. These "high" intakes are relevant to a small
percentage of the U.S. population. Fish intakes of 200 to 300 grams per day are
reported as 90th percentiie consumption figures in some of the surveys of native
American tribes in the United States.
page 3-4, second paragraph. EPA accepts FDA's suggestion to rewrite the second
sentence as follows: "Due to variability in the way individual process methylmercury in
the body and the lack of data on reproductive effects, an uncertainty factor of 10 was
used to derive the RfD from the benchmark dose."
*
page 3-4, third paragraph. EPA thinks that quantitative expression of the uncertainties
is better described as a "range of uncertainty of approximately an order of magnitude
around the RfD" than as "There are significant uncertainties in the RfD."
page 3-4, fourth paragraph. The New Zealand study reported by Kjellstrom et at. was
not the basis of the RfD. It was used only as confirming data. This report by Kjellstrom
et al. is available through public channels as an official report from the Swedish
government Dr. Kjellstrom has been contacted regarding peer-review of his document.
US EPA has also had these data reviewed by epidemiologists and statisticians who
regard its findings as defensible.
page 3-4, last paragraph. EPA agrees that it will be adequate to consider the Faroe
Islands study, the Seychelle Island's study, and the ATSDR-sponsored studies in the
Great Lakes region together under the broad statement:
"Data from these studies, when available, should be useful in decreasing the uncertainty
surrounding both the benchmark and the RfD."
page 3-5, Heading. "How do U.S. EPA's estimates of how much methylmercury is
harmful to humans compare with levels calculated by the U.S. Food and Drug
Administration, WHO, and the states?"
EPA will look at the FDA overview for details regarding revised wording in this section.
EPA will provide more detail on what the states include in their fish-consumption
advisories. EPA can reference the NAS/NRC's volume seafood Safety and a recent
-------
to have high levels of mercury (FDA Consumer. 1994). FDA believes, however, that
given existing paatterns of fish consumption that few women eating such high mercury
fish (less than 1%) will experience slight reductions in the margin of safety. However,
due to the uncertainties associated with the Iraqi study, FDA has chosen not to use the
Iraqi study as a basis for revising its action level. Instead, the FDA has chosen to wait
for findings of prospective studies of fish-eating populations in the Seychelle Islands and
in the Faroe Islands.
pg. 3-1. "What effects....", second paragraphs. EPA will use the wording: "there were
two major epidemics involving industrially contaminated fish." This wording does not
detract from epidemics that involved fewer persons than were identified in Niigata and
Minamata.
pg. 3-2, second paragraph.
EPA has reservations about using vague terms such as "large amounts of fish". Review
of these data indicate that consumption probably exceeded 200 to 300 grams per day,
but this is poorly documented in the literature with which we are familiar. Harada
(Critical Reviews in Toxicology, 1995) cited an average intake of approximately 336
grams of fish per day for a fisherman (pg. 16).
pg. 3-2, fourth paragraph. This paragraph represents the same issue of use of imprecise
terms. We would prefer statements that include estimates of the mercury concentration
in parts per million. Otherwise we find the statement "contaminated with methylmercury"
to be more descriptive of the situation.
pg. 3-2, third paragraph. We agree to delete "from fish" from line 4. It is replaced with
"from food".
Although the comment continues with "From line 4 ad down the page", checking
the report suggests that FDA is referring to page 3-3. We agree with FDA that the
consumption in Minamata and Niigata was long-term (e.g., decades) in duration.
Looking at data on mercury concentrations in fish indicates that in early years of the
epidemic fish contained mercury in excess of 10 parts per million wet weight. After
approximately 1965 to 1966 there was a substantial decline in the mercury
concentrations in fish to concentrations under 1 part per million in Minamata and Niigata
(please see pages 44 to 50; pgs. 82 to 86, and page 263, in Tsubaki and Irukayama's
volume entitled Minamata Disease).
The most severe cases of neurological damage (Minamata disease) occurred
early in the epidemic when fish with mercury concentrations over 10 ppm were frequent.
However, subtle cases of neurological damage appeared as long as 5 years after
6
-------
Table 3-1
WHO Data on Mercury in Hair
Fish Consumption Frequency
No unusual mercury exoosure*
Less than one fish meal per month
Fish meals twice a month
One fish meal a week
One fish meal each day
Average Mercury Concentration in
Hair (ug mercury per g of hair)
2
1.4 (range 0.1 to 6.2)-
1.9 (range 0.2 to 9.2)-
2.5 (range 0.2 to 18.2)-
11 6 (range 3.6 to 24.0)-
General estimate presented in the Criteria Document on Methylmercurv
(WHO, 1990).
from Airey, 1983
9
-------
summary of advisories prepared by EPA.
Differences in approach between EPA and FDA are noted. In particular the FDA levels
are based on effects in adults. The EPA reference dose is based on developmental
effects in infants whose mothers experienced methylmercury exposure.
page 3-5, third paragraph. EPA will add the following wording to this paragraph.
"Cross-comparison of WHO recommendations regarding risk associated with hair
mercury concentrations is facilitated by data reported in the WHO Environmental Criteria
Document Methylmercury on mercury concentrations in 559 samples of human head hair
from 32 locations in 13 countries. The WHO report found that mercury concentrations
in hair increased with increasing frequency of fish consumption (see Table 3-1)."
Unpublished data (submitted to U.S. EPA by the U.S. Food and Drug Administration)
from the early 1980s on a group of United States 1431 women of child-bearing age
indicate hair mercury concentrations of 0.48 ppm for the overall sample of 1431 women
and a mean concentration of 0.52 ppm for the 1009 women who reported consuming
some seafood. These mercury concentrations correspond to hair mercury
concentrations associated with fish consumption at the level of one meal per month to
one meal per week based on data shown in Table 3-1.
The small magnitude of the increase in hair mercury concentration suggested in these
unpublished data for those women reporting seafood consumption differs from the
patterns observed in most surveys identified in the world literature on hair mercury
concentrations. This difference between these unpublished data and most literature
underscores the need for survey data on hair or blood mercury concentrations from a
representative sample of the United States population to estimate body burden of
mercury in the general United States population.
8
-------
estimating the magnitude of exposure to methylmercury.
The discussion about linking the risk of 100 grams/day consumption of fish with the 5%
risk of neurologic disorder is interpretation. These recommendations are presented in
different places in the IPCS/WHO 1990 volume. It is not entirely clear that the 100 gram
number refers only to a 5% risk.
EPA also does not accept the idea that "very few women in the U.S. are likely to be
exposed to mercury at levels associated with increased risk to the fetus." EPA agrees
that the percent of women with mercury exposures producing risk to the fetus is small
(e.g., < 5% of the population reporting fish consumption), but this does not translate to
a small number of women. Our estimates are that the number is approximately 80,000
to 85,000 maternal/fetal pairs.
page 3-15, paragraph 3. EPA agrees that it will clarify this statement to indicate that
children are at greater risk because of higher exposure to methylmercury compared with
adult exposures. Children do develop methylmercury intoxication. It is not known if the
dose-response curve for young children differs from that of the adult. Regarding the
question of whether or not it is reasonable to assume that the developmental endpoint
applies to young children: the effect is on the developing visual-motor system. This
system is not fully mature until approximately 4 years of age. EPA has followed the
research indicating that methylmercury has adverse affects on broad areas of CNS
function and cellular metabolism. Consequently the idea that methylmercury produces
effects only with in utero exposures is at variance with the data.
page 3-18, third paragraph. The issue of hair mercury has been dealt with above.
page 3-18, fourth paragraph. All of the issues raised by FDA in this comment have
already been discussed by EPA in the Risk Characterization Volume (Volume VI). EPA
finds that for the Summary Volume, it is adequate to acknowledge the variability and
uncertainty in the RfD. The reader interested in the basis for this uncertainty and
variability is urged to consult this volume.
11.
-------
page 3-5. paragraph 5.
This sentence refers to recent reanalysis of the Niigata episode. The estimated range
of 40 to 70 ppm hair mercury is from the abstract. Careful reading of the body of the
manuscript and the tables supports the EPA interpretation of this paper. Specifically
FDA interpreted the paper to mean that hair mercury concentrations under 20 ppm were
not considered reliable because of the analytical technique below 20 ppm. This is not
the situation. The analytical method has a detection limit of 0.5 to 3.0 ppm. The
investigators (Kinjo et al., 1995) note that the limitation of 20 ppm is based on the
"normal range of mercury concentration reported to be under 20 ppm for Japanese
people on a common fish diet." This is not a reflection of the analytical method, but of
was is assumed to be "background" from a "normal fish diet". The investigators
acknowledge on page 97 that "Our exclusion of certain subjects with a low hair mercury
concentration and the hair length determination as stated above were somewhat
arbitrarv: we examined the effects of variation of these parameters on the threshold dose
estimate....The exclusion of subjects under 20 pom of maximum hair mercury
concentration resulted in only slight alteration of the threshold value for males, but a
remarkable alteration for females. Thus, the threshold value was chances from 24.7 to
49.3 ppm for females." Details of these modelling results are presented in Table 2 on
page 97. These data do not support limiting threshold values to 40 to 70 ppm as implied
in the abstract.
pages 3-6 through 3-15. "What is the estimated size of the population at risk??
EPA will address some additional comments to the issue of ambient mercury as it
contributes to marine and fresh-water fish species. In the Mercury Study Report to
Congress, the EPA will continue to rely on 3-day fish consumption data for the general
population to predict exposures for methylmercury. For a developmental toxin such as
methylmercury short-term exposures are considered the relevant period of exposure
rather than long-term exposure patterns. Longitudinal data on fish consumption have
their own problems that EPA has identified in Appendix H of Volume III of the Mercury
Study Report to Congress. Problems with longitudinal data include subjects modification
of diet in response to recording these data. Interestingly this was noted in the Project
III report by Smith et at. supplied by FDA. The fourth summary point from the
investigators indicates that "reported dietary intake declines systematically throughout
the reporting year, being approximately 30% lower at the end than at the beginning of
the year. This may represent a systematic under-reporting of seafood consumption of
the panel as a whole."
EPA will refer to the need for population-based data on blood or hair mercury
concentrations. Such data would, help to provide a "cross-check" on other methods of
10
-------
MERCURY STUDY REPORT TO CONGRESS
VOLUME I:
FINDINGS AND RECOMMENDED ACTIONS
DRAFT
October 20, 1995
Office of Air Quality Planning and Standards
and
Office of Research and Development
U.S. Environmental Protection Agency
-------
• identification of the types of human health consequences of mercury exposure and the
amount of exposure likely to result in adverse effects:
• evaluation of mercury exposure consequences for ecosystems and for non-human
species:
• identification of populations especially at risk from mercury exposure (by virtue of
innate sensitivity or due to increased exposure);
• efficiencies and costs of control technologies; and
• the costs to society on account of mercury contamination.
The Report used the above types of information to assess the impact of emissions to air of
mercury from a variety of sources. This assessment included judgements as to the potential hazard to
humans and wildlife of methylmercury exposure which (as is described in succeeding sections) is
largely through the consumption of contaminated fish. There was no attempt in this Report to do a
risk benefit analysis of fish as an important source of protein and calories in the diet of U.S.
populations. Such an analysis would be*beyond the scope of the CAAA mandate. As emphasized in
succeeding sections, it is feft that generally the U.S. fish supply is safe at typical levels of
consumption. .
The various analyses documented in this Repon were designed and conducted in accordance
with accepted guidelines and procedures. For example, the human health risk assessment performed
for this Repon follows published Guidelines for Risk Assessment (including guidelines on Exposure
Assessment, Developmental Toxicity, Carcinogenidty and Germ Cell Mutagenicity) and uses
established methodologies for quantitative assessment of general systemic toxicity (e.g., in the
calculation of reference doses (RfDs) and reference concentrations (RfCs)). Moreover, the assessment
of ecological effects, presented in Volume V, follows U.S. EPA's Framework for Ecological Risk
Assessment. Criteria values for protection of piscivorous wildlife were developed using the
methodology developed for the Great Lakes Water Quality Initiative.
In 1994, the National Research Council of the National Academy of Sciences, in Science and
Judgment in Risk Assessment, recommended several areas in which U.S. EPA could improve its risk
assessment and risk characterization practices. These recommendations are listed below along with a
description of how they were implemented in this Repon.
• Provide an understanding of the type and magnitude of an adverse effect that a
specific chemical or emission could cause under particular circumstances. The Repon
characterizes both the type and magnitude of health and ecological effects associated
with airborne emissions of mercury from anthropogenic sources.
• Validate methods and models. All models used for the Repon were critiqued by
scientific experts and model predictions were compared to measured mercury levels
using the most appropriate data available.
• Describe the basis for default options. All assumptions are described and justified
based on available data. Where appropriate, exposure models were modified to
improve assumptions and to focus on areas of prediction wherein use of model
assumptions is most justified.
October 20, 1995 1-2 DRAFT: Do Not Cite or Quote
-------
LIST OF TABLES AND FIGURES
Page
Tables
3-1 WHO Data on Mercury in Hair . . 3.5
3-2 Estimated Population in the United States that Consumes Fish. Excluding Alaska and
Hawaii . 3.^4
3-3 Wildlife Criteria for Mercury 3-17
3-4 Mercury Sources With Sufficient Data to Estimate National Emissions 3-19
3-5 Source Categories and Pollution Prevention/Material Separation Opportunities . .... 3-23
3-6 Potential Mercury Emission Reductions and Costs for Selected Source Categories . 3-25
Figures
3-1 Estimated Exposure to Methylmercury from Fish and Shellfish: U.S. Women of
Child-Bearing Age Who Reported Consuming Fish and Shellfish, at Least Once
Owing a Three-Day Survey 3-8
3-2 Estimated Exposure to Methylmercury from Fish and Shellfish: U.S. Male Children
Who Reported Consuming Fish and Shellfish at Least Once During a Three-Day
Survey 3-9
3-3 Estimated Exposure to Methylmercury from Fish and Shellfish: U.S. Female
Children Who Reported Consuming Fisb and Shellfish at Least Once Daring a Three-
Day Survey 3-ro
3-4 Estimated Exposure to Methylmercury from Freshwater Fish: U.S. Women of Child-
Bearing Age Who Reported Consuming Freshwater Fish at Least Once During a
Three Day Survey '. 3-11
3-5 Estimated Exposure to Meihylmercury from Freshwater Fish: U.S. Male Children
Who Reported Consuming Freshwater Ffeh at Least Once During a Three Day Survey 3-12
3-6 Estimated Exposure to Methylmercury from Freshwater Fish: U.S. Female Children .3-13
3-7 Total Mercury Wet+Dry Deposition (Base Case Analysis) 3-21
October 20, 1995 vi DRAFT: Do Not Cite or Quote
-------
.Vro There \Varningq Against Fish Consumption in the United States?
Current levels of mercury in freshwater fish in the United States are such that advisories have
been issued in 40 states that warn against the consumption of certain amounts and species of fish that
are contaminated with mercury Nine states (current as of August 1995) have statewide advisories
(i.e.. advisories posted on every freshwater body in that state). The action levels chosen by the states
that trigger these advisories vary and in some cases are lower than the U.S. FDA action level of 1
ppm for fish in commerce. In some areas, freshwater ftsh can have mercury levels which exceed the
FDA action limit of I ppm. The concentration of methylrnercury in commercially important marine
species is, on the average, lower than the level that would trigger an advisory level.
Widespread contamination of the freshwater fisheries is a concern for several reasons
including the potential risk to fish consumers in the U.S., economic losses to commercial and
recreational fisheries from limitations oa catching or consuming affected fish, and potential health and
economic impacts to subsistence fishers. Native American cultural values are also diminished by a
contaminated fishery. Endangered non-human species, in particular the Florida panther, are at
significant risk from mercury contamination.
There was no attempt in th» Report to do a risk benefit analysis of fish as an important
source of protein and eateries in the diet of U.S. populations. Such an analysis would be beyond the
scope of the CAAA mandate. The focus of the Mercary Study Report to Congress has rather been
the assessment of potential hazards of mercury exposure including methylmercury m fish. The
U.S.EPA has not attempted to weigh the advantage of fish consumption vs. the risk of earing
methylmercury contaminated ftsh.
The economic magnitude of the impact of mercury contamination on freshwater fisheries
cannot be accurately predicted, but it is intuitive that fishing revenues could be impacted should
mercury contamination worsen. In 1991, 31 million adult Americans took advantage of freshwater
fishing opportunities, spending a total of S15.1 billion on freshwater sportfishing trips and equipment.
Mercury contamination poses a direct threat to the economic viability of these and other related
enterprises. In Arkansas alone, mercury fish advisories issued in an 8-counry area were estimated to
result in a loss of $5 million in revenues in 1993.
October 20, 1995 2-2 DRAFT: Do Not Cite or Quote
-------
2. MERCURY IN THE ENVIRONMENT
As a chemical element, mercury cannot be created or destroyed. The same amount has
existed on the planet since the earth was formed. Mercury, however, can cycle in the environment as
pan of both natural and human (anthropogenic) activities. Both measured data and the results from
global modeling have led to the understanding that anthropogenic mercury emissions (i.e.. emissions
attributable to human activities) equal or exceed those from natural sources (e.g., volcanic activity)
Human activity has the overall effect of making more mercury biologically available. Emissions of
mercury from human activity are thought to contribute from between 40 and 75 percent of the current
total annual input of mercury to the atmosphere. Measured data and modeling results indicate that the
amount of mercury mobilized and released into the biosphere has increased since the beginning of the
industrial age.
Once emitted to the atmosphere, mercury can deposit to the earth in different ways and at
different rates, depending on its physical and chemical form. The form of mercury emitted influences
its atmospheric fate and transport, as do conditions specific to its site of release (for example, stack
height and nearby terrain). The mercury bound to paniculate matter and the divalent form of
mercury can be deposited close to the emission source (e.g., within 10 kilometers). The elemental
form of mercury is known to circulate in the atmosphere for about one year before it is oxidized and
deposited. This relatively long residence time is sufficient for the emissions to be mixed into the
global atmosphere and deposited onto land or water bodies remote from the emissions source,
including outside of the United States. The result is that mercury deposition is a local, regional and '
global issue.
Mercury can be deposited directly to water bodies or can be transported from land by run-off
and enter ponds, streams, rivers, lakes, etc. The water bodies contain microorganisms that have the
metabolic capability to carry out chemical reactions which bind mercury to methyl groups, producing
methylmercury. Bodies of water without the microorganisms also appear to be capable of methylatmg
mercury. The chemical properties of methylmercury permit it to penetrate biological membranes.
Methylmercury is the form of mercury to which humans and wildlife are generally exposed, usually
from eating fish which have accumulated mercury in their muscle tissue.
Methylmercury is biologically concentrated or bioaccumulated. The concentration factor is
extreme, resulting in multiples in the hundreds of thousands. That is, an animal at a higher position
in the foodweb may have mercury concentrations thousands of times higher than an animal at a lower
position in the foodweb. The transfer of mercury in the foodweb or foodchain to progressively
higher concentrations in large fish is key to understanding how release of mercury to the atmosphere
results in exposure to high concentrations of mercury in fish, and ultimately humans. While the
overall process of mercury emission, deposition and bioaccumulation in fish is understood, there is a
great deal of variability and scientific uncertainty in the models used to describe the movement of
mercury from emissions to fish.
Methylmercury appears to accumulate in fish at different rates depending on the season of the
year and the type of water body the fish occupy. A number of chemical properties of water bodies
influence the methylmercury concentrations in fish as well. These include the acidity of the water
(fish in water with a low pH tend to accumulate more methylmercury), calcium content, dissolved
organic matter and other water chemistry parameters. Exactly which properties of water bodies and
the degree of influence these have over methylmercury concentrations in fish is not wefi understood at
this time. The models used to g*amin* the relationship between mercury emissions and
methylmercury concentrations in fish are a simplification of more complex and sometimes not well
understood processes that occur in nature.
October 20, 1995 2-1 DRAFT: Do Not Cite or Quote
-------
As in the Japanese poisoning epidemics, the signs and symptoms of disease were
predominantly those of the nervous system: difficulty with peripheral vision or blindness, sensory
disturbances, incoordination. impairment of walking, slurred speech and in some cases, death.
Children were affected, as well as adults. Of great concern was the observation that infants, born of
mothers who had consumed the methylmercury-contarrunated gram (particularly during the second
trimester of pregnancy) could show nervous system damage even though the mother was only slightly
affected herself. In both the Iraqi and Japanese epidemics, the levels of methyhnercury consumed
were much higher than the levels currently reported in the U.S. food supply.
Health endpomts other than neurotoxicity were evaluated by U.S. EPA using established risk
assessment Guidelines. Methylmercury has been shown to cause rumors in mice at high doses that
produce severe non-cancer toxicity. Low-dose exposures to methylmercury are not likely to cause
cancer in humans. Data on effects related to mutation formation (changes in DNA) indicate that
methylmercury could increase frequencies of mutation in human eggs and sperm. These data were
not sufficient, however, to permit estimation of the amount of methylmercury that would cause a
measurable mutagenic effect in a human population.
Research needs include studies which will delineate the most appropriate indicators of
neurotoxk effects for exposed adults, children and individuals exposed to metfaybnercury in ittero.
Well conducted studies are also needed to clarify critical levels at winch other task effects could
occur in humans. For all three forms of mercury, data are inadequate, conflicting or absent for the
following: adverse reproductive effects; impairment of iraiaane function; and genotaxic effects on
either somatic or germ cells (for elemental and inorganic mercury).
How Much Methylmercury is Harmful to Humans?
Information on the amount of methylmercury exposure producing particular combinations of
signs and symptoms in people has been analyzed to yield what are called quantitative dose-response
assessments. Both the Japanese and Iraqi epidemics are important to understanding how
methylmercury from food produces neurological disease in humans. In the epidemics in Minamata
and Niigata, the exposures were long-term, and the tissues of fish and shellfish were the sources of
methylmercury exposure. This establishes with highest scientific confidence that methylmercury in
fish can produce human disease. A limitation to these data is that many patients were severely
affected. The extent of methylmercury poisoning was so severe that finding subtle indications of
disease is difficult. Subtle indicators of poisoning are important in identifying levels of exposure
which will not cause any adverse effects.
The U.S. EPA calculates one such measure, called a reference dose or RfD. This is the
amount of material which can be consumed on a daily basis over a lifetime without harm, the RID is
based on th«l^t*wfcit*4ittt»ldefe m$ca» *sorcal!ed critical ef&ct; this & generally the first
uxiicatOT or mo* Sfcde indicate In calculating RfDs
U.S.EPA p*e*ilS& WfcMKMfcisrved adverse effect few* (J4OAEIX m fc fiamd fe^ either
inspection of or modeling of dose response data oa the critical effect. Ik is a means of determining
the threshold fer effect la the ipecies uafer snidy. Ute JfOAEL i* fitwt useftf wseo it is ffoa a
study wherein a deteonmatKBt
-------
3. FINDINGS OF THE MERCURY STUDY REPORT TO CONGRESS
This chapter integrates the findings and conclusions drawn from the analyses described in full
in the Report in order to present the risk manager with the information needed to make informed
decisions. A series of regulatory issues are presented and specific actions are recommended in
subsequent chapters. A brief listing of research needs follows each section defined by the questions
below; these are described in more detail in the Supplement to this volume .
The following questions are posed to summarize the risk characterization and guide the
discussion of regulatory issues:
1. What effects on human health can methylmercury have?
2. How much methylmercury is harmful to humans?
3. How do U.S. EPA's estimates of how much methylmercury is harmful to humans
compare with levels calculated by the U.S. Food and Dnif Administration , WHO
and the states?
4. What is the estimated size of the population at risk?
5. How much methylmercury exposure is harmful to wildlife and what are the effects?
6. What are some of the limitations of the risk analysis?
7. Which sources contribute to these impacts?
8. Is the problem getting worse or better?
9. What are the management alternatives for reducing mercury emissions?
What Effects on Human Health Can Methylmercury Have?
Data in both humans and experimental animals show that all three forms of mercury evaluated
in this Report (elemental, inorganic and methylmercury) can produce adverse health effects. Except
for people whose occupation brings them in contact with elemental mercury, humans will be primarily
exposed to methylmercury in fish. Methylmercury can produce a variety of adverse effects,
depending on the dose and time of exposure.
Neurotoxicity is the effect of greatest concern when adults and children are exposed to
methylmercury. Two major epidemics of methylmercury poisoning through fish consumption have
occurred. The best known of these two epidemics occurred among people and wildlife living near
Minamata City on the shores of Minamata Bay, Kyushu, Japan. The source of methylmercury was a
chemical factory that used mercury as a catalyst. A series of chemical methods identified
methylmercury in the factory waste sludge; this was drained into Minamata Bay. Once present in
Minamata Bay, the methylmercury accumulated in the tissue of shellfish and fish that were
subsequently consumed by wildlife and humans. Uteiaewied Iwdiof moony ins site Safe ranged
from 10 to 20 ppai in 1961. After 1969 average mercury coaceottatfom iafish fctdltlitn bdow 0.5
ppns fisfe was a routine part of thftdkt in these popuktwm. An as?cia^ fM coaraB^tkM iacicea
'
^mynmptt
3-1
-------
Nevertheless, the RfD for methylmercury is a reasonable estimate and is very likely to be protective
of human health.
This dose-response estimate is supported by additional studies in children exposed in utero
These include investigations among Cree Indians in Canada and New Zealanders consuming large
amounts of fish. In these studies the hair concentration of mercury is used to monitor mercury
exposure over time. Conclusions by the investigators in their official reports cite developmental
delays among the children born of mothers whose hair mercury concentrations during pregnancy were
6 to 18 ppm.
Currently a number of research studies are underway that further address the question of what
exposures to methylmercury in fish are associated with neurological disease. These studies include
more subjects than did die Iraqi study, are prospective in design, and utilize endpoims that are
anticipated to be more sensitive than the clinical signs and symptoms of methyirnercury
poisoning observed in Iraq. These studies are conducted in the Seychelles Islands in the
Indian Ocean, the Faroe Islands hi the North Atlantic Ocean, and in the United States: this
last study is sponsored by the Agency for Toxic Substances and Disease Registry (ATSDR).
Data from these studies, when available, should be useful in decreasing the uncertainty
surrounding both the benchmark dose and the RfD. The Food and Drug Administration has *
determined that revisions in the FDA action level for mercury concentrations of fish in
interstate commerce should wait until the new studies have reduced the level of uncertainty
surrounding the RfD.
One of the ATSDR oponoored otudieo io of infants whooc mothers consumed large quantitie
of Groat Lakco fish during pregnancy. In Spring. 1905 preliminary results indicated a higher
occurrence of primitive otartle reflexes among infants whose mothers reported consumption of high
levels of Great Lakes fish during pregnancy. There are many chemicals in fish that have been
reported to adversely affect child development (e.g.. PCBo). so camion must be exercised to avoid
over estimation of the likelihood that mercury expooures are causative to these results.
Uncertainties m tb!> rillc eaHm«£e8 wffi also tie reduced by addfr tonal research on die mode
actkm and toxicokinctics oi methy&nercury. This work sbouM be aecomp^cd by stodks on
pharmacokinetics leading to the development of applicable physiologically-based pbannacokinetics
models whk&iad^ft fetut,
How Do U.S. EPA's Estimates of How Much Methylmercury is Harmful to Humans Compare
With Leveb Calculated by the II& Food and Drag Admimstntftea, WHO and the States?
There are more than one thousand local and state warnings in the U.S. to limit intake of fish
because of chemical contamination. Warnings are issued because of a number of contaminants.
Methylmercury is most often included as one of the contaminants of sufficient concern to form the
basis for the warning. Often these warnings are issued based on local conditions.
Biological monitoring for mercury exposure can be done using hair, blood and/or
concentrations in other tissues as an index of body burden of mercury. These concentrations are used
as a surrogate for mercury concentrations in tissues such as brain. Recommendations staring limits on
methylmercury exposure have been expressed in the following units: micrograms/day; /ig'kg body
S'O"* 3-5
-------
weight/day; concentrations of mercury in tissues such as blood, hair, feathers, liver, kidney, brain.
etc.; grams of fish per day; and number of fish meals per time interval (for example, per month).
Table 3-1
WHO Data on Mercury in Hair
Fish Consumption Frequency
No unusual mercury exposure
Less than one fish meal per month
Fish meals twice a month
One fish meal a week
One fish meal each day
Average Mercury Concentration in
Hair (jig mercury per g of hain
2
1.4 (range 0.1 to 6.2)
1.9 (range 0.2 to 92)
2.5 (range 0.2 to 16.2)
11.6 (range 3.6 to 24.0)
Cross-comparison of
WHO recommendations
regarding risk associated
with hair mercury
concentrations is facilitated
by data reported by the
World Health Organization
(WHO) on mercury
concentrations in 559
samples of human head hair
from 32 locations in 13
countries. The WHO report
found that mercury
concentrations in hair
increased with increasing
frequency of fish
consumption (see Table 3-
1).
The scientific literature contains numerous references to increasing hair mercury
concentrations with increasing frequency of fish consumption. Hair mercury concentrations
depend on both the concentration of methyimercury in fish and the quantity offish consumed.
To date data on hair or blood mercury concentrations from a representative sample of the
United States population have; not yet been published. Unpublished date (submitted to U.S.
EPA by the U.S. Food and Drug Administration) from me early 1980s on a group of United
States 1431 women of chiM-bearing age indicate hair mercury concentrations of 0.48 ppm for
the overall sample of 1431 women and a mean concentrationof 0.52 ppm for the 1009
women who reported.'^acEM|n^.99iae-9ej^ciddL These meK^ concentrations correspond to
hair mercury <5oneentfadonstassociated wimvfiiahconsumptkm at the level of one meal per
month to one meal per
unpublished!
observed i»
This
survey
on data shown in Table3-U
in hair mercury concentcation suggested in these
reporting seaibod consumption differs from the patterns
m the world literature on hair mercury coocentrattons.
these unpublished data and most literatore underscores the need for
or blood mercory concentrations from a representative sample of the
United States popsiagon to estimate body burden of mercury m me general United States
population.
A number of different estimates exist for hair mercury levels that are associated with
low risks of neurological endpomts such as paresthesia. These estimates are sensitive.to
variables such as the half-life of mercury in the body (time to eliminate half the dose of
mercury). The range of half—lives usually uses 70 days as an average, with extreme of about
3-6
-------
35 to just over 200 days reported for different individuals. The half-lives for pregnant
women have not been directly measured. Half-lives for women during lactation are shorter.
possibly due to excretion of mercury into milk produced during lactation.
In 1969. in response to the poisonings in Mmamata Bay and Niigata, Japan, the U.S.
FDA proposed an administrative guideline of 0.5 ppm for mercury in fish and shellfish
moving in interstate commerce. This limit was converted to an action level in 1974 (Federal
Register 3_£> 42738, December d, 1974) and increased to 1.0 ppm in 1979 (Federal Register
44> 3990r January 19, 1979) in recognition that exposure to mercury was less than originally
considered. In 1994, the 1.0 ppm action level was converted from a mercury standard to one
based on methylraercury (Federal Register 49. November 19T 1984).
The action level takes into consideration the tolerable daily intake (TDI) for
methylmercury, as well as information on seafood consumption and associated exposure to
raemylmercury. The TDI is the amount of methylmercury that can be consumed daily over a
long period of time with a reasonable certainty of no harm, FDA (and WHO) established a
TDI based on a weekly tolerance of 0.3 mg of total mercury per person, of which no more
than 0.2 mg should be present as methylmercury. These amounts are equivalent to 5 and X?
jtgt respectively, per kilogram of body weight. Using the values lor methyhnercury, this
tolerable level would correspond to approximately 230 fig/week for a. 70 kg person or 33
ug/person/day. The TDI was calculated from data developed in part by Swedish, stadies of
Japanese individuals poisoned in the episode of Niigata whkri resulted from the consumption
of contaminated fish: and shellfish and the consideration of other studies offish-eating
populations.
Based on observations from the poisoning event later in Iraq, FDA has acknowledged
that the fetus may be more sensitive than adults to the effects of mercury (Federal Register
44l 3990, January 19,1979; Confie and Tollersoa*4934* FDA Consumer, September, 1994).
In recognition of these? concerns* FDA tu» provided advice te and women of
child-bearisg ag«to !in3&&circ^^ to have itigit levels of mercury
(FDA Consumer, 1994). FDA believes, however, that given existing paattems of fish
slight reductions m the inargin of safety. However, due to the uncertainties associated with the
Iraqi study, ^j^{^:.^o^.:i^-1»-:i^-^.li^ study as & basis lor revising its action level.
Instead, the F||p1*ar chosen to wait far findings of prospective studies of fish-eating
popufadons Iia^i^:;Si%!c|je|felslaads and is the Faroe Islands.
The World Health Organization's International Programme for Chemical Safety
(WHO/IPCS) concluded that the general population of adults (males and non-pregnant
females) does not face a significant health risk from methylmercury when hair mercury
concentrations are under 50 ng mercury/gram hair. Reports from the Niigata epidemic of
Minamata disease found lower thresholds with mean values in the range of 25 to
approximately 50 ug mercury/gram hair.
Clinical observations in Iraq suggest that women during pregnancy are more sensitive
to the "ffr^s of methylmercury with.fetuses at particularly increased, risk... The. WHO/IPCS
3-7 -f 3-o
-------
(1990) analyzed the Iraqi data and identified a 30 percent risk to the infant of abnormal
neurological signs when maternal hair mercury concentrations were over 70 ug/g. Using an
additional statistical analysis. WHCUPCS estimated a 5 percent risk of neurological disorder
in the infant when the maternal hair concentration was 10 to 20 ug mercury/gram of hair.
WHO1PCS recommended that as a preventive measure, in a subpopulation that
consumes large amounts of fish (for example, one serving or 100 grams per day), hair levels
for women of child-bearing age should be monitored for methylmercury.
The WHO/IPCS estimated (1990) that a daily methylmercury intake of 0.48 ug
mercury/kg body weight will not cause any adverse effects to adults and that a methylmercury
intake of 3 to 7 ug/kg body weight/day would result in a <5 percent increase in the incidence
of paresthesia in adults. Risk to this extent would be associated with hair mercury
concentration of approximately 50 to 125 ug gram hair. By comparison, the U.S. EPA's
reference dose, or the amount of methylmercury any person (including children and pregnant
women) can ingest every day without harm is 0.1 fig/kg body weight per day. This was based
on a benchmark dose equal to 11 ppm (jig/g) hair. Children have a higher exposure to
methylmercury (on a per kg body weight basis) than do adults. The U.S. EPA estimates in •
this Report that intake of methylmercury by children ages birth through 14 years residing in
the continental U.S. is 0.84 ug/kg body weight/day for males and 0.81 ug/kg body weight/day
for females.
A research need for this area is for biological monitoring (for exposure and effect) of
populations with either greater than U.S. average fish consumption (such as one serving or
100 grams per day) or consumption offish predicted or measured to have'higher than average
amounts of me&yimercury.
What is the Estimated Size of the Population at Risk?
The people in the U.S. who oonoumo fioh in cxocog of 100 groma'day (e.g., one
serving) are considered to be the population at rink. In particular, women of child bearing ago
and children aged 14 and younger arc of concern became of the adverse effects of
methylmeroury on the developing nervous system.
The general United States population that obtain their feh by purchasing them
commerciaily arnot believed'to be adversely affected by metbyimcrcury if they consume fish
in moderation m people in the U.S. woo routinely consume Ssh In excess of 100
grams/day (e,g,, one serving) are considered to be a population potentially «t risk, fa
particular, women of cluld-beanng age are considered of concern because of me adverse
effects of methyjmercuiy on the developing nervous system. Children ages 14 and younger
are of concern because they have higher exposures to medryknefcury from fisb, on a per
kilogram body weight basis compared with adults.
Women of child-besring age rather than only pregnant women are a group of concern for two
reasons. The first is that methylmercury persists in tissues. Measured half-lives of methylmercury in
range from about. 1 month to 9 mnmhs ^though half-lives of just over 2, mnnrhi are usually
3-8 3-^f
-------
observed. Thus, dietary intakes just prior to pregnancy are of concern rather than only
methylmercury intakes during pregnancy The second reason is that using estimates of the number of
pregnant women in the age group 15 through 44 years, in any given year 9 5 percent of women are
pregnant. Statistics to support a more detailed prediction of which groups of women in the United
States are high consumers of fish and which groups of women are the most likely to become pregnant
were not developed in this report.
There are two ways to estimate the size of the population at risk. The first evaluates the
quantity of methylmercury exposure on a body weight basis and compares this exposure to the U S .
EPA's reference dose. The second looks only at the quantity of fish consumed without making any
assumptions about the concentration of mercury in the fish.
The first method requires information on the amount of fish consumed and the concentration
of mercury in various species of fish. Data on fish consumption for a general population of women
in the United States were developed from the United States Department of Agriculture's Continuing
Surveys of Individual Food Consumption for the period 1989-1991 (CSFH 89/91). These surveys
assessed fish consumption over a three-day period. Data on mercury concentrations in fish were
available from the National Marine Fisheries Service and from published information on mercury in
freshwater fish. These sources have been used in this Report to estimate methylmercury intake fronv
fish.
The RfD for methylmercury is LO xlO*4 mg/kg body weight/day or 0.1 /xg/kg body
weight/day. Data on methylmercury consumption from fish for women of child-bearing age (ages 15
through 44 years) show the 50th percentile for adult females to be 0.08 Mg/kg body weight/day with
75th percentile exposure to be 0. 16 ng mercury/kg body weight/day. Consequently approximately
one-half of the general U.S. population of women of child-bearing age who consume fish at least
once within a 3-day period are predicted to ingest methylmercury at the RfD. The RfD Is expected to
be a level of exposure to be without advene effect.
These data are shown in Figures 3-1, 3-2, and 3-3, which describe the total quantity of fish
and shellfish consumed by children (ages 14 years or younger) or adult women of child-bearing age
(15 through 44 years) who eat fish often enough to be identified as fish consumers in the three-day
dietary survey, CSFII 89/91. Because freshwater fish may be of particular interest around
anthropogenic sources. Figures 3-4, 3-5, and 3-6 show the quantity of freshwater fish consumed by
children (aged 14 years or younger) and adult women of child-bearing age (15 through 44 years) wfco
aie fish one or mote tfesea daring the thwe-dstjr dietary survey period.
The second way the size of the population at risk is estimated is based on the WHO
recommendation that a preventive measure would be having hair mercury concentrations measured
among women of child-bearing age who eat a large amount of fish (for example, 100 grams/day or
higher). The WHO recommendation was based on an increasing risk of developmental delays in
infants whose mothers consume fish frequently during pregnancy. The level of risk associated with
consistent consumption of 100 grains or more of fish per day depends on the mercury concentration
in the fish eaten. Because persons who are consuming fish at 100 grams per day or higher may be
more reliant on site-specific sources than is the general population, the following estimates of the size
of the population judged to be at highest risk of methylmercury exposure are based on consumption of
fish in the amount of 100 grams per day or more.
3-9
-------
The States of Hawaii and Alaska were not included in the above estimates because CSFII
89/91 did not include these states. Because of their substantial coastal areas fish-consumption
patterns in Alaska and Hawaii are predicted to differ from those of the 48 contiguous states. The
number of women of child-bearing age (15 through 44 years) in Alaska is estimated to be
approximately 138.000 and in Hawaii is estimated to be approximately 284.000 persons. The
percentage of pregnant women for ages 15 through 44 years in Alaska was 9.7 percent and in Hawaii
this percentage was also 9.7 percent.
Children aged 14 years and younger are also a potentially at-risk population for
methylmercury exposure from fish. The population of children aged 14 years or younger in the 48
contiguous states is 53,463,000 with an additional 151.000 in Alaska and 240,000 in Hawaii. As
shown in Table 3-2. data from CSFII 89/91 identified 25 percent of children as consuming fish at
least once during the 3-day survey period. Based on these estimates approximately 13,306.000
children ages 14 and younger regularly consume fish, and the number of children consuming fish in
the quantity of 100 grams per day or more becomes approximately 665,000.
When methylmercury intake was expressed on a per kilogram body weight basis, children's
exposure was approximately 2-3 times that of adults. In the methylmercury poisoning epidemics in
Japan and Iraq children were affected as weli as adults. These effects were not seen only in children*
exposed to methylmercury in utero but included children exposed through ingesting methylmercury
from food. Whether or not children differ from adults in how much methylmercury exposure they
tolerate, without showing neurological effects is not known. Because the nervous system is not fully
mature until late in childhood and because methylmercury adversely affects the developing nervous
system, it appears imprudent to ignore the possibility that adverse neurological effects occur in
children at lower methylmeroury expoourec than in adults. It appears from U.S. EPA's estimates that
children are at increased risk of methylmercury exposure than are adults because of higher exposures
on a per kilogram body weight basis.
Fish consumption is modified by cultural, geographic, economic, social and health factors.
The estimated number of people from the CSFII 89/91 data who consumed 100 grams of fish or more
per day is for the general U.S. population. Whether or not dietary surveys of the general population
adequately represent subpopulations (such as recreational anglers, subsistence fishers, or Native
Americans) remains an issue.
A review of the published literature on quantities of fish eaten by groups that include anglers,
subsistence fishers and some Native Americans shows that average fish intake is higher for these
groups than for the general population. Some groups, for example native populations in Alaska,
consume on average, quantities of fish far higher than the general U.S. population. These reports
often did not provide sufficiently detailed, clear identification of the age and gender of their subjects
to determine if the women were of child-bearing age. Because many women and children in these
groups may obtain a substantial portion of their fish from a limited geographic region or from only a
few species of fish, the likelihood that they will exceed the benchmark dose for methylmercury
depends on the concentration of methylmercury in the fish they consumed. These particular groups of
women and children would benefit from a more specific evaluation. An example of such evaluation
would be biological monitoring based on mercury analyses of hair or blood.
Aa adfiribnri area Wia&m&SS&Wt* the effects of post-natil, pre-pubesce* methy
exposure OB development *i& the aim of ttarifging if
vttt'itti that^ It
October 20, 1995 3-18 DRAFT: Do Not Cite or Quote
-------
Wildlife Criteria for Mercurv
Organism
Mink
River otter
Kingfisher
Osprey
Bald eagle
Wildlife Criterion
(pg/L)
415
278
193
483
538
There is uncertainty and variability associated with each WC. These include lack of long-
term studies for mammals, lack of a no adverse effects level for birds, and extrapolation from one
species to another. It is not known if the species selected for WC development are the most sensitive
or appropriate species; also, it is not known if protecting individual animals or species will guarantee
protection of their ecosystem from harmful effects of mercury. There are uncertainties and expected*
variability in the BAF; it was the subject of a quantitative uncertainty analysis.
In restricted wildlife populations, effects of mercury originating from point sources have been
conclusively demonstrated and provide a residue basis for evaluation of mercury levels in animal
tissue as an indicator of risk to other populations. Although clear causal links have not been
established, mercury originating from airborne deposition may be a contributing factor to population
effects on bald eagles, river otters and mink. There is evidence to support the possibility of toxic
effects on the common loon and the Florida panther.
The information assessed in this Report suggests that ecosystems most at risk from airborne
releases of mercury exhibit one or more of the following characteristics:
• They are located in areas where atmospheric deposition of mercury is high;
• they include surface waters already impacted by acid deposition;
• they possess characteristics other than low pH that result in high levels of
bioaccumuiation; and/or
• they include sensitive species.
Research is seeded f& seven! areas to decrease uncertainty ia tie Msessmwfc of mercury
effects on non-human species. One general area of need are in basic infonna&m oft the fate aad
effects of mercury ia the eavkooisent. More applicable infonnatjoa coo&iead to reduced ancertainty
rathe wildlife criteria, tjkj«eaHJHiiadon factors* aswanptraw on feod cotBiraptiiOa Jad other factors.
A second area is improvement in the ability to detect ecological damage wte» it win fcct occurring
(at opposed to the ajnrtntnettf of ia&tsdtaa! specks which WAS done IB this Report).
What are Some of the Limitations of the Bisk
October 20, 1995
3-20
DRAFT: Do Not Cite or Quote
-------
A primary purpose of the Mercury Study Report to Congress was to assess the impact of U S
anthropogenic emissions on mercury exposure to humans and wildlife. The size of some populations
of concern have been estimated; namely women of child-bearing age and children who eat fish. The
adverse effects of methylmercury on wildlife have been described and quantified. For wildlife the
importance of site-specific effects of mercury exposure are anticipated to be greater than for humans
in the general population because wildlife obtain their fish from a much more limited geographic area
than do people. In the general population, people typically obtain their fish from many sources. The
question on whether or not the impact of mercury from anthropogenic ambient emissions can be
proportioned to the overall impact of methylmercury on wildlife is a much more difficult issue.
In Volume IQ of Mercury: Report to Congress, the U.S. EPA examined human and
wildlife exposure to mercury released from anthropogenic sources. The following human
exposure routes were included: inhalation, consumption of water, consumption offish, beef,
beef liver, cow's milk, poultry, chicken eggs, pork, lamb, green plants (eg., leafy vegetables,
potatoes, ihihs, grains and cereals) and ingestion of soil. Dermal exposures that resulted from
contact with soil and water, as well as exposure through inhalation of resuspended dust particles
and exposure through the consumption of human breast milk were not evaluated. After
evaluation of the results of the human exposure modeling, the only exposure route considered in.
Volume HI for wildlife was the consumption of freshwater fish.
The source attribution estimates in the Report are based on modeled data. To assess how well
the modeled data predict actual mercury concentrations in different environmental media at a variety
of geographic locations requires a data base against which to make these comparisons. Unfortunately
the extent of monitoring data is not sufficiently detailed to permit these comparisons. Substantial
additional monitoring data would facilitate such comparison.
As with environmental monitoring data, information on body burden of mercury in
populations of concern (blood and/or hair mercury concentrations) are not available for the general
U.S. population. Data on higher-risk groups are currently too limited to discern a pattern for the
U.S. population that is more predictive of methylmercury exposure than information on quantities of
fish consumed.
Finally there is the additional problem of predicting the impact of mercury in addition to
current exposures. The benchmark dose used in estimating the RfD depends on analyses of data with
the subjects clustered into dose groups. Most data on neurologically based developmental endpoints
are continuous; that is, not assigned to dose groups. For example, scoring on scales of IQ involves
points rather than a "yes/no" type of categorization. Measurements on the degree of constriction of
the visual field involve a scaling rather than a "constricted/unconstricted" type of variable. Although
arbitrary scales can be constructed, these groupings have generally not been done in current systems.
An additional difficulty occurs in estimation of benchmark dose when multiple endpoints have been
measured. Further research on appropriate methods for mathematical modeling is needed. For some
situations such information is known, but for methylmercury exposure and multiple endpoints
assessing the same system (i.e., developmentally sensitive neurological, neuromotor and
neuropsychological effects) the time-course/dose-response of such changes have not been clearly
established. Development of the mathematical models needs to be accompanied by understanding the
physiological/pathological processes of methylmercury intoxication.
October 20, 1995 3-21 DRAFT: Do Not Cite or Quote
-------
For wildlife, a separate problem exists. As for humans, mercury toxicity among wildlife
involves neurological effects. Available toxicology data from laboratory-based studies of wildlife
exposed to methylmercury have measured only gross clinical signs and symptoms of disease and death
or pathological changes accompanying these clinically evident changes. Physiologically based
evaluation of wildlife has not been done. The importance of more subtle endpoints of neurological
function is anticipated to be relevant to such practical questions as the ability of visual hunters such as
the loon to find food.
To improve the characterization of risk research needs highlighted in the preceding sections
should be addressed. Some additional work to decrease uncertainly should be directed toward the
exposure assessment. Evaluated local and regional atmospheric fate and transport models are needed.
This should utilize long-term national monitoring networks. Data to improve understanding of
movement of mercury through environmental media are also needed. The bioaccumulation factors are
major sources of uncertainty. This uncertainty will be decreased by unproved data to use in the
parameters of the bioaccunaaiation factor equations or by increased understanding of mercury
biogeoehemtstry in water bodies.
Which Sources Contribute to these Impacts?
In the emissions inventory (Volume II), it is estimated that anthropogenic emissions of
mercury in the United States are about 220 Mg (243 tons) of mercury per year. (This emissions
inventory is believed to be accurate to plus or minus 30 percent.) There are numerous and varied
sources of mercury in the U.S. that emit mercury virtually everywhere across the country. The
largest sources of mercury are those that combust fossil fuels (principally coal), mercury-containing
waste (e.g., municipal or medical), and smelting operations. There are also large contributors among
manufacturing sources (e.g., -hlor-alkali plants). Table 3-4 lists the source categories for which
there were sufficient data to estimate national mercury emissions.
Table 3-4
Mercury Sources With Sufficient Data to Estimate National Emissions
Anthropogenic Sources
Area
Point
Combustion
Manufacturing
Miscellaneous
October 20, 1995 3-22 DRAFT: Do Not Cite or Quote
-------
link between mercury emissions from anthropogenic sources and mercury concentrations m air. soil.
water, sediments and methylmercury concentrations in freshwater fish.
Current anthropogenic sources in the U S. are not the only contributors to increased mercury
concentrations in the environment. There are natural sources such as volcanoes and soils, and
emissions from previously deposited anthropogenic emissions that are cycling between the
atmosphere, land and water. There is a large reservoir of mercury already in the environment that
exists in the global atmosphere and sediments. Global modeling of the mercury cycle shows that
anthropogenic emissions rival or exceed those from natural sources. Anthropogenic emissions are
believed to contribute about 50 to 75 percent of the current total annual input of mercury to the
atmosphere. An estimate of the total annual input from anthropogenic sources globally is 3632 Mg
(4000 tons). The U.S. sources account for about 6 percent of the global inventory.
Anthropogenic mercury emissions from sources in the U.S. contribute to the global cycle and
likewise, global emissions contribute to deposition in the U.S. The long-range transport analysis
predicted that of the 220 Mg (243 tons) of U.S. emissions that were modeled, 78 Mg (86 tons), or 35
percent, were deposited in the continental U.S. while 145 Mg (160 tons) or 65 percent, were
transported outside of the continental U.S. The modeling simulation also predicted that 33 Mg (36
tons) are deposited annually in the U.S. from the global atmospheric reservoir.
The contribution of mercury to the environment from global sources, natural sources and
previous emissions complicates an assessment of the public health and environmental impact of
current U.S. emissions. Conclusions from the modeling of fate and transport of anthropogenic
emissions of mercury from industries identified in the Emissions Inventory (Volume II) of this
Report indicate that there is a plausible link between increases in mercury emissions from
these sources and increases in mercury concentrations in fish. Modeling-based efforts to
quantitatively assess the magnitude of this increase cannot be fuOy evaluated because of
limitations on monitoring data providing mercury concentrations in the aquatic food chain.
The weight of evidence supports an association between local mercury emissions and
increasing mercury levels in locally caught fresh water fish. Variables conaibotmg to this
linkage include the following: tfifc species of mercury that arc emitted; the overall amount of
mercury emitted from a source; the extent of mercury nwihytalionia the water botfy ; and the
climate conditions. For flslt cc4{ec^ at sites reinote &om the source, it is not possible to
assess the csontt&afkat of any particular soorce. A conclusion of this Report, however, is that
the majority of $* mercury deposited In the 0.SL is from anthropogenic sources.
Roiwi^lfe'nijprove the emissions inventory includes the acquisition of test data on
several iKJtabJiiourCes including the following: mobile sources* landfills, agricultural burning,
cote ovens, petroleum refining and some rnamrfacturing sources. Appropriate emissions data
from other countries \roald also be of great use. Speciated mercury test data ace needed from
all sources. Re*emissiott of mercury J&om deposited mercury is a major area of uncertainty*
Is the Problem Getting Worse or Better?
There are no national long-term monitoring data for either air or biota available to
address this question. The 1992 Study of Chemical Residues in Fish conducted by the U.S.
EPA should provide baseline data for fish levels nationally. A number of «niHfc« have
October 20, 1995 3-24 DRAFT: Do Not Cite or Quote
-------
implemented as proposed by the U.S. EPA (59 FR 48198 and 59 FR 48228 for municipal
\\aste combustors and 60 FR 10654 for medical waste incinerators). These two source
categories are currently estimated to comprise about 45 percent of the U.S. inventory.
What impact would a decline in deposition have on methylmercury concentrations in
fish? Because mercury is persistent in the environment, and such a global reservoir has
already accumulated, panicularly in sediments and soils, it is likely to take decades before fish
mercury levels would be comparable to preindustrial levels.
Longitudinal studies using ultra-clean sampling techniques and state-of-the-art
analytical methods are needed to help resolve questions of mercury concentration trends in
with time,
What are the Management Alternatives For Reducing Mercury Emissions?
Effective control of mercury emissions may require a mix of strategies including
pollution prevention, materials separation and conventional regulatory strategies to control
mercury emissions at the stack. Pollution prevention would be suitable for those processes or
industries where a mercury substitute is demonstrated and available (e.g., mercury cell chlor-
alkali plants). Material separation is an appropriate approach for processes where
mercury-containing products are disposed of by incineration, or where mercury can be
reduced in the fuel prior to the fuel being combusted (e.g., medical waste incineration and
coal combustion). The third approach, conventional regulatory strategies, may be necessary
when mercury is emitted to the environment as a result of trace contamination in fossil fuel or
other essential feedstock in an industrial process (e.g., smelting). Other non-traditional
approaches such as emissions trading or application of a use tax, or other market-based
approaches may also prove feasible for mercury control. However, these options are not
presented in detail hi this Report as the control technology analyses focused on what might be
achievable under the statutory language of Sections 1 12 and 129 of the CAA.
The analyses of control technologies and costs presented in this Report are not
intended to replace a thorough regulatory analysis, as would be performed for a rulemaking.
The information presented is intended to present the range of available options and provide a
relative sense of the extent of mercury reductions achievable and the general magnitude of the
cost of such reductions.
Table 3-5 characterizes the source categories which there were sufficient data to
estimate national emissions in this Report. The source categories are listed according to the
magnitude of their estimated emissions. Source categories where there are pollution
prevention or material separation opportunities are highlighted. As shown, the primary
opportunity for mercury reduction through pollution prevention lies in removing mercury from
products, or from switching to a low-mercury fuel. The costs of replacing mercury in specific
products were not estimated in this Report. The impact of replacing mercury with an
appropriate substitute would have to be analyzed for each affected source category. Switching
to a low-mercury fuel in the industrial/commercial and utility boiler sectors could achieve
October 20, 1995 3-27 DRAFT: Do Not Cite or Quote
-------
not be applicable in all circumstances, the estimates of potential reductions and costs should be used
only for relative comparisons among the source categories to give an initial indication as to where
mercury reductions would provide the most risk reduction for the least cost.
The estimates of cost for mercury reduction do not illustrate two important considerations.
One is that all of the cost of control is attributed to mercury removal. As described previously in this
Report, many of these controls achieve reductions of other pollutants as well. The benefits of these
additional reductions should also be considered. Second, the technologies available for mercury
control represent relatively new applications of these technologies. Thus, in the future, it is possible
that as new or emerging technologies develop, the cost-effectiveness of control may improve.
Table 3-6 also presents control costs in terms of dollars per U.S. citizen per year. This
information is presented to assess how the estimated national control costs on a per person basis
compare to estimates of what the average citizen might be willing to pay for reducing the mercury
pollution. The comparison is based on a survey of two thousand households in Minnesota who
indicated the willingness to pay about S100 a year to reduce the effects of toxic air pollution and
make more of their state waters fishable. Of this amount, about 5 to 10 percent was attributed to
mercury pollution, implying a value of about 7 to 13 dollars per year. Based on the rough analysis
presented in Table 3-6. the average cost per year to control these six categories comes to about S36
per person, with over 90 percent of that attributed to the cost of control of coal-fired utility boilers.
Of course, citizens elsewhere in the U.S. might be willing to pay either more or less than the
residents of Minnesota.
Research needs include the following. There is « need for fttfl-seaie testing of activated carbon
injection for utility beliefs. Pollution prevention measure* should be investigated. These could
include market-based sohrtion*, product labeling requirement*, tinaOt oa nevuses and incentives for
energy consumption.
3-«*>
October 20, 1995 3-32 DRAFT: Do Not Cite or Quote
-------
relevant to small percentages of the U.S. population. That
may be true, but I betieve that lack of such •magnitude* words leave readers with
incorrect u&pccssiou about the hasardi of mexhylmercury.
On page 3-3. third paragraph, FDA Identified hair mercury data from NO AA that would
be useful for a reality check on exposures in the U.S. Although EPA has referred to the
data, inrinded (ha> mean value for the study, and acknowledged the need for a broader
study in the U.S., they did not mention the maximum, value found (6.3 ppm is much less
-------
:C~--;E-95 *ED ":r _ 'JS EPA ECAO CINCINNATI = -4
tx)* VLB Jiyfa p.aa^L-
Pag« 2
than values suggested in the WHO table), or describe the shape of t£e distribution.
Instead, they questioned the reliability of tho dati since mean hair levels for consumers
vs. non-consumers showed little difference. The hair levels tor the entire group were
found to be distributed log-normally. Since food consumption amounts tend to be log- '
normally distributed, hair mercury data reflective of fish wasnmptlon. and mercury
cxpojurcs should alao be distributed log-normally. FDA's estimates of exposure to '
raethylmercury (computed using Monte Carlo analysis and 14-day fish consumption
figures) yield values that closely approximate exposure levels that would be predicted by
from the NOAA hair data. Although FDA's exposure analysis has not been published,
Alan Slwn'J analysis (published and peer reviewed) yields exposure figures that are
csscntiafly identical to oar own analysis. The bottom line is that NOAA's hair data are
a useful reality cheek on U.S. exposures and exposure estimates (published and
unpublished) developed using approaches that were more sophisticated than EPA's agree
with the rejionablenea of the observed hair data.
Ou page 3-5, paragraph 5, with regard to the reanalysis of (he Nlgaa poisoning. gDA
was not rtatiiig its *?cilViflfl WJtfl TCTUri tP Httfnfttinf the, fhreshcrfd doSCi TTie nncition
that pf ^h* aqihon. Tliis th<*"^ be COUTH inl
On page 3-6 (now J-8), last fall paragraph, line 5, ..in particular, women of child-
bearing ace who routinely **Mmn* la exeaM of 100 gram/day are considered .......
1
On page 3-8, (formally 3-f) second paragraph, EPA needs to provide exposure percentla
figures when discussing the intake oftmethyimercury by children, otherwise the reader
is left with the impression that children in general have these elevated exposures, not just
those eating at the highest percentflaa during 3-day survey periods.
On page 3-18, fourth paragraph (DOW flve), FDA was concerned that EPA adequately
delineate limitations of the risk characterization, significant uncertainties aimriafrd witb
nt» of the TT*TJ study for iralrtiH'fag an R£D, and uncertainties associated with their
exposure analysis (see above). EPA must indicate at the end of the first sentence that
"there are significant uncertainties associated with the- Irani study that were used to
establish the benchmark dose and the RfD. Among tfacao an the small population cue,
particularly at low exposure levels anndatgd with those in fish eating populations, and
differential MoavajtahiHty pf nifthyimercury in grain versus fUb.
cc:Bolger, Xsnse
TOTRL P. 03
-------
Memorandum
To: Addressees
From: Kate Mahaffey, Rita Schoeny
Subject: Response to comments from Greg Cramer dated 10/24/95
On 10/25/95 Rita Schoeny, Richard Jackson and Sheila Newton met by phone to
discuss responses to the comments on the revised Volume I of the U.S.EPA's
Mercury Study Report to Congress. Rita Schoeny and Kate Mahaffey subsequently
met on 10/30/95 to discuss proposed revisions and further disposition of comments.
Please note that changes described below and on the attached marked copies will be
made after the Report is reviewed by OMB and before it is submitted to Congress. It
is anticipated that the Report will be transmitted by EPA to OMB on or about 11/3/95.
Response to comments.
P 3-2, second par. We have revised text on the current page 3-1 to indicate that 300
g/day was an estimate offish consumption provided by Harada et at (1995) and that
this level is not typical of the U.S. population.
P 3-2 [3] third par. Magnitude of exposure is treated as for comment above with text
on current p 3-1 which gives measured levels of mercury in fish and estimate of
amount of fish consumed by the Japanese population.
P 3-5 third par. pg. 3-5, third paragraph. Regarding the summary of data on hair
mercury concentrations submitted by FDA, we previously reached (after much
discussion) a compromise under which we would provide the means for the data on
the persons who were said to consume fish and those who were said not to consume
fish. Our reservations about this data set can be reiterated.
These are unpublished data of unknown peer-review status.
These data are from an unidentified group who cannot be assumed to
be representative of the United States population.
The data show a very slight increase in hair mercury concentration in
response to consumption of fish. This observation varies from a
substantial body of peer-reviewed literature that associated increase in
hair mercury concentration with increased fish consumption for a large
number of populations in a number of different countries around the
world.
Under many circumstances it would be possible, as a further compromise, to provide
• an indication of variability in the data set by providing an error estimate. This would
\
^ \
"X •'•
^ V
\o\
-------
be standard practice in describing survey data. However, in the information we have
received from FDA no error estimates were provided.
We are highly reluctant to imply further acceptance of these data by adding the
maximum value reported for all of the above reasons, plus the following.
To assume that the maximum value reported in these data is
representative of a maximum for the United States population is
scientifically indefensible.
To provide this maximum value for hair mercury delivers an
inappropriate message to anyone concerned about persons, whether
from the general population or specific subpopulations, who consume
substantial quantities of fish.
To list a maximum value from an incompletely described data set that is
inaccessible to the general public is a disservice to risk communication.
P 3-5 fifth par. Attribution of the threshold estimate is made on current p 3-7. The
EPA description matches what is in the tables of the publication by Kinjo et at. 1995.
P 3-6, second par. Percentiles given on current p 3-8.
P 3-6, last par, line 5. Revision made on current p 3-8.
P 3-18, fourth par. We stand by our assertions that the uncertainties in the risk
estimate are described appropriately in our risk characterization volume (Volume VI.
These descriptions are reflected in Volume I. We have described in Volume VI that
use of a benchmark modeling procedure which uses the whole data set in to
determine the shape of the dose response curve removes the emphasis on the small
number of responders in the lowest dose grouping. We have also described the state
of the data on absorption of methylmercury. We have found no basis for assuming
that absorption from grain differs from absorption from fish; our external peer
reviewers agreed with this interpretation.
-------
3. FINDINGS OF THE MERCURY STUDY REPORT TO CONGRESS
This chapter integrates the findings and conclusions drawn from the analyses described in full
in the Report in order to present the risk manager with the information needed to make informed
decisions. A series of regulatory issues are presented and specific actions are recommended in
subsequent chapters. A brief listing of research needs follows each section defined by the questions
below: these are described in more detail in the Supplement to this volume .
The following questions are posed to summarize the risk characterization and guide the
discussion of regulatory issues:
1. What effects on human health can methylmercury have?
2. How much methylmercury is harmful to humans?
3. How do U.S. EPA's estimates of how much methylmercury is harmful to humans
compare with levels calculated by the U.S. Food and Drag Administration , WHO
and the states?
4. What is the estimated size of the population at risk?
5. How much methylmercury exposure is harmful to wildlife and what are the effects?
6. What are some of the limitations of the risk analysis? /_ H> s W^ fr A ~
7. Which sources contribute to these impacts? //f-t~ ln>*u"Ju, tliuit axe iiui eatgnsriTir fish
cons
3-1
-------
35 to just over 200 days reported for different individuals. The half-lives for pregnant
women have not been directly measured. Half-lives for women during lactation are shorter.
possibly due to excretion of mercury into milk produced during lactation.
In 1969, in response to the poisonings in Minamata Bay and Niigata. Japan, the U.S.
FDA proposed an administrative guideline of 0.5 ppm for mercury in fish and shellfish
moving in interstate commerce. This limit was converted to an action level in 1974 (Federal
Register 39. 42738, December 6, 1974) and increased to 1.0 ppm in 1979 (Federal Register
44, 3990, January 19> 1979) in recognition that exposure to mercury was less than originally
considered. In 1984, the 1.0 ppm action level was converted from a mercury standard to one
based on methylmercury (Federal Register 49. November 19, 1984).
The action level takes into consideration the tolerable daily intake (TDI) for
methylmercury, as well as information on seafood consumption and associated exposure to
methylmercury. The TDI is the amount of rnethylmercury that can be consumed daily over a
long period of time with a reasonable certainty of no harm. FDA (and WHO) established a
TDI based on a weekly tolerance of 0.3 mg of total mercury per person, of which no more
than 0.2 mg should be present as methylmercury. These amounts are equivalent to 5 and 3.3
u.g, respectively, per kilogram of body weight. Using the values for methylmercury, this
tolerable level would correspond to approximately 230 fig/week for a 70 kg person or 33
Hg/person/day. The TDI was calculated from data developed in part by Swedish studies of
Japanese individuals poisoned in the episode of Niigata which resulted from the consumption
of contaminated fish and shellfish and the consideration of other studies of fish-eating
populations.
Based on observations from the poisoning event later in Iraq, FDA has acknowledged
that the fetus may be more sensitive than adults to the effects of mercury (Federal Register
44: 3990, January 19. 1979; Cordle and Tollefson, 1984, FDA Consumer, September, 1994).
In recognition of these concerns, FDA has provided advice to pregnant women and women of
child-bearing age to limit their consumption of fish known to have high levels of mercury
(FDA Consumer, 1994). FDA believes, however, that given existing paatterns of fish
consumption that few women eating such high mercury fish (Jess than 1%) will experience
slight reductions in the margin of safety. However, due to the uncertainties associated with the
Iraqi study, FDA has chosen: not to use the Iraqi study as a basis for revising its action level.
Instead, the FDA has chosen to wait for findings of prospective studies of:
populations Jn^fne? SeycheUe Islands and in the Faroe Islands.
The World Health Organization's International Programme fof Chemical Saiety
(WHO/IPCS) concluded that the general population of adults (male/and non-pregnant
females) does not face a significant health risk from methylmercujy when hair mercury
concentrations are under 50 (ig mercury/gram hair. Reports-front the Niigata epidemic of
Minamata disease fewd^ower thresholds with mean values in the range of 25 to
approximately 50 ug me'rcuiy/giam hail. - I <, Tu<>/ fr\jTV^/^-S ftEPo&~£0
Clinical observations in Iraq suggest that women during pregnancy are more sensitive
to the effects of methylmercury with fetuses at particularly increased risk. The WHO/IPCS
3-7
-------
(1990) analyzed the Iraqi data ana identified a 30 percent risk to the infant of abnormal
neurological signs when maternal hair mercury concentrations were over 70 ug:g. Using an
additional statistical analysis. WHO 1PCS estimated a 5 percent risk of neurological disorder^ EM
in the infant when the maternal hair concentration was 10 to 20 us mercury/gram of hair.
WHO/IPCS recommended that as a preventive measure, in a subpopulation that
consumes large amounts of fish (for example, one serving or 100 grams per day), hair levels
for women of child-bearing age should be monitored for methylmercurv.
The WHO/IPCS estimated (1990) that a daily methylmercury intake of 0.48 ug
mercury/kg body weight will not cause any adverse effects to adults and that a methylmercury
intake of 3 to 7 ug/kg body weight/day would result in a <5 percent increase in the incidence
of paresthesia in adults. Risk to this extent would be associated with hair mercury
concentration of approximately 50 to 125 \ig gram hair. By comparison, the U.S. EPA's
reference dose, or the amount of methylmercury any person (including children and pregnant
women) can ingest eVlry day without harm is 0.1 ng/kg body weight per day. This was based
on a benchmark dose equal to 11 ppm (>g/g) hair. Children have a higher exposure to
methylmercury (on a per kg body weight basis) than do adults. The U.S. EPA estimates in
this Report that jntake of methylmercury by children ages birth through 14 years residing in
the continental 'U.S. is 0.84 ug/kg body weight/day for males and 0.81 ug/kg body weight/day
for females.r- V^. ^ /T^ ff// 7>£ftci/uT/<-£. J
A .research need for this area is for biological monitoring (for exposure and effect) of
populations with either greater than U.S. average fish consumption (such as one serving or
100 grams per day) or consumption of fish predicted or measured to have higher than average
amounts of methyirnercury.
What is the Estimated Size of the Population at Risk?
The people in the U.S. who consume fioh in occcoo of 100 gramo/day (e.g., one
serving) ore considered to bo the population at riolc. In particular, women of child 'bearing age
and children aged 11 and younger arc of concern became of the adverse effects of
methylmercury on the developing nervous system. i ._ S-O\JT/^^^ cotJ^uMtz
The general United Slates population fliapdoteintheir fish by purchasing them//o-o ^ *f ^
commercially: is iiot believed to be advecsely/jrrected by raethyunercury if they consume fis&j ^
in raoderatioik: Hie people hi the U.S. wh0 routinely consume fish in excess of 100
grams/day (e.g.» one serving) are considered to be a population potentially at risk. In
particular, women; of child-bearing age^te considered of concern because of the adverse
effects of methylmercury on the developing nervous system. Children ages 14 and younger
are of concern because they have higher exposures to raethyirnercury from fish on a per
kilogram body weight basis compared with adults.
Women of child-bearing age rather than only pregnant women are a group of concern for two
reasons. The first is that methylmercury persists in tissues. Measured half-lives of methylmercury in
adults range from about 1 month to 9 months, although half-lives of just over 2 months are usually
3-8
-------
page numbers refer to draft dated 31 October 1995
pg. 3-1, last paragraph.
Rewrite to read:
"Neurotoxicity is the effect of greatest concern when adults and children are exposed to
methylmercury. Two major epidemics of methylmercury poisoning through fish
consumption have occurred. The best known of these two epidemics occurred among
people and wildlife living near Minamata City on the shores of Minamata Bay, Kyushu,
Japan. The source of methylmercury was a chemical factory that used mercury as a
catalyst. Aa series of chemical methods identified methylmercury in the factory waste
sludge; this was drained into Minamata Bay. Once present in Minamata Bay,
methyimercury accumulated in the tissues of shellfish and fish that were subsequently
consumed by wildlife and humans. Fish was a routine part of the diet in these
populations. An average fish consumption for this population was reported to be in
excess of 300 g/day (reviewed by Harada, 1995). This is a greater level of fish
consumption than is typical for the general U.S. population."
pg. 3-2, 2nd full paragraph.
Rewrite to read:
"Over the next 20 years the number of people known to be affected with what became
known as Minamata disease increased to thousands. In time the disease was
recognized to result from methylmercury poisoning, and fish were subsequently identified
as the source of methylmercury. In 1961 measured levels of mercury in the fish ranged
from 10 to 20 ppm. As is often the situation with epidemics, the first cases noted were
severe. Deaths occurred among both adults and children. It also was recognized that
the nervous system damage could occur to the fetus if the mother ate fish contaminated
with high concentrations of methylmercury during pregnancy. The nervous system
damage of severe methylmercury poisoning among infants was very similar to congenital
cerebral palsy. In the fishing villages of this region the occurrence of congenital palsy
due to methylmercury was very high compared to the incidence for Japan in general.
After the source of mercury contamination was identified efforts were made to reduce
the release of mercury into the bay. After 1969 average mercury concentrations in fish
had fallen below 0.5 ppm."
pg. 3-4; line 7.
Rewrite to read:
"It should be noted, however, that these were substantial developmental delays. For
example, a child being unable to walk two steps without support at two years of age,
inability to talk based on use of two or three meaningful words by 24 months, or
presence of generalized convulsive seizures."
-------
pg. 3-21; fourth paragraph.
Rewrite to read:
"As with environmental monitoring data, information on body burden of mercury in
populations of concern (blood and/or hair mercury concentrations) are not available for
the general U.S. population. Data on higher-risk groups are currently too limited to
discern a pattern for the U.S. population that is more predictive of methylmercury
exposure than information on quantities of fish consumed. The selenium content of
certain foods has been suggestive as a basis for modifying estimates of the quantities
of mercury that product adverse effects of methylmercury. Currently data on this
mercury/selenium association form an inadequate basis to modify quantitative estimates
of human response to a particular exposure to mercury."
pg. 3-24, paragraph 3; beginning line 8.
Rewrite to read:
"limitations on monitoring data providing mercury concentrations in the aquatic food
chain. The results of modelling described in this report deal with freshwater bodies in
the contiguous United States. These findings do not estimate the source of mercury for
marine fish which is of international origin. The weight of evidence supports an
association between local mercury emissions and increasing mercury levels in locally
caught fish. Variables contributing to this linkage include the following: the species of
mercury that are emitted; the overall amount of mercury emitted from a sources; the
extent of mercury methylation in the water body; and the climate conditions. For fish
collected at sites remote from the source, it is not possible to assess the contribution of
any particular sources. A conclusion of this Report, however, is that the majority of the
mercury deposited in the U.S. is from anthropogenic sources.
pg. 6-1, first paragraph, line 8.
Rewrite to read:
"predicts that increased mercury deposition could lead to increased levels in fish, and
that increased levels in fish will lead to toxicity in fish-eating birds and mammals,
including humans."
-------
DEPARTMENT OF HEALTH & HLMAS SERVICES
^ — ^-
7 1995 A;ierv.aGA303-ii-3~2-i
MEMORANDUM
TO: The Assistant Secretary for Health
FROM: Chair, Ad Hoc Panel on Methyimercury
Environmental Health Policy Committee
SUBJECT: Mercury Stuffy Report to Congress from the
Environmental Protection Agency
In the Clean Air Act Amendments of 1990, a requirement was made of the Environmental
Protection Agency (EPA) to produce a study of mercury emissions, including the sources, rates
and amounts, control technologies and their costs, and health and environmental effects. The
study is completed and is due to be transmitted by court order to Congress on December IS,
1995.
As emphasized in the mercury emissions report, most humans' non-occupational exposure to
mercury from the environment will be in the form of methyhnercury consumed in contaminated
fish. The greatest potential risk to human* appears to be to the fetus in pregnant women
consuming more than one fish meal dairy, «j*p^«vting on the level of contamination in the fish
consumed. Issues raised by the nature and magn'tiy** of human health risk from fish containing
methylmercury are clearry within the province of the Food and Drug Administration (FDA).
Recognizing this, EPA brought these issues to the attention of FDA officials, who had some
concerns about the report. Specifically, FDA had made a decision to postpone review of their
action level for mercury in commercial fish pending publication of results from studies in progress,
and officials were concerned about the apparent disagreement embodied by EPA's decision to
proceed "«"g existing data that FDA had decided was insufficient. EPA's decision to use the
data was at least in part driven by the court-ordered release date of December IS. On a more
level, FDA officials feh that the EPA report, as written, might raise a level of public
concern that, in their view, was not justified by the magnitude of the apparent risk. In order to
address these difference*, EPA and FDA officials agreed to ask the HHS Environmental Health
Policy Committee (EHPC) for assistance in evaluating the studies upon which the EPA's risk
characterization was based and interpreting the results of those studies for the development of
appropriate public health policy.
-------
Page 2 - The Assistant Secretary for Health
A panel of scientists was formed, chaired by me and including members from NIH and CDC,
which met on several occasions with FDA and EPA scientists and policymakers as well as with
the authors of key studies in progress. The charge to the panel (attached) consisted of five
questions, all having to do with the quality and interpretation of data from human studies. The
first three questions focused on technical issues; after some deliberation, the panel members felt
that, in view of the extremely limited time available to them and the many months' work that had
already gone into the development of the report, a complete review of the health effects section
was beyond the panel's scope.
Most of the panel's work focused on question four of the charge, which dealt with the report's
characterization of risk associated with fish consumption. At the panel's request, numerous
revisions have been made in the executive summary (Volume I) of the EPA report reaffirming that
the U.S. fish supply is considered to be safe at typical levels of consumption and emphasizing that
the EPA's calculations of the benchmark dose and reference dose (RfD) for methylmercury
contain significant uncertainties which are unlikely to be resolved until additional data are
available. Several research studies that are currently underway are likely to provide such data.
Among these are a large study in the Seychelles Islands funded by FDA and NIEHS and another
NIEHS-funded study in the Faroe Islands of Denmark (both are study populations with high
levels offish consumption).
In recommending its revisions to the EPA Report, the PHS/EHPC panel does not mean to imply
that there are no risks associated with methylmercury exposure; rather, that the great majority of
the U.S. population is not in danger of consuming harmful levels of methylmercury and should not
be advised, directly or implicitly, to reduce their intake offish. On the other hand, the panel
supports EPA's goal of bringing to public notice, through this report, the potential risks to human
health of continued and increasing mercury emissions.
Question five of the charge, concerning the development of the public health message that should
accompany the release of the report, is still the *bcus of the panel's active participation in
collaboration with both EPA and FDA. The panel has taken on (from the Agency for Toxic
Substances and Disease Registry) an additional PHS member with risk communication expertise
and will be working with EPA and FDA officials in the development of public health messages
targeted to populations with different levels of risk and the formulation of answers to expected
questions from the public and the press. This continuing activity will be critical to achieve
understanding between agencies on the content and tone of the public communications and to
permit the Administration to present a consistent message on the human health risks of
methylmercury exposure to the U.S. population.
Throughout this process, both the EPA representatives and the FDA officials have acted in good
faith on the panel's recommendations. I and the other panel members have been most impressed
by the willingness of the EPA scientists to work with the PHS/EHPC panel to craft a report
summary that reflects in good measure the consensus views that emerged during this process.
While this remains an EPA report, the final version of the summary should be one that the PHS
-------
Page 3 - The Assistant Secretary for Health
can support and that can form the basis of a balanced public health/risk communication message
to the American public.
I will continue to keep you and the EHPC informed of our progress on this important public '-
health issue.
j)Q****- --
Richard J. Jackson; MD, MPH
Director
National Center for Environmental Health
Attachment
-------
SECTION TWO
REVIEW OF THE NOVEMBER 1995 DRAFT (OMB SUBMITTAL) OF THE
MERCURY STUDY REPORT TO CONGRESS
At the time of this review , the Mercury Study Report was to have been submitted to
Congress by December 15. 1995. A number of comments were addressed by the EPA during
the time just prior to December 15th. The comments and responses included in this section
represent the comments on and changes made to the OMB review draft. This section
contains the following in order:
1. National Marine Fisheries Service comments dated November 28, 1995.
2. Department of Energy comments dated December 7, 1995.
3. Council of Economic Advisors (CEA) comments dated December 11, 1995.
4. EPA response to December 11, 1995 CEA comments.
5. Office of Science and Technology Policy (OSTP) comments dated December 11, 1995.
6. EPA responses to December 11, 1995 OSTP COMMENTS.
7. FDA comments dated December 14, 1995.
8. NOAA comments dated December 21, 1995.
-------
,AA
UNITED STATES DEPARTMENT OF COMMERCE
National Oceanic and AtmoapnaHc Administration
NATIONAL MARINE F'SHERIES SERVICE
"335 East-West M
-------
The promulgation of a lower RfD, with the attendant
perception of an increased health risk from methylmercury in
fish, could have serious and unnecessary adverse consequences for
the fisheries industry and seafood consumers. Furthermore, EPA
has used data from a U.S. Department of Agriculture consumption
survey that covered only a 3 -day period to estimate consumer
exposure to methylmercury from fish, although it is widely
recognized that such short-term surveys do not appropriately
capture fish consumption. EPA's report does not define the
population subgroups that may be at risk, and does not
distinguish between, for example, occasional consumers of
commercially caught fish and subsistence fishermen consuming
freshwater fish that might be impacted by mercury emissions.
Additionally, we note that EPA has extrapolated the impact of
mercury emissions on freshwater fish to marine species with no
obvious supporting information.
Although the report to Congress contains many caveats and
qualifiers relative to the uncertainty surrounding its
conclusions, NMFS remains concerned that the public may become
unduly alarmed and perceive the report as recommending against
the consumption of fish. Some of the data and methodology are
sufficiently suspect that further examination and clarification
should occur prior to submission to the Congress. We believe
that an interagency review of outstanding issues is imperative.
Sincerely,
Rolland A. Schinitten
Assistant Administrator
for Fisheries
-------
Department of Energy
Washington, DC 20585
CEC 7 1995
Mr. Kevin Neyland
Office of Management and Budget
725 17th Street N.W.
Washington, D.C. 20503
Dear Mr. Neyland:
Thank you for me opportunity to review the draft "Volume 1: Findings and Recommended
Actions from the Mercury Study Report to Congress."
The Department's Offices of Policy and Fossil Energy were provided Volume I of the seven-
volume study for review. We recognize that information gaps which we highlight in the attached
comments may be addressed in other volumes. However, based on the Volume I summary, we
believe that important uncertainties remain about the estimations of human health risk and that
data for measuring emissions by source can be significantly improved. Moreover, concrete
options for improvement are available to EPA within a limited time period and budget. We rote
that:
1) relevant peer-reviewed and published information on human exposures and effects will
be available from the Seychelles Islands study in one to two months and
2) EPA did not ask either the Clean Air Science Advisory Committee or an interagency
peer review group to review the current draft
This study is important to consumers, health providers, natural resource managers, fisheries, and
industry. It interacts with the forthcoming Electric Utility Study and the Great Lakes Initiative.
For these reasons, it is important to produce a respected study. We urge EPA to delay this
report for the purpose of incorporating imminent information and providing more thorough
review.
If you have questions regarding the attached comments, please call Beth Campbell at 586-5590.
>an W. Reicher
Acting Assistant Secretary for Policy
-------
• O2
Comments on EPA's Mercury Study
Section 2. Mercury in the Environment
a) Page 2-1 indicates that anthropogenic emissions of mercury contribute from 40 to 75 percent
of the current total annual input of mercury to the atmosphere, while page 3-24 states from 50 to
75 percent. What is the correct value?
b) An article has recently appeared (November 12,1995) in the Detroit News regarding levels of
pollutants in Lake Superior. Specifically, the Chippewa-Ottawa Treaty Fishery Management
Authority indicated that the contamination by mercury has declined by 36-65 percent since 1992.
EPA should verify these data in light of the mercury levels reported in the Mercury Study, sources
of mercury, and subsequent predictions about the future levels. Also, more recent information
should be obtained on mercury concentrations in the other great waters.
For this section as well as a subsequent discussion in Section 3 about whether the problem is
getting worse or better, data from EPA's Great Lakes Initiative may be useful.
Section 3. Findings of the Mercury Study Report to Congress
a) As indicated on page 3-5, more relevant information on human exposures to methylmercury is
forthcoming from controlled studies (Seychelles Islands, Faroe Islands, and the U.S.). Results
from these studies have been presented at national meetings, and peer reviewed journal
publications are anticipated in January-February of 1996. The EPA should utilize the data
presented in 1995, or delay release of the Mercury Study until the results are formally published.
The data from the controlled studies will significantly reduce some of the uncertainties
surrounding the RfD and the benchmark dose for human exposure to methyimercury. Precedent
for such a decision exists. Unpublished results from the FDA are quoted (page 3-6) in Supplement
A to strengthen associations between fish consumption and mercury concentrations in human hair.
EPA should also use similar unpublished data from the Seychelles and Faroe Islands studies to
verify the RfD for methylmercury.
b) What is the methylmercury threshold for humans? Toxicologkal exposure levels must be
paired with population exposure data to be meaningful. That is. to establish the dimensions of the
problem it is necessary to look at the intersection of the percent of the population that actually
encounters the conditions defined to be exposure exceeding the safe level. EPA has attempted to
do this in Table 3-2, p. 3-17, but the data presented there are not in the same format as the
discussion of RfD, NOAEL, and LOAEL. Instead the data are presented in 3-day exposure
format rather than one-day exposure. Moreover, it is not explained how the highest 5% of the
fish-consuming population relates to the RfD estimates. A better analysis of the data will be
necessary to get a clear picture of the extent of the methylmercury exposure problem.
c) EPA identifies women of child-bearing age, children 14 years and younger, and certain
populations with high exposure levels as three high risk categories. On p. 3-9 (end of first
-------
P.O3
paragraph) EPA states that"... approximately one-half of the general U.S. population of women
in child-bearing age who consume fish at least once within a 3-day period are predicted to ingest
methyl mercury at the RfD. The RfD is expected to be a level of exposure to be without adverse
effect." Does EPA mean to sav that nne-htilf will exceed the RfP^ The argument for concern is
different in that case, but even then we do not know to what extent the exposure exceeds LOAEL
and we do not know whether the RfD is the actual threshold for fetal adverse impacts. Again, use
of a 3-day exposure estimate with RfD is inconsistent with the estimate EPA males about one-
half the population of women.
d) In three places (Section 2. Mercury in the Environment, p. 3-9, and p. 3-19) the report states
that fresh waters, and particularly surface waters already impacted by acid deposition, appear to
be greater sources of exposure to methylmercury. The map in Figure 3-7 shows a mercury
deposition pattern which is most serious east of the Mississippi. Does the geographic
consumption of freshwater fish correspond to the geographic pattern of acid and mercury
deposition? The section "What is the Estimated Size of the Population at Risk" mixes information
on all fish and shellfish consumption and freshwater fish consumption in a manner which is
confusing. A table with population numbers (corresponding to Table 3-2) for freshwater fish
consumption would be useful at the national level but regional freshwater fish consumption data
would be even more useful.
Overall. EPA's document could do a better job of clarifying the distinction between high exposure
categories of the population (due perhaps to exposures to high-dose fish or the pattern of fish
consumption) and higher susceptibilities (due to characteristics of the individual such as age or
fetal development). There would be great value, as EPA states, in additional information about
the susceptibility of children and fetuses and information about groups with higher exposures,
such as subsistence fishers.
Section 3. Subsection: Which Sources Contribute to these Impacts?
a) On this issue the Department of Energy has an obvious interest Since no comprehensive
program was established for mercury under section 112 pre-1990, it is our suspicion that the data
are at a minimum incomplete and consequently misleading. It is likely that the inventory of
emission sources is more complete and accurate for some sources that for others - thus very
possibly presenting a misleading overall picture of sources. Highly regulated sources such ax
electric utilities can reliably estimate their emissions using emission factors, energy input data, and
coal seam type information. Less regulated sources such as area sources or even large sources
with inconsistent or untraceable materials inputs (such as municipal waste combusters) are likely
to underestimate their emissions. Together information from this mix of sources would present a
biased picture. We note that if data are obtained from the Toxic Chemical Release Inventory,
where the enforce ability of under-reporting is questionable, a similar systemic error is likely.
Since TRI data begin in 1989, they cannot be used for data over a long period.
Moreover, knowing the emissions by source category does not indicate the actual sources within
that category. It may be that utility mercury emissions are distributed in a rough equality among
utility units, or it may be that a small number of units emit a disproportionate amount of mercury.
-------
L2.13.93
! 006
U
.39 FAX 202 280
2023954S33-
::. s
EXECUTIVE OFFICE OF THE PRESIDENT
COUNCIL OF ECONOMIC ADVISERS
WASHINGTON. Q.C. 20606
December 11, 1995
MEMORANDUM FOR T.J. GLAUTHIER, ASEQCIATB DIRECTOR
NATURAL RESOURCES.«>/SNER6Y, AND SCIENCE, OMB
FROM: ALICIA MUNNELI$ns£MBER-NOMINZE
SUBJECT: Need for Delay of EPA Clean Air Act Mercury Report
EPA has produced a huge report that involves a. number of
significant scientific and economic issues of concern to a number
of agencies. As currently written, the report could ba
interpreted as suggesting that the nation's fish supply poses a
significant public health threat, notwithstanding the health
benefits of increased fish consumption. The report also could be
read as providing a rationale for a set of very expensive end-of-
pipe control measures to reduce risks whose magnitudes are not
yet well understood. A sixty-day delay would provide time for
further agency review and interagenoy discussion to address these
issues. My understanding is that ZPA has the ability under its
draft consent agreement with the. Sierra Club to obtain such a
delay without the need for Sierra Club approval.
I also understand that OSTP is developing a detailed set of
recommendations for addressing some of the outstanding scientific
questions. The recommendations include the use of new
information that may affect the setting of the "reference dose,"
and recommendations for presenting the risk information in a vay
that indicates Its reliability and its applicability to the
general population or specific cubpopulations. CZA concurs with
the need to address these scientific questions. Squally
important is the need to put the conclusions of the report in a.
proper economic and policy context, even though the report
correctly notes that its purpose is not to fully analyse or
recommend policy actions. I have attached some suggested
revisions prepared by Mike Toman that address these concerns.
I believe that a number of agencies favor this delay. If
EPA is opposed to it, than we need to have a deputies-level
meeting Jto discuss the dispute. Please let ma know how this
issue develops, and'do hot hesitate to contact us if you have any
questions about our views. Thank you.
-------
21007
FAX 202 260 9768^ ^ 2_g
2- 13-95 *HD U:59 FAX 20 ^ ^ 2_g5 , ^^ _ 2023954339-
Comments on Summary Volume of EPA Mercury Study (Mika Toman, CEA)
(NOTS: While these comments are directed at the summary ^volume,
they also apply to the corresponding material in the rest of the
report.)
(1} page 3-25: this eaetion is titled "Management Alternatives
for Reducing Mercury Emissions," presumably in keeping with
section 112 (n) (l) (B) . The discussion does note that the analysis
of control strategies and coats is not intended to replace a
thorough regulatory analysis. In keeping with that point, the
discussion also should note that emissions control is only one
possible means for risk control. Measures that reduce human
exposure is another alternative for reducing threats to human
health (this measure would not ameliorate negative ecosystem
effects, but these do not seen.to ba that serious). Another
approach would involve reduced use of high-mercury coal. The
report needs to avoid becoming inadvertently disingenuous by
focusing on control strategies because that's what 112(n)(l)(B)
askad for, when in the larger context other ri*k management
strategies need to be evaluated as wall.
(2) page 3-28: the discussion of dollars per citizen in
connection with Table 3-6 needs 'to be completely revised.
contrary.to what is suggested later in the text, coat per ton
removed is a perfectly adequate description of cost-effectiveness
and should be presented. SPA then can argua that even high-cost
reductions are justified, but that argument has no place in this
report given its scope (or if it is to be included, it needs a
great deal more attention). Even if one were looking to compare
costs and benefits, costs per citiaan ia not very informative.
There is no information provided in Volume l with respect to
the Minnesota contingent valuation survey of willingness to pay
for reduced mercury, and the discussion elsewhere in the report
ia minimal. Was it peer reviewed? Axe thara other studies that
can be compared? At this juncture, EPA should simply report
costs per ton removed and leave benefit valuation for another
study.
The discussion of research needs at the bottom of page 3-28
also focuses solely on prevention measures. It needs to be
deleted or made more balanced by addressing, for example, the
nead for research oh how exposure limits might be structured to
reduce the number of potentially exposed parsons falling through
the cracks.
(3) page's-1: "Consistent with the above comments, point #3 in
this list should be delated. The same probably ia true at |4 —
not because the points are necessarily wrong, but because they
arise out of context here and go well beyond the charge.
-------
2-,2-M
2023354333-
32805155IS
(4) page 6-1 j Hera EPA states action* it intenda to take as a
result of this study, having previously acknowledged that this
study itself does not provide a full analytical basis foV drawing
policy conclusions. This whole section is problematic in terns
at going beyond the charge. It would be better if the report
simply summarized the health risk conclusions and caveats reached
(to the extent this ia not already done previously in the voluae)
and then discussed hew this analysis fits together with other
ongoing efforts, rather than couching the discussion in terns of
recommended actions.
Copy to:
Jack Gibbons, OSTP
Xatie MeGinty, CEQ
Sally Xataen, OXRA
-------
12/11/95 MON 10: 17 FAX 919 541J)j»40 EPA QAQPS REAG
to«42
EXECUTIVE OFFICF OF THE PRESIDENT
COUNCIL OF 6CONOMIC
SENIOR ECONOMIST
DBCEMHER4, 199S
MEMORANDUM FOR: KEVIN NEYLAND, OMB
FROM:
SUBJECT: EPA HG STUDY
I wffl not have much time to review the icience in the report, so these comments
ere dbvcted moitiy «t the coaduriooi end policy issue* raised fa the nmnuuy vohune.
However, I would note that some of the identified by Rick Belief at the briefing do seem
troubling. Why, for example, does EPA.uie an extreme measured risk value ead then
apply ilOx safety fector? A* Rick's question points out the RfP U Bgrrififiantiy leae
eotuerrdxve, even accepting the safety ftctoi, if ooe starts with tbeMLEmk. Iwouid
urge thnt eapebk p«jple in the white HMue do a sciub of the risk analysis in order to
enure **m it is fiifly cooaistent with our riik ptiDciplet, i&cbidfaig both good icigtK* eod
avoidance of uiqasdfled conservulsra. We do not mmidus report coming out ntsacKo
tUiuon tut it adds fuel to the fire about toe Afiniui'*^*^lrti*H commitment to baia&ced nsk
aasaumcnt. I would suggest that you ask for input from both OIRAaodOSTPud
IXtrmcg now to the conclusions, I have the fbDowing comment*
(1) page 3-25: thia sectioo is titled "Majagoneot Altenatives fcr Redwing Mercury
,' presumably in keeping wkfa Motion ll^nX^XB). The riimunon does note
that the aaatyns of control strategies tod costs is not iatoidtd to rephwtt thorough
regulatory anaiysii. Inteq^vgtehthcpcMtlgsten If one were looking to compare costs and benefits, cons per
citizen are not very informative. And there is no information provided in Volume 1 with
-------
L2m/95 ttON 10:17 FAX 918 341 0340 EPA OAQPS
Df?c:-frtii-tQq5 20:42 rca • I4J014
rwpect to the Minnesota survey. I* this survey discussed elsewhere in the report? Wash
peer reviewed? Are there other studies that can be compared? It seems to me that the
way out of this morass is tor EPA simply to repui i costs per toa removed and leave
benefit valuation for another study.
Th« discussion of research needs at the bottom of page 3-28 also focuses solely on
prevention measures. It needs to be deleted or made more balanced by addressing, for
example, the need tor research on how exposure tiimia might be structured to reduce the
number of potentially exposed persons fining through the cracks.
(3) p4g«S-l: as indicated abovt, point 03 in this list should be ddet«rl. Tharame
probably is true of #4 — not because the points are necessarily wrong, but because they
arise out of context here and go weft beyond the charge.
t
(4) page 6-1: Here EPA. states actions it intends to take as a result of this study, having
previously, acknowledged that this ttudy ttsclf does oot provide * fiifl analytic*! basis fcr
drawing policy conclusions. This whole section is problematic in terms of going beyond
the charge. It would be better if the repon simply summarized the health risk uoucluaions
and caveats reached (to the extent this is not already done previously in the volume) and
then discussed how this analysis fits together with other ongoing efforts, rather than
oouohiag th* diaeuscioa ia terms of recommended actions.
TOTflL P.003
-------
EPA Responses to CEA Comments (12/11/95) on the Draft Mercury Study
This memorandum is in response to Alicia Munnet's comments on the draft
mercury Study which were transmitted to T.J. Glauthier dated December 11, 1995.
Specific comments offered by the Council of Economic Advisers (CEA) are addressed
below. CEA notes that while their specific comments are directed at the summary
volume, the comments also apply to the corresponding material in the rest of the report.
The EPA notes that where changes have been made in the Executive Summary, the
corresponding changes have also been made elsewhere in the report.
1. The discussion of management alternatives has been revised to note that emission
controls are not the only means for reducing human exposure. CEA also suggests
reducing the use of high mercury coal. Unlike sulfur, different types of coal do not vary
significantly in mercury content and as a result, this is not a considered a control option.
However, other management measures such as emissions trading, emissions
averaging and fuel switching are described in Volume VII of the report. The discussion
of management alternatives in the Executive Summary will be broadened to include
these other alternatives.
2. The discussion of dollars per citizen has been removed from the report as well as
the section in Volume VII on social costs of mercury pollution which included the
Minnesota contingent valuation study. The discussion of research needs will be
expanded - probably included as a separate section. However, the EPA is unclear
about CEA's suggestion about research on exposure limits and requests clarification.
3. Points #3 and #4 have been deleted.
4. The section on Recommended Actions has been deleted. It is now incorporated in
the Management Alternatives section and is characterized as ongoing activities rather
than recommended actions.
-------
21002
,-o -a F\T ''O0 280 9788 .
12 13 95 «ED LI.=3 FAi -u~ ;<2-'2-s5 I S:CiPW . 2023954635-
EXECUTIVE OFFICE OF THE PRESIDENT
OFFICE OF SCIENCE AND TECHNOLOGY POLICY
WASHINGTON, oc, zosoo
December 11. 1995
MEMORANDUM FOR: T.I. GLAUTHIER^
FROM: BOB WATSON '
SUBJECT: Comments on EPA Report on Mercury
The Office of Science and Technology Policy, Envirooment Division, was asked to review
a draft of the EPA Mercury Study Report that was produced at the request of Congress.
Section 112(n}(B) of the Clean Air Amendment* of 1990 requires a study of mercury
emissions from electric utility steam generating uniu, municipal waste combustion units, and
other sources, including area sources. The study wu to consider the rate and mass of such
emissions, the health and environmental effects of such emissions, technologies which are
available to control such emissions, and the costs of such technologies. The OST? is deeply
concerned of the release of this eight volume document at this tine. Our concerns stem
largely from the minimal peer review of the report and from substantial daca gaps that may
result in an incorrect depiction of human risk. The OSTP strongly recommends the following
courses of action, in order of preference. The preferred approach is option one, which calls
for a 60-day delay for an extended intaragency review. The Council of Economic Advisers
concurs with ihe recommendation to seek this delay in release of the report.
OPTION ONE; 6Q»pav Delay for an Extend^ latetagenev Review
Several agencies haye significant stakes in the outcome of this study; have contributed
essential data, or have undertaken parallel efforts. They should be included in the
extended comment/review process. It was clear at the intengency meeting on Friday,
' that agencies had hardly read any of the latest document and universally complained
about inadequate time to review. Further, most comments at the meeting focused on
the health aspects of the study because that is the area all agencies skimmed first.
However, we suxpect agencies have not had time to even begin to adequately review
the other chapters of the report; serious review of these chapters must take place.
Considering the importance and potential impacts of the Report, a review by the EPA
Science Advisory Board may have been quite appropriate. The absence of this kind of
peer review highlights the need for a concerted and meaningful review by at least the
other affected agencies. A delay would allow agencies to critique the complete report
and be satisfied they had a chance to be heard ud accommodated. This kind of buy-
in is important.
Given that the Seychelles data is in page proof*, the two month delay would even
ailo-wa preliminary consideration of that data. The University of Rochester
researchers could be invited to present their data at a workshop session and EPA could
examine the data for consistency with their proposed reference dose. We understand
-------
2003
12/13.95
202
2023I5463S-
there are also concerns about the emissions data. A delay of 2 month* would alao
allow an interagcncy workshop to discuss this chapter.
QPTTQN TWO: A Three to Six Month Delay ta Consider Important PendinP S^diej ^
One of the major concerns regards the adequacy of external participation and peer
review of the EPA study. Several leading researchers whose publications support or
refute basic assumptions of the report were not asked to participate. We recommend
that a more extensive external peer review of the Study is required. This review
should include Clarkson. Crump, and others who have made significant contributions
to the literature or have published studies which conflict with the Report's assumptions
or methods. This delay would allow a more complete assessment of the Seychelles
study, with the estimation of a reference dose based on all available scientific data.
Resonse Comment
the Peeee
Deadline
EPA should include responses to the issues raised by the other agencies including the
specific recommendations of OSTP provided below. However. OSTP believes that
this process cannot be completed and the response* returned to the agencies, reviewed,
and signed off on in the short time period available.
EPA should present the risk characterization in terms which are easily comprehensible
Co the general public. Because of significant uncertainties in the risk assessment and
the supporting data, risks may be expressed as categorical rather than numerical point
estimates, and should recognize the variability of the risk by demographics and
location. EPA could present its reference dose as a range which accounts for
uncertainty and the conflicting experimental and observational data (e.g., 0.1-0.4
ug/kg/day). Because of the relative consistency of date regarding mercury, this range
would be narrow enough for the reference dose concept to remain useful. EPA should
consider site and population specific risk assessments in the final report. Comparative
risks, and risks posed by typical exposure scenarios should also be included.
SPECIFIC RECOMMENDATIONS UNDER OPTION THREE
(To be addressed in the final report if the December deadline is not extended.)
The report should explicitly state that due to the limitations of the available data and
the range of uncertainty it should be regarded as a best estimate given limited date,
and clearly point out the need for additional research to more accurately quantify
anthropogenic mercury emissions.
Estimates should be provided to the number of significant digits to which reasonable
certainty exists. Some measure of dispersion (i.e., confidence limits) should
accompany these estimates. Alternatively, ranges should be provided.
-------
21004
FAI 202 280
The report identifies the difficulty of linking airborne emissions to environmental
mercury levels in a quantitative way. At best, anthropogenic emission combine with
naturally occurring "background" levels ID contribute to the total environmental
burden, but the proportion is unknown. This must be clearly reflected in the report
summary. fc
Because of the global transport of mercury in the environment, it is likewise
impossible to describe what proportion of environmental mercury in the United States
originates within the United States and what comes from sources abroad.
Some of the data on mercury levels in fish are twenty years old. Despite limited
indications that some of this data is comparable to contemporary levels, not all of
these data will reflect current levels of mercury and may be inaccurate because of
problems with older analytic techniques. Newer information is likely to more
accurately reflect current mercury levels and be more reliable because of improved
analytical techniques.
Mercury levels vary widely among fish species and geographic locations. Attempting
to estimate health risks while ignoring these differences may result in a significant
error in the riak estimation and a disruption of fish consumption and commerce in the
United States. These differences must be noted explicitly in the report.
The magnitude of the public health risk is a function of the mercury levels in fish and fish
consumption patterns. Because some U.S. jubpopularions are probably at significantly greater
risk:
EPA should make some attempt to identify vulnerable tubpopulations and analyze the
f^~\ risks to these populations including coastal populations, residents of the Northern and
^} Eastern U.S., residents of Alaska and Hawaii, Native Americans, and subsistence and
commercial fishers.
Many of the concerns about this report could be addressed1 by organizing the data in a
different manner.
EPA could present its reference dose as a range which accounts for uncertainty and
the conflicting experimental and observational data. Because of the relative consistency
of data regarding mercury, this range would be narrow enough for the reference dose
concept to remain useful. Based on this new reference dose range, the health risk from
methyl mercury could be stratified into categories of low, medium, or high.
The proportions of the US population felling into these categories, including the
differential assignment of sensitive subpopulatioot, could be calculated allowing some
rouglfcalcuiatioir of the numbers of individuals at risk. Similarly, fish species, stocks,
and geographic ranges could be likewise designated as low, medium, and high risk
based on tissue levels and consumption patterns.
-------
11005
12-i3.-95 *EI> U:59 FAX 202 280 97^2-12-S5 . 3;Q8pM , 2023954533- 32905'55;s
Maps of the United States could be Included in the summary report graphically
fummariziflg the risk from methyl mercury. One map could show the distribution of
mercury risk based on mercury levels in fish tissue independent of consumption
patterns. A second map could graphically depict fish consumption pattern} independent
/^ of fish levels. The third map would depict the risk hosed on the product of mercury
J^/ levels in fish and rates of consumption, and would effectively depict the populations
and location* at risk, This method of data presentation would allay the stated concerns
that all fish species and fish stocks would be unfairly characterized aa contaminated or
dangerous by the release of the Report. This type of risk characterization would also
avoid unnecessarily alarm among the majority of US consumers who are at low risk
while providing adequate notice to those populations which are at increased risk.
Acknowledging basic risk communication techniques .
EPA should include several scenarios which depict the risk in terms that are
comprehensible to. most individuals. These should be scenarios based on usual
consumption patterns and should describe the risk posed by eating various common
seafood on a daily, weekly, and monthly basis.
EPA should also provide a range of comparative risks, including consequences of
reduced fish consumption and increased beef consumption. This serves to put the risk
in a perspecdve from which the public can make informed decisions regarding dietary
intake.
cc: Jack Gibbons
Sally Katzsn
Katie McGinty
Joe Stiglitz
Brad Campbell
Mike Toman
Kevin Neyland
-------
EPA Responses to OSTP Comments (dated 12/11/95! on the Mercury Study Report to
Congress
This memorandum is in response to Bob Watson's comments on the Mercury Study
which were transmitted in a memorandum to TJ Glauthier. dated December 11. 1995. For
simplicity, the comments being addressed have been numbered sequentially in the original
memorandum, beginning with the section entitled "Specific Recommendations Under Option
Three".
1. The emissions inventory has been revised to label the emissions estimates as best point
estimates. The degree of uncertainty in the estimates is labeled in Volume II as "high".
"medium" or "low". This characterization will be included in the executive summary. In
addition, while Volume II repeatedly point out that there are data gaps in the inventory and
includes a separate section on research needs, we agree that more of this language is needed in
the executive summary and will add it.
2. Estimates of emissions are now provided to two significant digits accompanied by the
characterization of uncertainty as described above.
3. Background levels of mercury in the environment (which result from natural emissions, past
emissions that are still re-circulating and sources of mercury to the environment other than air
(e.g., water discharges) also contribute to environmental loading. This point will be highlighted
in the executive summary. In addition, this comment states that it cannot be described what
proportion of environmental mercury in the U.S. originates in the U.S. and what comes from
overseas. Actually this was estimated for airborne mercury in Volume III of the report. Section
5.2.1 includes a mass balance of mercury deposition. This discussion (or parts thereof) will be
included in the executive summary, including Table 5-3 which summarizes a mercury mass
budget for mercury deposition in the U.S.
4. While it is true that some of the fish data are 20 years old, these data from the National
Marine Fisheries Service are believed to be representative of current levels with two exceptions.
The first is for catfish; the old number does not reflect the current practice of farm-raising catfish.
We are not aware of any new data for farm-raised catfish, but it is foot-noted in the document
that current farm-raised catfish are likely to have lower mercury levels. Second is the fish
species called "pollack". This term actually refers to a different species than what used to be
called pollack in the 1970's. The data we have from the National Marine Fisheries data base
refers to the "old" characterization of pollack. We are checking for any current mercury levels
and also to see how a different level would affect our fish consumption analysis.
5. The seafood data that EPA used came from several sources. Data on mercury concentrations
-------
in various species of freshwater fish were based on reports by Bahnick et al.1 and Lowe et al.:
These two sets of data were based on specimens of fish collected at a number of geographic
locations within the L'.S. These data sets were chosen because they provided as good an estimate
of nationally-based data as we are aware of at this time. Data on mercury concentrations in
marine fish and shellfish were obtained from databases maintained by the National Marine
Fisheries Ser\ ice. It is our understanding that these data bases are updated continually and
represent the best Federal estimate of mercury concentrations in fish and shellfish species.
Unfortunately, once fish and shellfish enter interstate commerce, information on the
location from which they were obtained is often no longer available. The mercury concentration
data used were based on the concentration reflecting the central tendency of data from the
national database maintained by the National Marine Fisheries Service. This estimate of central
tendency is regarded as the best estimate for the U.S. as a whole. Certainly within the range of
data for methylmercury in a panicular type of fish there will be both higher and lower values than
the one used in the calculations that are part of the draft report.
The draft Mercury Study describes, in considerable detail, the geographic and local
conditions surrounding mercury contamination that modify the concentration of methylmercury
in fish and shellfish. The report specifically urges individuals to consult state and local
authorities regarding advisories on locally caught fish and shellfish.
We will seek to highlight in the executive summary that there are regional differences in
both fish levels and fish consumption patterns that affect exposure. There have also been
specific inquiries about Alaska species of shellfish, in particular the Alaskan king crab. Table 2-
17 in Volume III as well as Table 2-1 in Volume I are both foot-noted with respect to how the
average values are obtained. In the case of crab, the average mercury level is 0.117 ppm. This
value, as illustrated in both tables, is the average of the mean concentrations found in 5 crab
types: blue crab (0.140 ug/g), dungeness crab (0.183 ug/g), king crab (0.070 ug/g), tanner crab
(C. opilio) (0.088 ug/g) and tanner crab (C. baird) (0.102 ug/g). As shown, there is not a large
difference in mercury levels between these different types.
6. We are trying to obtain consumption data for the States of Alaska and Hawaii. Also, there is
another analysis underway to break the CSFII 89/91 fish consumption survey down into age
groups. If this analysis can be completed in time, it will be included.
o
'Bahnick, D., Sauer, C., Butterworth, B., and KuehlJD. (1994). A National Study of
Mercury Contamination of Fish. Chemosphere 29(3):537-546
:Lowe. T.P., May, T.W., Brumbaugh, W.G., and Kane, D.A. (1985). National
Contaminant Biomonitoring Program: Concentrations of seven elements in fresh-water fish,
1978-1981. Arch. Environ. Contamin. Toxicol. 14: 363-388.
-------
7. The reference dose is currently presented as a range in the report which illustrates the order of
magnitude uncertainty around this \alue. as well as other conflicting anaKses. The bod\ of the
Report provides extensive discussions of variability and uncertainty in the risk estimates.
Quantitative uncertainty anaKses were provided when feasible: these include the uncertainty
analysis for the bioaccumulation factors in Volume V and the analyses of the human RfD in
appendix D to Volume IV. The references to these in Volume I will be expanded. It shouU oe
noted that Volume I and Volume VI both refer to the human RfD as an estimate with a range of
uncertainty spanning an order of magnitude. Charts in Volume VI have already been re-drawn to
reflect this range.
8. Several comments related to the depiction of regional risk and suggested that various maps of
fish consumption patterns and fish mercury levels be combined. This is an interesting idea, but
there simply is not the consumption data to support this type of analysis. The national data we do
have on fish mercury levels is for freshwater fish only and even these measured data were so few
that it was necessary to model fish concentrations near industrial facilities. In addition, most of
the fish consumed by people in the U.S. is from interstate commerce. As described above, once
fish and shellfish enter interstate commerce, information on the location from which they were
obtained is no longer available. Thus, matching consumption to the national data that we do
have on freshwater fish levels is not consistent with what the average person is likely to be
eating.
There is a possible solution to this issue. In Volume V, the Ecological Assessment, we
have presented overlays of high mercury deposition and areas of the U.S. with acidified lakes.
These are the regions of the country we predict to be most impacted by mercury deposition (the
acidified water is correlated with mercury methylation). This type of map could be brought
forward into the Executive Summary to illustrate regional differences.
9. EPA has been working with the other federal public agencies on a public health message. We
previously thought to provide the risk associated with eating common seafood on a daily ,
weekly, or monthly basis as the comment suggests. However, this type of analysis would clearly
highlight that canned tuna (the most commonly consumed seafood) if consumed on a daily basis
would cause the RfD to be exceeded. It was thought that this analysis would be alarming to the
consumer.
10. The draft report does not suggest that the average consumer should reduce fish consumption.
The draft transmittal letter from the Administrator to Congress states clearly that the typical U.S.
consumer is not being advised, either implicitly or explicitly, to reduce their fish consumption.
General Comments
1. Peer review.
Ensuring"the quality of scientific underpinning actions by U.S. EPA has been a major
thrust of Agency policy. In order to ensure top quality science, U.S. EPA has been advised by
review bodies including its Science Advisory Board (SAB) to consider peer review of its
-------
programs, methods and products to be a top priority. To this end Administrator Carol Browner
issued a memorandum requiring expert review wherever appropriate and requinna L'.S. EPA
Programs and the Office of Research and Development (ORJD) to develop specific guidelines for
NCientific review. The ORD final guidelines for review of scientific products were issued in
November 1994. These were used as the basis for the peer review plan for this Report. The
Mercury Studv Report to Congress was considered by ORD and the Office of Air Quality
Planning and Standards. Office of Air and Radiation (OAQPS/OAR) to be one of U.S. EPA's
major and most visible outputs. As such the Report was considered to fall into category I.
requiring the highest level of scientific peer review. The components for category 1 review
include the following: approval of the peer review plan by the Assistant Administrator of ORD:
review of the product by appropriate U.S. EPA scientists: review of the product by appropriate
scientists external to the Agency: convening a peer review meeting; and stringent recordkeeping
on all phases of the review process.
Because of the wide scope of the Report and the interest in mercury by many
stakeholders, it was felt that the process of generating the Report should be open to external
input. Meetings with U.S. EPA Report authors were held with members of the public at their
request: for example, during early stages of Report generation, U.S. EPA staff met on a quarterly
basis with scientists and engineers representing the Electric Power Research Institute (EPRI).
Meetings were also held with the Portland Cement Association and with other requestors. The
Agency accepted and reviewed submissions of data and mercury assessment material throughout
the study period; these were used as was considered appropriate by U.S. EPA scientists.
In order to gather input and critiques on preliminary assessments, several of these were
presented at conferences and scientific meetings. Early results of the emissions inventory (found
in Volume II) were presented at both regional and national meetings. Draft health assessments
were also shown for purposes of discussion at scientific meetings on mercury. In January of
1994 a review draft of the emissions inventory was made publicly available.
Internal scientific review of a draft of the entire Report (minus Volume I, Executive
Summary) was begun in November of 1994. Following procedures for review and clearance
established for the Office of Health and Environmental Assessment (now the National Center for
Environmental Assessment, NCEA) within ORD. the draft was reviewed by four scientists from
that Office. In addition, the draft was reviewed by a U.S. EPA Mercury Study Work Group
consisting of staff from the following U.S. EPA offices: Office of Science. Planning and
Regulatory Evaluation (OSPRE/ORD); Office of Health Research (OHR/ORD); Office of Policy
Analysis and Review (OPAR/OAR); Office of Water (OW); Office of Solid Waste and
Emergency Response (OSWER); Office of Prevention, Pesticides and Toxic Substances
(OPPTS); Office of Policy, Planning and Evaluation (OPPE); and Region V. Scientists
representing the State of New York and the State of Michigan also participated in the Work
Group and in this phase of review.
Included as part of the Report are summaries of human health risk assessments which
comprise parts of the Agency's Integrated Risk Information System (IRIS). IRIS is a publicly
available computerized data base which provides U.S. EPA consensus health risk assessment
-------
information. IRIS files must undergo specific forms ot internal and external review before they
are made available on the system. The IRIS documents on mercury were reviewed as pan of the
Mercurv Study Report to Congress. The following six IRIS summaries were distributed to
reviewers as Appendix B of Volume IV of the Report: reference concentration for elemental
mercury : cancer assessment for elemental mercury, reference dose .for inorganic mercury, cancer
assessment for inorganic mercury, reference dose for methylmercury. and cancer assessment for
methy[mercury. Internal review for the IRIS documents consisted of the appropriate Agency
Work Group discussion and closure (referred to as "verification"). The Work Groups charged
with reviewing IRIS information and achieving consensus on its validity are comprised of U.S.
EPA scientists from a variety of disciplines relevant to human health risk assessment and who
represent ORD, the Regions and the Program Offices. The two Work Groups have been
organized around either the assessment of carcinogenic effects (the Carcinogen Risk Assessment
Verification Endeavor (or CRAVE) or the assessment of general systemic toxicity (RfD/RfC)
Work Group). To enhance the mercury expertise of the Work Groups and to allow for discussion
of alternate risk assessment approaches, scientists from FDA. ATSDR and the State of New
Jersey were invited to participate in the RfC/RfD Work Group discussions; they were not part of
the consensus process, however, and did not participate in Agency decisions on the assessments.
After consensus on the assessment was achieved. IRIS documents were revised and received
external review (see below) as part of the external review draft of the Report. Following external
review and revision the IRIS documents were either reviewed and cleared by the Work Group
chair (RfDs and RfC) or given a pass-around review by the whole Work Group (CRAVE).
External review of the Mercury Report to Congress and the appended IRIS documents
was done in two steps: a Federal interagency review and a non-Federal external review. A
meeting of Federal reviewers was held at the U.S. EPA, Washington DC on January 9. 1995 to
discuss scientific issues in the Report. Representatives of the following Agencies were invited to
attend and to submit written comments at the time of the meeting: ATSDR, NIEHS, NOAA,
USDA, DOE, FDA and USFWS. The names and addresses of the reviewers can be found at the
beginning of each volume of the Report. Written comments were received from all Agencies
participating in the review. A summary of reviewer comments, consensus opinion ot" reviewers
and U.S. EPA's response to comments can be found in this Appendix.
The second phase of external review included comments from non-Federal experts.
Reviewers were chosen based on scientific expertise and availability. An attempt was made to
include representatives of a spectrum of groups with interest in mercury: academe, research
groups. State agencies, industrial concerns and environmental groups. The names and affiliations
of reviewers can be found at the beginning of each volume. All reviewers were required to
submit written comments on the report including the IRIS documents. A public review meeting
was held January 25-26, 1995 at U.S. EPA in Cincinnati, OH. A notice of the meeting was
published in the Federal Register, and there was time set aside each meeting day for members of
the public to comment. The pre- and post-meeting reviewer comments, synopses of meeting
discussions and conclusions of the meeting comprised the external review report. Examples of
external review comments are in Appendix A to this Volume. In response to reviewer
comments, specific changes were made in the External Review Draft. As the Risk
Characterization (Volume VI) was revised, a second review of that volume was done.
-------
One revised component was reviewed in advance of the remainder of Volume VI; this
the estimate of population size, amount of fish consumed and measured amount of mercury
in marine and freshwater fish. This assessment (now included in Volume III. Appendix H and
summarized in Volume VI) was sent to two external reviewers expert in statistics and
demographics. These reviewers were selected by a U.S. EPA contractor who was provided with
criteria for reviewers and a list of potential candidates. The entire revised Risk Characterization
was subjected to internal and external review. Scientists in ORD. OAQPs and Office of Water
were sent the volume and requested to submit comments. Four external reviewers were selected
by a U.S. EPA contractor based on criteria provided by U.S. EPA. Among these criteria were
that two reviewers be included who had commented on the External Review draft. Written
comments on the risk characterization were provided by these four external scientific reviewers.
Copies of all review comments and external input are archived at the National Center for
Environmental Assessment in Cincinnati. Ohio.
-------
191=5013640
FDAJOFFICE OF SEAFOOD
Note
December 14, 1995
Note to: Phil Spiller
From: Gregory Cramer, Ph.D.
Subiecr Comments on EPA December, 1995 redraft of Mercury Study- Vol. 1 Executive
Summary
Given the short time frame for providing comments, a few hours, my comments are brief.
Page 2-2, par 1, line 7: "The concentration ...is on average, lower than the FDA action level. "
par 1, last line: "or who eat fish from couUuiiiiaual lucul waters."
Page 2-2, Table 2-1 The mercury levels in table are factually inaccurate and should be changed
based on information provided to Glen Rice at EPA by Dr. Greg Cramer on 12-13-95.
Page 3 4, par 2, last full line: "emissions complicates an assessment of possible emission-
reduction
Page 3-4, par. 3, line 4: concentrations in fresh water fish.
par 3, line 7: "These findingi do not estimate the source of mercury for marine fish."
(Delete discussion of international sources-simply marine species).
3-1Q, 1st full par., next to last line: "new studies have reduced the level of uncertainty
associated with low-dose methybncrcury exposures."
Page 3-10, What Are Lereb of Mffchybnercury ...... World Health Organization (delete and
State Recommendations — these are never discussed by EPA).
Page 3-10, par 3, line 4-5. Delete entire sentence "The «p«*"ig advice and action level of the U.S.
FDA is ........ ". This statement is incorrect and misleading.
Page 3-10, par. 4: Delete entire paragraph since the basis of state warnings is never discussed in
this section.
Page 3-11, para 2: EPA continues to discount the value of existing hair mercury data for U.S.
consumers. The small difference between consumers and non-consumers cited by EPA is reflected
-------
=*CM IFF ICE IF SEAFCCD ~3 91513=412640 =.;
Page 2
.1 other less highly exposed populations as well, e..g, "Report on the Great Takes Anglers Pilot
Exposure Assessment Study" (Draft Report April 1995), provides the same kind of data as the
U.S. hair survey data. Collectively, these and other data from U.S. studies show EPA's exposure
assessment exaggerate the size of population and extent of risk among fish consumers.
Page 3-13: line 1: Delete discussion of childrens exposures and size of population at risk. This
subject should be presented elsewhere'—not in section on levels of exposure addressed by FDA,
WHO...etc.
Page 3-13, par. 3: The general United States population that obtain is not believed to be
adversely affected by methylmercury. Delete-"if they consume fish in moderation''- it is vague and
alarming to those who don't know what is moderation or who might consume a fish every week
and mistakingly think that amount exceeds moderation.
Page 3-13, par. 3, lines 8-10: Delete entire sentence "As an example, the U.S. FDA limit has
issued an advisory for the general population". FDA's advice was cautionary and was not
intended as a statement that persons might be at risk.
Page 3-14, 1st and 3rd full par, last sentences for each: Delete these sentences. Ingestion of
MeHg at the RfD says nothing about risks associated with MeHg exposures at that level. These
two sentences, therefore, offer the reader not much by the way of information and are merely
—nentially confusing.
.age 3-14, line 2: "...on mercury in freshwater fish". It is unclear whether data from
contaminated sites were used to estimate consumer exposure from commercial species harvested
from uncontaminated areas. A clearer statement of what was done with these data is needed.
Page 3-14, par. 1 and 4: EPA's estimates of consumer exposure are dependent on 3-day survey
results. As a consequence, EPA's estimates of consumer exposure are inconsistent with what is
known about long term fish consumption in the U.S. as well as other estimates methylmercury
exposure among U.S. fish consumers.
Page 3-17, line 3: "...consumption of less than 100 grams of fish per day raises concern for
children". This phrase is scary because it doesn't give any sense of how much less than 100
grams i* likely to be harmful. Puther, it isn't very useful without talking about body weights,
extent of contamination, etc. This entire sentence starting with "When the smaller " needs to
be removed.
Page 4-1, last par, 2nd sentence: "...reduced human exposure is another alternative that would
need to be explored for reduing risks to human health..."
Delete, unless willing to add caveat that such consideration would need to incorporate
extensiTe risk/benefit analysis focussing on risks associated with reduction of
availability of a highly vahiabk protein source.
on revise to match up with the paralr, p4-7, for example,
-------
rE5-ir-i=56 17=35 =~Cf IFF ICE Cr SEnFGGD '1 919155412S40 =.34
Pace 3
"reducing buman exposure through focus on consumption controls for local fish from
contaminated water bodies is another alternative that would need to be explored..."
-------
ID:
DEC 21 '95 17:46 No .012 P
UMrreo STATm
NATOMAL IWWMNE
Mr. Kavin Nayland
Office of Managaaant end Budget
735 17th Street, v.w.
Washington, D.C. 20503
Dear Mr. Neyland:
Ke have reviewed the revised Voluaa I, Executive flumary, of
the B* A Maroury Study Report to Congress. In ooapering this
version with the previous draft (dated October 2S, 1995), we
noted that language haa baan added to further qualify tne
report's findinga and associated uncertainties. However, MX7J
eontlnuea to hava t&e concerns ve previously identified, i.e.,
SPA'a uee of a lower reference doee for methyljaercury, and of the
3-day USDA eurvey dete to eetiaate fian eoneuaption. Furtheraore,
we have not received a recent varaion off the oovpleta report, and
so are unaware of the reviaione, if any, aado to other chapters
besides the executive Susmary.
He eppreciate the i-*enth delay in the report*e release, and
we veiooae the opportunity to provide ooveente on further
reviaione to the document.
Sincerely,
Rollendl. fehaittea7'
THS.
on lUtycM fape»
-------
SECTION THREE
REVIEW OF THE DECEMBER 14.1995 DRAFT OF THE MERCURY STUDY
In February. 1996 the decision was made to proceed with the release of the Mercury
Study by April 15. 1995. The latest draft of the study as of December 14. 1995 was again
distributed to Federal agencies for review and comment. Extensive changes to the document
were made as a result of this review. This section contains the following in order:
1. EPA's "Summary of Federal Comments on the Mercury Study" dated February 22. 1996.
This piece represents an overview of all of the comments that had been received and how various
issues were to be resolved.
2. CEA comments dated February 8. 1996.
3. EPA responses to February 8, 1996 comments.
4. USD A comments dated February 16. 1996.
5. EPA responses to February 16,1996 USD A comments,
6. OSTP comments dated February 15, 1996,
7. EPA response to February 15, 1996 comments,
8. CEA comments dated February 14, 1996,
9. EPA responses to February 14, 1996 CEA comments,
10. FDA comments dated February 16, 1996,
11. EPA responses to February 16, 1996 FDA comments,
12. OMB comments dated February 16, 1996,
13. EPA responses to February 16, 1996 OMB comments,
14. NOAA comments dated February 16, 1996.
15. EPA responses to February 16, 1996 NOAA comments.
16. NIEHS comments dated February 15, 1996,
-------
17. EPA responses to February 15. 1996 NIEHS comments.
18. CDC comments dated February 28. 1996.'
19. CDC comments dated February 16. 1996.
20. Department of Health and Human Services comments dated February 19. 1996.
21. EPA response to CDC comments dated February 28. 1996: CDC comments dated February
16. 1996: Department of Health and Human Services comments dated February 19, 1996.
22. Department of the Interior (DOT) comments dated March 1. 1996,
\
23. EPA response to March 1, 1996 comments by DOI.
24. Department of Energy (DOE) comments dated February 21. 1996.
25. EPA responses to DOE comments.
-------
Summary of Federal Comments on the Mercury Study -- February 22,1996
This summary represents EPA's general response to Federal reviewers comments on the
December 14. 1995 draft of the Mercury Study. These comments were received in mid-
February. 1995. The purpose of this document was to provide a summary of all the comments
that had been received and an assessment of how the comments would be addressed.
Volume I
General Nature of Comments
The majority of comments point to the inclusion of more information from the technical
volumes in Volume I. Commenter wanted more description of transport and exposure (DOE)
and control technology and costs of control (DOE. ). One commenter stressed that the risk
characterization discussions in Volume I should be easily comprehensible to the general public
(OSTP). The same commenter felt that variability of risk should be discussed by demographics
and location(OSTP). Commenters from HHS said that there should not be recommendation as to
the safe number of fish meals. Another HHS commenter remarks that the study does not provide
a basis for changing current dietary advice to the public with respect to the consumption of fish.
Comments (OMB) indicated the need to include more discussion of uncertainties and
assumptions. OMB also wanted more information.and analyses of non-air sources of mercury
including natural sources: these are addressed below. OMB also suggested the following for
Volume I: revision (to include costs) or deletion of material on pollution prevention: discussion
of uncertainties of estimates of costs and effectiveness of controls in the context of discussion of
regulatory alternatives; deletion of factors to be considered in evaluating options.
Changes that Are Planned
We had, in an earlier draft of Volume I. included a discussion of study design and
summaries of conclusions of all technical volumes. We will put this text back into Volume I.
U.S. EPA agrees that this Agency should not recommend a number of safe fish meals: it
was not our intention to provide this. We will review text on recommendations made by other
Agencies to ensure that there is no confusion as to the source of these recommendations. We
likewise agreed that the Report should not comment on a change in the FDA advice. We will
review the text.
All parties agree that the public health agencies have a large role in the communication or
risk from food consumption.
EPA agrees to provide a more balanced executive summary in that more information u ii!
be provided on emissions (and their uncertainties) and control technologies and costs.
The pollution prevention material will be reviewed so that the most useful information :>
-------
presented. Regulatory factors will be deleted as well as other changes consistent with changes
made to Volume VII (see Volume VII comments).
Note: Supplement A has been revised to«only include the summary of the scientific peer review
comments. All other material in Supplement A has been deleted.
Comments Still under Consideration
Volume I had originally been written with the objective of comprehensibility to an
informed lay public. In the course of discussions mediated by EHPC and involving FDA much
language was changed or inserted as negotiated. To make changes in much of the risk
characterization text (in current chapters 1. 2 and 3) would be to go back on agreements which
have already been reached among Agencies.
There is some discussion of variability of risk as a function of demographics (focussed on
recreational anglers and high end fish consumers). As discussed there are very few data which
can be used to describe variability of risk with location. We propose to include some of the
deposition plus low pH water-body information from Volume V. We will also evaluate the use
of a map from EPA's Office of Water which provides location of measure mercury values in
environmental media. It should be noted that all available data on measured mercury levels in
media including biota were presented in text and tables in Volume III.
Comments that EPA Is Unable to Address
None identified at this time. The executive summary will reflect the changes made in the other
volumes. Changes that we are unable to make in other volumes are described below.
-------
Summary of Federal Comments
Volume II
General Nature of the Comments
DOE and OSTP provided the bulk of the comments on Volume II. Comments suggested
that EPA should quantify natural emissions, global emissions and direct water discharges. Also.
commenters noted that estimates were missing for a number of source categories that could be
important. DOE also suggested that EPA provide a boiler-by-boiler emissions estimate for
utilities and should predict future utility emissions.
Changes Planned
Addition, as least qualitatively, of water discharges as another important source of
mercury to the environment.
A more complete description of the uncertainties in the emissions inventory will be added
to the executive summary.
Comments Still Under Consideration
EPA will review the Argonne report that DOE mentions that pertains to coke oven and refinery
emissions. However, DOE did not provide a full reference. A copy of this report is requested.
Comments EPA is Unable to Address
Global and Natural Emissions
EPA believes that the best available data on global and natural emissions is already
included in the report. These data are believed to be the best available assessment of global and
natural emissions and their relationship to anthropogenic emissions. This information represent*
the consensus opinion of an international panel of mercury experts which was convened by the
Electric Power Research Institute.
Data Gaps in Inventory
EPA admits that there are data gaps in the inventory and lists the source categories we
were unable to quantify. The need for additional emissions data for a number of these sources >.
listed as a research need. Some commenters felt it surprising that there were no data for certain
source categories such as coke ovens. A thorough review of these source categories was made
during the development of the mercury emission factors. No data were found that were sufficiev
to calculate emission factors for the source categories listed. As emission tests are conducted .1
data are obtained, EPA routinely updates it's emission factors in the AP-42.
-------
Mercury Speciation
Virtually all of the measured mercury data for mercury sources are reported as total
mercury, except in the case of utility boilers where the samples were speciated. The importance
of speciated mercury data is noted in Volume II as a research need and is discussed in depth in
V olume III. In Volume III, various speciation scenarios were assumed for the modeling
analyses. These analyses demonstrated how mercury speciation affects the dispersion of mercury
through the atmosphere.
Strengths, assumptions and weaknesses in the data.
In the discussion of each source category there is a discussion of the source of the data
and how the data were used in the calculation. An estimate of the quality of each of the estimates
is also made. EPA agrees that more of this information should be added to the executive
summary.
Boiler Specific Estimates
EPA does have boiler specific emissions estimates. They were not included in the report
because we felt that the data would be mis-used (e.g., certain boilers or companies would be
"targeted" for reductions).
Trends in Emissions
DOE refers to the 1995 estimates for certain source categories as predictions based on
future regulations. These are not predictions, these are updated emissions e -Lmates for the year
1995 and do not reflect predicted regulation. A trend for utility boilers is calculated in the Utility-
Study. It was not included In this report because the supporting documentation was not included
(e.g., energy forecasts).
-------
Summar} of Federal Comments on Mercury Study
Volume III
General Nature
Most of the comments received for Volume 3 pertain to the use of the fish consumption
data and central tendency estimates of mercury concentrations in fish to estimate human
exposure to mercury through this route. The fish consumption comments have focused on the
appropriateness of applying U.S.D.A. consumption survey results to estimate long-term fish
consumption rates. The comments on data sources used to estimate mercury concentrations in
seafood have focused on the differences between the U.S. FDA estimates and those of the U.S.
EPA. the geographic variabilit) of mercury concentrations in fish, and the use of mercury
concentrations from wild catfish vs. those from farm-raised stocks.
Other comments pertain to the global fate and transport of Hg: specifically, whether
mercury emissions from local sources effect methylmercury concentrations in oceanic fish.
comparing the effects of water discharges of mercury to those from air emissions, providing
more information on human exposure through dental fillings, validation of computer mqdels
used and uncertainties in emissions estimates.
Changes Planned
Fish Consumption Rates
External reviewers specifically indicated that it was appropriate to utilize U.S.D.A.
consumption survey results to estimate long-term fish consumption rates. Since specific
comments have focused on the use of data from the upper percentiles (i.e., 95th and above), we
have focussed on these. We have analyzed longitudinal study data (1973/74 NPD data) and plan
to compare the fish consumption results of this survey with those of the U.S.D.A. We also plan
to more clearly highlight the uncertainties in these survey methods. Local or regional surveys
will also be presented to better depict the range of fish consumption rates seen in groups that may
consume higher quantities of fish. Also results of fish consumption surveys from Alaska and
Hawaii will be presented because these groups were not included in the U.S.D.A. or NPD
surveys. We may also attempt to indicate differences in fish consumption rates across U.S.
regions as shown by the NPD survey.
Concentrations of Mercury in Fish
We will attempt to show more clearly that the freshwater fish data were from wild
populations and that most of the freshwater fish in commerce are farm-raised. If data on mercur>
concentrations in the farm-raised fish are obtained, these will be presented: if not, then this will
be highlighted as a source of uncertainty. We will present the discrepancies between the NMFS
-------
data (used in the analysis) and those of FDA and attempt to explain the differences.
Comments Under Consideration
More dental amalgam information.
Effects of anthropogenic source mercury emissions on marine fish.
Comparing the magnitude of water discharges relative to air emissions..
Comments that EPA is Unable to Address
1. Geographic differences in mercury levels in fish. EPA has attempted to present a great deal of
local data and presented the predictions of models. Given the current data bases and the inter
water body variability this seems inadvisable further application or analysis seems inadvisable.
2. EPA nas highlighted the range of estimates of natural and anthropogenic mercury emissions.
No other measured data are known to exist at this time.
3. Validation of the computer models.
These models have not been validated but then no other mercury fate and transport
models have (at least to our knowledge) been validated at this time. EPA is currently conducting
research on some of the more uncertain areas. Model validation will take a much longer period
of time and will require a great deal of money.
EPA judged that the current measured data were inadequate to conduct an exposure
assessment around atmospheric emission sources. As a result of this judgement, EPA opted to
model the fate, transport and resulting exposures from emissions. EPA attempted to explain the
uncertainty in the results and attempted to convey the uncertainty in the modeling in the
conclusions.
In the document a comparison of typical measured values with model predictions was
presented (Ch.4, Vol. 3). The data input to the model were those measured at remote sites and the
results were comparable to those measured at remote sites.
4. Comment stated that use of models is inappropriate for prediction of U.S. emissions impact
on methylmercury in marine fish (NOAA). U.S. EPA agrees: in fact no prediction of mercury
levels in marine fish consequent to emissions was done in the Report. The Report explicitly
states in Volume III that mercury levels tied to either long range transport or local deposition 01
U.S. emissions can only be done for the 48 contiguous states and thus apply only to fresh water
fish. All discussion of mercury levels in marine fish and other seafood are based on reported
mercury measurements, largely from NMFS.
-------
Summary of Federal Comments on the Mercury Study
Volume IV
General Nature of Comments
All comments dealt with the U.S.EPA reference dose (RfD) for methylmercury. There
were no comments on the other forms of mercury or on any health endpoints not related to the
RfD for methylmercury. Some commenters felt that EPA either was not required to or should
not calculate a mercury RfD (DOE and NIEHS). Commenters either said that there was low
confidence in U.S. EPA RfD or that confidence, uncertainty or assumptions were not described
(DOE. FDA). Commenters said that data from the study published on the Seychelles should
either be used to substantiate the U.S.EPA RfD or should supplant the RfD (OSTP, DOE.
NOAA, FDA). One commenter stated that the measures in the Iraqi study were subjective and
should be not be used (USDA). Comments said that the recent analyses by Cox et al 1995 and
Crump et al 1995 indicate that the Iraqi data should not be used (DOE. NOAA. FDA). One
comment said that the Seychelles and Iraqi data are not comparable (NOAA). One comment said
that the derivation of the RfD depended only on 4 data points (FDA). One comment had
questions as to the statistical methods used in the benchmark dose (BMD), specifically as to the
effect of grouping the data on the outcome of the model (OSTP). said that the BMD is more
conservative than use of a no adverse effect level (NOAEL) (USDA). One commenter said that
consumption of mercury in fish is substantially different that consumption of mercury in grain
(FDA) and that there are no data on health effects of mercury reported from fish consumption.
(NOAA). Two commenters allude to a study on Peruvian fishers as being an indication that the
Iraqi data should not be used (NOAA, FDA). One comment refers to the not yet published study
in the Faroes (FDA).
Changes that Are Planned
Additional discussion of the Kjellstrom study of the high fish-consuming populations in
New Zealand will be included in Volumes IV (Health Effects) and Volume VI (Risk
Characterization). A recent analysis by Crump and co-workers provides a BMD of 17 ppm Hg,
maternal hair, which is comparable to the Iraqi BMD of 11 ppm, serving as the basis of the U.S.
EPA RfD. Discussion of this paper (including the PBPK model), the endpoints in Kjellstrom
and the peer review of Kjellstrom will be included.
A discussion of the tests employed and endpoints measured in the Seychelles study will
be included in Volumes W and VI. This will include the following: discussion of the sensitivitv.
discussion of appropriateness (both in terms of power and expectations regarding age at testing).
points of comparison and discontinuity with the endpoints reported in Kjellstrom, the Iraqi stuck
and the Peru study.
Analysis and discussion of mercury hair levels will be expanded to include the Seychelles
and Peru studies.
-------
Comparisons of benchmark doses derived from the Iraq. New Zealand and Seychelles
studies. A BMD for DDST results of the pilot Iraq study was done using the data groupings as
published in Neurotoxicology, midpoint of the hair mercury range, abnormal plus questionable
responses combined and either a Weibull or polynomial model is 16 ppm Hg maternal hair, in
close agreement with the BMD fro Iraq and New Zealand. We are attempting a BMD for the
four positive measures in the main study
Discussion of the study of Peruvian fishers conducted and written in 1985 and published
this year will be included.
Comments Still under Consideration
U.S.EP.-T. has individual data from the Iraqi study. In order to use incidence of effects .
some grouping of data must be done to provide n>l. The grouping of maternal hair
concentrations published in Seafood Safety was done to provide consistency with other analyses.
We have not tested the effect of changing groupings on the BMD. If one were to use all
published hair mercury levels with corresponding effects, scores on tests would have to be used
(instead of incidence of effects); this would preclude a BMD on combined effects.
We hope to discuss the papers of Cox et al 1995 and Crump et al 1995. A preprint of the
Crump paper was made available to us, and Dr. Crump wrote text for insertion to the discussion
of the RfD, which is found in Volume IV.
We have referred to some unpublished data on hair mercury in Volume I; FDA provided
data. We would need to see the data from NOAA for evaluation.
At this point we have no more published information from the study of the Faroes
population. We will request reports or published abstracts from the study directors.
Comments that EPA Is Unable to Address
1. EPA has gone beyond it's mandate by revising the RfD and the RfD derived is inappropriate:
U.S.EPA is required in the CAAA of 1990 to conduct and provide
a study of mercury emissions from electric utility steam generating units, municipal waste combustion units, and
other sources, including area sources. Such study shall consider the rate and mass of such emissions, the health and
environmental effects of such emissions, technologies which are available to control such emissions, and the costs
of such technologies.
In order to determine health effects of an environmental agent, U.S.EPA applies the 1983
NAS paradigm for risk assessment. EPA's published Guidelines for Risk assessment and
established procedures for the evaluation of systemic non-cancer effects. This includes the
calculation of an RfD when data support it. The U.S.EPA RfD/RfC work group reached
consensus on the advisability of using available data as the basis for a scientifically supportable
That data were sufficient for calculation of an RfD was the opinion of all but one external
-------
reviewer (including Federal reviewers) in the context of our peer review held in early 1995
No change in the RfD at this time is anticipated. Analyses so far of the New Zealand data
and the available data from the Seychelles provide BMD which are consistent with the RfD and
certainly within the order of magnitude range defined for other RfD.
Uncertainties in the RfD, ranges by definition and by comparisons to other analyses.
assumptions and defaults used have been described at length in Volumes IV and VI. A brief
discussion of some salient points is in Volume I. There are quantitative analyses of uncertainty
in Appendix D to Volume IV.
The reliance of the RfD on four data points is obviated to a large extent by the application
of the curve fitting procedures and calculation of the BMD.
The BMD used as the basis for the RfD is the lower limit of a 10% effect level. This was
chosen because there are studies published which indicate that this is a reasonable predictor of a
NOAEL for developmental effects. These studies have indicated that use of this level is no more
conservative than estimating NOAELs from studies by more typical means.
2. EPA did not consider selenium interaction.
There are not sufficient data to assume differences in methylmercury effect as a function
of food type. There is some discussion of selenium interaction in Volumes IV and VI; our peer
review expert panel were adamant that the interactions between selenium and mercury are not
well characterized and not quantifiable. There is no evidence that the Iraqi population was
selenium deficient or starved at the time of consumption of mercury contaminated grain. There
are data that show that methylmercury is well-absorbed in the gut from foods. There are data
from the poisoning incidents in Japan that health effects are consequent to consumption of
methylmercury in fish; FDA used some of these data as the basis for their ADI. More recent
analyses and follow-up of the Japanese populations provide bounding for lower level exposure.
The study of Peruvians done in 1985 and published this year does not conclusively refute an
association between methylmercury in fish and health effects. The recently published studies in
the Seychelles are not negative. The authors publish some correlations between increased
maternal hair mercury and abnormal plus questionable scores in the DDST administered in the
pilot study. The authors also show correlations between maternal hair mercury and four
measures of developmental neurotoxicology in the main study for testing at 26 months of age.
When outliers are removed the relationship remains significant for one measure (decreased
activity levels in male children). The authors state in several of their papers that results are not
yet conclusive and that results from testing at 66 months of age must be examined.
-------
Summary of Federal Comments on the Mercury Study
Volume V
General Nature of Comments
The only comment (NOAA) dealt with use of bioaccumulation factors in the estimation
of mercury levels in freshwater fish consequent to U.S. mercury emissions: the comment also
mistakenly described use of models to link emissions to mercury in marine fish. There were no
comments on estimation of ecologic hazard from mercury emissions. FWS wanted discussion of
differences between Great Lakes Initiative water criteria and those in the Report added to
Volume I and also inclusion of a table from Volume VI that depicts human and wildlife
exposures in a common metric.
Changes that Are Planned
Discussion of GLI and Mercury Study differences in water criteria development will be
discussed in Volume I.
Comments Still under Consideration
Inclusion in Volume I of human and wildlife table from Volume VI.
Comments that EPA Is Unable to Address
The uncertainty in the BAF is described at length in Volumes III and V and is quantified to some
extent in the appendix to volume V. While some external reviewers had critiques of the BAF as
simplistic, no alternatives were available or .offered. It was agreed that the limited empirical data
available supported the use of the linear BAF.
-------
Summary of Federal Comments on Mercury Study
Volume VI
General Nature of Comments
None of the comments specifically referred to Volume VI although obviously, the
comments directed towards the RfD and the exposure analysis have bearing on the risk
characterization.
Changes that Are Planned
Changes to the risk characterization will reflect the changes made in all volumes, ranging
from uncertainties in the emissions inventory, to RfD issues and fish consumption.
It is expected that Volume VI will be shortened. In addition, the following changes are
contemplated for Volume VI (these changes being dependent on the outcome of the 2/23/96
principal's meeting). Note that these changes would also be reflected in the executive summary.
• The quantitative risk characterization will focus only on populations who eat fish from
fresh water bodies that are impacted by mercury deposition. In addition, the quantitative
assessment will likely only focus on those subpopulations that are on the high end of the
fish consumption distribution (i.e., not the general population).
• Mercury exposure from seafood would be described as an exposure pathway but would
be characterized as part of the total exposure that people have to mercury i.e., a general
background exposure. The executive summary would highlight that we are unable to
model the deposition and subsequent methylation and uptake of mercury by marine
species.
• In order to calculate the total fish exposure (freshwater fish plus marine fish), the fish
consumption data will still be used (with the additional data as described above). This
fish consumption analysis defines the exposure only. The number of people at or above
the RfD or any other health benchmark, due to seafood exposure would not be estimated.
nor would any exposure to a particular marine species be illustrated.
• The assessment of risk to the human population would focus on the message that
incremental increases in mercury levels in freshwater fish due to mercury emissions from
anthropogenic sources (as modeled in Volume III) have the potential to increase the total
mercury exposure in those populations that are at the high end of the consumption
Comments Still under Consideration
Volume VI changes will reflect changes that are finally made to all volumes.
-------
Comments that EPA Is Unable to Address
Comment stated that FDA used the EPA RfD to develop an action level of 1 ppm.
Comment :hen states that EPA uses this action level to determine populations at risk (USDA).
This is not how the analyses were done. The FDA action level was not based on the EPA RfD .
but was ser through action of courts: it considers FDA's own ADI which is four-fold higher than
EPA's. The EPA analysis of population at risk was done using two points of comparison: the
EPA RfD and the range around it; and the WHO recommendation of 100 g fish per day as an
upper limn of consumption.
-------
Summary of Federal Comments on the Mercury Stud}
Volume VII
General Nature of the Comments
Many comments disagreed with the "social costs" section. DOE provided numerous
suggestions for the technical analysis. OMB thought that a quantitative cost assessment should
be performed for pollution prevention options. Comments reflected the opinion that the study
should not recommend any particular regulatory actions.
o
Changes Planned
\
The social costs section will be removed in its entirety.
"Recommended actions" will be consolidated v. ith management alternatives and "regulatory
factors' will be deleted.
CEA's comments have been addressed essentially in their entirety. A draft disposition memo has
been provided to OMB on 2/21/96.
DOE's technical comments will be addressed essentially in their entirety. More of the
uncertainties regarding mercury control will be discussed in the executive summary.
Comments Still Under Consideration
A sensitivity analysis of HgCL and this impact on control costs may be performed or else the
comment will be addressed qualitatively.
Activated carbon costs: this analysis will be compared to DOE's. EPA requests that DOE
provide their cost analysis and its assumptions that is being compared to EPA's.
The "pollution prevention" table will be reviewed so that it illustrates the most useful
information.
Comments EPA is Unable to Address
EPA is unable to predict or calculate the cost of the numerous alternatives available for pollution
prevention. A analysis of this type is beyond the scope of this report. Individual manufacturing
processes would need to be examined as well as an entire range of substitution options. The
Mercury Study states that it is not an in-depth regulatory analysis and additional control options
would need to be considered in the course of a rulemaking for any one source category.
-------
P.302 CC3
EXECUTIVE OFFICE OF THE PRESIDENT
, COUNCIL OF ECONOMIC ADVISERS
WASHINGTON D C 23500
MEMBER
February 8, 1996
MEMORANDUM FOR MARTHA KEATING, EPA
CC: KEVIN NEYLAND, OMB
FROM: ALICIA MUNNELL
SUBJECT: Comments on EPA Draft Mercury Study
General Comments
While CEA has not had the opportunity to review the entire
seven-volume report, we have reviewed the summary volume and on
the basis of that material we have a number of concerns with the
study. The report involves a number of significant scientific
and economic issues of concern to a number of agencies. As
currently written, the report could be interpreted as suggesting
that there is a significant public health threat posed by the
nation's fish supply, notwithstanding the health benefits of
increased fish consumption. The report also could be read as
providing a rationale for a set of very expensive end-of-pipe
control measures to reduce risks whose magnitudes are not yet
well understood. It is definitely not clear that the facts
support these conclusions.
I also understand that White House Office of Science and
Technology Policy is developing a detailed set of recommendations
for addressing some of the outstanding scientific questions. The
recommendations include the use of new information that may
affect the setting of the "reference dose," and recommendations
for presenting the risk information in a way that indicates its
reliability and its applicability to the general population or
specific subpopulations. CEA concurs with the need to address
these scientific questions. Equally important is the need to put
the conclusions of the report in a proper economic and policy
context, even though the report correctly notes that its purpose
is not to fully analyze or recommend policy actions.
Specific comments on December summary Volume
(NOTE: While these comments are directed at the summary volume,
they also apply to the corresponding material in the rest of the
report.)
(1) page 3-25: this section is titled "Management Alternatives
for Reducing Mercury Emissions," presumably in keeping with
section 112(n)(1)(B)- The discussion does note that the analysis
of control strategies and costs is not intended to replace a
-------
-E3--je-l356
5.303, Q03
thorough regulatory analysis. In keeping with that point, the
discussion also should note that emissions control is only one
possible means for risk control. Measures that reduce human
exposure is another alternative for reducing threats to human
health (this measure would not ameliorate negative ecosystem
effects, but these do not seen to be that serious) . Another
approach would involve reduced use of high-mercury coal. The
report needs to avoid becoming inadvertently disingenuous by
focusing on control strategies because that's what 112(n)(i)(B)
asked for, when in the larger context other risk management
strategies need to be evaluated as well.
(2) page 3-28: the discussion of dollars per citizen in
connection with Table 3-6 needs to be completely revised.
Contrary to what is suggested later in the text, cost per ton
removed is a perfectly adequate description of cost-effectiveness
and should be presented. EPA then can argue that even high-cost
reductions are justified, but that argument has no place in this
report given its scope (or if it is to be included, it needs a
great deal more attention) . Even if one were looking to compare
costs and benefits, costs per citizen is not very informative.
There is no information provided in Volume l with respect to
the Minnesota contingent valuation survey of willingness to pay
for reduced mercury, and the discussion elsewhere in the report
is minimal. Was it peer reviewed? Are there other studies that
can be compared? At this juncture, EPA should simply report
costs per ton removed and leave benefit valuation for another
study.
The discussion of research needs at the bottom of page 3-28
also focuses solely on prevention measures. It needs to be
deleted or made more balanced by addressing, for example, the
need for research on how exposure limits might be structured to
reduce the number of potentially exposed persons falling through
the cracks.
(3) page 5-1: consistent with the above comments, point #3 in
this list should be deleted. The same probably is true of #4 —
not because the points are necessarily wrong, but because they
arise out of context here and go well beyond the charge.
(4) page 6-1: Here EPA states actions it intends to take as a
result of this study, having previously acknowledged that this
study itself does not provide a full analytical basis for drawing
policy conclusions. This whole section is problematic in terms
of going beyond the charge. It would be better if the report
simply summarized the health risk conclusions and caveats reached
(to the extent this is not already done previously in the volume)
and then discussed how this analysis fits together with other
ongoing efforts, rather than couching the discussion in terms ot
-------
EPA Responses to CEA Comments (2/8/96) on the Draft Mercury Study
This memorandum is in response to Alicia Munnel's comments on the draft
mercury Study which were transmitted to Martha Keating, EPA. dated February 8.
1996. Specific comments offered by the Council of Economic Advisers (CEA) are
addressed below. CEA notes that while their specific comments are directed at the
summary volume, the comments also apply to the corresponding material in the rest of
the report. The EPA notes that where changes have been made in the Executive
Summary, the corresponding changes have also been made elsewhere in the report.
1. The discussion of management alternatives has been revised to note that emission
controls are not the only meaYis for reducing human exposure. CEA also^suggests
reducing the use of high mercury coal. Unlike sulfur, different types of coal do not vary
significantly in mercury content and as a result, this is not a considered a control option.
However, other management measures such as emissions trading, emissions
averaging and fuel switching are described in Volume VII of the report. The discussion
of management alternatives in the Executive Summary will be broadened to include
these other alternatives.
2. The discussion of dollars per citizen has been removed from the report as well as
the section in Volume VII on social costs of mercury pollution which included the
Minnesota contingent valuation study. The discussion of research needs will be
expanded - probably included as a separate section. However, the EPA is unclear
about CEA's suggestion about research on exposure limits and requests clarification.
3. Points #3 and #4 have been deleted.
4. The section on Recommended Actions has been deleted. It is now incorporated in
the Management Alternatives section and is characterized as ongoing activities rather
than recommended actions.
-------
ID
February 16. 1996
Review of EPA Mercury Study Report to Congress, Volume HI, Exposure Assessment
from USDA/Offlce of Risk Asaeiiment and Cost-Benefit Analysis
INTRODUCTION
USDA was asked to review Volume III of the EPA Mercury Study Report to Congress
(draft), TWs report was mandated by the Clean Air Act Amendments of 1990 which asked the
EPA to study the impacts of mercury air pollution. This review is limited to the portion of the
document covering human exposure. Volume ID. This review also cites related portions of
Volume 1: Executive Summary which refer to human exposure.
REFERENCE DOSE OF METHYLMERCURY
The model use^ for estimating the reference dose (RfD) of methylraercury was explained
in Volume 1, Chapter 3. Discussed were the uncertainty factors used in order to account for
aenaitivt subpopulations, extrapolation of animal data, lack of chronic exposure data, etc. They
then presented the human exposure incidents (Iraq, Cree Indians, etc) which were used to
estimate the reference dose based on effects seen in children whu had prenatal exposure. It is
difficult to objectively measure effects to adult humans because so many of the effects are
subjectively reported. Tests of nervous system and mental development in growing children
were U3od aa indicators of toxicity. Tha mothers of the children were reported as having little or
no effect from their direct exposure.
This strongly suggests that the children (especially, developing fetuses) me « more
wwiiive subpopuiauon than adults. The measured level from this sensitive population is then
used to set the benchmark which is extrapolated downward (as allowance for the uncertainty
factors) Lu establish the RfD, Dy starting from a benchmark based on An apparently sensitive
population, the RfD dose may be set at an even more conservative (lower) level than their normal
procedure.
This reference dose is then used in the development of the FDA action level of 1 ppm.
The EPA report uses this action level, which is very conservative, to estimate the populations at
risk (about 5% based on fish conttuuipliou). If this reference doae is too conservative, the
population at risk would be overestimated.
MERCURY LEVELS AND US PISH CONSUMPTION
Table 2-1 in Volume I lists the mercury coucautration In the most frequently consumed
fish in the United States. These top ten species account for about 80% of the fish marketed in the
US. The mercury levels were averages based on several different studies of wild caught fish and
market studies. (No indication is givan here or in Volume ITT as to how the averages were
determined or what were the ranges and standard deviations associated to these estimates.) It is
interesting to note that none of these ten fl$h groups had mcrvuiy levels anywhere near the FDA
-------
.04
action level of 1 ppm. While the mercury level in tuna fish was 0.2 ppm, all of the others were in
iLe rouge of 0.1 ppm and below,
In addition, the mercury level of catfish is based on wild caugX1 catfish. Essentially all Of
the catfish marketed in the US are firm raised. FDA repcrLs UiAt mercury Icvcla in form raised
fish may be much lower than wild caught catfish (personal communication). The testing
methods usftd by the FDA have a detection limit of 0.1 ppm so most of the results are reported u
<0.1 ppai.
A compilation of figures from USD A CREES and ERS estimate that 12 or more percent
of the fiih eaten in the US are farm raised catfish. This is a considerable portion of the total fish
exposure used hy The EPA to estimate mercury exposure to humans. If the mercury levels In
farm-raised catfish are significantly lower than wild caught (as the FDA reports indicate) then the
human exposure to mcmuy would be overestimated.
The report states that game fishermen, certain native Americans and other groups may eat
a higher level offish than the reported averages and thus increase their level of cxposmc to
marcury. Most of these, fish with the very high levels of mercury come from lakes mostly
concentrated in the Northeast quadrant of the US. These lakes have been studied and identified;
the results have been well publicised. The best evidence suggests that fish from the. affected
areas are not consumed by people. The risk message has been successfully communicated.
The basic conclusion in Volume in of EPA1 s Mercury Study Report was that flsh arc the
primary »ouroc of human exposure to methylmercuiy. Estimates of human exposure ait based
on the average person (man, woman or child) eating an average amount of the avenge fish.
While this method does provide an estimate of average exposure, it flula to consider other factors
which could provide additional information, especially when considering subpopuJations and
regional differences.
On page ES-S of Volume m in Conclusion 7. there axe assumptions about the relative
magnitudes of exposures between different species. These assumptions include the fish
consumption rate, relative consumption rate compared u> total body weight, and Cn» rate of
biomagnifir*iion between trophic levels within the aquatic food chain. These same assumptions
were considered valid for humans. It teems dear that the missing factor here is the level of
mercury in the fish that are Actually eaten by humans.. Fish come from various sources and
production methods which each have different range levels of mercury contamination.
AN ALTERNATE FISH CONSUMPTION MODEL
Following is a proposed model to look at actual fiali that are being eaten in the US. This
model is designed to suggest an alternative approach to evaluating human exposure by a more
careful evaluation of the agent of that exposure, fish consumption and the origin of the fish so
ujusvunod. As such, it is not meant to he RYhaustive in considering all of the possibilities and
outcomes. It is a simplified fault tree model without the Boolean operators (* and gates' and' or
gates' etc.) By laying out all of the possibilities, the routes of significant exposure CAT* he more
easily identified-
The diagram of the model is attached. To maintain clarity, only one iteration of ike
branch point is shown at cash level of me model. Tn concept, the branch points are potentially
repeated for each of the branches at the higher (previous) level. (For example, the branch point at
F2 is repeated for each branch of t-,.) A brief explanation of the model follows.
01
-------
FE3-SQ-1SSS ,15:52 FROM ~0 • 5191954106-10
The 'output event' at the top of the model is the total amount of fish consumed in the US each
year, here represented by <£.
At Fl3 the question is whether the fish are foreign or domestic,
At F2, the source of the fish is subdivided into farm-raised or not. (In the fully drawn model, this
branch point would appear under both branches of F,.)
At F3, the fish are further subdivided into the 2 basic categories of finfish and shellfish
(crustaceans are included with shellfish).
At F«, the categories would be divided by marine or freshwater species.
At F5, this branch point could be divided two different ways . You could consider groups offish
based on trophic level, ie fish having similar levels of mercury contamination due to similar
place on the food chain; or, you could subdivide into species or groups based on exposure
similarities.
At F6, this branch looks at the geographic source of the fish. This may be less pertinent for the
marine fish due to movement offish populations and the movements of the ocean currents.
At F7, this branch looks at the basic habitat of the fish. This branching would not be applicable
across the complete model since its focus is on the freshwater species.
When data are applied to this model, you can estimate what the actual grorv offish are being
eaten and make estimates of the mercury contamination levels in these fish. While much of this
model may not seem at first to be practical, it does provide a logical way to separate out places
where there is little exposure such as farm raised fish and shrimp. It also identifies places of
greatest exposure such as higher trophic level marine fish.
DISCUSSION
This EPA Mercury Study Report to Congress provides an enormous amount of
information about the sources and movements of mercury in our environment. However, the
models used to estimate the reference doses and the human exposure appear to be overly
conservative (underestimating the exposure level that causes toxic effects and overestimating the
human exposure). There is no doubt that mercury can be toxic and that it is highly desirable to
limit human and environment exposures as much as is reasonable and feasible. However, the
conservative estimates presented in this report may provide a rationale for even stricter control oi
mercury emissions than has already been achieved through other measures. We must consider
whether there would be any noticeable public health improvement by further restricting mercury
emissions. We already have strict controls on the use of mercury and especially wastes resulting
from manufacturing incineration. Use of mercury in manufacturing has decreased markedly in
the last two decades. More restrictive measures would cause increased public costs in a time
-------
39:53 ^SOM | TQ 51S1S5412£4^' 0.33
when there are many other areas of public health concern where such investments could provide
greater benefits to the general public. There needs to be a stronger burden of proof that there are
significant public health problems due to present mercury contamination levels, and that further
restrictions in mercury emissions would provide a detectable improvement in public health.
Any warning about consuming fish because of presumptive methyl mercury
contamination must be carefully crafted. There are large numbers of individuals who are
required for health reasons to severely limit fat intake. This places most animal products (ev .1
poultry) off limits. The alternative of market purchased fish is an important addition to the diet
of such individuals, providing animal protein and tasty choices. To unduly frighten such
individuals into avoiding market purchased fish carries with it the alternative risk that they return
to consuming more meat and poultry products. This carries consequences which could reverse
public health gains made in the prevention of heart and blood vessel disease. It would be
preferred that EPA more fully characterize the risk of human exposure and craft its message on
fish consumption more appropriately.
-------
-EB-ZQ-13S6 39=53 -ROM
31913541CE-10 = .24
Model for Human Exposure to Methylmercury via Fish Consumption
Total US fish consumption
Source of fish: imported or domestic
Source of fish: wild caught or farm raised
Category of fish: finfish or shellfish
Category of fish: marine or fresh water
F5 Trophic level: bottom feeders to top carnivores
N
F5 Geographic source
F7 Habitat: lake, river, estuary, bay, ocean
TC"
-------
EPA Response to USDA Comments (dated February 16. 1996) on the Draft Mercury Study
RfDfor Methylmercury
Uncertainties in the RfD, ranges by definition and by comparisons to other analyses.
assumptions and defaults used have been described at length in Volumes IV and VI. A brief
discussion of some salient points is in Volume I. There are quantitative analyses of uncertainty
in Appendix D to Volume IV. It is acknowledged that a certain degree of subjectivity is to be
found in some of the measures used in the Iraqi studies. The qualitative and quantitative analyses
of uncertainty In the Report indicate that paresthesia has a larger are of uncertainty attached to it
than any of the measures taken in children.
Uncertainties around some of the endpoints measured in the Iraqi children are described
in recent papers by f Cox et al 1995 and Crump et al 1995: these will be discussed in the text on
description of the RfD. A preprint of the Crump paper was made available to us. and Dr. Crump
wrote text for insertion to the discussion of the RfD. which is found in Volume IV. Both Cox et
al and Crump et al.. as well as the U.S. EPA analysis in Appendix D of Volume IV show-
considerable uncertainty in thresholds estimated from the data on late walking. Birth date
uncertainty also would have an impact on exposure uncertainty if correspondence of exposure
and gestation was estimated (by Marsh et al., 1978) from birth date to any great extent. That is,
exposure may have occurred to a lesser extent (or not at all) than assumed during the critical
period of gestation. The result would be a lower exposure associated with the observation.
depending on the width of the critical time window during gestation and on the importance of
duration of exposure in the elicitation of the particular effect. If the exposure occurred after the
critical period, any observation of an effect would be attributed to causes other than
methylmercury and be included in the background.
The expectation that late walking would be an endpoint with some uncertainty was one of
the considerations that led the U.S. EPA to base its RfD on the combined incidence of four
developmental endpoint-in children reported in Marsh et al. (1987). Given the differences in
methodologies and data sets used, an eight fold range in quantitative risk estimates is not
considered large by risk assessors.
The objection has been raised that the RfD based on studies in Iraqi children is based on
only four data points. The reliance of the RfD on four data points is obviated to a large extent by
the application of the curve fitting procedures and calculation of the benchmark dose (BMD).
U.S. EPA has individual data from the Iraqi study.. The grouping of maternal hair concentrations
published in Seafood Safety was done to provide consistency with other analyses. Many months
ago we tested the effect of changing groupings on the BMD. Results of that analysis showing that
no there is no effect on the calculation of the BMD will be included in Volume IV in discussion
of the RfD.
The BMD used as the basis for the RfD is the lower limit of a 10% effect level. This w as
chosen because there are studies published which indicate that this is a reasonable predictor of a
-------
NOAEL for developmental effects. These studies have indicated that use of this level is no more
conservative than estimating NOAELs from studies by more typical means. An RfD is defined
as an estimate (with uncertainty spanning perhaps an order of magnitude) of a daily exposure to
the human population (including sensitive subgroups) that is likely to be without an appreciable
risk of deleterious effects during a lifetime. Note that this definition acknowledges applies to
sensitive subpopulations. which for methylmercury are expected to be developing human
organisms. Setting an RfD on effects in the sensitive population in the absence of an additional
uncertainty factor for protection of sensitive groups is the U.S. EPA 's established methodology.
This was done for methylmercury.
One USDA comment stated that FDA used the L'.S. EPA RfD to develop an action level
of 1 ppm. The comment then states that U.S. EPA uses this action level to determine populations
at risk. This is not how the analyses were done. The FDA action level was not based on the U.S.
EPA RfD but was set through action of courts: it considers FDA's own ADI which is four-fold
higher than U.S. EPA's. It is agreed that the Mercury Study should make no comment on the
FDA action level.
Mercury Levels and U.S. Fish Consumption
The discussion of the data sources for mercury content of freshwater fish will be modified
to indicate that the data are from wild fish stocks and not from farm-raised fish.
Data describing methylmercury concentrations in marine fish were predominantly based
on the National Marine Fisheries' Service (NMFS) data base, the largest publicly available data
base on mercury concentrations in marine fish. In the early 1970s, the NMFS conducted testing
for total mercury on over 200 seafood species of commercial and recreational interest (Hall et a!..
1978). The determination of mercury in fish was based on flameless (cold vapor) atomic
absorption spectrophotometry following chemical digestion of the fish sample. These methods
are described in Hall et ai. (1978).
Although the NMFS data were initially compiled beginning in the 1970s, comparisons of
the mercury concentrations identified in the National Marine Fisheries Service's data base with
compliance samples obtained by the United States Food and Drug Administration indicate that
the NMFS data are appropriate to use in estimating intake of mercury from fish at the national
level of data aggregation. In 1994, Greg Cramer of the Office of Seafood of the Center for Food
Safety and Applied Nutrition of the US Food and Drug Administration reported on "Exposure ot
U.S. Consumers to Methylmercury from Fish". He noted that recent information from National
Marine Fisheries Service (NMFS, 1990) indicated that the fish mercury concentrations reported
in the 1978 report do not appear to have changed significantly. The US FDA continues to
monitor methylmercury concentration in seafood. Cramer (1994) observed that results of recem
US FDA surveys indicate results parallel to FDA's and NMFS's earlier findings. To illustrate.
Cramer estimated the mean methylmercury content of the 1973 samples of canned tuna at 0.21
ppm Hg, whereas a recently completed survey of 245 samples of canned tuna was 0.17 ppm H^
These data are considered to be comparable, although the small decrease reported between these
two studies may reflect increased use in canned tuna of tuna species with slightly lower average
-------
methylmercury concentrations. The National Academy of Sciences' National Research Council's
Subcommittee on Seafood Safety (1991) also assessed the applicability of the NMFS1 1970's
data base to current estimates of mercurv concentrations in fish. This subcommittee also
* ._ *
concluded that the 1978 data base differed little in mercury concentration from FDA compliance
samples estimating mercury concentrations in fish. The averages were obtained from the NMFS
public data base. The averages were obtained by determining the mean of the means for each
species considered for a particular type of fish. For example, the CSFII data only note the
consumption of freshwater catfish; they do not indicate the type of freshwater catfish. The entire
group was averaged.
There are other marine fish pollutant data bases (e.g.. one maintained by the FDA). This
data base and another will be described in the analysis. The average concentrations will be
compared to those utilized in the assessment in Appendix H.
The USDA comments also indicates that fish from contaminated lakes are not consumed
by humans. While EPA aggrees that consumption of fish from these water bodies will be
decreased, it will probably not be eliminated. Studies indicate that, despite the issuance of health
advisories, anglers will fish and consume fish caught from a posted water body. Consumption of
fishing from posted waters occurs both through ignorance of an advisory (Knuth, 1994) or
disregard of an advisory (Connelly et al., 1990). It also should be noted that not all water bodies
that are fished have been tested for high methylmercury levels in fish.
Discussion of cross-sectional food consumption surveys vs. longitudinal surveys will be
enhanced. Comparison of advantages and disadvantages of both types of surveys will be included
as well as the opinion of the NAS on this subject.
Data from a longitudinal survey will be included. The analyses of these data are in
. substantial agreement with the analyses done on the CSFII 89/91 survey; that is between 2 - 5%
of the general U.S. population consumes on the average 100 g or more per day of fish/shellfish.
Included in this high-end group are subpopulations such as recreational anglers, people who
augment their diet with self-caught fish for economic reasons and people who consume large
amounts of fish from person preference or cultural mores. A chart on comparison of high end fish
consumption values from various surveys will be added.
USDA asserts that estimates for "average" people consuming typical concentrations of
methylmercury are the only ones considered. The analysis of methylmercury exposure through all
fish utilizes the distribution of seafood consumption among respondents to the CSFII 89/91. This
includes those who consume smaller amounts of fish as well as those who consume qualities
greater than average. One of the unique advantages of the CSFII 89/91 data base is that it
includes measures of body weights. These were utilized to provide exposure estimates on a per
kilogram body weight basis.
To confirm the that the high daily consumption rates measured in the upper percentiles of
the cross- sectional CSFII 89/91 study persisted. EPA examined fish consumption rate studies ot
other high fish consuming groups: anglers and certain Nativa American Tribes, who traditional 1\
-------
consume large quantities offish. These studies (some of which are different in design or purpose)
corroborate that people consume high quantities of fish over longer periods of time. EPA also
examined longitudinal consumption studies, specifically, the XPD 1973/74 assessment.
Individuals at the upper end of the distribution again were shown to consume over.100 grams per
day of fish.
It should also be pointed out that the analysis of atmospheric anthropogenic mercury
sources did not focus on typical or average individuals. These individuals were assumed to be
fishing near a large emission source.
The fish consumption patterns of children were also examined. As with the adults, the the
distribution of these consumption patterns was utilized. As a result not only was the mercury
intake of a typical child assessed but those of children who consume larger amounts were also
were also estimated.
Volume III describes many of the local and regional studies of mercury content of fish.
These local studies demonstrate the wide range of fish tissue contamination in the U.S. These
data were not utilized in the assessment because of the assessments national focus. If the
U.S.D.A. is interested in a state-wide assessment, EPA suggests that the recent paper by Stern et
al., 1996 may be informative.
An Alternate Fish Consumption Model
Thank you for sending the brief discussion of this interesting approach. Unfortunately it
is too late in the process of report delivery to include a major novel assessment methodology.
There will be opportunity for further evaluation of this and other new methods at a later time
during the anticipated Federal re-evaluation of data on methylmercury.
The U.S. EPA acknowledges that FDA is responsible at the Federal level for the safety of
the commercial seafood supply for human consumption. In order to characterize the potential
risk to humans, however, it is necessary to describe exposure from various routes. The majority
of exposure to humans (outside of occupational settings) is through consumption of
methylmercury-containing fish. An early draft of the Mercury Study Report to Congress focused
solely on fresh water fish in the contiguous forty-eight states. Our peer review (including both
Federal and external scientists) were firm in their critique that this focus was unacceptable for a
nation-wide characterization of risk consequent to mercury emissions. In the course of
evaluating the role of total fish consumption by various segments of the U.S. population, the
graphs alluded to in your memo were generated. U.S. EPA acknowledges that the graphs, while
scientifically correct, were liable to misinterpretation. To prevent this from occurring, the risk
characterization (described in Volume VI and summarized in Volume I) has been changed in the
following ways.
1. All graphs will be removed from Volume VI.
2. The population consuming more than 100 grams of fish per day will be calculated. This
-------
analysis was already included in the risk characterization based on cross-sectional consumption
data (CSFII 89/91), but will also be done using longitudinal data.
3. The text will discuss the population consuming > 100 grams fish/day, focusing only on
women of child-bearing age and children.
4. The types of fish eaten by people consuming > 100 grams day will be described. For top
species, the mercury levels of the fish they are eating will be listed (i.e.. if they are eating
primarily commercial fish). The outcome of this is that a consumer can see that most species they
are likely to eat are low in mercury. The exceptions to this last statement are shark, swordfish
and barracuda: that FDA and some fishing industry groups have issued advice on consumption of
these species by pregnant women will be included.
5. For freshwater fish, exposure to methylmercury will be discussed in the context of the RfD
but no graphs will be shown. The marine species contribution to the overall exposure for
populations consuming large amounts of fish will be discussed qualitatively (that is. in terms of
proportion offish obtained from commercial sources).
6. Scenarios of high end fish consumers are being described.
In addition, to facilitate sending an appropriate message about fish in commerce U.S.
EPA has agreed to include the transmittal letter from Administrator Browner at the beginning of
the Mercury Study report to Congress. In addition, language from this letter will be included in
two other places in Volume I
References
Connelly, N., T. Brown and B. Knuth. 1990. New York Statewide Angler Survey, 1988, New-
York State Department of Environmental Conservation, Albany, 1990.
Cramer GM (1994) Exposure of U.S. consumers to methylmercury from fish. pps. 103-118. In:
DOE/FDA/EPA Workshop on Methvlmercurv and Human Health Eds: Moskowitz PD, Saroff L.
Bolger M, Cicmanec J. and Durkee J. Conference Number 9403156. Published through:
Biomedical and Environmental Assessment Group, Brookhaven National Laboratory, Upton,
New York.
Hall RA, Zook EG, and Meabum GM (1978) National Marine Fisheries Service Survey of Trace
Elements in the Fishery Resource. NOAA Technical Report NMFS SSRF-721, National
Technical Information Service No. PB 283 851, March. U.S. Government Printing Office,
Washington. DC cited in Seafood Safety. Chapter 6. "Chemical Health Risk Assessment -
Critique of Existing Practices and Suggestions for Improvement", pp. 187-281. National
Academy of Sciences Press, Washington, DC, 1991.
Knuth, B. 1994. Policy Implications of Communicating Health Risks from Fish Consumption in
Mercury Pollution Integration and Synthesis. Eds. Watras. C. and J. Huckabee. pp 643 -648.
-------
SENT BY:OFFC SCI & TECH POLICY: 2-16-36 : 5:13PM :ENVIROIMMEST DIVISION- 313 541 0840:2 2/ 5
EXECUTIVE OFFICE OF THE PRESIDED"
OFFICE OF SCIENCE AND TECHNOLOGY POLICY
WASHINGTON, D C ' -
February IS, 1996
MEMORANDUM FOR TJ. GLAUTH1ER, OMB
CC: JACK GIBBONS, OSTP
SALLY KAT7.EN, OIRA
KATIE MCGINTY, CEQ
AIJCIA MUNNELL, CEA
BRAD CAMPBELL, CEQ
KEVIN NEYLAND, OMB
FROM: » BOB WATSON
SUBJECT: OSTP Comments on the EPA Mercury Report
Mercury is a major environmental issue. It is therefore critical that this EPA report
utilize the best science to he both credible and transparent so that alt stakeholders can huy-in to
its analysis and conclusions.
OSTP continues to have major concerns with the report as it stands. Like us, it in
quite clear that other agencies also have major technical difficulties with the science presented
in the report. While we have seen some agency comments, we are still awaiting comments
from FDA and the health section of DOE who we know have concerns and expertise relevant
to specific sections of the report.
Because of this report's likely level of visibility, it is critical that ail comments raised
hy the agencies be adequately addressed. OSTP believes that interagency meetings devoted to
a item by item discussion of the comments be conducted to resolve issues in order that
appropriate revisions can be made. Prior to final agency review and sign-off, a time-line for
publication can be determined. The interagency meeting being planned for next week should
review where we stand.
OSTP' comments follow. Some aro new points from further examination of the report
and review of other agency comments, while others reiterate points made earlier by OSTP.
Volume I (Executive Summary)
The analysis and conclusions presented in the Executive Summary must accurately match those
in the body of the report. As discussed in the December interagency meeting, this is not
entirely die case in the present draft. In particular, the executive summaiy does not adequately
convey the health risk information conveyed in the larger report. The risk characterization
should be presented in terms which are easily comprehensible to the general public, perhaps as
categorical rather than numerical point estimates. The executive summary should also discuss
the variability of risk by demographics and location.
-------
QY-UfFC SCI i- TECH POLICY. 2-16-36 : 5- 14PM : E.NV J Ro.NMEYT DIVISION 913 541 0840,* 3/
Volume II (Inventory of Anthropogenic Emissions)
A number of potential contributors of anthropogenic emissions of mercury listed on
Table ES-3 (agricultural burning, landfills, primary mercury production, and refineries) are
listed as having no available data. In the case of primary mercury production and refineries,
which might be expected t» he significant sources, this is surprising given present reporting
requirements. For instance, oils from parts of California contain globules of mercury which
must be extracted during refining. Lack of data from these sources aggravates the relative
contribution of those sources whose emissions are measurable. If the data for these sources do
not exist, the report should forcefully advocate research to determine their emissions,
As shown in Table ES-8, the degree of uncertainty for ail sources of anthropogenic
mercury emissions is moderate to high. However, the table also reports emission numbers
with several significant figures which do not appear to be justified based on available data.
The report notes (on 5-1) that one peer reviewer estimates that missing sources could
contribute as much as 20% more mercury to the U.S. total. The report should explicitly state
that due to ihe limitations of available data and the range of uncertainty, the numbers should he
regarded only as best estimates. Based on the available data, the report should also forcefully
and clearly point out the need for, and types of, additional research to more accurately quantify
anthropogenic mercury emissions.
Though natural mercury emissions are not the subject of the report, a more iliorough
discussion of these is needed, together with best estimates of total natural emissions so that
valid comparisons can be made with anthropogenic emissions. "Natural" mercury emission;
may actually be misleading since much of the mercury now geologically bound in soils or
marine sediments may have been deposited from historical metals smelting. Re-emission of
this historic anthropogenic mercury may explain why fish mercury contents appear to remain at
relatively constant levels despite significant decrease in mercury use. The report, and the
executive summary needs to address the relevance of older, anthropogenic emissions which
have now been fixed in the "natural" system.
Volume III (Exposure Assessment^
The EPA models of mercury deposition processes are used to estimate mercury in the
environment. However, differences between the model results and measured mercury
concentrations in fish suggest that there are problems with the models or model inputs. This
problem needs to be explicitly discussed.
The report implies that the primary mechanism for human exposure is through fish
consumption. The report should also address the relative contributions of human risk from
both air emission sources and water discharge sources as well as from fish ingesiion. If fish
consumption is the principal pathway for human exposure, the report should explicitly address
the different mercury pathways to water. Additional discussion is needed in the report of other
potential pathways that have been suggested such as mercury vapor nutgassing and methylation
from dental fillings which account for usage of 41 tons of mercury per year (compared with an
estimated 40 toas of mercury emissions per year from all utility sources).
-------
ol'OrrC NCI & TECH POLICY: 2-16-26 ; 5.14PM -ENVIRONMENT DIVISION- 313 541 0840;# 4-' 5
csr? @
The report should make some attempt to identify vulnerable subpopulations and
analyze the nsks to these populations including coastal populations, residents of the northern
and eastern U.S., Native Americans, and subsistence and commercial fishers.
Mercury levels in fish: Some of the data on mercury levels in fish is twenty years old,
calling into question analytic techniques and data accuracy. The report should deal with the
issue ot' data accuracy with regard to data age and source. The report does not take into
account variability in mercury content offish by species and geographic location. It is known
that some shellfish and finfish (sharks, swordfish) concentrate mercury. This variability
should be discussed and evaluated in the report. Regional variations in fish mercury content
also must be addressed. Fish from relatively polluted eastern estuaries probably contain
significantly more mercury than those from the Gulf of Alaska. Fish species, stocks, and
geographic ranges should be designated as low. medium, and high risk.
OSTP continues to believe that the report should include maps of the United States that
graphically summarize the risk from methyl mercury. One map could show die distribution ot
mercury levels in fish tissue or some other surrogate measure, independent of consumption
patterns. A second map could graphically depict fish consumption patterns independent offish
mercury levels. A third map combining the data from the first two would effectively depict
populations and locations at risk.
Volume IV fHealth Effects')
OSTP has serious concerns about the statistical methods used to estimate the
benchmark dose, in particular, the grouping of incidence data from the Iraqi study. Putting
data from 81 individuals into five groups to estimate 3 model parameters can lose statistical
validity. Do different groupings affect model parameter values? The most defensible approach
is to fit a model to ail of the data points.
Studies on health effects of mercury associated with fish consumption in the Republic
of Seychelles were recently published as a series of 11 papers in Neuroto.rtcology. That study
was much larger than the study of high mercury exposures in Iraq, which was used by EPA to
estimate benchmark and reference doses for methylmercury. It is essential that EPA compare
these two studies by examining whether the model developed with data from the Iraqi study is
predictive of observed data from the Seychelles study. Consistency between the two studies
would substantially strengthen die conclusions of the EPA study; inconsistency would indicate
that either the model is incorrect or that other factors are influencing the dosc-rcsponsc in one
or both of these studies.
Volume VII (Control Technologies)
This volume needs to more directly state that there are significant uncertainties with
technologies lor mercury control and that a number of different control technologies will he
required for different mercury forms and sources. Paniculate control and carbon filter devices
have Iiighly variable success ratios based on a variety of physical parameters and should not be
seen as a "quick fix."
OSTP defers tc> the CEA comments in regards of the social cost analysis of the report.
OSTP concurs with UOE's comments in regard of control technologies and in particular notes
-------
<-" i ir.\ i JiViiiO.N-' "313 541 0840:;? 5/ 3
the DOn estimates of utility costs for mercury control which are several orders of magnitude
higher than sulfur dioxide controls.
-------
EPA Responses to OSTP Comments (dated February 15. 1996) on the Mercury Stud\.
(Memorandum from Bob Watson to T.J. Glauthier.i
These comments from OSTP are summarized and addressed in the "Summary of Federal
Comments on the Mercurv Studv -- February 22. 1996" located in the front of this section
-------
EXECUTIVE OFFICE OF THE PRESIDENT
COUNCIL OF ECONOMIC ADVISERS
WASHINGTON. O.C 20500
February 14, 1996
Q l(D
MEMORANDUM FOR TJ GLAUTHIER, OMB
CC: SALLY KATZEN, OIRA
KATIE MCGINTY, CEQ
BOB WATSON, OSTP
FROM: ALICIA MUKKELL,
SUBJECT: Further Comments on EPA Draft Mercury Study
CEA has had a chance to examine the August 1995 version of
Volume vii of the report, "An Evaluation of Mercury Control
Technologies and Costs." Notwithstanding the title of the
volume, Chapter 3 includes an effort to assess the damages
associated with mercury contamination. CEA has serious concerns
with this discussion and believes that it needs a major rewrite
before it can be acceptable. Detailed comments are attached to
this memo. Our basic concern is that the discussion is both
conceptually and empirically flawed in ways that convey a sense
of much greater potential loss from mercury contamination than
the evidence can support. The evidence supports only the
following conclusions about damages: (a) some effect on
recreational fishing values from having to eschew consumption of
catch, but this damage at the very least is uncertain; (b) some
monetized losses from adverse health effects, assuming that the
healtft effects themselves are well substantiated (which may not
be the case), though these losses also are uncertain; (c) very
uncertain ecological effects, though selected species (one
endangered) appear to be affected; and (d) some lost cultural and
subsistence values, though again the magnitude is very unclear.
These conclusions suggest a need for concern, but they do not
constitute a call for action.
We also have major problems with Table ES-5 in the Summary.
This table flags a number of technologies as having "High"
emission reduction potential, but these technologies have not
been subjected to any cost-effectiveness analysis in the study.
Some of these, like fluorescent tube breakage, battery
composition, and the capacity to separate materials from waste
streams before combustion, are in fact subjects of controversy in
the RCRA context. Others, like fewer teeth being burned in
crematories, involve pretty small mercury quantities and just
seem silly in any event. This table needs to be broken up and
scrubbed to indicate which technologies have high technical and
economic (coat-effectiveness) potential and which have high
-------
i 5 ' 'r
technical potential but unknown or high costs, with an emphasis
on the more meaningful options.
»
We would largely defer to DOE on the review of the coat and
cost-effectiveness calculations for controlling mercury emissions
from utility boilers and other combustion sources. We would,
however, note the confusion that arises when the Executive
Summary indicates a. large particulate load produced by carbon
injection in utility boilers (page E-15), but Appendix B seems to
indicate no increase in disposal costs (Table B-7, page B-ll).
My understanding from the meeting, we had in your office was
that after you had a chance to collect and examine the agency
comments on the study, we would have another meeting to take
stock of what to do next. I believe that we need to have such a
meeting before a release strategy for the report is finalized.
After you have reviewed the comments, please let me know where
matters stand. Thank you.
-------
Detailed comments on "social cost" analysis in Chapter 3,
Volume VII of EPA mercury study:
(1) The discussion of recreational fishing on pages 3-13 through
3-15 and 3-20 presents lots of estimates of the total value of
this activity and its importance to the economy, but no estimate^
-------
N 0 . • j C 4 P , 1 0
C£i\
4
the Minimata incident, raising questions about comparability to
the US case).
Recommended fix: delete entire discussion. It would be poaajblf.
to replace it with valuation information drawn from other
contexts £e.o.. willingness to pav to avoid IP effects of lead
exposure^ to the extent that there are reliable dose-response
relationships that can be used.
(4) The discussion of ecological effects on pages 3-22 through
3-27 seems to be mostly speculation and estimates of total
potential value that do not describe the marginal cost -of mercury
contamination (as in (1) above). As the report notes, the
affects at a community or ecosystem level remain very uncertain?"
the only reasonably solid material is the potential harm to the
endangered Florida panther and some localized loon and other
populations (not involving endangered species).
Recommended, fix: replace gntire jj[i3cusgion with a_ brief
statement that there could be harmful ecological effects, but
these are un^cerjp^in othpr^han a couple of examples.
(5) The discussion of subsistence and cultural values on pages
3-16 and 3-20 touches upon issues that are important to consider,
but it suffers from woeful lack of information and documentation.
Only two pieces of information are offered — a report by the
National Wildlife Federation, and a report EPA carried out with
affected Tribes in Wisconsin. Even if one accepts the difficulty
of attaching monetary values in these cases, neither reference
provides quantitative information on the scale of the problem and
both could be criticized as being less than completely objective.
Without more information about the extent of subsistence and
culturally motivated angling, and how disruptive its curtailment
is, the reader of this volume is left with a political argument
but no analysis.
Recommended fix: acknowledge the importance of these Issues. _but
put more documentation in this discussion of tfce scale of the
problems and be clearer about the basic uncertainties surrounding
these assessments.
-------
EPA Responses to CEA Comments (dated February 14, 1996) on the Draft Mercury Study
This memorandum is in response to Alicia Munnell's comments on the draft Mercury
Study which were transmitted to T.J. Glauthier. OMB. dated February 14. 1996. Specific
comments offered by the Council of Economic Advisers (CEA) are addressed below.
Social Costs
The CEA provided detailed comments regarding the "Social Costs" section of Volume
VII. This section (attached) was deleted in it s entirety.
Pollution Prevention
Concerns were raised by both CEA and OMB about Table ES-5 (see attached) in
the Executive Summary which listed source categories with varying potential for pollution
prevention alternatives to end-of-pipe control technology. The objection was that the alternatives
mentioned had not been subjected to a cost analysis. This table was removed and language was
substituted that stated that when considering opportunities for pollution prevention, it is
important to consider both the achievable reductions and the costs of these options.
-------
3.2 Social Costs of Mercury Pollution
The "social costs" of mercury pollution can he defined as the negative impacts ot
environmental mercury contamination on human society.1 These costs can take a variety ot
torms. They can include negative impacts on market goods with a well-defined economic value.
such as reductions in the catch. Social costs also can include negative effects on other items that
people care about, but are more difficult to value because they are not traded in markets. An
example is damages to populations ot" endangered species. Other social costs may have some
component of measurable economic value, but also have other, less tangible values. Lost
recreational and existence values associated with healthy and abundant fish and wildlife, as well as
adverse health effects associated with environmental contamination, fall into this latter category.
This section summarizes available information on the social costs of mercury pollution. It
focuses on four categories of social costs: (1) the cost associated with fisheries, especially
potential economic losses to the freshwater commercial and recreational fishing industry and
diminished cultural values. (2) the costs of human health effects, (3) the cost associated with
ecological effects, including potential damages to endangered species and (4) the public's
willingness to pay to reduce toxic air contamination.
3.2.1 Costs and Values .Associated with Fisheries
Economic and social values associated with fisheries can be grouped into five general
categories: commercial, recreational, subsistence, cultural and non-use values. These values can
be significantly reduced by mercury contamination. To illustrate the potential magnitude of such
losses, the first section describes values in each of the five categories, providing dollar values or
other quantitative measures when available. The second section discusses fish consumption bans
and advisories due to mercury contamination, the primary cause of economic and social losses
related to fisheries. The third section describes the cost implications of such fishery
contamination and advisories, including examples where fishing values have been diminished by
mercury contamination.
3.2.1.1 Values Associated with Fisheries
Present values associated with fisheries are summarized below, as they appeared in the
literature collected and reviewed for this study. The emphasis in this section is on the value of
freshwater fisheries since this study focuses on the impact of mercury deposition in freshwater
bodies. Commercial values are presented in pounds of fish (landings) and U.S. dollars (market
values). Recreational values are presented as fishing-related expenditures, or values based on
contingent valuation or travel cost studies. Subsistence and cultural values are discussed
qualitatively, and the non-use values presented are based on contingent valuation studies.
Commercial Fishing Values. In 1992, U.S. commercial fisheries in the Great Lakes alone
landed over 30 million pounds of finfish with a total value of approximately $22 million (NOAA.
1993). Notably, native .American Tribe-licensed commercial fisheries comprised almost half of the
total dollar value of Great Lakes commercial catch (FWS. 1994). In states of the Great Lakes
region, substantial revenue is generated by commercial fishing. Volumes of fish caught in 1992
1 These costs of pollution could also be called the benefits of pollution reduction. This report.
however, uses only the term social cost.
September 7. 1995 3-12 DRAFT: Do Not Cite or Ou<
-------
and approximate values of landings in the top seven U.S. freshwater commercial fishing states are
indicated in Table 3-5 (NOAA. 1993). Similar national statistics for non-Gjeat Lakes commercial
treshwater fisheries (e.'g.. the Mississippi River) are unavailable (NOAA. 1993).
Table 3-5
1992 Commercial Fish Landings in Great Lakes States
State
Michigan
Wisconsin
Ohio
Indiana
Pennsylvania
Minnesota
Total
Thousand Pounds
of Fish Landed
15.057
8.484
0
4.985
1.358
4X5
269
30.638
Total Dollar Value
S10.337.0UO
5.914,000
2.555.000
2.550.000
395.000
101.000
521,852.000
Source: NOAA, 1993
Recreational Fishing Values. In 1991. 31 million Americans adults took advantage of
freshwater fishing opportunities, spending a total of $15.1 billion on sportfishing trips and
equipment. American freshwater anglers took 390 million fishing trips (an average of 13 trips
each) and fished 440 million days (and average of 14 days each). Anglers spent an average of
$491 annually, or approximately $43 per fishing day (one fishing person going fishing on one day
is referred to as one fishing day) (U.S. Fish and Wildlife Service, 1993a).
Information'specific to non-Great Lakes freshwater anglers and Great Lakes freshwater
anglers is provided in Figure 3-1. For non-Great Lakes anglers, total fishing-related expenditures
equalled 13.8 billion (for 431 million fishing days), and Great Lakes fishing related expenditures
equalled $1.3 billion (for 25 million fishing days). Non-Great Lakes fishers averaged $458 per
participant in 1991, and Great Lakes fishers spent an average of $524 per angler. Per fishing day
Non-Great Lakes participants spent less than Great Lakes participants (i.e., $32/day vs S53/day)
(U.S. Fish and Wildlife Service, 1993a).
Table 3-6 summarizes and compares examples of reported recreational fishing day values
(by Americans engaged in sportfishing) based on travel cost methods, contingent valuation
methods, or expenditures. For the purpose of this comparison, all values have been normalized i
1992 dollars. Overall, the values range from $16 to $69 per day, with a mean of about $38 per
day.
September 7. 1995
3-13
DRAFT: Do Not Cite or Qu<
-------
Figure 3-1
1991 Recreational Fishing Expenditures
Freshwater Trip and
Equipment Expenditures
freshwater, except
Great Lakes
SI 3.3 billion
Freshwater Fishing
(in millions)
440
Total
Freshwater, except
Great Lakes
Great Lakes
Fresnwater
Anglers
Detail does not add to total
bocausa of muitiole responses.
Freshwater
Days
Freshwater
Trips
Source: FWS 1993a
September 7. 1995
3-14
DRAFT: Do Not Cite or (X
-------
Table 3-6
Examples of Values Reported for Recreational Fishin"
Type of Value
Mean benefit/day of anadromous fishing
Mean benefit/day of warm-water fishing
Mean benefit/day of cold-water fishing
Average value of a fishing day for trout.
including resource costs (travel cost methods)
Average value of a fishing day for trout,
including resource costs (contingent valuation
method)
Average value of a fishing day for catfish,
including resource costs (travel cost methods)
Average value of a fishing day for catfish.
including resource costs (contingent valuation
method)
Total expenditures (including memberships,
magazines, etc.) per day for sportfishing in
general
Trip and equipment expenditures per day for
sport fishing in general
Trip and equipment expenditures per day for
non-Great Lakes freshwater sportfishing
Value
(1992 S)
S67
S24
$38
$23-35
S3 1
$16-23
S22
S48
S42
S33
Source
Walsh et al. 1988 (in
U.S. EPA. I993c)
Walsh et al. 1988 (in
U.S. EPA. 1993c)
Walsh et al. 1988 (in
U.S. EPA. 1993c)
Vaughan & Russell.
1982"
Charbonneau & Hav
1978 (in Vaughan &.
Russell. 1982)
Vaughan & Russell.
1982"
Charbonneau & Hay
1978 (in Vaughan &.
Russell. 1982)
U.S. Fish and Wildlife
Service. 1993a
U.S. Fish and Wildlife
Service. I993a
U.S. Fish and Wildlife
Service, 1993a
The freshwater sportfishing industry has far reaching effects on the U.S. economy, as
indicated in Table 3-7 (SFI, 1992). Anglers' direct expenditure of over S15 billion generated A
total economic output of over $46.5 billion." The industry provides over 633,000 jobs and
worker earnings of more than $13.1 billion. Annual recreational fishing expenditures for 1991 ot
the top three inland states which support a substantial freshwater fishing industry are Michigan
($1.1 billion) Wisconsin ($0.85 billion) and Minnesota ($0.80 billion). Recreational fishing
expenditures, however, are considerable in every one of the 50 states. For example, the states
with the smallest recreational fishing expenditures in 1991 were North Dakota and Delaware.
" Includes the value of retailers' direct sales to anglers, suppliers' sales of goods and services to
retailers, and of wholesalers' and manufacturers' sales to suppliers. Also includes other indirect effects
such as expenditures by the employees of all these organizations.
September 7. 1995
3-15
DRAFT: Do Not Cite or
-------
Table 3-7
Economic Impacts of Sportfishing by Water Type
Water Type
Freshwater
(excluding Great
Lakes)
Great Lakes
Total
Expenditures
Output
Earnings
Billion dollars
14.0
1.4
15.4
42.4
4.1
46.5
11.9
1.2
13.1
Jobs
577.000
56,000
633.000
Source: SR IW2
where a total ot S46
million and $60 million.
respectively, was spent
on sporttishina (SFI.
1992).
Subsistence
Fishing and Cultural
Values. The value ot
subsistence fisheries is
not captured by avail-
able recreational
fisheries data. For
subsistence fishing
individuals, an added
benefit is the avoided
cost of buying food for
which the captured fish substitutes. Unfortunately, very little information is available on the value
ot subsistence fishing. Information that is available focuses mainly on Native American
subsistence anglers. For example, members of the Mille Lacs Band of Chippewa Indians in
Wisconsin catch 90 percent of the fish they eat themselves (U.S. EPA. 1992b).
Fishing also is an integral part of the cultural and/or spiritual practices of many U.S.
subpopulations. The scant information available on this topic focuses mainly on Native
Americans. The National Wildlife Federation, for example, cites the Final Report: Economic.
Social, and Psychological Impact Assessment of the Exxon Valdez Oil Spill in describing the cultural
importance of fish, other animals and the environment in general to traditional Native American
tribes:
Damage to the subsistence resources—damages the whole [Native]
culture.... Embedded in the activities of hunting, fishing, and
gathering is a way of life, a set of values, a way of seeing the world
that values bears, salmon, eagles, and water as spiritual and social as
well as economic resources. Threats to the resources and activities
that are so fundamentally embedded within the Native culture thus
threaten that very culture itself. (Impact Assessment, Inc.. 1990 as
cited in NWF, 1993.)
Other subpopulations where fishing and/or fish consumption are an important part of the
culture and traditions include some Asian American communities, and long-time subsistence and
commercial fishing communities (e.g., watermen and associated communities on the Chesapeake
Bay).
Cultural and spiritual values do not lend themselves to quantification. Nonetheless, the\
are important values that must be noted when considering the potential costs of mercury
contamination to society.
Non-Use Values. Surveys show that some Americans who do not fish are nonetheless
willing to pay over S100 a year (1992 dollars) to make their state or national waters fishable.
Non-users surveyed in the early 1980s were willing to pay S180/household (1992 dollars) to ha\c
September 7, 1995
DRAFT: Do Not Cite or Qt
-------
U.S. waters made fishable (Mitchell and Carson 1984 as cited in U.S. EPA. 1993c). Mitchell and
Carson ( 1986 as cited in U.S. EPA. 1993a) estimated that about two-thirds of this amount. -
roughly $120/household. represented the amount respondents were willing to pay to make their
in-state waters fishable. More current survey data, since the Mitchell and Carson study in the
early 1980s, have not been located. U.S. EPA (1994a), however, has suggested that .Americans
may be willing to pay more for clean water today than a decade ago because of (1) greatei
awareness of water quality conditions and (2) stronger preferences for water quality
improvements.
3.2.1.2 Fishing Bans or Advisories Due to Mercury Contamination
Mercury and other toxic contaminants can decrease the values discussed above by reducing
U.S. fishing opportunities. One concrete way in which fishing opportunities may be reduced is
through fish consumption bans or advisories, enacted by the Food and Drug Administration
(FDA) (bans on commercial fishing) or by state governments (advisories and bans on recreational
and subsistence fishing).
Mercury is the only metal that the FDA regulates in fish (WSDE. 1989). FDA presently
bans the commercial sale of finfish or shellfish when mercury levels in tissue exceed 1 /u.g
mercury/g tissue (i.e., 1 ppm). A 1979 court decision raised the FDA mercury action level from
0.5 to the present I ppm because of economic impacts of the 0.5 ppm standard on the
marketability of several marine fish species (GLFCATF, 1990).
Some states use the FDA's limit of 1
ppm mercury in fish tissue as a trigger for
issuing consumption advisories, but many
other states have lower trigger levels.
commonly 0.5 ppm (Table 3-8). Though fish
consumption advisories and bans vary from
state to state, they generally:
(1) Provide information to recreational
and subsistence anglers about
contaminant levels in finfish and
shellfish caught from specific water
bodies; and
(2) Recommend how much fish tissue
from particular fish species the general
population and specific subpopulations
can consume.
State mercury fish advisories often
provide more stringent consumption
recommendations for sensitive subpopulations,
such as children. Because mercury can pass
through the placenta or mother's milk to
sensitive fetuses or infants, stringent
consumption recommendations are often
Table 3-8
Example Action and Trigger Levels for Fish
Consumption Bans or Advisories
Nation or
State
U.S. (FDA)
Canada
Arkansas
California
Illinois
Indiana
Ohio
Pennsylvania
New York
Michigan
Wisconsin
Action or
Trigger
Level
1 ppm
0.5 ppm
1 ppm
0.5 ppm
1.0 ppm
1.0 ppm
1.0 ppm
1.0 ppm
1.0 ppm
0.5 ppm
0.5 ppm
Source
40 FR 45663
GLFCATF 1990
DOH 1993
WSDE 1989
GLFCATF 1990
GLFCATF 1990
GLFCATF 1990
GLFCATF 1990
GLFCATF 1990 •
GLFCATF 1990
GLFCATF 1990 !
September 7. 1995
3-r
DRAFT: Do Not Cite or
-------
provided for pregnant women, nursing women and women who plan to have children.
As of 1994. 1.ZS4 consumption advisories due to toxic contamination of fish were issued in
the U.S. (U.S. EPA, 1994b). Sixty percent of these advisories were due to mercury contamination
(Figure 3-2) (U.S. EPA. 1994c). As shown in Figure 3-3. there are 37 states that have at least
one waterbody under mercury advisory and another seven states with statewide mercury advisories
(U.S. EPA. 1994b). In U.S. waters in the Great Lakes region, mercury is the only metal for
which advisories have been necessary (GLFCATF. 1990). Note that different states cannot be
compared based simply on the number of
advisories because (1) the approaches for
determining trigger levels for advisories vary
from state to state and (2) more water
bodies are tested in some states than in
others.
Figure 3-2
Contaminants in U.S.
Fish Consumption Advisories
PCB« 21%
Chlordarve 6%
Oloxln 4%
Mercury 60%
Source: U.S. EPA. 1994c.
3.2.1.3 Cost Implications of Fishery
Contamination and Advisories
Mercury contamination of fish clearly
can reduce the commercial, recreational.
subsistence, cultural and non-use values of
fisheries. Mercury fishing bans can reduce
dollar values associated with commercial
fisheries directly, when fish are illegal to sell.
or indirectly, when public perceptions that
fish are undesirable to buy and eat can lower
the market value of fish not covered under a
ban (NWF. 1993). Both of these effects lead
to a decline in the economic benefit derived
from a commercial fishery. Mercury advisories can reduce the recreational value associated with
fishing because anglers must:
(1) Cease fishing, thus relinquishing the total value of the fishing experience;
(2) Change to new fishing locations or new species of fish, presumably reducing the
value of the fishing experience;
(3) Catch and release fish, possibly reducing the value of the fishing experience; or
(4) Ignore the advisory, potentially resulting in some health risk.
Advisories can decrease the value of subsistence fisheries by the amount spent on substitute
foods (or by nutritional losses if substitute foods are not purchased). Advisories can reduce
cultural values by limiting or banning traditional fishing and fishing-related activities. Non-use
values may be reduced by the willingness of non-users to pay simply to know that waters are
fishable.
September 7, 1.995
3-18
DRAFT: Do Not Cite or Quo
-------
Figure 3-3
Mercury Fish Consumption Advisories by Slate
S..UIH- US I'l'A.
STATES WITH AT LEAST ONE WATERBODY UNDER MERCURY ADVISORY
STATES WITH STATEWIDE MERCURY ADVISORIES
STATES WITH NO MERCURY ADVISORIES
3-19
DRAM': Do Not Cite or Quote
-------
While the potential for such reductions in tishing values is clear, the national magnitude ot
economic and social costs associated with mercury contamination in fisheries has not been
estimated. In tact, very' little information exists on fisheries-related costs that are specifically
attributable to mercury. Provided below, however, are three examples where costs attributable to
mercury contamination have been addressed either qualitatively or quantitatively.
Arkansas Recreational Fisheries. The Arkansas Game and Fish Commission is the oniv
agency contacted (of six different states and three federal agencies contacted for this report) that
has attempted to quantify fishing-related dollar losses due to mercury contamination. The
Commission estimated a loss of fishing expenditures due to mercury fish consumption advisories
of over S5 million dollars from 1991 to 1992. This loss was estimated based on decreases in
tishing license purchases in counties where mercury advisories were issued, multiplied by the
average number ot trips an angler takes per year, and by the average per-trip expenditures. The
Commission has not published its findings (Armstrong. 1994).
New Jersey Quality of Life. The New Jersey Department of Environmental Protection
and Energy concluded in its 1994 report Task Force on Mcrcwy Emissions Standard Setting: Final
Repon on Municipal Solid Waste Incineration that reducing municipal solid waste mercury
emissions will improve the "quality of life" in New Jersey by decreasing the accumulation of
mercury in aquatic systems. The Department, however, did not quantify these benefits.
Wisconsin Tribes. In 1991 and 1992. U.S. EPA and the eleven Native American Tribes in
the State ot Wisconsin conducted a comparative risk project to evaluate environmental risks faced
by the Tribes (U.S. EPA. 1992b). The Ag'ency attested that the common methods for evaluating
economic and social damages by converting them to dollar values were inadequate when dealing
with non-market and difficult-to-price activities such as damages to subsistence fishing and
associated cultural losses. Instead, a qualitative approach was used to evaluate losses. U.S. EPA
made two important observations relevant to mercury contamination from air emissions:
(1) Nonpoint source contamination was the most important source of social and
economic damages: and
(2) Increases in fish consumption advisories cause substantial damage to cultural
values. (Note that 219 of the 244 fish consumption advisories in Wisconsin are for
mercury) (U.S. EPA 1994b).
U.S. EPA also noted that the top three social and economic damages to the tribes were (1)
diminishment of cultural and religious values: (2) damage to subsistence activities (e.g., subsistence
fishing); and (3) damage to natural resources in commercial use. Fishing plays a role in all three
of these areas. With respect to cultural values, for centuries the Wisconsin Native Americans
have built traditions around spearing fish and sharing the catch. Growing concerns about limiting
fish consumption and limiting the locations where fish may be caught can seriously impact the
Tribe's traditions.
3.2.2 Costs of Human Health Effects
Adverse effects on human health are another potential cost of mercury pollution. The
reduction in welfare caused by an adverse health effect usually can be translated into an economu
value. However, available data are insufficient to estimate mercury levels in the U.S. population
currently, or whether adverse health effects are occurring due to current levels.
September 7. 1995 3.20 DRAFT: Do Not Cite or Qu. •
-------
Volume IV of this Mercury Study Report to Congress reviews a large number ot human
and animal studies on the health effects ot mercury and mercury compounds. Among other
adverse effects discussed in detail in that volume, methylmercury exposures have been found to
cause a variety of developmental effects in children. These effects include delays in speech and
motor development, mental retardation, reflex abnormalities and seizures. A study of a
population in New Zealand observed an inverse correlation between IQ in children and hair
mercury levels in their mothers (Kjeilstrom et al.. 1989). Although the costs of these effects
associated with mercury exposures have not been quantified, lower IQs and more developmental
problems could lead to decreased happiness for children and their families, increased health care
and educational expenditures and reduced overall productivity later in life, imposing a cost not
only on the affected individual but also on society. Other adverse health effects associated with
mercury exposure could result in similar costs.
Settlements in lawsuits over the health effects of mercury provide tangible evidence of the
value society places on these effects. Two examples are presented below.
Minamata. Japan. Between the 1930s and the late 1960s, the Chisso Corporation's
chemical plant on Minamata Bay in Japan discharged mercury-containing wastewater into the Bay.
Fish and shellfish in Minamata Bay concentrated the mercury to acutely toxic levels, leading to
the first case in history of widespread mercury poisoning. In the 1950s, symptoms of mercury
poisoning began appearing in the residents of Minamata who had been consuming the fish.
Because of the extremely high contamination levels, many of the victims of "Minamata disease"
died; others were paralyzed or showed other severe neurological symptoms. Several thousand
people were confirmed to be^ suffering from the disease in the late '80s. with almost 15,000
additional people seeking to be confirmed as victims as well.
The Kumamoto District Court has ordered Chisso to pay damages to victims confirmed to
be suffering from mercury poisoning. The damages cover medical expenses and transportation to
hospitals, and living expenses for the disabled. In one trial, the Court ordered the State,
Kumamoto Prefecture and Chisso to pay damages totalling 674 million yen (about S4.7 million-')
to victims ("Government". 1990, Kudo, 1992).
Native Americans in Ontario. In 1985, two Native American Tribes were awarded
damages for mercury contamination of the Wabigoon-English river system in northwestern
Ontario ("Settlement". 1985; "GLFP", 1986; Blair, 1987). The mercury pollution was traced to a
paper mill owned by Reed. Inc. until 1979 and then sold to Great Lakes Forest Products, Ltd.
The Tribes relied on the river system for subsistence fishing, and their suit claimed that mercury
pollution in fish had caused health problems in the communities. Social and economic costs also
occurred when the government banned fishing in the water system.
The total settlement was for $16.7 million (Canadian dollars). GLFP was to pay $6
million. Reed. $5.8 million, the Canadian government, $2.8 million, and the province of Ontario.
$2.2 million. Of the total settlement, $2 million was to be paid into a fund to address health
problems from mercury poisoning. About 40 to 50 members of the Tribes were estimated to be
affected by methylmercury poisoning; the Mercury Disability Fund provided a source of
compensation for these people. The remainder of the settlement would be used to address the
social and economic consequences of the contamination.
3 Based on a 1990 average exchange rate of 144.79 yen per U.S. dollar ^International Monetary Fund.
International Financial Statistics. August. 1993).
September 7. 1995 3-21 DRAFT: Do Not Cite or Quui
-------
"• --3 Costs ot Ecological Effects
Mercury contamination in the environment may cause adverse effects that diminish non-
human populations, communities and ecosystems and their associated values. These can include
nonconsumptive values, such as the values associated \\ith observing, photographing and feeding
\vildlite. Although less tangible and more difficult to quantify, adverse ecological effects caused by
mercury contamination also may diminish the value placed on preserving natural environments,
either tor future generations or simply for the knowledge that such environments exist, separate
trom any personal use of those resources (i.e.. non-use values).
This section discusses ecological values that may be diminished by mercury contamination.
It tirst summarizes ecological values and then discusses how mercury contamination may diminish
those values. It also provides case study examples where ecological values may have been
diminished by mercurv contamination. »
6
3.2.3.1 Values .Associated \vith Wildlife
Nonconsumptive values of natural populations, communities, or ecosystems are difficult to
quantify and are generally not presented in the literature. Sparse qualitative information on the
values associated with threatened and endangered species is available. Most quantitative
information available is on the nonconsumptive uses of wildlife in general.
Consumptive values associated with fishing and hunting are not discussed here. Values
associated with fishing are discussed in Section 3.2.1. Hunting values in general are not examined
because the literature retrieved for this study did not identify mercury as a significant factor in
reducing robust hunted populations. Piscivorous furbearers are mentioned, however, because (1)
they are particularly likely to accumulate high levels of mercury; (2) in certain areas, they have
reduced populations, possibly due in part to overhunting and loss or degratl-uion of habitat; and
(3) poisoned furbearers may be easy prey for raptors, thus passing high concentrations of mercury
up the food chain to other vulnerable populations.
Threatened and Endangered Species. Effects of mercury on individual members of
threatened and endangered species are of particular concern. Population numbers of these
species are so low that effects on an individual animal or plant can translate into an important
effect on the species population and even the community as a whole. For example, the death tit
one reproductive female in a population including only three reproductive females could be a
major step toward extinction from the region or global extinction of that species. Extinction of
the species (e.g., a top predator species) could in turn change the community's species
composition dramatically.
Although extinction is a natural process, human activities have been accelerating this
process at rapidly increasing rates. Since Europeans colonized this continent 400 years ago, more
than 500 species, subspecies and varieties of North American animals and plants have become
extinct (in contrast to an average of only 12 species lost every 400 years during the 3,000 years oi
the Pleistocene Ice Age) (U.S. Fish and Wildlife Service, 1993b).
Benefits of preserving species from extinction fall into four categories (U.S. Fish and
Wildlife Service, 1993b):
September 7. 1995 3-22 DRAFT: Do Not Cite or Qu.
-------
Protection of ecosystem integrity. Protecting natural diversity protects the integrity
of ecosystems. Moreover, the decline of a sensitive species can alert us to the
decline of the entire system. For example, the rapid decline in raptors such as
bald eagles and peregrine falcons served to alert humans to previously unforeseen
adverse effects of the pesticide DDT.
• Preservation of human uses. Plants and animals can provide a wide variety of
known and as-yet-undiscovered uses for humans, including agricultural, medicinal
and industrial uses. Some plants and animals also provide the service of cleaning
our environment. For example, filter feeders, such as the endangered Higgins' eye
pearly mussel, improve water quality.
• Maintenance of human aesthetic pleasure. Many people prize threatened and
endangered species for their aesthetic values, to be enjoyed by themselves as well
as generations t& come.
• Moral and responsible stewardship. Some believe that humans have a stewardship
obligation to prevent our activities from exterminating other forms of life.
Wildlife in General. With respect to non-threatened or endangered wildlife species,
individual effects are less important than population effects. For example, if one bird in a
population of 1,000 is adversely affected by mercury contamination, no population effects would
be expected. If 500 out of 1,000 birds experienced an adverse effect such as decreased
reproductive success, however, the reproductive rate of the population as a whole would decline,
and the population could be substantially affected.
One way to quantitatively assess the value of wildlife in general is to consider Americans'
expenditures for nonconsumptive activities (Figure 3-4). In 1991, 76.1 million Americans
participated in nonconsumptive recreational activities including viewing, feeding and
Figure 3-4
Nonconsumptive Wildlife-Related Expenditures in 1991
Nonconsumptive Expenditures Trip-Related Expenditures
(Total expenditures S18.1 billion) (Total expenditures 37. 5 billion)
Other [no costs
S-U8 million
6";
Trio-related
S7 S billion
41*.
Fooa
S3 billion
10".
Source: U.S. Fish and Wildlife Service, 1993a.
September 7. 1995 3-23 DRAFT: Do Not Cite or Qui
-------
photographing wildlife (U.S. EPA. 199?a). This population spent a total $18.1 billion on such
activities, ot which 53 percent (S9.6 billion) was tor equipment (e.n.. binoculars). 41 percent ($7.5
billion! was for trip-related expenses (e.g.. transportation, food, lodging) and 6 percent ($1.1
billioni was tor other related expenditures (e.g.. magazines, membership dues) (U.S. EPA, 1993a).
Table 3-9 provides other measures trom the literature regarding the %aiue or
nonconsumptive wildlife activities, such as wildlife viewing. The table also summarizes
information on non-use existence values of wildlife, such as the value of an eagle to someone who
will never see that eagle.
Table 3-9
Nonconsumptive Use and Non-use Values of Wildlife
Type of Value
Value
Source
Use Values
Mean consumer surpius/dav for
nonconsumptive wildlife activities
Mid-value of a specialized recreation day
Nonconsumptive wildlife recreation (value
for one day)
Nonconsumptive use value, including game
waterfowl
- S25/dav
S9.39/day (1986 S)
Sl8.71/dav(l992S)
S26.34/day (.1992 S)
S7.88/day (1986S)
Walsh et al. 1988 (U.S. EPA.
1994a)
Water Resource Council
1979 (DOI, 1987)
Loomis 1988 (U.S. EPA.
1993c)
Walsh et al. 1990 (U.S. EPA.
1993c)
DOI 1987
Non-use Values
Non-users' (typical individuals in Wisconsin
who do not watch eagles) willingness to pay
to preserve bald eagles in Wisconsin, which
are federally listed as threatened in the state
Non-users' willingness to pay to preserve
striped shiners (fish) in Wisconsin, which
are state-listed as endangered in the state
S 16/year (1987S)
S 5/year (1987 S)
Boyle and Bishop 1987
Boyle and Bishop 1987
3.2.3.2 Mercury Contamination of Populations, Communities and Ecosystems
Mercury contamination has been documented in endangered species such as the Florida
panther and the wood stork, as well as populations of loons, eagles and furbearers such as mink.s
and otters (Schneider. 1991). These species are at high risk of mercury exposure and effects
because they either are piscivores (fish eaters, which tend to accumulate high levels of mercur> >
or eat piscivores.
Mercury effects on higher levels of organization (i.e.. communities and ecosystems) are r,
well studied. Some indicators of ecological health in mercury-contaminated ecosystems., howew
were mentioned in the reviewed literature.
September 7..1995
DRAFT: Do Not Cite or Oi.
-------
Florida Panther Populations. Mercury has been confirmed as the cause ot" death and a
likely contributor to population declines of the endangered Florida panther. In 1989, one of only
three breeding age females in the Everglades National Park was found dead. Upon autopsying
the panther, toxic levels of mercury (110 ppm) were found in her liver. (Lethal mercury liver
concentrations ranged from 37 to 145 ppm in feral domestic cats that have died from mercury
poisoning.) (Roelke et al.. 1991).
In addition to increasing mortality, mercury contamination can decrease reproductive
success in the Florida panther. Mercury ingested by a pregnant panther passes through the
placenta to the developing fetus, potentially causing abortions, stillbirths, congenital defects and
behavioral modifications that result in early neonatal death. Figure 3-5 shows the significant
inverse correlation between mercury concentrations in the mother panther and survivorship of the
young (Roelke et al.. 1991).
Figure 3-5
Panther Survivorship as a Function of Mercury Exposure
2.0 i
Number of
Offspring 1.0
Surviving
0.0
n«8
n«4
n«5
>0.00 • £0.25 >0.25. £0.50
>0.50
itration (ppm)
Average number ot" offspring (+. standard error) surviving to the age of 6 months per
female per year in relation to mercury concentration in the whole blood of the female
parent, n = number of females.
Source: Roelke et al. 1991
Florida panthers are exposed to mercury principally by ingesting aquatic-fed prey such as
raccoons and alligators (Roelke et al., 1991). Mercury contamination is clearly reducing the
reproductive fitness and increasing the mortality of the Florida panther population. Because so
few Florida panthers remain (only 30 to 50 in the wild) (Jordan, 1990), mercury contamination
may contribute to the extinction of this endangered species (Roelke et al., 1991).
Wood Stork Populations. Mercury has been detected in feathers of the endangered
stork, though detected levels have not reached toxic levels. Young wood storks in Florida had
mercury levels of 1.870 Mg/kg (ppb) dry weight; higher mercury levels would be expected for
September 7. 1995
3-25
DRAFT: Do Not Cite or Qu.
-------
adults trom the same area (Burner et aL 199?). For comparison. Eisler ( 19S7) suggested a
mercury concentration in leathers ot <5.()()0 an kg \\et weight as a criterion protective ot birds.
Fleming et al. (1984 as cued in Burger et aL 1993) reported mercury levels of 660 Mg-'kg in \vood
siork eggs, again some-what less than Eisler's (1987) recommended criterion of <90()-2.00() ug,kg
xvet weight in eggs.
.Although measured mercury levels in the endangered wood stork have not exceeded
Eisler s (1987) recommended criteria, elevated mercury concentrations have been observed in the
storks. Such elevated levels may pose a concern if the mercury contamination interacts with other
stressors in a way that results in adverse effects to the wood storks.
Loon Populations. A study of the Wabigoon drainage in northwestern Ontario indicates
that mercury contamination has adversely affected populations of the common loon, a piscivorous
bird. In the early 1970s, relatively few loons were observed in high-mercury areas of the drainage
basin. Mercury levels were negatively correlated with successful breeding by loons for 160 km
downstream from a methylmercury source. In high-mercury areas, loons initiated fewer nests.
produced rewer eggs, produced abnormally colored eggs, achieved low hatching success and lost
more eggs to predation (Barr. 1986).
These adverse effects are known manifestations of mercury contamination in. birds.
Failure to produce eggs, abnormally pigmented eggs, teratogenic effects, high mortality in
hatchlings and reduced nest site fidelity have previously been observed in mercury-contaminated
birds. Mercury also may make loons more susceptible to secondary infections, especially during
stressful activities such as molting and migrating. Investigations in response to the die-off of over
2.500 loons in the Gulf of Mexico in 1983 found that elevated levels of mercury were associated
with abnormally high infestations of parasites (Barr. 1986).
Mercury was the only toxicant found in prey species in the Wabigoon study area that
existed at levels high enough to produce the observed adverse effects on loon populations. A
chlor-alkali plant was the main source of the mercury contamination, but high levels of mercury
were also found in lakes unaffected by the plant, possibly due to airborne and/or natural mercury
sources (Barr. 1986).
Eagle and Furbearer Populations. Mercury is suspected of reducing reproductive success
in three piscivores (eagles, minks and otters) in the Great Lakes Region (Schneider. 1991).
Levels of up to 43,000 Mg/kg of mercury have been found in raptor feathers (Spronk and Hartog.
1970 as cited in Burger et al., 1993). For comparison. Eisler (1987) recommends 5.000 /xg/kg
fresh weight in feathers as a criterion protective of birds.
In one Ontario incident, an eagle was found scavenging on a mercury-poisoned dead otter
at Clay Lake (Wren. 1985). Mercury levels in the otter were very high (liver - 96.000 Mg/kg: '
kidneys - 58,000 /ig/kg; brain - 30,000 Mg/kg) and well above known lethal doses for otters. ~The
primary source of the mercury was a chlor-alkali plant that discharged mercury directly into the
river. Although population data were not collected, an Indian trapper had discontinued trapping
in the area because furbearers such as otter and mink had disappeared in recent years, providing
anecdotal evidence of population declines. However, because other stressors could also be
present (such as hunting pressure and habitat loss or degradation), the role of mercury in
contributing to any furbearer population declines is uncertain.
September 7. 1995 3-26 DRAFT: Do Not Cite or Qu.
-------
In a separate incident, a mink died from methylmercury poisoning near the mercurv-
contammated,Saskatchewan River i^Wren. 1985). Effects on raptor (e.g., bald eagie) individuals
or populations in these area were not studied, but could prove to be important. The role of air
deposited mercury may or may not be significant in this case.
Communities and Ecosystems. Strong; evidence of mercury damage to community and
ecosystem parameters (e.g., productivity, species diversity) was not found in the literature. On the
contrary, studies that did mention community and ecosystem parameters noted that no substantial
ecological effects were obvious in field studies (i.e.. not laboratory simulated conditions). Two
examples follow that illustrate how indicators of ecosystem and community health and related
variables have not declined correspondingly with increased mercury levels in fish.
Although fish in Lake Roosevelt. Washington exhibited high levels of mercury (i.e.. whole
body levels of 80 to 250 Mg/kg). indicators of ecological health suggested that ecological effects
were not occurring. Zooplankton populations increased as the reservoir aged. Recruitment.
growth and survival remain high for walleye, the major sportfish in the Lake. Of the sportfish in
the La-ke, only sturgeon were not doing well, possible attributable to overfishing or competition
for food sources with the walleye, although there was insufficient information to rule out metal
toxicity as a factor affecting the sturgeon population (WSDE. 1989).
Although high concentrations of toxic substances including mercury were found in
Wisconsin lake fish, ecological damage is less apparent (U.S. EPA, 1992d). Sportfish populations.
including walleye, northern pike and muskellunge, have remained stable in most of their native
ranges despite increased fishing pressure in recent years. Populations of herbivorous fish (e.g.,
whitefish, perch), however, generally are healthier than predator fish populations. This suggests
that toxic substances with a high propensity to bioaccumulate (such as PCBs and mercury) are
having some adverse effects on predator populations. Bald eagles living inland have much greater
reproductive success than those living near Lake Michigan, but PCB contamination rather than
mercury contamination has been suggested as the probable cause of this difference. Mink and
otter populations have declined, but toxins other than mercury, as well as factors other than
toxins, may be contributing to this decline (U.S. EPA, I992b).
3.2.4 Willingness to Pay to Reduce Toxic Air Pollution
The contingent valuation method was used to measure willingness to pay to reduce toxic
air contamination in Minnesota. Two thousand households in Minnesota were surveyed, and the
results showed that the average resident would be willing to spend $0.35 per day to reduce the
effects of toxic air pollution. This suggested a total annual value of $550.7 million for the state.
In reporting these results, the Minnesota Pollution Control Agency (MPCA) suggested that about
5 to 10 percent of this value might be attributable to mercury, implying a value of $28 to $55
million per year for eliminating mercury air pollution (MPUC. 1993).
September 7, 1995 3-27 DRAFT: Do Not Cite or Oi
-------
Table liS-5
Source Categories and Pollution Prevention/Material Separation Opportunities
Source Category
Nom her of
Sources
Tons/year
Mercury
(''missions
Potential for Pollution Prevention or
Materials Separation Options to Achieve
Significant Mercury deductions
Comments
Medical waste incinerator
37()0
64.7
Medium
Materials separation provides vanahlc
reductions depending on efloit
Municipal waste comhuslor
150
55
Medium
Materials separation piovides vaii.ible
reductions depending on eflort
Coal-fired utility boilers
1043
, 51
High for fuel switching, currently Low
for other options
Options include fuel switching, eneigy
conservation and coal cleaning.
Coal-fired commercial/
industrial boilers
l(X)0's
22.8
High for fuel switching, currently Low
for other options.
Options include fuel switching, energy
conservation and coal cleaning,
Primary lead smellers
Low
Mercury is trace contaminant ol ore
Secondary mercury production
7.4
Low
Mercury is the primary component ol the
process.
Chior-alkali plants using the mercury
cell process
6.5
High
A mercury-free process is a demonstrated
technology.
Portland cement kilns
III)
6.5
Low
Mercury is trace contaminant ol luel and
raw materials.
Oil-fired commercial/industrial
boilers
lOOQ's
High for fuel switching, currently low
for other options
Options include fuel switching, energy
conservation and coal cleaning.
Coal-fired residential boilers
KXKJ's
3.5
High for fuel switching, currently low
for other options
Options include fuel switching, energy
conservation and coal cleaning.
Sewage sludge incinerators
200
1.8
Low
Mercuiv is liacv contaminant
Fluorescent lamp breakage
5(M)
million
1.5
High
Industry is working to lower the mercury
content of fluorescent lumps.
HS 10
DKAIT: Do Not ('He 01 Q
-------
Table ES-5 (continued)
Source Categories and Pollution Prevention/Material Separation Opportunities
N
Source Category
Geothermal power plants
Laboratory use
Dental preparation
Lime manufacturing
Primary copper smelters"
Electrical apparatus manufacturing
Instrument manufacturing
Crematories
Wood-fired boilers
Carbon black production
Oil-lired utility boilers
Battery production
Number of
Sources
18
113
8
1000
1600
24
227
16
Tons/year
Mercury
(•'missions
. 1.4
0.8
0.8
0.7
0.7
0.6
0.5
0.4
0.3
0.25
0.25
0.02
Potential for Pollution I'revt-nlion or
Materials Separation Options to Achieve
Significant Mercury Reductions
Low
High
High
Low
Low
High
High
High
High for fuel-switching
Low
Low
High
Comments
Mercury is trace contaminant.
General use of mercury in labs cuuld be
curtailed.
A mercury-free filling is available.
Mercury is a trace coni.imin.ini
Mcicury is trace contaminant ol me.
Solid state circuitry is an alternative to
mercury switches.
Alternatives exist to using mercury in
various types of instrumentation.
Related to use of mercury denial
amalgams.
Probably not a viable option as the
purpose of woo
-------
DEPARTMENT OF HEALTH &. HUMAN SERVICES Public Health Service
*-*-
Food and Drug Administration
Rockville MD 20857
February 16, 1996
Ms. Martha Keating
U.S. Environmental Protection Agency
North Carolina Mutual Life Building
411 West Chapel Hill Street
Durham, North Carolina 27701
Dear Ms. Keating:
I am writing to you in response to the February 7, 1996 memorandum on
the Mercury Study Report to Congress from T.J. Glauthier of the Office of
Management and Budget (OMB). In that memorandum, representatives from
a number of agencies were advised that, in order for OMB to complete its
review of this report and determine appropriate next steps, final Agency
comments on the draft report were to be delivered directly to you and to
OMB by February 16, 1996.
We are pleased to have an opportunity to provide comments on EPA's
Mercury Study Report to Congress. Our comments are based on the
following drafts of the EPA report which, as far as we know, are the current
drafts: A copy of the Executive Summary dated December, 1995, and the
remainder of the report dated August, 1995. Our comments focus on those
portions of the report that address health effects from the consumption of
seafood. Many of these concerns were expressed during FDA's review of
previous drafts of the report.
Our perspective derives from the fact that, at the Federal level, the safety of
the commercial seafood supply for human consumption, including the safety
of fish that may contain toxic chemicals from both natural and human
sources, is the regulatory responsibility of the Food and Drug Administration
(FDA) under the Federal Food, Drug and Cosmetic (FD&C) Act. The public
health consequences of methylmercury in the U.S. seafood supply have
been monitored and studied by FDA for decades.
-------
Page 2
Based on data from two methylmercury poisoning events1, FDA established
in the late 1960's an "Allowable Daily Intake" (ADI) for methylmercury of 30
ug/day for a 70 kg person. FDA's ADI was set at a level ten times lower
than the lowest observed effect level for methylmercury. The purpose of
the ADI was to provide a basis for establishing a regulatory "action level" for
methylmercury. An action level guides the Agency in determining whether
an amount of methylmercury found in seafood renders that seafood
adulterated within the meaning of the FD&C Act. The action level for
methylmercury, set in 1979, is 1 ppm for all seafood.
Important new data on the effects of methylmercury exposure from studies
(partly funded by Federal agencies) of fish-eating populations in the
Seychelles Islands, the Faroe Islands, and Peru2 have recently become
available or will be available soon. FDA has been advising the public for
several years that it would analyze these data when they became available
and make adjustments to its action level should this analysis indicate that
change is warranted. FDA's reassessment is now beginning. The
reassessment will also take into account an expert review of methytmercury
toxicology to be carried out over the next 6-8 months by the World Health
Organization. The results of FDA's reassessment will be presented to FDA's
Food Advisory Committee and will be published for public comment.
Although we understand that EPA is under an obligation to issue its report,
we are concerned about the report as it relates to seafood. The following
discussion presents FDA's concerns about specific technical aspects of the
methylmercury report.
CONCERNS ABOUT EPA'S RfD
EPA has established its own "Reference Dose" (RfD) for methylmercury and
has recently revised it. This RfD is essentially an ADI (see above) in that it
is an estimated exposure to methylmercury that is intended to provide a 10-
1 Seafood poisonings tied to severe industrial pollution in Minimata, Japan
(1953-60) and Niigata, Japan (1965).
2 Early childhood development studies in the Seychelles Islands, a series of
articles published in NeuroToxicology, 16, 1995; Faroe Islands (Grandjean et al.,
NeuroToxicology, 16(1), 27-34, 1995); and in Peru (Marsh et al.,
NeuroToxicology, 16(4), 717-726, 1995).
-------
Page 3
fold margin of safety/uncertainty. Consequently, there are, in effect, two
Federal ADIs for methylmercury in fish: EPA's RfD and FDA's ADI.
The actual values for EPA's RfD and FDA's ADI are different. EPA's RfD is
four times lower than FDA's ADI. The primary reason for this discrepancy is
that EPA has based its RfD on data from a study of a poisoning event in
Iraq, while FDA has long felt that this study contains significant
uncertainties that make it inappropriate for setting policy or for setting
regulatory limits for methylmercury. It is worth noting that the studies
mentioned above (Seychelles Islands, Faroe Islands, Peru) were specifically
designed to address uncertainties raised by the Iraqi study about the risks of
low-level methylmercury exposure in fish-eating populations.
Most importantly, in the Iraqi study only the four mother-infant pairs with
the lowest exposures provided meaningful data. In FDA's view, this sample
size is so small as to render the drawing of valid conclusions almost
impossible. In addition, FDA has had longstanding reservations about the
considerable uncertainty in the measurement, of effects reported in some
children. For example, lateness in walking was estimated many months to
years later by mothers, even though the exact age of the children was never
known because, as reported by the study's investigators, birth dates are not
events that are remembered or recorded in the subject population. Knowing
the date of birth is essential to determining whether a child experienced
lateness in walking. Finally, several recently published reanalyses of the
Iraqi data by the authors of that study and others3 illustrate the pitfalls
associated with relying on a meager data set with significant uncertainties.
In the draft report, EPA defines an RfD as an estimate of daily exposure to
humans that is likely to be without appreciable risk during a lifetime (p.3-8).
The report then suggests that the RfD can provide a basis for estimating the
size of the U.S. population at risk from methylmercury in seafood (p.3-13
and 14). Although the report points out that the RfD need not be the only
basis for making this determination, the body of the report contains graphs
that show that large segments of the fish-consuming population exceed the
RfD. We understand that ranges of uncertainty have been added to the
graphs since we last saw them, but the strong implication remains that there
is a linkage between exceeding the RfD and risk of adverse health effects.
3 Cox et al., NeuroToxicology 16(4), 727-730, 1995; Crump et al., J. Risk
Analysis 15, 523-532, 1995.
-------
Page 4
Consequently, we are concerned the public will perceive the RfD as
presented in the EPA report as a dividing line between safe and unsafe.
For example, according to the EPA report, nearly 50 percent of American
women of child-bearing age are exposed to methylmercury above the EPA
RfD. Without understanding many complex caveats, some may draw the
erroneous conclusion that such women who are exposed above the RfD are
in danger while those who are exposed below the RfD are completely safe.
Such a misperception would seriously and needlessly erode public
confidence in the safety of seafood.
IMPORTANT NEW DATA
The Seychelles study is prospective in design and is focused on the effects
of prenatal exposure to methylmercury as evidenced in the developmental
progress of children. The published results of the study through the first 29
months of life -- covering the critical developmental years -- is comparable to
the period studied in the Iraqi poisoning event upon which EPA relies for its
RfD. Moreover, the Seychelles study population includes 762 mother-infant
pairs, and thus has much greater sensitivity than the Iraqi study.
Although FDA has not completed its evaluation of the Seychelles study, the
study's authors report that they found no evidence of significant
methylmercury-related developmental effects in children up to 29 months of
age. These children had been prenatally exposed in mothers whose
methylmercury exposures were well in excess of FDA's ADI (as evidenced
by hair levels). Similar results have been reported in a just-published study
of 131 mother-infant pairs in a fish-eating population in Peru that found no
evidence of effects in young children who had been exposed in. utero at
maternal exposure levels comparable to those observed in the Seychelles.
Although we understand that EPA is under a tight deadline, we believe that
any report to Congress that discusses the potential public health risks in
seafood should be based on a complete and thorough analysis of all available
data. FDA is concerned that failure to address these data that bear directly
on the question of risks from methylmercury in seafood will lead to public
misunderstanding about seafood safety.
-------
Page 5
METHYLMERCURY EXPOSURE MODELING FOR SEAFOOD CONSUMERS
EPA has used data from a three-day U.S. Department of Agriculture
consumption survey to estimate consumer exposure to methylmercury from
fish. Although such surveys can be used appropriately to estimate
consumer exposure, we believe that, in this case, the approaches taken by
EPA could result in exaggerated estimates of exposure and estimates of
greater risk for seafood consumers than actually exist. Such exaggerations
could occur because seafood is consumed on an average of less than once
every three days.
EPA's estimates of exposure based on three-day modeling do in fact exceed
levels that would be predicted based on available hair mercury data for U.S.
consumers. A study of hair mercury levels among U.S. seafood consumers
sponsored by the National Oceanic and Atmospheric Administration (NOAA)
of the Department of Commerce4 found average mercury levels in hair
among U.S. seafood consumers to be lower than would be expected based
on EPA's calculations. Literature references to three EPA-sponsored hair
mercury studies5 yield a similar picture of mercury exposure among U.S.
citizens.
EMISSIONS AND OCEAN FISH
As a final scientific point, EPA's report concludes that the weight of
evidence supports a connection between local mercury emissions and
elevated mercury levels in locally caught fish. FDA does not debate this
point. FDA believes, however, that there is insufficient evidence to support
EPA's extrapolation of those conclusions to open-ocean fish. FDA suggests
that EPA differentiate between estimated elevations in freshwater species
due to point-source emissions and the putative elevated levels in commercial
open-ocean species due to anthropogenic emissions.
Final Report, December, 1985: Study to Determine the Possible Hazard of
Methylmercury in Seafood to the Fetus in Utero: A Joint Project of the National
Fisheries Institute and Tuna Research Foundation: Supported by NOAA Grant
NA80AAD-00132.
5 Cited in Marsh et al., NeuroToxicology, 16(4), 717-726, 1995.
-------
Page 6
CONCLUSION
We understand that EPA is under an obligation to issue its report on
methylmercury emissions. However, the serious public misunderstanding
about the safety of seafood that could result from the issuance of the report
in its current form would be a disservice to the public and could lead to a
further loss of confidence in government, if EPA can meet its obligation to
analyze the health effects of methylmercury without including its analysis of
seafood safety, deletion of that discussion is one option that EPA may wish
to consider.
A seafood analysis, if it is deemed necessary, could also be issued later,
after an appropriate reanalysis that includes the new data, a revision of the
report, and an adequate peer review process. If EPA must issue its entire
report immediately, EPA could issue it as an interim report with a clear
statement that important new data on methylmercury in seafood are being
evaluated and are therefore not included in this report.
In any event, FDA recommends that the transmittal letter from Administrator
Browner to the Congress that accompanies the report contain language
similar to the letter drafted last December. That letter provided reassurances
about the safety of commercial seafood and the current reg'.io.tory regime
for methylmercury in commercial seafood.
Sincerely yours,
A.
*•*•
William B. Schultz ^
Deputy Commissioner for Policy
cc: T.J. Glauthier
-------
EPA Responses to Comments (dated February 16,1996) From Mr. William B. Schultz.
Food and Drug Administration, Regarding the Draft Mercury Study Report to Congress
Responses to specific comments made by the FDA are presented below.
Concerns with the RfD published by U.S. EPA.
A discussion of the tests employed and endpoints measured in the Seychelles study will
be included in Volumes IV and VI. This will include the following: hypothesis tested,
population size'and type, endpoints measured, times of testing. There will be no EPA analyses of
these data in the Mercury Study Report to Congress.
Analysis and discussion of mercury hair levels will be expanded to include the Seychelles
and Peru studies. Discussion of the study of Peruvian fishers conducted and written in 1985 and
published this year will be included.
The objection has been raised that the RfD based on studies in Iraqi children is based on
only four data points. The reliance of the RfD on four data points is obviated to a large extent by
the application of the curve fitting procedures and calculation of the benchmark dose (BMD).'In
addition, we have individual data from the Iraqi study.. The grouping of maternal hair
concentrations published in Seafood Safety was done to provide consistency with other analyses.
Many months ago we tested the effect of changing groupings on the BMD. The results of that
analysis which show that no there is no effect on the calculation of the BMD will be included in
Volume IV in discussion of the RfD.
The papers of Cox et al. 1995 and Crump et al. 1995 will be discussed in the text on
description of the RfD. A preprint of the Crump paper was made available to us, and Dr. Crump
wrote text for insertion to the discussion of the RfD, which is found in Volume IV. Both Cox et
al and Crump et al., as well as the U.S. EPA analysis in Appendix D of Volume IV show
considerable uncertainty in thresholds estimated from the data on late walking. The peculiar
nature of the uncertainty, in this case, makes it difficult to distinguish between 7 ppm maternal
hair mercury ) and 114 ppm as a best.(maximum likelihood) estimate for the threshold. Crump et
al. (1995) show that changing the definition of late walking from greater than 18 months to 18
months or greater eliminates the bimodal uncertainty with a best threshold estimate of 230 ppm.
The implication the analyses is that the background incidence of late walking, as reported in
other studies, is not consistent with the lower thresholds. While this is true, the use of historical
controls for this analysis may not be appropriate, given the relatively large number of
observations at low exposure levels in the Iraqi cohort; 33% of the observations were at hair
mercury concentrations considered to be background levels (3 ppm or less).
Late walking, as assessed in the exposed Iraqi population (Marsh et al., 1978) is almost
certainly a valid indicator of methylmercury toxicity but may well be unreliable as the sole basis
for detailed dose-response analysis. The primary reason for this may be the uncertainty in
maternal recall for both birth date and date of first walking. The uncertainty, in this particular
-------
case could be quite large, given the lack of recorded information. The primary impact of this
kind of uncertainty would be on the response classification of individuals at the upper bound of
normal (18 months for first walking) and at the lower bound of abnormal. The lowest abnormal
first walking times presented in Marsh et al. (1978) 20 months. The impact of assuming
uncertainty in the classification of the observations in these two groups is large given the large
number of observations in the two groups (19 data points at 18 months and 8 data points at 20
months). The analysis in Appendix D to Volume IV of the Report to Congress shows that
thresholds estimated for late walking are unstable when classification uncertainty is considered.
The same kind of subjective uncertainty is applicable to the late talking endpoint, as well. The
thresholds for late talking, however, are much more stable, statistically, as there are fewer
observations that are near the normal/abnormal threshold value of 24 months.
Birth date uncertainty also would have an impact on exposure uncertainty if
correspondence of exposure and gestation was estimated (by Marsh et al., 1978) from birth date
to any great extent. That is, exposure may have occurred to a lesser extent (or not at all) than
assumed during the critical period of gestation. The result would be a lower exposure associated
with the observation, depending on the width of the critical time window during gestation and on
the importance of duration of exposure in the elicitation of the particular effect. If the exposure
occurred after the critical period, any observation of an effect would be attributed to causes other
than methylmercury and be included in the background.
The expectation that late walking would be an endpoint with some uncertainty was one of
the considerations that led the U.S. EPA to base its RfD on the combined incidence of four
developmental endpoint in children reported in Marsh et al. (1987). Given the differences in
methodologies and data sets used, an eight fold range in quantitative risk estimates is not
considered large by risk assessors.
Important New Data
The EPA agrees to refrain from expressing a position on the recently published papers
describing part of the results of the studies conducted on fish-consuming populations in the
Seychelles. The EPA and other Federal agencies will participate in a process to evaluate these
data, other emerging and existing data on health effects of methylmercury.
Comparisons of the methylmercury RfD and mercury exposure from seafood.
The EPA acknowledges that FDA is responsible at the Federal level for the safety of the
commercial seafood supply for human consumption. In order to characterize the potential risk to
humans, however, it is necessary to describe exposure from various routes. The majority of
exposure to humans (outside of occupational settings) is through consumption of methylmercury-
containing fish. An early draft of the Mercury Study Report to Congress focused solely on fresh
water fish in the contiguous forty-eight states. Our peer review panel (including both Federal and
external scientists) was firm in their critique that this focus was unacceptable for a nation-wide
characterization of risk consequent to mercury emissions. In the course of evaluating the role of
total fish consumption by various segments of the U.S. population, the graphs alluded to in
-------
FDA's comments were generated. We acknowledge that the graphs, while scientifically correct.
wrere liable to misinterpretation. To prevent this from occurring, the risk characterization
(described in Volume VI and summarized in Volume I) has been changed in the following ways.
1. All graphs will be removed from Volume VI.
2. The population consuming more than 100 grams of fish per day will be calculated. This
analysis was already included in the risk characterization based on cross-sectional consumption
data (CSFII 89/91), but will also be done using longitudinal data.
3. The text will discuss the population consuming > 100 grams fish/day, focusing only on
women of child-bearing age and children.
4. The types offish eaten by people consuming > 100 grams day will be described. The
outcome of this is that a consumer can see that most species they are likely to eat are low in
mercury. The exceptions to this last statement are shark and swordfish: that the FDA and some
fishing industry groups have issued advice on consumption of these species by pregnant women
will be included.
5. For freshwater fish, exposure to methylmercury as predicted by the modeling analyses will be
discussed in the context of the RfD but no graphs will be shown. The marine species
contribution to the overall exposure for populations consuming large amounts of fish will be
discussed qualitatively.
6. A description of the subpopulations likely to be high end fish consumers will be included
Methylmercury exposure modeling for seafood consumers
The discussion of cross-sectional food consumption surveys vs longitudinal surveys will
be enhanced. Comparison of advantages and disadvantages of both types of surveys will be
included as well as the opinion of the NAS on this subject.
In addition, data from a longitudinal survey will be included. The analyses of these data
are in substantial agreement with the analyses done on the CSFII 89/91 survey; that is between 2
- 5% of the general U.S. population consumes on the average 100 g or more per day of
fish/shellfish. Included in this high-end group are subpopulations such as recreational anglers,
people who augment their diet with self-caught fish for economic reasons and people who
consume large amounts of fish from personal preference or cultural mores. A chart on
comparison of high end fish consumption values from various surveys will be added.
We have referred to some unpublished data on hair mercury in Volume I; FDA provided
these data. Also to be added are results from older studies done by U.S. EPA, and some more
recent references from other sources.
-------
Emissions and ocean fish.
The EPA did not state any conclusion linking U.S. anthropogenic emissions and
increased mercury in ocean fish. The text will be reviewed to ensure that no such impression is
given.
Conclusion
The EPA has agreed to include the transmittal letter from Administrator Browner at the
beginning of the Mercury Study report to Congress. In addition, language from this letter will be
included in two places in Volume I.
References
Cox, C, D. Marsh, G. Meyers, and T. Clarkson. 1995. Analysis of data on delayed development
from the 1971-72 outbreak of methylmercury poisoning in Iraq: assessment of influential points.
NeuroToxicoloev 16:727-730.
Crump, K., J. Viren, A. Silvers, H. Clewell, J. Gearhart, and A. Shipp. 1995. Reanalysis of dose-
response data from the Iraqi methylmercury poisoning episode. Risk Analysis 15:523-532.
Marsh, D., M. Turner, J. Crispin Smith, P. Allen and N. Richdale. 1995. Fetal methylmercury
study in a Peruvian fish-eating population. NeuroToxiclogy 116:717-726.
-------
ID: FEB 20'96 13:54 No.004 P.02
EXECUTIVE OFFICE OF THE PRESIDENT
OFFICE OF MANAGEMENT AND BUDGET
WASHINGTON. D.C. 2O3O3
rrp I c IQ
umccor rtB I 0 '»
INFORMATION AND
R ISOLATOR*
MEMORANDUM FOR TJ, GLAUTffiER
From: Sally
Subject: EPA' s Draft Mercury Report
EPA'* draft Mercury Report has important implications for future EPA rulemakings.
Control of mercury emissions is/will be an important element, for example, in the Hazardous
"Waste Combustion MACT proposal (currently under E.O. 12866 review), the Medical Waste
Incinerator MACT, the MACT for primary lead smelters, and any regulatory actions that may be
taken to address air toxics emissions from electric utilities.
The Report as currently drafted includes an assessment of the control technologies and
costs associated with actions to reduce or control mercury air emissions. We believe that
assessment is incomplete and, in particular, does not include any analysis of the costs associated
with substitute materials and fuels,
In addition, the Report presents an exposure assessment in terms of (1) the contribution of
anthropogenic sources to mercury contamination offish and (2) human and wildlife exposure
through fish consumption. We believe this assessment also is inadequate in that it does not state
clearly the strengths, assumptions, and uncertainties of the assessment. Such additional basic
information is especially important because this exposure assessment is likely to be incorporated
in subsequent analyses prepared to support future regulatory actions by EPA and other Federal
agencies.
We have not addressed the health and ecological issues in the risk characterization
volumes; we understand that others will do so.
Our detailed comments are attached. These are, we believe, matters that can and should
be addressed before the April 15 consent agreement deadline.
Attachment
-------
ID: FEB 20'96 i: 35 No. 004 P. 03
Attachment
Volume I: Executive Summary
•
1. The discussion on p.3-4 on the contribution of anthropogenic emissions should be
expanded to provide a more comprehensive description of the sources of mercury
contammaiioa This discussion is important because the primary pathway for both human
and wildlife exposure is through fish consumption. Air deposition from anthropogenic
U.S. sources is one of the sources of mercury contamination in fish. However, other air
deposition sources (including global and naturally-occurring sources) and such non-air
sources as industrial and municipal water discharges and the release of mercury from
stream and lake sediments are also important sources of mercury contamination in fish.
The discussion in this Section should address explicitly these other sources of mercury
contamination in fish in order to place in a more appropriate context the contribution of
the sources of mercury air emissions that are the focus of this Report. Conforming
changes should be made to other parts of the Report, e.g., Volume U. and Supplement A.
2. The discussion beginning on p.4-1 (Management Alternatives) identifies industries or
source categories where pollution prevention or material separation is a "suitable" or
"appropriate" approach. Table 4-1 identifies 13 of 24 listed source categories (including
dental amalgam and crematories) as having a "high" potential for pollution prevention or
material separation. The Report states, however, that no estimate was made of the costs
of replacing or substituting away from mercury in these product or source categories and
that such an analysis should be done in evaluating the merits of replacing mercury for each
source category.
We believe that any judgment on the potential "opportunity" for pollution prevention or
material* separation should properly consider the cost of using substitute materials. In the
absence of such information, EPA should revise or delete this material to avoid making
premature judgments on the potential opportunity for pollution prevention or material
separation. Conforming changes should be made in other Sections of this Report (.e.g.,
Section 4.2.1 of Volume VH and Table 9-2 of Supplement A).
In addition, any discussion of regulatory alternatives and costs in Volume I should include
a discussion of the uncertainties associated with the Agency's estimates of the cost and
effectiveness of the controls discussed.
-------
ID: FEB 20'96 13:56 No. 004 P. 04
3. Section 4 (p.4-2) lists a set of factors that are to be considered in evaluating options for
mercury control. These factors outline some of the difficulties of carrying out a careful
evaluation of regulatory actions.
While we understand that it is difficult to conduct a careful evaluation of regulatory
actions to reduce mercury emissions, we do not understand the context for these
observation* or the implications that this discussion may have for future evaluation
regulatory options. We believe this discussion should be reserved for a later evaluation of
regulatory actions and deleted from this Report. Conforming changes should be made in
other volumes of the report (the only other place we could find this list was in the
Supplement A summary of the study).
Volume H: Inventory of Anthropogenic Sources of Mercury Emissions
in the United States
1 . The Report largely focuses discussion on anthropogenic U. S. sources of mercury
emissions. Potential risks to human health and the environment arise from exposure to
mercury from all sources. The Report should describe and attempt to quantify other
mercury sources (e.g., natural sources). The Report should discuss the strengths,
assumptions, and weaknesses in the estimates of mercury sources.
2. The Report does not adequately distinguish among mercury compounds; some mercury
compounds (e.g., roethylmercuiy) pose greater risks than other compounds.
Volume HI: An Assessment of Exposure from Anthropogenic Mercury Emissions
in the United States.
1. The Report relics on exposures predicted using atmospheric fate and transport models.
The Report should clearly state the strengths, assumptions, and uncertainties associated
with modeling the predicted levels of exposure and indicate the effect of these factors on
the overall assessment. The Report should also indicate the extent to which the predicted
exposure levels have been validated by empirical data. Conforming changes should bo
made in other Sections of the Report (e.g., Volume I and Supplement A).
-------
ID: FEE 20 '96 13:56 No. 004 P. 0
2, The estimates of total fish consumption and of consumption by species in Appendix H
appear to overstate both total consumption and consumption of species with higher levels
of mercury, The Report should clearly state the strengths, assumptions, and uncertainties
associated with these estimates. The Report should also indicate the extent to which these
estimates are consistent with other sources of empirical data on fish consumption.
Volume VH: An Evaluation of Mercury Control Technologies and Costs
1 . There are important differences between Section 2.3 (Activated Carbon Injection for
Utility Boilers) of the Report and the corresponding discussion hi the draft Utility Report
(Sec pages 1 1-41 to 1 1-46). In particular, the draft Utility Report appears to offer a more
cautious appraisal of the potential effectiveness of activated carbon injection foz mercury
control.
The discusadon in both Rqports should be reviewed to assure that they provide an accurate
assessment of the effectiveness and the uncertainties of using activated carbon control
technology. Conforming changes should be made in other parts of the Report, e.g.,
Supplement A.
Section 3%- Sofijfll C"«ta pf Mercurv Pollution - This section discusses "the negative
impacts of environmental mercury contamination on human society." This section
provides both aggregate measures of the value of certain activities that may be adversely
affected by mercury emissions (such as total expenditures for sportfishing trips and
equipment) as well as providing "case studies" such as the settlement of two major
lawsuits (Canada and Japan) arising from methyl mercury poisoning (although neither
incident involved air emissions).
Section 3.2 does not address, however, the incremental effects of anthropogenic mercury
emissions (the focus of this report) in terms of the resulting damages or adverse effects to
the resources, values, and attributes (e.g., the effects on commercial fisheries or
recreational fishing) identified in this discussion. In the absence of such an analysis, it is
impossible for the reader to assess the potential magnitude of the social cost of mercury
pollution from these sources. In fact, the discussion of "costs" offered by this Section may
even be misleading to the extent that it substantially overstates the resources at risk and
the likely damages arising from mercury air emissions. Given the substantial problems
with the current presentation, EPA should delete this Section and leave benefit valuation
for later studies. Conforming changes should be made to other parts of the Report (e.g.,
Volume I (p.2-4) of the Executive Summary, and Supplement A. including the peer review
discussion at p. A-33).
-------
ID: FEB 20'96 13:57 No. 004 F.Ob
Supplement A: Summary of toe Study and Scientific Peer Review
Section 9 (p.9-15) sets out Recommended Actions. This Section identifies areas targeted
for action, including source categories subject to the air toxics provisions of the Clean Air
Act, evaluation and regulation of certain source categories under Section 112(c)(6), and
regulation of utility boilers.
In light of the fact "the analyses of control technologies and costs in this Report are not
intended to replace a thorough regulatory analysis," we believe it is premature to specify a
set of recommended regulatory actions. We believe EPA should delete or substantially
revise this Section. Conforming changes should be made in other parts of the report (e.g.,
Volume I: Executive Summary, p-4-9).
-------
EPA Responses to OMB Comments I dated February 16. 1996) on the Mercury Study.
(Memorandum from Sally Katzen to T.J. Glauthier.)
These comments from OMB are summarized and addressed in the "Summary of Federal
Comments on the Mercurv Studv -- February 22. 1996". located in the front of this section.
-------
UNITED STATES DEPARTMENT OF COMMERCE
National Oceanic and Atmospheric Administration
NATIONAL MARINE FISHERIES SERVICE
*335 East-West M cnwav
•Silver Soring, MD 2O91Q
THE DIRECTOR
Ms. Martha Keating
U.S. Environmental Protection Agency
North Carolina Mutual Life Building
411 West Chapel Hill Street
Durham, North Carolina 27701
Dear Ms. Keating:
In a February 7, 1996, memorandum on the Environmental
Protection Agency's (EPA) Mercury Study Report to Congress, the
Office of Management and Budget (OMB) asked our agency to provide
EPA and OMB with any final aaency comments on the draft report by
February 16, 1996. We understand that OMB is anxious cc complete
its review of the report and determine appropriate next steps.
I welcome this further opportunity to address what we
consider to be serious deficiencies in the report. However, I
believe that the interagency review process may not be brought to
full and effective closure if subjected to the pressures of an
accelerated publication schedule. Failure to resolve the
significant differences of opinion between EPA and its sister
agencies will cause confusion and lessen the utility of an
otherwise important report on the sources and fates of
anthropogenic mercury. I offer the following comments with the
hope that you will give them serious consideration before moving
the report forward to OMB.
I remain greatly concerned about those parts of the report
that'deal, directly or indirectly, with mercury contamination of
marine fish (seafood) and the estimates of health risk to
U.S. consumers. Several broad subject areas require your careful
attention:
Reduced RfD for Methylmercury
The lowering of the Reference Dose (RfD) for methylmercury
and its use in estimating health risks to seafood consumers,
based upon the Iraqi grain poisoning episode, is no longer
defensible in the face of the now published Seychelles study of
neurodevelopmental effects in children. It is apparent that the
two sets of data are simply not comparable; nor is there any
reason to be surprised by this finding. The Iraqi data were used
as a stopgap measure in the absence of reliable human illness
data related to fish consumption. Cox et al. (the same group of
Rochester School of Medicine researchers involved in the
THE ASSISTANT ADMINISTRATOR
FOR RSHEHIES
Printed on Recycled Paper
-------
Seychelles study), in a recently published paper concluded that
"...while this (Iraqi) study was important for establishing a
dose-response relationship in humans, the uncertainty in the
parameters of this relationship is large, and further human
studies are needed." The longitudinal, main study of Seychellois
children with a cohort of 736 mother-infant pairs reveals no
associations between maternal hair mercury concentrations and
neurodevelopmental outcomes at 29 months of age. While the
results of the full 66-month study are yet to be reported, these
carefully designed and conducted measurements indicate that the.
health risks to fish consumers and their offspring are
negligible. Given these facts, we conclude that the decision to
reduce the RfD for methylmercury was ill advised.
Mercury Toxicity in Seafood x
The results of the study of a fish-eating population in the
Seychelles, where maternal hair methylmercury levels ranged from
0.5 to 26.7 parts per million (ppm), indicate that processes
controlling toxicity from this dietary source differ
significantly from acute or subacute poisoning episodes (e.g.,
Minimata and Iraq). This conclusion is supported by the newly
published data from a fetal methylmercury study of a Peruvian
fish-eating population. There, a cohort of 131 mother-infant
pairs with maternal methylmercury hair levels similar to the
Seychellois group (range of 1.2 to 30.0 ppm) was studied during
the period 1981-84, again involving the Rochester scientists. No
increase in the frequency of neurodevelopmental abnormalities in
early childhood could be found. The authors go on to discuss the
possibility that marine fish contain elements, selenium being one
of special interest, that reduce the toxicity of methylmercury in
humans. These and other documented observations of dietary
effects should be more carefully considered in any attempt to
assess human health risk from mercury in seafood.
Accumulation of Anthropogenic Mercury in Marine Fish
In its response to the Congressional request under the terms
of the 1990 Clean Air Act Amendments, EPA chose to use
accumulation of methylmercury in marine fish as a critical
endpoint in evaluating the potential health effects of mercury
emissions from sources such as electric utility steam generating
units and municipal waste incinerators. By so doing, the report
is vulnerable to severe criticism due to the progressive weakness
of the arguments used to link the environmental and ecological
processes which determine the ultimate fate of the original
mercury emissions. For example, the models used to describe the
atmospheric transport and chemical transformation of mercury are
-------
entirely defensible, as are hypotheses for transport
into aquatic systems. However, the simplistic model used to
describe the transfer of methylmercury through increasing trophic
levels of freshwater fishes is not substantiated in quantitative
terms. Furthermore, there is a complete breakdown in argument
and logic when the transition is made to discussion of effects on
marine fishes (especially wide ranging oceanic fishes). The
report provides no demonstrable association between emissions
from U.S. terrestrial mercury generators and the accumulation of
methylmercury in marine fishes. There is only an assertion that
during the past hundred years of industrialization there has been
an increased rate of redistribution of mercury throughout the
environment, including the oceans. However, except for instances
of local industrial pollution, there is no evidence that mean
mercury levels have measurably increased within recent history.
EPA is urged to reconsider its argument for linking the
accumulation of methylmercury in marine fishes with atmospheric
emissions and stay within the confines of a defensible
hypothesis.
Consumer Exposure to Methylmercury in Marine Fish
The validity of a particular consumption database used to
determine dietary intake of seafood could be argued.ad infinitum.
For its purposes, EPA chose to use the results of a 3-day recall
survey conducted by the U.S. Department of Agriculture during the
period 1989-1991. Surveys of this type are known to be
inaccurate for making benchmark measurements of foods, like
seafood, that are consumed relatively infrequently compared to
other staples of the U.S. diet. The use of consumption survey
data with all of their inherent uncertainties, when coupled with
contaminant (mercury) data for fish and shellfish, yields only an
indirect measure of exposure to methylmercury from seafood.
Biomarkers of mercury exposure, such as blood and hair levels in
individual consumers, provide a direct and much truer measure of
human exposure.
The EPA report failed to consider some very useful
unpublished data from a study sponsored by the National Oceanic
and Atmospheric Administration on hair methylmercury levels in
U.S. consumers, again generated by the Rochester group and
collaborators. In that study, approximately 1400 women of child-
bearing age, residents of all 50 states and demographically
balanced, provided hair samples for analysis. The geometric mean
hair methylmercury level was 0.32 ppm, while only four
individuals exhibited levels above 3.9 ppm. These are "real-
world" numbers that should be considered in any attempt to
-------
estimate exposure of U.S. consumers to mercury from fish, either
marine or freshwater.
Perception of Unsafe Seafood
*
In spite of any caveats to the contrary, the report leaves
an informed reader with the notion of an undue risk to
U.S. consumers, particularly women and children, as a result of
eating marine fish. This, in spite of the fact that no evidence
of methylmercury related illness from seafood consumption can be
presented. Unfortunately, the estimates of exposure for these
population groups, and the attempt to relate them to the RfD, are
likely to be construed as a basis for more restrictive controls
on fishery commerce or as a signal to the general public to avoid
eating seafood. In view of the great uncertainties involved in
the determinations of both exposure and health risks from
anthropogenic mercury, a fact acknowledged in the report, we ask
chat EPA reexamine the utility of including estimates of
comparative levels of risk for U.S. consumers of seafood.
I have outlined our major concerns with the draft report,
i.e., where we believe further dialog between our agencies would
be beneficial and agreements could be reached before publication
of the final report. I do this in full recognition of the
magnitude of the reporting task that you were required to
undertake, the complexity of the subject matter, and the time
constraints under which you are placed. Please be assured of our
continuing cooperation and assistance in bringing the report to
completion.
Sincerely,
Holland A. Schmitten
Assistant Administrator
for Fisheries
-------
EPA Responses to Comments (dated February 16,1996) From Holland M. Schmitten,
National Oceanic and Atmospheric Administration, Regarding the Draft Mercury Study
Report to Congress
The EPA's responses to specific comments from NOAA are presented below.
Reduced RfDfor Methylmercury
The EPA has agreed to refrain from expressing a position on the recently published
papers describing part of the results of the studies conducted on fish-consuming populations in
the Seychelles Islands. The EPA and other Federal agencies will participate in a process to
evaluate these data and other emerging and existing data on the health effects of methylmercury.
A discussion of the tests employed and endpoints measured in the Seychelles Islands
study will be included in Volumes IV and VI. This will include the following: hypothesis tested.
population size and type, endpoints measured, and times of testing. There will be no EPA .
analyses of these data in the Mercury Study Report to Congress.
The papers of Cox et al. (1995)' and Crump et al. (1995): will be discussed in the text on
description of the RfD. A preprint of the Crump paper was made available to us, and
Dr. Crump wrote text for insertion to the discussion of the RfD. which is found in Volume IV.
Both Cox et al. and Crump et al.. as well as the EPA analysis in Appendix D of Volume IV,
show considerable uncertainty in thresholds estimated from the data on late walking. Birth date
uncertainty also would have an impact on exposure uncertainty if correspondence of exposure
and gestation was estimated (by Marsh et al., 1978)3 from birth date to any great extent. That is,
exposure may have occurred to a lesser extent than assumed, or not at all, during the critical
period of gestation. The result would be a lower exposure associated with the observation,
depending on the width of the critical time window during gestation, and on the importance of
duration of exposure in the elicitation of the particular effect. If the exposure occurred after the
critical period, any observation of an effect would be attributed to causes other than
methylmercury and be included in the background.
1Cox, C, D. Marsh, G. Meyers, and T. Clarkson. 1995. Analysis of data on delayed
development from the 1971-72 outbreak of methylmercury poisoning in Iraq: assessment of
influential points. NeuroToxicology 16:727-730.
2Crump, K., J. Viren, A. Silvers, H. Clewell, J. Gearhart. and A. Shipp. 1995. Reanalysis
of dose-response data from the Iraqi methylmercury poisoning episode. Risk Analysis 15:523-
532.
3Marsh, D., M. Turner, J. Crispin Smith, P. Allen and N. Richdale. 1995. Fetal
methylmercury study in a Peruvian fish-eating population. NeuroToxiclogy 116:717-726.
-------
occurred after the critical period, any observation of an effect would be attributed to causes other
than methylmercury and be included in the background.
The expectation that late walking would be an endpoint with some uncertainty was one of
the considerations that led the EPA to base its RfD .m the combined incidence of four
developmental endpoints in children reported in Marsh et al. (1987). Given the differences in
methodologies and data sets used, an eight-fold range in quantitative risk estimates is not
considered large by risk assessors.
Mercury Toxicity in Seafood
Analysis and discussion of mercury hair levels will be expanded to include the Seychelles
Islands and Peru studies. Discussion of the study of Peruvian fishers conducted and written in
1985 and published this year will be included in Volume IV.
There are not sufficient data to assume differences in methylmercury effect as a function
of food type. There is some discussion of selenium interaction in Volumes IV and VI; our peer
review expert panel was adamant that the interactions between selenium and mercury are rrot
well characterized and not quantifiable. There is no evidence that the Iraqi population was
selenium-deficient, or starved, at the time of consumption of mercury-contaminated grain. Our
external peer reviewers were very strong in their opinions that the data on selenium are not
sufficient to warrant any more emphasis than had already been made in the Report.
There are data that show that methylmercury is well absorbed in the gut from foods.
There are data from the poisoning incidents in Japan that health effects are consequent to
consumption of methylmercury in fish; the Food and Drug Administration (FDA) used some of
these data as the basis for their acceptable daily intake. More recent analyses and follow-up of
the Japanese populations provide bounding for lower level exposure. The study of Peruvians
done in 1985 and published this year does not conclusively refute an association between
methylmercury in fish and health effects. The recently published studies in the Seychelles
Islands are not negative. The authors state in several of their papers that results are not yet
conclusive and that results from testing at 66 months of age must be examined.
Accumulation of Anthropogenic Mercury in Marine Fish
The Mercury Study does not state any conclusion linking U.S. anthropogenic emissions
and increased mercury in ocean fish. The text will be reviewed to ensure that no such impression
is given.
Consumer Exposure to Methylmercury in Marine Fish
The discussion of cross-sectional food consumption surveys vs. longitudinal surveys will
be enhanced in the risk characterization. Comparison of advantages and disadvantages of both
-------
types of surveys will be included as well as the opinion of the National Academy of Sciences on
this subject.
In addition, data from a longitudinal dietary survey will be included. The analyses of
these data are in substantial agreement with the analyses done on the CSFII 89/91 survey; that is.
between 2 - 5% of the general U.S. population consumes, on the average, 100 g or more per day
of fish and/or shellfish. Included in this high-end group are subpopulations such as recreational
anglers, people who augment their diet with self-caught fish for economic reasons, and people
who consume large amounts of fish from personal preference or cultural customs. A chart on
comparison of high-end fish consumption values from various surveys will be added.
We have referred to some unpublished data on hair mercury in Volume I: FDA provided
these data. Also to be added are results from older studies done by the EPA, and some more
recent references from other sources. To our knowledge we have received no data on hair
mercury from NOAA.
Perception of Unsafe Seafood
It is clear from the poisoning episodes in Minimata and Niigata that ill effects occur in
humans as a consequence of consuming mercury-contaminated fish. Recent re-evaluations of
these incidents are described in volumes VI and I. Information presented at scientific meetings
on results of the studies in the Faroe Islands indicates that long-term lower-level mercury
exposure via fish and marine mammal consumption also has adverse effects. As the data from
the Faroe Islands study have yet to be published and subjected to external and inter-agency
review, they are not described in the Report (except for study design presented in published
abstracts).
The EPA acknowledges that the FDA is responsible at the Federal level for the safety of
the commercial seafood supply for human consumption. In order to characterize the potential
risk to humans, however, it is necessary to describe exposure from various routes. The majority
of exposure to humans (outside of occupational settings) is through consumption of
methylmercury-containing fish. An early draft of the Mercury Study Report to Congress focused
solely on fresh water fish in the contiguous forty-eight states. Our peer reviewers (including both
Federal and external scientists) were firm in their critique that this focus was unacceptable for a
nation-wide characterization of risk consequent to mercury emissions. In the course of
evaluating the role of total fish consumption by various segments of the U.S. population, the
graphs alluded to in your memo were generated. We acknowledge that the graphs, while
scientifically correct, were liable to misinterpretation. To prevent this from occurring, the risk
characterization (described in Volume VI and summarized in Volume I) has been changed in the
following ways.
1. All graphs have been removed from Volume VI.
-------
2. The population consuming more than 100 grams of fish per day has been calculated. This
analysis was already included in the risk characterization based on cross-sectional consumption
data (CSFII 89/9IX but it has also been done using longitudinal data.
3. The text discusses the population consuming more than 100 grams of fish per day. focusing
only on women of child-bearing age and children.
4. The types of fish eaten by people consuming more than 100 grams per day have been
described. The outcome of this is that a consumer can see that most species they are likely to eat
are low in mercury. The exceptions to this last statement are shark, swordfish and barracuda; the
wording that FDA and some fishing industry groups have issued advice on consumption of these
species by pregnant women has been included.
5. For freshwater fish, exposure to methylmercury. as predicted by the modeling analyses, is
discussed in the context of the RfD. but no graphs are shown. The marine species contribution to
the overall exposure for populations consuming large amounts of fish is discussed qualitatively.
6. A description of the subpopulations likely to be high-end fish consumers has been included.
To facilitate sending an appropriate message about fish in commerce, we have also agreed
to include the transmittal letter from Administrator Browner at the beginning of the Mercury
Study Report to Congress. In addition, language from this letter will be included in two other
places in Volume I.
-------
SENT BY'.NIEHS RTF ; 2-18-36 ; 9:36AM : NIEHS- 313 541 0840;# I/ 2
DEPARTMENT OF HEALTH & HUMAN SERVICES Public Health Service
National Institutes of Health
National Institute ot
KJ I rM C f~ I j fV\ >• Environmental Heart* Sclencas
February 15, 1996 I^ICR/ {^ - K \^J P.o.Boxi2233
R"»«»rch Triangle Park, N.C. 277O9
TO: Ms. Martha Keating, EPA
FROM: Director, NffiHS
SUBJECT: Mercury Study Report to Congress
This memo is in response to a request from Mr. T.J. Glauthier, Office of Management
and Budget (OMB), for comments on the Mercury Study Report to Congress prepared
by the Environmental Protection Agency (EPA) per Section 112 (n) (1) (B) of the 1990
Clean Air Act Amendments.
The following comments focus on Volume 1: The Executive Summary of the Report
and pertain to the Health Risk Assessment.
1. There is considerable uncertainty regarding the quality of the science that
EPA has reHed on in its mercury study to determine the health risk from the
consumption of methyl mercury contained in dietary fish. The EPA report relied largely
on data in the early 1970s from ingestion of contaminated seed grain. The most
sensitive population were women in the reproductive age group consisting of 81
maternal-child pairs. The dose response data were described by a broken line or
hockey-stick, dose-response model. There were only four points at the lower end of this
curve.
There is considerable uncertainly regarding the validity of dose-response
calculated from the data derived from subjects at the lower end of the model because, i)
the health end-point was parcsthesia, a subjective clinical feature with a 5-8% occurrence
in people in the general population, ii) the health end-point for children was based on
age of walking and talking at two years of age. Rural Iraqi children do not have
recorded dates of birth, iii) the study population was very small from which to base an
estimate of risk affecting millions of people and, iv) the exposure to methyl mercury was
acute, and high dose in grain seed, which may not be an appropriate model for
estimating the health effects from chronic ingestion of fish containing low levels of
methyl mercury.
2. The currently available scientific data is probably not adequate to establish a
new threshold level for mercury exposure or for mercury concentration in fish. This
Printed on Recycled Paper
-------
SENT BY'NIEHS RTF ; 2-18-96 ; 3:37AM : NIEriS- -'-3 541 0840;* 2/ 2
•IU(€H£
Page 2 - Ms. Martha Keating, EPA
was stated in the National Institute of Environmental Health Sciences (NIEHS) Report to
Congress on Methyl Mercury, Section 4,2, page 13, (July 1995). The two reasons
provided for this conclusion were, "(1) the limited amount of data, both from humans
and from animals, available in the lower range of exposure and (2) the sensitivity of the
outcome measures (health end-points) used in this evaluation."
3. The presently available data are not adequate to go forward with establishing
new thresholds for mercury at this time. The NIEHS has recognized the need for new
and more appropriate data and has been funding two research studies, (fish eating
populations in the Seychelles Islands and Faroe Islands), that are designed to fill existing
gaps in knowledge regarding health effects from chronic low level exposure to methyl
mercury. Results of portions of the Seychelles study were published in the February
issue of the journal, Neurotoxicology. However, there has not yet been critical review
of the results reported in this publication so it is not possible at this time to comment on
whatever impact the study might have on the recommendations of the EPA Report.
There is no question that methyl mercury is a very important neurotoxin,
particularly as it affects fetal and neonatal development. The above comments are
intended to assist in providing responsible and informed information for the protection of
the health of the public.
I hope these comments are helpful.
Kenneth Olden, Ph.D.
cc:
Mr. T.J. Glauthier, OMB
-------
EPA Responses to Comments (dated February 15.1996) from Dr. Kenneth Olden. National
Institutes of Environmental Health Sciences
This memorandum is in response to comments which were submitted to Martha Keating ,
U.S. EPA, dated February 15, 1996. The comments expressed focus on uncertainty and
variability in the data that support U.S. EPA's Reference Dose (RfD).
Concern regarding the Iraqi data base
The objection has been raised that the EPA's reference dose (RfD) based on studies in
Iraqi children is based on only four data points. This issue has been dealt with in at least three
ways. First, multiple end-points rather than only delayed walking and talking have been used in
development of the RfD. Second, the analysis is not based exclusively on grouping of subjects.
Third, the reliance of the RfD on four data points is obviated to a large extent by the application
of the curve-fitting procedures and calculation of the benchmark dose (BMD). The EPA has
individual data from the Iraqi study. The grouping of maternal hair concentrations published in
Seafood Safety was done to provide consistency with other analyses. Many months ago we
tested the effect of changing groupings on the BMD. Results of that analysis showing that there
is no effect on the calculation of the BMD will be included in Volume IV of the Mercury Study
Report to Congress, specifically in the discussion of the RfD.
With regard to the RfD, the health end-points considered were developmental delays in
children whose mothers ingested methylmercury during critical periods in gestation. Your
comment notes that the health end-point for children was based on the age of walking and talking
at two years. The RfD was based not just on these end-points, but also was based on
neurological scores determined from clinical evaluation of the children for cranial nerve signs,
speech, involuntary movement, limb tone strength, assessment of various reflexes (including
deep tendon reflexes, primitive reflexes, plantar responses), coordination, sensation, posture, and
the ability to sit, stand and run. The evaluation used also assessed mental symptoms and the
occurrence of seizures.
Using late walking as an example of the underlying concerns of variability and
uncertainty: late walking, as assessed in the exposed Iraqi population (Marsh et al., 1978) is
almost certainly one valid indicator of methylmercury toxicity. However, it has not been used as
the sole basis for detailed dose-response analysis. Among the reasons for uncertainty and
variability is maternal recall for both birth date and date of first walking. The primary impact of
this kind of uncertainty would be on the response classification of individuals at the upper bound
of normal (18 months for first walking) and at the lower bound of abnormal. The lowest
abnormal first walking times presented in Marsh et al. (1978) was 20 months. U.S. EPA has
provided an analysis (in Appendix D to Volume IV and Volume VI, Risk Characterization of the
Mercury Study Report to Congress) which shows the effect of variability on these threshold
-------
estimates.
Birth data uncertainty also would have an impact on exposure uncertainty if
correspondence of exposure and gestation was estimated (by Marsh et al.. 1978) from birth date
to any extent. That is. exposure may have occurred to a lesser extent (or not at all) than assumed
during the critical period. The result would be a lower exposure associated with the observation.
depending on the width of the critical time window during gestation and on the importance of
duration of exposure in the elicitation of the particular effect. If the exposure occurred after the
critical period, any observation of an effect would be attributed to causes other than
methylmercury and be included in the background.
The expectation that late walking would be an end-point with some uncertainty was one
of the considerations that led the U.S. EPA to base its RfD on the combined incidence of four
developmental end-points in children described by Marsh et al. (1987).
With regard to whether the exposure vehicle was methylmercury from fish or
methylmercury from grain, available data indicate both sources are highly bioavailable (greater
than approximateiy 95% of methylmercury is absorbed across the gastrointestinal tract). As to
the duration of the exposure, it is evident that there were cases of severe methylmercury
poisoning among Iraqi infants whose mothers had high hair mercury concentrations.
Consequently, although the outbreak in Iraqi was of short duration (compared with chronic
exposures observed among fish-consuming populations), the duration was long enough to
produce the severe developmental deficits identified.
Concern about the ability to establish a threshold
NIEHS expressed concern that the currently available scientific data are probably not
adequate to establish a new threshold level for mercury exposure or mercury concentration in
fish. First, it is critical to note that the U.S. EPA is not attempting to set a'n acceptable
concentration of mercury in fish. That is done by U.S. FDA for fish in interstate commerce and
by various State and local health departments. The reference in the comments to Section 4.2 of
the NIEHS Report to Congress on Methylmercury appears to focus on the importance of
supplementing evaluations in humans with data from supporting animal studies because of the
difficulty of identifying the most subtle adverse effects of methylmercury exposure using clinical
testing procedures. In Section 4.2 of the NIEHS Report it is recommended that "Assessment
procedures for supportive animal studies should include those discussed above, as well as
procedures to assess visual disturbances and long-term and/or latent, as well as late-appearing
effects of methylmercury exposure." Consequently, it appears that this section refers to the
difficulty of establishing a level of methylmercury exposure below which no effects occur rather
than indicating there are not data to identify effect levels. That the ability to identify effects is
dependent on the type of testing used is consistent with the philosophy behind the U.S. EPA
Reference Dose. The EPA's RfD is a dose at which adverse effects have not been identified
based on the available data utilized in formulation of the RfD.
-------
Use of new data
As the majority of these new data are either not yet published or have not yet been subject
to rigorous review, it was decided that it was premature for U.S. EPA to make a change in the
methylmercury RfD at this time: i.e. U.S. EPA will use its published reference dose for
methyl mercury of 1 x 10-4 mg/kg-day or 0.1 ug/kg-day. An inter-agency process, with external
involvement, will be undertaken for the purpose of review of these new data, their evaluation.
and evaluations of existing data. An outcome of this process will be assessment by U.S. EPA of
its RfD for methylmercury to determine if change is warranted.
The EPA concurs with the NIEHS that methylmercury is a very important neurotoxin,
particularly as it affects fetal and neonatal development. Comments intended to assist in
providing responsible and informed information for the protection of the health of the public are
indeed welcome.
-------
4
DEPARTMENT OF HEALTH & HUMAN SERVICES
Centers 'or L
and Prevention CDC'
Atlanta GA 30333
28 1996
Ms. Martha Keating
U.S. Environmental Protection Agency
North Carolina Mutual Life Building
411 West Chapel Hill Street
Durham, North Carolina 27701
•\
Dear Ms. Keating:
The Centers for Disease Control and Prevention (CDC) is pleased
to provide comments on the Environmental Protection Agency (EPA)
Mercury Study Report to Congress. Staff from CDC's National
Center for Environmental Health (NCEH) have reviewed Volume 1 of
the report.
Drs. Richard Jackson, Director, and Thomas Sinks, Associate
Director for Science, NCEH, were members of an Ad Hoc Panel of
the Environmental Health Policy Committee, Department of Health
and Human Services (HHS), that reviewed Volume 1 of the report.
The panel heard from researchers conducting investigations on
methylmercury exposures in Iraq and the Seychelle Islands, staff
from the Food and Drug Administration (FDA) , and authors of the
EPA report. The panel focused on the report's characterization
of risk associated with fish consumption. At the panel's
request, numerous revisions were made in the executive summary
(Volume 1) reaffirming that the U.S. fish supply is considered to
be safe at typical levels of consumption and emphasizing that
EPA's calculations of the benchmark dose and reference dose for
methylmercury contain significant uncertainties which are
unlikely to be resolved until additional data are available. In
recommending its revisions, the panel did not mean to imply that
there are no risks associated with methylmercury exposure;
rather, that the great majority of the U.S. population is not in
danger of consuming harmful levels of methylmercury and should
not be advised, directly or implicitly, to reduce their intake of
fish. However, the panel does support EPA's goal of bringing to
public notice, through this report, the potential risks to human
health of continued and increasing mercury emissions. A copy of
the findings of the Ad Hoc Panel is enclosed.
Additional recommendations include:
1. EPA should remove any recommendation concerning a safe number
of fish meals that a person should consume unless the
recommendation reflects a current guideline issued by the FDA
or that of a State or local government. Any new national
-------
Page 2 - Ms.-Martha Keating
guidelines should be developed by the Department of Health and
Human Services in an effort to reach a consensus with other
involved stakeholders, such as EPA.
2. Risk communications regarding the human health implications
of the report should be referred to Dr. Richard Jackson,
CDC's NCEH, who chaired the Ad Hoc Panel described above.
Drs. Jackson and Sinks have reached agreement on these two
recommendations with Dr. Robert Huggett, Assistant
Administrator for Research and Development,, and Ms. Mary
Nichols, Assistant Administrator for Air and Radiation, EPA.
A
In addition, changes to Volumes 2-7 should reflect any
changes already made to the Executive Summary.
We appreciate the opportunity to comment on this important
report. If you have any questions, please contact Dr. Thomas
.Sinks at (770) 488-7001.
Sincerely,
David Satcher, M.D., Ph.D.
Director
Enclosure
-------
.' DH'AKIVH M Of HFM TH \ HUMAN SEKV
MEMORANDUM
TO: The Assistant Secretary for Health
FROM: Chair, Ad Hoc Panel on Methylmercury
Environmental Health Policy Committee
SUBJECT: Mercury Study Report to Congress from the
Environmental Protection Agency
In the Clean Air Act Amendments of 1990, a requirement was made of the Environmental
Protection Agency (EPA) to produce a study of mercury emissions, including the sources, rates
and amounts, control technologies and their costs, and health and environmental effects. The
study is completed and is due to be transmitted by court order to Congress on December 15,
1995.
As emphasized in the mercury emissions report, most humans' non-occupational exposure to
mercury from the environment will be in the form of methylmercury consumed in contaminated
fish. The greatest potential risk to humans appears to be to the fetus in pregnant women
consuming more than one fish meal daily, depending on the level of contamination in the fish
consumed. Issues raised by the nature and magnitude of human health risk from fish containing
methylmercury are clearly within the province of the Food and Drug Administration (FDA).
Recognizing this, EPA brought these issues to the attention of FDA officials, who had some
concerns about the report. Specifically, FDA had made a decision to postpone review of their
action level for mercury in commercial fish pending publication of results from studies in progress,
and officials were concerned about the apparent disagreement embodied by EPA's decision to
proceed using existing data that FDA had decided was insufficient. EPA's decision to use the
data was at least in part driven by the court-ordered release date of December 15. On a more
fundamental level, FDA officials felt that the EPA report, as written, might raise a level of public
concern thyt in their view, was not justified by the magnitude of the apparent risk. In order to
address these differences, EPA and FDA officials agreed to ask the HHS Environmental Health
Policy Committee (EHPC) for assistance in evaluating the studies upon which the EPA's risk
characterization was based and interpreting the results of those studies for the development of
appropriate public health policy.
-------
Page 2 - The Assistant Secretary for Health
A panel of scientists was formed, chaired by me and including members from NIH and CDC,
which met on several occasions with FDA and EPA scientists and policymakers as well as with
the authors of key studies in progress The charge to the panel (attached) consisted of five
questions, all having to do with the quality and interpretation of data from human studies The
first three questions focused on technical issues; after some deliberation, the panel members felt
that, in view of the extremely limited time available to them and the many months' work that had
already gone into the development of the report, a complete review of the health effects section
was beyond the panel's scope.
Most of the panel's work focused on question four of the charge, which dealt with the report's
characterization of risk associated with fish consumption. At the panel's request, numerous
revisions have been made in the executive summary (Volume I) of the EPA report reaffirming that
the U.S. fish supply is considered to be safe at typical levels of consumption and emphasizing that
the EPA's calculations of the benchmark dose and reference dose (RfD) for methylmercury
contain significant uncertainties which are unlikely to be resolved until additional data are
available. Several research studies that are currently underway are likely to provide such data.
Among these are a large study in the Seychelles Islands funded by FDA and NIEHS and another
NIEHS-funded study in the Faroe Islands of Denmark (both are study populations with high
levels offish consumption).
In recommending its revisions to the EPA Report, the PHS/EHPC panel does not mean to imply
that there are no risks associated with methylmercury exposure; rather, that the great majority of
the U.S. population is not in danger of consuming harmful levels of methylmercury and should not
be advised, directly or implicitly, to reduce their intake offish. On the other hand, the panel
supports EPA's goal of bringing to public notice, through this report, the potential risks to human
health of continued and increasing mercury emissions.
Question five of the charge, concerning the development of the public health message that should
accompany the release of the report, is still the focus of the panel's active participation in
collaboration with both EPA and FDA. The panel has taken on (from the Agency for Toxic
Substances and Disease Registry) an additional PHS member with risk communication expertise
and will be working with EPA and FDA officials in the development of public health messages
targeted to populations with different levels of risk and the formulation of answers to expected
questions from the public and the press. This continuing activity will be critical to achieve
understanding between agencies on the content and tone of the public communications and to
permit the Administration to present a consistent message on the human health risks of
methylmercury exposure to the U.S. population.
Throughout this process, both the EPA representatives and the FDA officials have acted in good
faith on the panel's recommendations. I and the other panel members have been most impressed
by the willingness of the EPA scientists to work with the PHS/EHPC panel to craft a report
summary that reflects in good measure the consensus views that emerged during this process.
While this remains an EPA report, the final version of the summary should be one that the PHS
-------
Page 3 - The Assistant Secretary for Health
can support and that can form the basis of a balanced public health/nsk commurucauon message
to the American public
I will continue to keep you and the EHPC informed of our progress on this important public
health issue
Richard J. JacksorCMD, MPH
-------
cc
Dr Jo [vey Boufford, P-DASH
Mr. William Farland, EPA
Dr Rita Schoeny, EPA
Dr Martha Keating, EPA
Dr Kate Mahaffey, EPA
Dr Kaye Wachsmuth, FDA
Dr. Richard Levine, NICHD
Dr. Robert Goyer, NIEHS
Dr. Thomas Sinks, NCEH
Dr. Tim Tinker, ATSDR
Dr. David Satcher, CDC
Dr. Claire Broome, CDC
Members, Environmental Health Policy Committee
-------
SENT BYJXerox Telecopier 7021 ; 2-16-96 ; 3!29PM I NCEH/OD-* 919 541 Q840;» 6
Februaiy 16, 1996
Chair, Ad Hoc Panel on Methylmercury
Environmental Health Policy Committee
Comments on the EPA Mercury Study Report to Congress
Mr. William Raub
Senior Advisor, Office of the Assistant Secretary
for Planning & Evaluation, HHS
Thank you for the opportunity to provide comments on the EPA Mercury Study Report to
Congress. My staff, in my capacity as Director of the National Center for Environmental
Health (NCEH), Centers for Disease Control and Prevention (CDC), have reviewed Volume
1 of the report.
Dr. Thomas Sinks, Associate Director for Science, and I, were members of an Ad Hoc Panel
of the Environmental Health Policy Committee, Department of Health and Human Services
(HHS), that reviewed Volume 1 of the report. The panel heard Scorn researchers conducting
investigations on methylmercury exposures in Iraq and the Seychelle Islands, staff from the
Food .and Drug Administration, and the authors of the EPA report. The panel focused on the
report's characterization of risk associated with fish consumption. At the panel's request,
numerous revisions were made in the Executive Summary (Volume I) of the EPA report
reaffirming that the U.S. fish supply is considered to be safe at typical levels of consumption,
and emphasizing that EPA's calculations of the benchmark dose and reference dose (RfD) for
methylmercury contain significant uncertainties which are unlikely to be resolved until
additional data are available. In recommending its revisions to the EPA report, the panel did
not mean to imply that there are no risks associated with methylmercury exposure; rather,
that the great majority of the U.S. population is not in danger of consuming harmful levels of
methylmercury and should not be advised, directly or implicitly, to reduce their intake offish.
On the other hand, the panel supports EPA's goal of bringing to public notice, through this
report, the potential risks to human health of continued and increasing mercury emissions. A
copy of the findings of the AdHoc Panel are enclosed.
At this time, I would like to make the following additional recommendations:
1. EPA should remove any recommendation concerning a safe number offish meals that
a person should consume unless the recommendation reflects a current guideline
issued by the Food and Drug Administration or that of a State or local government. I
believe that any new national guidelines should be developed by HHS working to gain
a consensus with ofher involved stakeholders, like EPA.
-------
SENT BYtXerox Telecopier 7021 ! 2-16-96 ; 3:30PM I NCEH/OD^ 919 541 0840;$ 7
Page 2 - Mr. William Raub
2. Risk communicationa regarding the human health implications of the report should be
referred to me, since I chaired the Ad Hoc Panel described above.
9
Dr. Sinks and I have already reached agreement on these two recommendations with
Dr. Robert Huggett, Assistant Administrator for Research and Development, and
Ms. Mary Nichols, Assistant Administrator for Air and Radiation.
3. Additional changes to Volumes 2-7 should be made to reflect any changes already
made to the Executive Summary.
Thank you again for the opportunity to comment on the report. If you have any questions,
please contact Dr. Thomas Sinks at (770) 488-7001.
Richard J. Jackson, MD, MPH
Attachment
cc:
Mr. Kevin Neyland/OMB
OD
NCEH
Doc.:Raub.rjj
Preparer & ContactrTSinks (770) 488-7001
Spelling Verified:aj 2-16-96
-------
,lf«KV, ,,
j
/fr
ID: !^EB 16'96 18 = 00 No'.005 P.10
DEPARTMENT OF HEALTH & HUMAN SESVICES office of tha Secretary
, O,C, ' 20201
TEB I 9 1996
Mr. TJ.Glauihier
Executive Office of the President
Office of Management ?PJ Budget
Washington, D.C. 20503
Dear Mr. Qlauthier:
Thank you for the opportunity to join my colleagues fiptu the Fuud anil Drug Administration,
the Centers for Disease Control, and the National Institute of Environmenul Health Sciences
in commenting on the Mercury Study conducted by the U.S. Environmental Protection
Agency. They will be responding directly
From the overall perspective of Uic DepHrunem of Health and Human Services, I emphasize
two points:
1 The Mercury Study does not provide a basic for changing current dietary advice to the
public with respect to the consumption of fish. With the exception of specific U.S.
waters that are identified in flsh advisories, the U.S. fish supply is safe at the levels
typically consumed by Americans. Fish aie a critical part of a healthy diet. Any uurs
of or reference* to the Mercury Study to discourage consumption of fish would be
contrary to prudent public-health practice,
2. The Department of Health and Human Services, through the Food and Drug
Administration and the Centers for Disease Control, is particularly well-positioned to
communicate with the public about human-health risks associated with mercury in fish.
1 urge that the OMB and other agencies defer tu Uic Department on such mailers Should
they arise in the wake of transmitting the Mercury Report to the Congress.
Sincerely,
William F. Raub. Ph.D.
Science Advisor
-------
EPA Response to Comments on the Mercury Study from Dr. David Satcher. Centers for
Disease Control (dated February 28. 1996). Dr. Richard Jackson. Centers for Disease
Control (dated February 16. 1996). and William F. Raub, Department of Health and
Human Services (dated February 19.1996).
Regarding the "Safe " Number of Fish Meals
As you know, the EPA has been in continual dialog with scientists and policy experts in
the various Federal agencies. There have been a series of meetings and conference calls at which
a number of points of agreement have been reached. Within both the Executive Summary
(Volume I of the Report) and the Risk Characterization (Volume VI), the EPA supports the broad
statement that members of the general population consuming (in moderation) fish from mixed
fish species are not at risk of harm from methylmercury. With respect to advice concerning a
safe number of fish meals that a person may consume, the EPA only cites the Food and Drug
Administration recommendations regarding consumption of fish with higher than average
concentrations of methylmercury. and refers persons who consume fish from noncommercial
sources to their local and State health departments for additional advice.
Risk Communication Issues
With respect to risk communication issues, the EPA continues to rely on the message
prepared by the ad hoc panel on methylmercury.
-------
United States Department of the Interior
FISH AND WILDLIFE SERVICE
Washington. D.C. :0240
FWS/AES/DEC
Ms. Martha H. Keating
Environmental Protection Agency
Office of Air Quality Planning and Standards
411 West Chapel Hill Street
Mail Route 15
Durham, North Carolina 27701
Dear Ms. Keating:
The Office of Management and Budget requested that the U.S. Fish and Wildlife Service
provide you with our comments on the draft "Mercury Report to Congress," which is the
subject of a court order. This comprehensive report will serve to provide the Congress
and the general public with valuable insight into the dilemma of mercury pollution and
the risks to and effects on public health and the environment. Although no formal
regulatory actions are proposed, it is important to know that the risks and effects of
mercury pollution on the environment are becoming increasingly evident, especially in
poorly buffered waters of the Midwest, Northeast, and in South Florida. It is for these
reasons that we offer our comments to help you develop a final set of documents to the
Congress.
Pursuant to the Fish and Wildlife Coordination Act as amended (16 U.S.C. 661-667), the
enclosed comments are provided to assist in the development of national understanding
and progress in removing ongoing threats to wildlife resources. The term "wildlife" as
used herein includes birds, fish, mammals, and all other classes of wild animals and types
of aquatic and land vegetation upon which wildlife are dependent. The FWS's review
under the FWCA does not relieve Federal agencies of their substantive obligation to
comply with applicable provisions of the Endangered Species Act of 1973 as amended
(16 U.S.C. 1531-1544). Section 7 of the ESA mandates that Federal agencies consult
with the FWS or the National Marine Fisheries Service, as appropriate, to ensure that
Federal actions are not likely to result in jeopardy to threatened or endangered species
or the adverse modification of their critical habitat. Thus, actions taken as a result of
this report should be coordinated with the FWS under the above mentioned authorities.
-------
These comments supplement and update informal staff comments on the Executive
Summary of the report, sent to EPA in mid-December of last year. Should you have any
questions regarding our comments, contact the Assistant Director. Ecological Services
(Attn: Chief, Division of Environmental Contaminants) at (703) 358-2148.
Sincerely,
DIRECTOR
Enclosure
-------
U.S. Fish and Wildlife Service
Comments on the Mercury Report to Congress
as Proposed by the U.S. Environmental Protection Agency
Page 3-18 of the Executive Summary, Table 3-4 lists wildlife criteria for water for each of
five species. For four of these species (mink, otter, kingfisher, and bald eagle) the
numbers are different by factors of 3.5 to 6.9 from those calculated by EPA in its final
Great Lakes Water Quality Guidance of March 23, 1995. As mentioned at the
coordination meeting called by OMB, FWS staff stated that EPA should clearly state
why the numbers have changed and provide a range of risk as opposed to a single
number/criteria approach. The basis of presentation for the Great Lakes was a
geometric mean calculated for birds compared with mammals. A comparative range
based on the range between the Great Lakes numbers and those of the summary report
would be a good presentation strategy. This seems reasonable given the significant
uncertainty associated with confounding issues of mercury methylation factors, water pH,
hardness and buffering capacity, as well as water trophic status and bioaccumulation and
magnification factors. While Volume V of the Report does include the documentation,
some statement of divergence from the Great Lakes Water Quality Guidance should be
included in the Executive Summary to avoid reader confusion. We believe EPA should
briefly summarize, in the Executive Summary, the changed assumptions as mentioned in
Volume V of the Report. EPA should consider substituting Table 5-9 of Volume V
which shows these differences.
On page 3-20 of the Executive Summary, the statement is made that wildlife exposure
was based on consumption of freshwater fish. The eagle and mink will be eating birds
and mammals with potentially greater mercury contamination. This is especially true for
the federally endangered Florida panther (mentioned on page 3-19) where it feeds on
raccoons which magnify mercury. This issue was recently addressed in a peer reviewed
paper by Charles Facemire (now retired) of the FWS and co-authors: Facemire et. al.
(1995) Environ. Health Perspect. 103(Suppl 4): 79-86 entitled, "Reproductive Impairment
in the Florida Panther: Nature or Nuture?" EPA presented wildlife risks in the Executive
Summary and the entire Report whicli characterize fish as the only risk-assessed food
source of contamination to higher predators. This should be clearly stated in a
discussion of how this affects the changed water criteria numbers for at least the mink
and bald eagle and preferably the Florida Panther as well. EPA should make the
clarification statement relative to this issue after the statement concerning fish
consumption at the top of page 3-20.
Statement number 4 on page 4-7 of the Executive Summary misses the mark. What
drives the choice between wildlife or human criteria would first require a comparative
assessment of the risk-based numbers. At the OMB coordination meeting, we suggested
that the Report would be strengthened by including a table which provides comparative
data among wildlife species and humans. The table would include the water quality
criterion relative to the critical, respective endpoints, the oral dose No Observable
Adverse Effects Level (NOAEL), dietary concentration or tissue concentration that
-------
corresponds to the other parameters and corresponding sediment numbers, if available.
A range would be preferable for this type of report to diffuse any contentious
misconstruction of the Reference Dose concept and the proposed changes for humans.
The risk management decision on what to base regulation follows such a comparative
risk assessment. Obviously, endangered species concerns would be part of this risk
management decision. It should not be characterized as an "either/or" situation beiore
the comparative risk assessment in the Executive Summary. We note that recent review
of Volume V contains information along the line of thought, as suggested above, in
Table 5-6. This table compares the RFD and LOAEL for several wildlife species as
well as for three different types of human consumers. All the data are normalized into
comparative risks expressed as RFD or LOAEL fish tissue concentrations. We strongly
recommend that this table be included in the Executive Summary to clarify two likely
important points of concern. First, the information demonstrates that the fish
concentration-associated wildlife risks and human risks are quantitatively very similar. If
we clean up emissions that cause mercury contamination in wildlife, it would be a far
more easy selling point with the knowledge that it is the overall environment that is at
stake, not just human consumers of fish. Secondly, this presentation would be helpful
regardless of the choice of human-related exposure data that finally end up being used
by EPA to show comparative differences in human risks. Because there is concern over
input data that revise the human RFD being proposed, perhaps EPA can get the Report
out by showing the differences, expressed as fish concentrations, between the old and the
proposed RFDs and include both in Table 5-6 of the Executive Summary as well as in
Volume V.
-------
EPA Response to Comments (dated March 1.1996) From Mr. Bruce Blanchard. U.S. FWS.
Regarding the Draft Mercury Study Report to Congress
The specific points raised in Mr. Blanchard's letter concerning the assessment of
mercury hazards to wildlife are addressed below Please note that in the draft circulated for
review on March 22. 1996. we were not able to make all of the changes requested.
Differences Bet\veen the Mercury Study And Great Lakes Water Quality Guidance
We agree that the Executive Summary should explain the variations between the Great
Lakes Water Quality Guidance and the Wildlife Criteria in the Mercury Study Report to
Congress. Our Office of Water in consultation with their legal experts have supplied clarifying
language which is being inserted into Volume I (Executive Summary), Volume V (Ecological
Risk Assessment) and Volume VI (Risk Characterization).
Wildlife Exposure and Fish Consumption
As the comments suggest, it will be made clear in the discussion of Wildlife Criteria
(WC) that fish only were considered as sources of mercury ; data limitations preclude the
consideration of other, higher trophic level food sources. Volume V of the Report discusses the
potential for mercury exposure in the Florida panther through the consumption of contaminated
racoons. The discussion of data limitations to the calculation of a WC for panthers will be
excerpted for Volume I.
Factors to Consider in Addressing Mercury Reductions
In response to comments by other Federal reveiwers, the section in Volume I, Chapter 4
entitled Factors to Consider in Addressing Mercury Reductions has been removed. The table to
which you refer and an accompanying analysis can be found the Risk Characterization volume.
Volume VI. We agree that estimated no effect levels for humans and for mammalian and avian
wildlife species appear to be very similar. There are some points of discontinuity which do not
permit a direct comparison: these are discussed in Volume VI. The discussion will be
summarized in Chapter 3 of Volume I.
-------
Department of Energy
Washington, DC 20585
February 21, 1996
Ms. Martha Keating
U.S. Environmental Protection Agency
North Carolina Mutual Life Building
411 West Chapel HiU Street
Durham, N.C. 27701
Fax:(919)541-0840
Dear Ms. Keating:
Thank you for the opportunity to review the draft Mercury Report to Congress. The Department
of Energy recognizes the importance of this study in its own right and as part of the Report to
Congress on electric utility air toxic emissions ("the Utility Report").
Sec. 112 (n)(l)(B) of the Clean Air Act Amendments of 1990 requires that the Mercury Report
shall address a number of topics; specifically, the Report...
shall consider the rate and mass of such emissions, the health and environmental effects of
such emissions, technologies which are available to control such emissions, and the costs
of such technologies.
The Department believes that the Report to Congress, including the Executive Summary, should
discuss all four elements outlined in the law in a balanced fashion. Specifically, we believe the
Executive Summary should present more information about rate and mass of emissions, control
technologies, and their cost.
As requested in T.J. Glauthier's memo of February 7, 1996, the Department is providing written
comments. The comments include materials given Assistant Administrator Mary Nichols on
February 2, 1996, and a small number of additional comments. The Department supports your
effort to produce a balanced, clear Mercury Report to Congress, and stands ready to help you in
this effort. In this regard, we urge that EPA incorporate the concerns of all agency reviewers.
Finally, the Department believes the forthcoming Utility Report should reflect the changes made
in the Mercury Report.
Sincerely,
Jan W. Reicher
Acting Assistant Secretary for Policy
Printed witn soy ink on recycled paper
-------
Department of Energy Comments on
Mercury Report to Congress
Executive Summary Should Address All Required Study Elements
The Department of Energy believes the Mercury Report to Congress and its Executive Summary
should address all study elements required by law in appropriate balance. The Executive
Summary (dated December 1995) did not present an adequate summary of the seven volumes of
the study or the topics requested by Congress. About four out of forty pages are devoted to
sources of mercury in the environment and the methods of transport and exposure (compared to
two report volumes), one table to possible control technologies for all emission sources, and one
table to point estimates of costs for one control technology for selected source categories. The
bulk of the Executive Summary is focused on health risk studies and risk characterization.
EPA is not Required to Establish a Revised RfD
Recognizing the uncertainties inherent in the data currently available to establish safe levels for
mercury exposure for human health effects, and that EPA is only required to consider "...the
health and environmental effects of such (Hg) emissions..." (Sec 112 (n) (1)(B) of the Clean Air
Act), the Mercury Report to Congress does not need to contain recommendations for a revised
reference dose (RfD). More complete material for an RfD calculation should be available this fall
in time for the regulatory determination authorized for the Utility Report
The National Institute of Environmental Health Sciences (NIEHS), part of the National Institutes
of Health, was asked under Section 301, Hazardous Air Pollutants, of the Clean Air Act "to
conduct, and transmit to the Congress by November 15,1993, a study to determine the threshold
level of mercury exposure below which adverse human health effects are not expected to occur."
They have funded studies of fish-eating populations in the Seychelles Islands and Faroe Islands
and these studies arc in the process of being analyzed and published now. Data that will provide
useful information for evaluating the effect of low-level chronic exposures to methylmercury will
be available in the summer of 1996 from the studies conducted on the Seychelles Islands. Part of
this data has been released recently but further data is being analyzed. This data could be used,
outside of the scope of the Mercury Report, to reassess the RfD for mercury.
Improving Confidence in Reference Doses
Confidence is low for estimates of reference doses (the "threshold level of mercury exposure
below which human health effects are not expected to occur") that are based on Iraqi data (1971-
72 outbreak of methylmercury poisoning in Iraq). These data are for 84 mother-child pairs
involved in an acute poisoning incident The reason for establishing reference doses is to prevent
low-level chronic exposures to methylmercury through the ingestion offish.
-------
In a recent article, the research group that published the first analysis of the Iraqi data commented
that "The best way to estimate effects at low-exposure levels is to have data at low-exposure
levels..." (Cox, et. al., NeuroToxicology, 16(4), p. 729).
The Seychelles Island study is acquiring data at low-exposure levels. This study involves 1,450
children whose mothers eat an average of 12 fish meals per week. These mothers have an average
hair mercury concentration of 6 ppm (the average hair mercury concentration in the U.S. is 1
ppm). The development of a group of 750 of these children is being monitored from the age of
six months to five and one-half years. Full results of this study arc expected to be available in the
summer of 1996. Partial results arc available and contradict studies of the Iraqi data.
Deposition Models Have Not Been Rdeased tn the Public for Comment and Peer Review •
EPA models COMPMERC and RELMAP are used in Volume HI to estimate transport and
deposition of mercury. The results of these models form the bases for predicting exposure and
health effects. Without public access to and proper review of these models, the findings of the
Mercury Report are open to question and doubt. One result that raises doubts about the accuracy
of these models is the concentrations of mercury in water reported in the Mercury Report. They
arc greater (by an order of magnitude) than any measured in U.S. surface waters (except where a
point source was discharging into the body of water). This apparent overcstimation of mercury
deposition on surface waters would lead to subsequent overestimation of mercury levels in fish.
The deposition results predicted from models need to be validated with actual observable
measurements and correlated with actual measured mercury levels in fish.
-------
Volume TT: Inventory nf Anthropogenic Emissions
Data. Although EPA did not provide plant-by-plant emission estimates for electric utilities,
EPA should assemble such data for this source category. Plant-by-plant information on electric
utilities is important, because cost effectiveness of mercury control options can vary by almost an
order of magnitude when applied to various combinations of plant size (250-1,000 MW) and coal
burned (low-meirury, 0.05 ppm, Wyoming bituminous coal to high-mercury, 0.39 ppm, Pennsylvania
Appalachian bituminous coal). This observation will be explored more thoroughly in comments
below directed at Volume VII: Evaluation of Mercury Control Technologies and Costs.
The Department disagrees with EPA's contention that estimates of mercury emissions from iron and
steel manufacturing (coking operations) and refineries are difficult to assemble. Following the same
method of emissions factors that EPA employed extensively throughout Volume II, Argonne National
Laboratory previously was able to generate emissions inventories for both these source categories
located in the 11-state Great Lakes region. Argonne found that collectively these two source
categories accounted for almost 10% of the mercury inventory in the Great Lakes region.
Furthermore, that same Argonne study found that pulp and paper mills contributed over 10% of the
mercury inventory in the Great Lakes. Based on the relative size of EPA's general commercial and
industrial combustion source category, it is possible that EPA may have included pulp and paper mills
in that source category. Although on a plant-by-plant basis, pulp and paper mills do not produce
large mercury loadings, there are a large number of such plants, at least in the Great Lakes. This
raises the policy question of the threshold of a source category that would be subject to EPA
regulation.
Level Plaving Fiffrj|, On page ES-14 of the Executive Summary, EPA cites expected declines
in mercury emissions from certain source categories, based upon future scheduled regulations, and
other trends in the industries' operations. If presented for some industries, EPA should extend that
argument to account for expected decreases in electric utility mercury emissions resulting from the
implementation of controls for Title IV of the Clean Air Act Amendments of 1990. Equity
considerations should play a key role when EPA is considering the contribution of, and possible
regulation of, multiple source categories as in the case of mercury. Confounding this exercise is the
unsynchronized regulatory schedules that the respective source categories face. Before a regulatory
determination, EPA should fully account for progress source categories have already made, and can
reasonably be expected to make in the near future, to control mercury. This approach creates a
baseline, and provides an avenue through which emission trades could be executed, as suggested in
our previous comments.
Volume TTT: An Assessment nf Exposure from Anthropogenic Mercury Emission in the United States
EPA recognizes the fact that the mechanism of mercury revolatilization is still uncertain.
Consequently, EPA should be careful to distinguish between direct (inhalation) and indirect
(ingestion) pathways of mercury to sensitive populations. The separation may elevate the importance
of water discharges relative to air emissions, and is important because some source categories (e.g.,
pulp and paper mills, chlor-alkali plants, iron and steel coking operations, refineries, Portland cement
plants) produce considerable water effluent contributing to the mercury inventory. Geographic
distance from susceptible populations, siich as Native Americans subsisting on a predominantly fish
diet, may also play a role and constrain the range of source categories that EPA might consider for
regulation.
-------
Volume VTI: Fvalnafjpn nf Mercury Control Technologies and Costs
DOE's cost effectiveness estimates for carbon filter bed technology applied to electric utilities
correspond to those that EPA provides in this volume. However, we believe that EPA has seriously
underestimated costs for activated carbon. Our estimates of cost effectiveness range 1.5-8 times
higher than the upper end estimate that EPA provides in this volume. Further checking revealed that
we had used the same references (e.g., Felsvang) as EPA to derive our cost and benefit information.
We- believe, based on EPA's assumption of 95% mercury removal, that EPA may have
underestimated the high feed rate (73 mg/nr) required to achieve at least 80% mercury removal.
Since the cost of activated carbon is the major cost element for such an application, a high feed rate,
necessary to achieve high mercury removal, can have a dramatic effect on the cost of this technology.
-------
Highlights of Department of Energy Comments
on the
Draft Mercury Report to Congress1
Vol. T: Exceptive Summary
The Executive Summary may be the most important volume of the report for decision makers,
making it essential that this volume represent the entire Report and be thoroughly reviewed and
considered before publication. The Executive Summary contains many generalizations and
material summaries which do not adequately describe the limitations and uncertainties contained
in the body of the report. Two examples are described below.
Example 7. In the Executive Summary, EPA estimates that 65 percent (160 tons) of the
anthropogenic U.S. emissions (243 tons) were transported outside the continental U.S. and that
36 tons are deposited annually in the U.S. from the global atmospheric reservoir. EPA then states
(p. 3-4), "The weight of evidence supports an association between local mercury emissions and
increasing mercury levels in locally caught freshwater fish.'1 EPA should present highlights of this
evidence in the Executive Summary since the global inventory contributes about one-third of U.S.
deposits and EPA's long-range transport models predict significant long-range deposition.
Example 2. Page 3-4 indicates there is a "plausible link between increases in mercury emissions
from industries identified in the emissions inventory and increases in mercury concentrations in
fish." However, industrial uses of mercury in the U.S. have been declining since the 1970's. The
report hypothesis, if correct, needs better documentation. An alternative hypothesis might focus
on exposure pathways from past emissions.
Volume H! Inventory of Anthropogenic Sources
This is an impressive survey and the Department acknowledges that the task was enormous.
Nonetheless, the Report should include more information about natural sources and other current
anthropogenic sources, as well as estimates of re-emissions of historic anthropogenic mercury.
EPA estimates that natural sources may account for approximately 40 percent of total global
mercury emissions. Anthropogenic sources such as fungicides and industrial waste discharges
may be particularly important It is necessary to know the relative contribution of current
anthropogenic sources to total U.S. and global mercury emissions to determine what difference it
would make to total risk if these sources were controlled.
It is also important to convey clearly that a number of the estimates for emission sources are
estimated or extrapolated from a few samples, hypothetical plant constructs, or emissions factors
and operating statistics. The weakness of methods used to make inventory estimates should be a
point highlighted both in this Volume and in the Executive Summary of the overall Study.
'Department comments refer to copies of Volume I dated December 1995 and Volumes
Il-Vn dated August 1995. These arc the versions provided by OMB in January 1996.
-------
Volume TTT! Assessment of FTpnsure from Anttiropf)gf.r^jjc Emissions
This volume focuses on exposure as the result of deposition from local sources and long-range
atmospheric transport EPA models of these deposition processes are used to estimate the
mercury deposited in the environment Differences between model results and measured mercury
concentrations in fish suggest that the models are flawed - either in their hypotheses about
methods of mercury exposure or the estimation of the model forms. Uncertainty and need for
further research in this area should also b*. dghlighted in the Executive Summary and more fully
addressed in mis Volume of the Report
Volume VIl! Mercury Control Technologies and Costs
Two of the four items which Congress mandated be part of this Report are "technologies which
are available to control such emissions, and the costs of such technologies." This volume is the
only one addressing these topics. Of the volume's four chapters, one is largely addressing social
costs of mercury pollution and another summarizes ongoing and potential federal and state
control strategies. The Department believes the discussion of social costs of mercury pollution
should not be presented before the human health effects have been assessed with the forthcoming
studies. The Department also believes that this volume should more fully discuss the complexity
of control efforts for different forms of mercury and different sources and, given these problems,
the preliminary nature of technology development for mercury control.
-------
Department of Energy
Washington, DC 20585
January 30, 19%
MEMORANDUM
TO:
FROM:
DanReicher
Acting Assistant Secretary
Patricia Fry Godley
Assistant Secretary rfbr Fossil Energy
SUBJECT: Fossil Energy Comments on EPA Mercury and Utility Studies
Attached, per your request, are FE's comments on the Mercury Study. We have prepared
our comments in two sections—general and specific—and have limited our review to
Chapters 2 and 7, which deal with emissions as well as control technologies and costs.
We hope these comments are helpful to you as you prepare the Departmental response to
this important study.
If you have any questions regarding these comments, please address them to Bob Kane of
my staff on 586-4753. Bob and Chuck Schmidt from the Pittsburgh Energy Technology
Center have been providing staff input to Beth Campbell of your staff since the release of
this draft last week.
Regarding your other question about the release of the Utility Study; we agree that it
should be delayed until the completion of the Mercury Study.
Attachment
-------
EPA MERCURY STUDY REPORT TO CONGRESS
FOSSIL ENERGY REVIEW COMMENTS
General Comments
Puhlic Comment Pe
A formal public comment period, extending at least for 60 days, and preferably longer, should
be provided on the entire mercury report, given the complexity and length of the report.
Initially, the draft report was available to the public on January 25, 1995. EPA originally
requested comments from industry and interested parties to be submitted no later than March
15, 1995. However, the Executive Summary was not available for review at that time. This
significant report, along with the Executive Summary, should be offered for public review so
that parties from ail perspectives can examine and comment on it Only by making the repon
as comprehensive and accurate as possible will Congress and others be fully informed about
mercury emissions.
Mercury Controls Would Add Significant CVtsts to me Utility Tndustrv
• The total cost to the utility industry and electricity consumers is estimated to be
between $1.2 and $10.4 billion per year nationwide (based on total mercury emissions
from utility sources of 40 tons per year).
i
• If a utility had to add a wet scrubber for mercury control, cost are projected to range
from $116,000 to $174,000 per pound of mercury removed for low sulfur coal and
$76,000 to $1 14,000 per pound of mercury removed for high sulfur coal. In
comparison, the cost to remove a pound of sulfur dioxide from flue gas is about &
cents..
Significant Uncertainties Still Remain
If the report is to be of value to Congress and the public, it must contain the best scientific
information, including current data, models that have been subjected to wide review, and
model results that have been validated against real-world measurements. If time constraints
do not allow the inclusion of these types of information, then information inadequacies must
be clearly identified and future research needs specified. The repon should serve as a state-
of-scknce and mgognw- that more research is needed on mercury emissions, exposure routes,
health effects, and methods of control. The report, in its present state, should not be used for
decision-making purposes due to these significant uncertainties and data gaps. (
I
Tone of Executive Summary Not Consistent With Body of Report
The Executive Summary may be the most important volume of the repon for decision makers,
making it even more important that this volume be available for public review.* It contains
many generalizations and material summaries mat do not adequately present sufficient
V
-------
information on the limitations and uncertainties that are contained in the body of the report.
Below ire listed two examples of statements contained in the Executive Summary that do not
convey the appropriate uncertainties and limitations, and which could result in
misinterpretation of the TPP**"?IS given in the other volumes of the report
Example No. 1 In the Executive Summary, EPA estimates that 65 percent (160 tons) of the
anthropogenic U.S. emissions (243 tons) were transported outside the continental U.S. and
that 36 tons are deposited annually in the U.S. from the global atmospheric reservoir. EPA
then states on page 3-4 of the Executive Summary, The weight of evidence supports an
association between localfemphasis added) mercury emissions and increasing mercury levels
in locally caught freshwater fish". EPA should present highlights of this "weight of evidence"
in the Executive Summary since about one-third of the mercury that* deposits annually in the
U.S. comes from the global inventory, and EPA's long range transport model predicts
significant deposition of mercury at distances between 25 and 1000 Kilometers from point
sources. Thus, these seemingly contradictory statements need to be reconciled.
Example No. 2 On page 3-4 of the Executive Summary, the report indicate* there is a
"plausible link between increases in mercury emissions from industries identified in the
emissions inventory and increases in mercury concentrations in fish". Industrial uses of
mercury in the U.S. have been declining since the 1970s, yet the mercury concentrations in
fish have increased since then. If this hypothesis is correct, more supporting documentation is
necessary.
Analyses in the Mercury Report are Bevond the Scope of Clean Air Act (CAA) MsndfltTfi
Section 112(n)(B) requires the EPA to conduct a study of mercury emissions from electric
utility steam generating units, municipal waste combustion units, and other sources, including
area sources. Specifically, "Such a study shall consider the rate and mass of such emissions,
the health and environmental effects of such emissions, technologies which are available to
control such emissions, and the costs of such technologies".
Volume VII: An Evaluation of Mercury Control Technologies and Costs contains a section
(3.2) that evaluates the social costs of mercury pollution. This section of Volume Vn
attempts to compare the costs of controlling mercury to the societal benefits of reduced
mercury emissions. An analysis of this type was not mandated in the CAA, and should not
be included in the report to Congress. The CAA mandated EPA to determine only the costs
of mercury control technologies, which are contained in Section 3.1 of Volume Vn.
-------
Specific Comments
EXECUTIVE SUMMARY
Table 3-1, "Best Point Estimates of National Mercury Emissions in 1990", contains a column
indicating predicted 1995 mercury emissions for four source categories. As footnoted, the
predicted 1995 emissions were not used in any of the modeling analyses. Since these data are
highly speculative based on penHinf regulations, and varying assumptions, they appear
unrelated to the description of 1990 point sources of mercury emissions, and should not be
included in Table 3-1.
Figure 3-1 is missing.
VOLUME n- INVENTORY OF ANTHROPOGENIC EMISSIONS IN THE UNITED
STATES
The mercury report should include more accurate estimates from natural sources and other
current anthropogenic sources, as well as estimates of re-emissions of historic anthropogenic
mercury. EPA estimates that natural sources may account for approximately 40 percent of
total global mercury emissions. This is a significant percentage that should be considered by
EPA in any comprehensive study of mercury emissions. Anthropogenic sources such as
fungicide and industrial waste discharges may be particularly important Without these
estimates, the extent of past, present, and future contributions to mercury risks from current
anthropogenic sources cannot be adequately understood. It is necessary to know the relative
contribution of current anthropogenic sources to total U.S. and global mercury emissions in
order to determine what difference it would make to total exposure from mercury emissions
of these sources were controlled.
VOLUME VII: AN EVALUATION OF MERCURY CONTROL TECHNOLOGIES AND
COSTS
Mercnry Controls Would Add Significant Cost to the Utility Industry
• The total cost to the utility industry and electricity consumers is estimated to be
between $1.2 and $10.4 billion per year nationwide (based on total mercury emissions
from utility sources of 40 tons per year). Applying control technology to reduce
mercury f«ms^o"< from power plants would result in a reduction of mercury emissions
of less than 1 pound per day from a 500 MW power plant (assuming a cost of $25,000
per pound of mercury removed)
• The utility industry would inequitably bear almost all of the costs for emissions
reduction referred to in the draft EPA Mercury Study Report to Congress (six source
categories, including utilities, comprise 70% of all mercury emissions considered).
• Technology currently being tested at a small scale for use on power plants has
estimated costs of $14,400 to $38,200 per pound of mercury removed.
-------
• If a utility had to add a wet scrubber just for mercury control, costs are projected to
range from $116,000 to $174,000 per pound of mercury removed for low sulfur coal
and $76,000 to $1 14,000 per pound of mercury removed for high sulfur coaL In
comparison, the cost to remove a pound of sulfur dioxide from flue gas is about 8
cents.
Cost Effectiveness of Control Technology is I Inlrnnwn
• Mercury is ubiquitous and truly a global pollutant. One can sample the atmosphere
anywhere in the northern hemisphere (near point sources or in remote areas) and find
concentrations of mercury on the order of 1 to 2 nanograms per cubic meter.
• Reducing mercury emissions from domestic sources is unlikely to reduce the mercury
' concentration in fish in a reasonable amount of time. The Executive Summary (page
3-5) accurately states:
"What impact would a decline in deposition have on methylmercury
concentrations in fish? Because mercury is persistent in the environment, and
such a global reservoir has already accumulated, particularly in sediments and
soils, it is likely to take decades before fish mercury levels would be
comparable to preindustrial levels."
o Control M^rcurv Emi«inrK HTP Pnnrl
No full-scale demonstrations of mercury control technology for utility boilers have
been conducted in the U.S.
The amount of mercury that would be expected to be captured by a control technology
at a specific power plant cannot be predicted at this time (i.e., the same technology
might capture 20% of the mercury at one power plant and 80% at another).
Removal effectiveness depends on the species or compound forms of mercury present
at a specific power plant To date, no single control technology has been identified
that removes all forms of mercury. Each form of mercury requires its own control
technology.
The forms of mercury produced vary significantly from one power plant to another.
It is important to note that control technologies currently used for municipal waste
combustors will not remove mercury as effectively or at comparable costs when
applied to utility boilers.
1
-------
The effectiveness of mercury removal decreases dramatically with lower
mercury concentrations. Mercury concentrations in the flue gas of power plants
are very low (between 1 and 10 micrograms per normal cubic meter of flue
gas).
Mercury concentrations found in municipal waste combustors are one hundred
times greater than those found in utility boikrs.
Mercury tends to be present in a more easily removable form in municipal
waste combustors, primarily as soluble mercuric chloride.
-------
Methods to Measure Mercurv Concentrations arc
• Before cost effective mercury controls can be developed, methods must be developed
that measure and distinguish the various forms of mercury emitted by power plants.
Currently available methods for sampling and analysis of mercury emissions from
power plants are only capable of accurately determining the total amount of mercury
present
Additional Specific Comments (Volume VII)
Page Comment
1-3 The activated carbon injection (ACT) efficiencies reported are based primarily on pilot-
scale data (Appendix A). This means that ACI is not yet utilized by the U.S. utility
industry, and that all conclusions drawn regarding the removal efficiency and cost of
ACI have a high degree of uncertainty.
2-10 The report indicates an average reduction in Hg of 21% using coal cleaning equipment
as currently practiced by the U.S. coal industry. It is expected that significantly
higher Hg reductions can be achieved with the application of emerging coal
preparation processes. Proof-of-concept scale testing of selective agglomeration and
advanced column flotation have indicated Hg removals as high as 80%. Bench-scale
testing is also being carried out to investigate the use of naturally-occurring microbes
to reduce the Hg (and other trace elements) from coal. If adopted by the coal
industry, these advanced technologies have the potential to remove an additional 10-12
tons of Hg from the annual utility emission inventory.
2-13 It is acknowledged that the collection of Hg via paniculate control devices is highly
variable and dependent upon flue gas temperature, the method of collection (e.g., ESP,
fabric filter), mercury speciation, etc. Further, the presence of HC1 can result in the
formation of HgCl2 (mercuric chloride), which can readily adsorb onto carbon-
containing paniculate matter (PM). Conversely, SO] can act as a reducing agent,
converting soluble oxidized Hg to elemental Hg, which is more difficult to collect All
of these points highlight the uncertainty associated with the control of Hg from electric
utilities, and should be highlighted in the Executive Summary.
2-16 EPA considers carbon filter beds as one of two possible control techniques (along with
ACI) in its assessment of the cost effectiveness of control technologies for utility
boilers. . Although carbon-filter beds have been installed on five German power plants
(for controlling residual SO2 downstream of PGD to prevent the formation of
ammonium sulfate), none has been put into commercial practice in the U.S. However.
EPA concludes that "based on European applications, carbon-filter beds are applicable
to MWCs and utility boilers...". This conclusion may be accurate for the European
utility sector, but it is much too premature for the U.S. electric power industry, and
should be deleted from the report
-------
2-17 In discussing the potential negative impact of producing activated carbon, EPA
references a report by Rester (1993). In that report, Rester uses a coal-Hg content of 1
ppm in calculating the release of Hg during the coal-charring process to produce
activated carton. Although EPA concludes that the Hg-release risk is negligible -
compared to the Hg subsequently adsorbed by the activated carbon, the reference to
the Rester data is perplexing. The 1 ppm coal-Hg content is one-to-two orders of
magnitude higher mat any of the raw coal data presented in Table 4-3 of Volume n of
the EPA report
@2-21 There are only four applkations of selenium filters in the world (all in Sweden, three
on smelters and one at a crematory). However, me EPA concludes (Section 2.15.3)
that the technology is applicable to other point sources, including "flue gases from
power plants." Given that the technology has not been commercially demonstrated in
the U.S. or elsewhere on a utility boiler, this conclusion is somewhat speculative, and
should be Dieted,
2-22 Adding to the uncertainty associated with AC technology, EPA reports a study by
Fksvang et aL (1993) that showed 25-45% greater Hg capture when using sulfur- and
iodide-impregnated carbon compared to using activated carbon. However, in studies
by Miller (1995) and Krishnan (1994), better Hg reduction resulted with activated
carbon than with iodated carbon. This highlights the need for further research in the
control of Hg from utility boilers. This point should be emphasized.
2-25 EPA, based on two pilot-scale test programs, point out that the effectiveness of ACI
improves dramatically at lower flue gas temperatures (at or below 250°F). However,
they correctly state that this is below typical utility flue gas temperatures upstream of
PM control equipment For high-sulfur coals, lowering the flue gas temperature below
250oF can cause acid condensation, and, consequently, equipment corrosion. Again,
this points to the high degree of uncertainty regarding the «**ft»iml and economic
readiness of Hg control technology, and the need to continue laboratory, bench-scale,
and pilot-scale testing.
2-27 EPA states the "No validated methods are available at mis time for quantifying Hg
speciation in utility flue gas..." This is an extremely important point and needs to be
included in the report's Executive Summary.
2-31 EPA states that me "effectiveness of ACI in removing different forms of Hg is still
being studied". Given this uncertainty, it is premature to consider ACI as a
commercially-ready technology. Moreover, EPA must be careful in citing the ACI
technology as the basis for estimating-cost impacts of Hg control on the utility
industry.
2-31 There is uncertainty as to whether ACI removes elemental Hg. Studies by Felsvang et
aL (1993) showed that ACI ahead of a spray dryer/ESP resulted in 46.4% elemental
Hg removal, compared to essentially zero reduction in elemental Hg without ACI.
However, this represents only one test at the pilot scale. Work by Joy/Niro and
-------
Northern States power at pilot- and full-scale indicated that only 10-20% total Hg
removal was achieved with a spray dryer/fabric filter system without ACL With ACI,
total Hg removal increased to 55%, and when iodine- and sulfur-impregnated carbon
were used, total Hg reductions increased to 90%. This suggests that carbon
impregnated with sulfur or iodine is required to remove elemental Hg.
In discussing emerging technologies for utility boilers, EPA identifies impregnated AC,
sodium sulfide injection, and AC circulating fluidized bed. Several developing
technologies such as condensing heat exchanger, wet PGD additives, gold-impregnated
monoliths, or advanced coal cleaning should also be mentioned.
3-1 In assessing die cost impact of Hg control EPA assumes that electric utilities will
continue to be able to pass on the cost of control to their customers because consumers
do not have other options for acquiring the same service. However, with deregulation
of the electric-power industry, the added cost of Hg control could significantly impact
the competitiveness of certain utilities.
3-2 In describing the model used to calculate the cost of utility Hg control, EPA assumes
that there is sufficient chlorine (0.1 percent) in the model coal to convert all Hg in the
flue gas to HgClj. This is a major assumption in that results from laboratory, bench-,
and pilot-scale tests indicate that the oxidized.form of Hg (e.g., HgCl,) is easier to
remove from utility flue gas than is the elemental form of Hg. Thus, the model will
predict a higher Hg were removal (and, therefore, a lower cost of removal), than if
there were elemental Hg present in the flue gas. However, research has shown that
coal-fired utility boiler flue gas can contain a fairly significant percentage of elemental
Hg. The cost effectiveness of utility boiler Hg control presented in Table 3-1 should
be recalculated using a more realistic ratio of oxidized versus elemental Hg.
(37) 4~15 EPA recognizes that its analysis of the costs and effectiveness of controlling Hg
^-^ emissions from utility boilers is preliminary and states The information presented is
intended to present a range of available options and provide a relative sense of the
extent of mercury reductions achievable and the general magnitude of the cost of such
reductions" (bold and italics added for emphasis). Further study is needed to Mil in the
information gaps before a thorough assessment.of the specific impacts to the utility
sector can be determined.
-------
Part II of Department of Energy Comments on
Draft Mercury Report to Congress
Vol. 1: Executive Summary
It should be noted that the data in table 3-1 conflict with the data in Table 3-6 from, the
prior report We would assume that this reflects improvements in the estimates from the prior
report Also, Table 3-1 is not identical in format with Table ES-3 of the current Vol. n inventory
report
On page 3-5 EPA discusses the decline of mercury levels in fish over decades. The
question is for the reader, what is the mechanism by which it becomes no longer biologically
available? Even if it is discussed elsewhere, to complete the thought here the report should
discuss how that happens.
Also on page 3-5, EPA asserts that the primary mechanism is through human consumption
of fish. It would be more informative to give a rough estimate as to whether this is merely the
largest plurality of exposure, or whether it constitutes virtually the entire exposure. There is no
way to know from this description the balance of factors, and therefore no way to know whether
inhalation in the ambient air is a moderate factor or no factor at all.
Vol. II: Inventory of Anthropogenic Emissions
nf data and quality at data. EPA has done an impressive and admirable job
of surveying the potential sources of mercury. But many of the source categories have no data
assigned as emission estimates. In some cases where one should expect some information to be
available, such as hazardous waste incinerators, primary mercury production, mercury compound
manufacturers, byproduct coke production and petroleum refineries, no estimates are given at all.
See Table ES-3, at ES-6. These are not area sources but large facilities -which one would expect
to have the resources to have made some estimates of emissions.. This, of course, can only
exaggerate the relative contribution of those source categories whose emissions are measurable.
For example, it focuses attention on coal-fueled sources whose mercury content can be
established based on analysis of the particular seams used. EPA reports at 4-59 that the mercury
content of coal can be as high as 8 ppm by weight If coal combustion sources are major
contributors, would not coke production sources also be? Without data, these sources are written
off entirely.
In addition, looking below the surface, a great number of emission sources are estimated
on extrapolations from a few samples or hypothetical plant constructs. Or they are aggregated
from emission factors and industry operating statistics (e.g., aggregate production, energy
consumption, etc.) Emissions of mercury from individual sources is not well known. The study
would be greatly improved by obtaining more reliable data for each of these source categories.
A peer reviewer estimates that the missing sources could contribute as much as 20 percent
more mercury to the U.S.. total (Vol n, at 5-1) Perhaps it was beyond the scope of the statutory
l^
-------
mandate to produce basic scientific data where none existed, but there are too many source
categories for which the estimates are either unavailable or not calculable to be assured about the
report.
The information-instruments srflle. Of course, this presents a dilemma: on the one
hand, more information could potentially produce a better product by enabling EPA to tailor
regulations to controlling the sources more cost-effectively; on the other, more information would
delay action and have a cost that (given today's budgetary battles) would be hard to find. In
other words, failure to spend money now might cost the private sector more in compliance costs
or result in a poor regulatory product that is indefensible.
For example, mercury seems to be a good candidate for emissions trading. It is probable
that there are great divergences in marginal control costs among sources, given that some
pollution prevention alternatives listed by EPA actually save energy and therefore cost less. A
market would accelerate the adoption of these alternatives. On the downside, the hot spots
problem does not exist in significant amounts because: (1) the emissions are usually in trace
amounts: (2) the sources are widely dispersed; and (3) the emissions remain in the atmosphere for
long periods and are transported great distances. The other side is that some sources are more
likely than others to affect deposition to waters from which consumed fish are caught (such as
utilities surrounding the Great Lakes). In sum, a national mercury market like acid rain market is
an economically viable approach, given sufficient information on inventories to construct
reasonable and equitable credit allocation scheme.
omments.
Page 3-1 says that data from 1977 predate the use of catalytic
converters and unleaded gasoline. Unleaded gasoline was mandated to be available in July 1974.
Catalytic converters were mandated for the 1975 model year ( fall of 1974), although a petition by
the automaker, agreed to on judicial review, delayed full implementation until later, catalysts were
required on 1975 model year vehicles sold in California. Although the requirements were not
fully implemented until 1980, it is mistaken to say that mere was no unleaded and no catalyst cars
in 1977.
The description of mercury removal due to coal cleaning at 4-8 is not clear. It would
seem from EPA's description that it has not computed the amount of emissions correctly. EPA
identifies two factors that contribute to the total amount of reduction from coal cleaning, the
percent of the coal cleaned and a percent reduction due to cleaning. One way to look at it is that
there are two factors, a percent of coal cleaned, and a percent removal for amounts cleaned., If
that is so, the calculation is incorrect In the alternative, if the 21 percent reduction in mercury
cited accounts for both factors, that is not clear from the description. This should be rewritten to
clarify.
On 4-44 EPA states that the useful life of an electric switch is 20-30 years. In the same
paragraph it states mat 50-percent are still in use 50 years after their introduction. This needs to
be fixed.
Table 4.21, page 4-60 omits the location of the Doe Run Smelter.
-------
Significant missing data. EPA has stated that the primary mechanism for human
exposure is through fish consumption. Thus, an important consideration is to compare emissions
to the air with water emissions in order to find out the relative extent to which these result in
increases in mercury and the pathway by which fish uptake increases human exposure.
EPA states at 2-1 that 95 percent of mercury emissions since 1890 have been deposited to
terrestrial soils, 3 percent to ocean surface waters, and 2 percent remain in the atmosphere. It is
not clear from this whether the 95 percent in terrestrial soils includes deposition of water. It is
also not clear the extent to which runoff from soils into water occurs.
In any case, it becomes clear that the cost-effectiveness of controls would be significantly
affected by the fact that emissions into the air are deposited to a large extent to soils rather than to
waterways. If the 95 percent figure "is accurate, than a hypothetical reduction of mercury
emissions of 100 percent would only be 5 percent effective in the short term, since that is the only
amount that affects the water directly. In short, reductions of air emissions are mostly ineffective.
It would appear in theory that the cost-effectiveness of water reductions would be higher.
Given that water discharges affect fish mercury levels directly, while air emissions affect
them only to a limited extent or indirectly, it becomes imperative in any evaluation of mercury
risks to treat the issue as a multimedia problem. It was EPA's mandate to evaluate air emissions,
but EPA could have at least mentioned its voluminous findings from its recent Great Lakes
Initiative as it mentioned natural emissions in its inventory. If fish consumption is indeed the
principal pathway to exposure, the study should orient around the various mercury pathways to
water, and an appropriate evaluation must include the relative contribution of water discharges.
Policy perspectives. EPA has categorized the emission sources into area sources and
three categories of point sources. The highest ranking point sources are four combustion sources.
It follows closely from the organization mat one may characterize the inventory as dividing
emissions into: (1) four categories of combustion sources that together emit the bulk of emissions
(see 6-1), and (2) other sources.
However, one can see a different logic by reorganizing the combustion sources and the
manufacturing sources: MWIs and MWCs are actually the combustion of manufactured products,
not fossil fuels. When reorganized according to manufactured products versus fossil fuels, so that
the manufactured category would include the emissions from both the manufacture and disposal
and combustion, you establish a total life cycle of products. Viewed from this perspective, the
contribution from products to total U.S. emissions significantly increases relative to fossil fuel
sources.
The data are very instructive: battery manufacturers emit a trivial amount of mercury, 0.02
tons per year, which EPA rounds to 0 percent; electrical apparatus and instrument manufacture
emit 0.46 tpy and 0.5 tpy, respectively, which round to .2 percent But these products are
discarded on a regular basis into municipal waste, whose combustion emits 55 tpy, 22.7 percent
Similarly, MWIs emit 26.7 percent of the total
-------
Thus, there emerges a different perspective on the data than appears in EPA's report.
Whereas in the existing method it is clear that combustion sources are die predominant category,
leading one down the path to the easy next step of arguing that emissions controls on combustion
sources are the solution, an alternative would be to divide the world into manufactured product
sources and fossil fuel sources. In this alternative division, fossil fuel sources (utility at 21.2
percent, commercial/industrial at 12.0 percent, and residential at 1.4 percent) only total 34.6
percent, whereas the products from a whole life cycle are roughly 50 percent
If one takes as given that the aggregate of mercury in products constitute the single largest
use, one focuses attention on these categories. This sends the policy analysis in a whole new
direction. Obviously, the emissions from manufacture are trivial, and the problem is disposal in
combustion. There are a variety of alternatives to controlling MWC emissions: (1) one would be
to end MWC entirely: then the opportunity for vaporizing the mercury is less, and more of it stays
in the manufactured materials; (2) combustion with emissions controls; (3) source reduction
through material substitution (pollution prevention). It would seem that the most effective means
of reducing emissions is in reduction of mercury use in products and processes, i.e., pollution
prevention.
VoL ED: Assessment of Exposure from Anthropogenic Mercury Emissions
In VoL HI EPA estimates the mercury exposure to mercury based on its inventory of
emissions from the six categories it identified in the emissions inventory. 'Long-range transport
modeling was used to estimate the regional and national impacts.
Given time constraints, our comments are simply that EPA did not address the question of
relative contribution of risks to humans from air emissions sources and water discharge sources.
For example, in section 2.3 EPA examines mercury exposure from various sources. Section 2.3.5
at 2-8 covers mercury in freshwater. But nowhere in these discussions is there an adequate
treatment of the relative contribution of the various sources to water. This supports the
assessment given earlier, that EPA needs to consider this aspect of the study.
VoL VE: Evaluation of Mercury Control Technologies and Costs
Because of the short turnaround, this review of VoL Vn has focused on the applicability
and costs of the retrofit controls for industrial and utility sources.
General Cnmn^iiU. Overall, the evaluation of technologies represents a significant
effort by EPA staff and reflects a great deal of expertise in these technologies. Therefore, the
comments made here are raised as questions in need of clarification.
The approach seems to achieve the staled goal of providing an order of magnitude
indication of what the reduction potential and associated costs are (for industrial and utility
sources). The results are achieved by scaling model plant estimates to the national level.
However, white these results provide a "feel" for controllability and costs, they lack the site-
specific factors that would be required in a more rigorous analysis. It is hoped, therefore, that
-------
these estimations will not be used as the basis for policy, but rather as an indicator of the need for
a more rigorous analysis, which would include site-specific factors.
Muiriripfll Waste Combustors. On page B-3, Table B-l, the use of a scaling exponent
of 0.6 in footnote b is not appropriate. The use of a scaling exponent of 0.6 is usually appropriate
only for rather simple equipment, such as tanks and pumps. If the equipment is modular or more
complex, a higher scaling factor is more appropriate. Most emissions controfequipment scales at
0.8 or higher.
On page B-4, the additional carbon may have an important effect on the ash if it is
destined for utilization as a pozolonic building material. Strict limits are imposed on ash for such
applications. For landfill it does not matter. '
On page B-6^Table B-3, these estimates of capital costs of carbon bed technology are far
higher than has been suggested by an independent analyst in the literature. William Vatavuk in
Estimating Costs of Air Pollution Control at 167, cites, a rough estimates from a vendor source as
$20-25 per scfm for carbon steel and twice that for stainless, which may be required in high
chlorine cases. See comment below. What lifetime is reflected in the carbon replacement cost?
Comments on Utility Boiler Estimates. An analysis of a model utility plant is described
in Appendix B.3. We have the following comments:
One page B-l9, correct the statement that an increaa*in capital of 86% results from an
increase in size from 100 to 975 MW. See the page for the actual estimate.
On page B-21, Table B-14, the fabric filter capital costs seem low by a factor of about 2.
The capacity factor is listed on the wrong line of the table. The same comments apply to Table B-
15, on B-22.
On page B-25, Table B-17, again, Vatavuk cites a rough estimate from a vendor source as
$20-25 per scfm for carbon steel and twice that for stainless, which may be required in high
chlorine cases. This estimate is for up to 500,000 scfm. The 975 MW model plant corresponds
to about 2.4 million SCFM. accepting a value of $40/scfra (assuming stainless steel) and applying
a scaling exponent of 0.85, the resulting cost is about $75 million. The estimates in Table B-17
are about $250 million.
Also, generally, there is no mention that the cost estimates were derived from retrofit
application. Retrofit costs could easily be 50% higher than new installation costs.
-------
MRR-13 96 10:39 FROM:922 W 4128925917 TO:919 541 0B40 PPGE:02
Comparison of FPA and POP Costs of Using Activated Carbon for Mercury Control
in Coal-FiredJElectric Utility Boilers
Discussion of EPA Costs
The cost analysis for mercury control in coal-fired boilers list in Appendix B in Volume VII of the
draft EPA Mercury Report to Congress contains a number of assumptions that are questionable,
and in some cases, unrealistic.
Questionable Assumptions
\
1. There are no data available to suggest that 90% of the flue gas mercury, in any form at
157 C ( 320 F), can be removed in coal-fired power plant applications. Since the only
reference given is a memo from RTI to W. Maxwell of EPA, it is not possible to evaluate
the assumptions and extrapolations made to calculate the activated carbon to mercury
ratios listed in Table B-12 of Volume VII. At this time, it is unknown whether or not that
up to 90% mercury removal can be accomplished by merely increasing the amount of
activated carbon injected.
2. Spray cooling requires large amounts of water, and ducts sufficiently long to allow enough
residence time to evaporate the injected water. Few coal-fired power plants have long
duct lengths suitable for spray cooling. Also, the impacts of spray cooling on the
downstream equipment is not know. Considerable additional costs associated with
corrosion and plugging of the duct may be associated with spray cooling to lower the flue
gas temperature.
3. All of the example cost analyses using activated carbon were made on low-sulfur coal
applications. Many utilities bum medium- to high-sulfur coal. Spray cooling to reduce
the flue gas temperature to increase the mercury capture capability of activated carbon will
not be feasible due to the large amounts of acid formed when the flue gas temperature is
lowered below 120 C. Also, many units employ flue gas conditioning (addition of sulfur
trioxide) to enhance ESP collection efficiencies when low sulfur coals are burned. Spray
cooling would exacerbate downstream corrosion problems in these applications.
Unrealistic Assumptions
1. The form of mercury was assumed to be entirely HgCl2 to increase the "conservatism" of
the cost analysis. At this time, the distribution of mercury species in the flue gas is
unknown. However, the data obtained to date suggest that as the sulfur content of the
coal decreases, the amount of elemental mercury emitted increases. At normal flue gas
temperatures, it may be more difficult to capture elemental mercury. Also, to be
consistent with other sections of the Mercury Report, a flue gas consisting of 50%
elemental, 30% oxidized, and 20% paniculate-bound mercury should be used.
Laboratory- and bench-scale data indicate that different mechanisms are associated with
-------
3R-13 96 10:39 FROM:922 Ul 4128925917 TD:919 541 0840 Pf=)GE:05
the capture of elemental and oxidized forms of mercury by activated carbon and that
different temperatures may be required for optimum removals.
2. The activated carbon to mercury ratio for the spray cooled system listed in Table B-12 is
460 to one, based on the results of only one small-scale test. It is too premature and
inappropriate to calculate costs based on this carbon to mercury value. Later tests have
shown that much larger amounts of carbon are needed.
DOE Cost Analysis
The DOE estimated costs were based on carbon to mercury ratios similar to those used by EPA in
the ACI only mode. The main differences are that instead of assuming 90% mercury removal, we
assumed about 50% (the less the removal, the higher the 5/11*of mercury removed). For spray
cooling with ACI, we used a higher carbon to mercury ratio than 460 to ojie.
Recommendations
EPA discusses the present state-of-knowledge pertaining to using ACI for mercury control in
Volume VTI and lists the uncertainties in measuring mercury in flue gases and lack of full-scale
testing. A more complete discussion of the potential adverse impacts of ACI with or without
spray cooling on the operation of coal-fired utilities is recommended. However, this present
state-of-knowledge is not transferred to both Executive Summaries (Volume VTI ES, and Volume
I). Both summaries contain the same paragraph describing "two important considerations", both
implying ancillary benefits or increased cost-effectiveness of yet unknown mercury control
technologies. What needs to be added are statements describing the uncertainties of trying to
apply mercury control technologies developed for one source category (municipal waste
combustors) to another (coal-fired utilities). This is especially important for the uninformed
reader or others who might only read the Executive Summaries. It is equally appropriate to add
the sentence "Also, it is possible that potential national costs for mercury control may increase as
more is know about the technological issues associated with effectively controlling mercury
emissions from various source categories."
-------
EPA Responses to DOE Comments (Dated February 21, 1996) on the Mercury Study
Report to Congress
To facilitate the review of EPA's responses to DOE. the DOE comments have been
numbered sequentially. The responses to specific comments presented below correspond to the
numbered comments in the February 21, 1996 letter to Martha Keating, EPA from Dan W.
Reicher, DOE.
1. EPA agrees to provide a more balanced executive summary in that more information will
be provided on emissions (and their uncertainties) and control technologies and costs.
2. And 3.
The EPA's reference dose (RfD) for methylmercury was.developed independently from
the development of the Mercury Study. It has been publicly available since May 1995 on
the EPA's Integrated Risk Information System. For the Mercury Study, the RfD was (in
late versions of the Report) not used to characterize risk. The Agency has also agreed to
incorporate the data from the Seychelles and Faroe Islands studies into the RfD
calculation
4. EPA chose to model mercury emissions from anthropogenic sources because the existing
data measuring mercury concentrations around these sources were judged to be
insufficient to estimate the resulting mercury exposure to humans or wildlife. EPA
reported the publicly-available measured mercury data in Chapter 2 of Volume 3; data
reported included: fish concentrations, atmospheric concentrations, soil concentrations,
water column concentrations and sediment concentrations. Some of the measurement
data collected in remote sites was also used for a comparison with modeled results.
While the model results are not validated at this time, they are comparable to
measurement data. For example, EPA cites measured concentrations of mercury in fish.
These data are the compiled results of national, state and local surveys. They include both
freshwater and saltwater species. The measured mercury fish concentrations range from <
0.01 to 8.94 ug/g (ppm). The highest mercury concentration reported in Volume 3 was
measured in a largemouth bass in New Jersey. Although EPA has focussed on the range
of measured data, other measured concentrations that exceed 1 ppm are also reported. If
there are other peer-reviewed data concerning environmental mercury or methylmercury
concentrations that should be included, please inform the EPA.
Using the hypothetical settings described in Chapter 4 of Volume 3, the methylmercury
concentrations predicted in fish depended on the overall deposition rate of mercury from
the atmosphere. Several factors affected mercury deposition rate near the model plants:
distance from the source, form of mercury emitted, and the effective height of the stack.
A range of methylmercury concentrations was predicted by the models used in the Report.
The highest methylmercury concentration predicted in fish was 4.9 ppm. This was
-------
predicted to occur in a hypothetical lake 2.5 Km from a small municipal vsaste combustor
in the eastern half of the U.S. Please see the Report or contact EPA for more details
regarding the modeling.
The methylmercury concentrations in fish were predicted using a bioaccumulation factor
(BAF) that was derived from several sets of measurement data. EPA attempted to
characterize the uncertainty of the BAF both in Volume 3 and Volume 5. Throughout the
Volume EPA has attempted to describe the uncertainty associated with the modeling.
EPA feels that this is reflected in the conclusions of the Volume. EPA will attempt to
illustrate uncertainty and convey it more clearly in Volume 1 of the Report.
5. EPA does have boiler specific emissions estimates. They were not included in the report
because we felt that the data would be mis-used (e.g., certain boilers or companies would
be "targeted" for reductions). This commenter felt it surprising that there were no data
for certain source categories such as coke ovens. A thorough review of these source
categories was made during the development of the mercury emission factors. No data
were found that were sufficient to calculate emission factors for the source categories
listed. As emission tests are conducted and data are obtained, EPA routinely updates it's
emission factors in the AP-42. EPA agreed to review the Argonne report that DOE
mentions that pertains to coke oven and refinery emissions. However, DOE did not
provide a reference for, or copy of, this report.
6. DOE refers to the 1995 estimates for certain source categories as predictions based on
future regulations. These are not predictions, these are updated emissions estimates for
the year 1995 and do not reflect predicted regulation. A trend for utility boilers is
calculated in the Utility Study. It was not included in this report because the supporting
documentation was not included (e.g., energy forecasts).
7. EPA was unable to distinguish between the inhalation and ingestion pathways for
sensitive subpopulations except as noted in the quantitative exposure assessment. In
addition, EPA also did not assess the impact direct discharges might have on the water
these groups may be consuming. The assessment focused only on anthropogenic air
emissions.
8. We could not find the commenter's reference in Felsvang to high required feed rates for
80 percent efficiency. However, the point seems to be that a minimum quantity of carbon
exists that is required to get adequate contact between flue gas and adsorbent. We
examined the carbon injection models to find if they met the commenter's criterion of 73
mg/m3 (carbon to flue gas). The models operating above 200 °F met the criterion, but the
200 °F models did not (models Ib, Ic, and 2b). A sensitivity analysis was performed on
those models with the results shown in the following table. Text to accompany the table
and to follow the sensitivity analysis for percent HgCl2 was added as follows:
. 2
-------
For the models operating at 200 T, the small amounts of carbon injected nmv be
insufficient for proper mixing with the gas stream. Table B-18 shows the results of a sensitivity-
analysis for carbon feed rates at the amount used for estimating cost effectiveness (3.5 mg
carbon/m of flue gas) and a suggested value of 73 mg/m3. As shown in the table, increasing the
carbon injection rate by a factor of 20 increases the cost of operating the control svstems
(mills/kWh) by about 40 to 120 percent depending on the model. Cost effectiveness values in $/lb
rfHg removed increase by similar percentages. TJte cost increases are caused by increased
values for carbon usage and disposal. Carbon injection system costs for the 100 MW boiler have
relatively modest increases because a minimum cost for the injection system is built into the
model, i.e., the system will not cost less than about $110,000 for any of the models.
The following table was added to the text:
Table B-18
Sensitivity Analysis for Carbon Injection Feed Rates
on Utility Boilers
Model
lb(975MW)
lc (975 MW)
2b(WOMW)
Feed Rate, mg
carbon/actual m3 flue
gas
3.5
73
3.5
73
3.5
73
Cost Effectiveness,
$/lb ofHg Removed
17,300
24,400
4,970
10,990
27,700
34,800
Cost Effectiveness,
mills/kWh
' 1.43
2.01
0.41
0.90
2.26
2.83
9.
10.
The executive summary will be reviewed to ensure the tone and content are consistent
with the entire report.
EPA believes that the best available data on global and natural emissions is already
included in the report. These data are believed to be the best available assessment of
global and natural emissions and their relationship to anthropogenic emissions. This
information represents the consensus opinion of an international panel of mercury experts
which was convened by the Electric Power Research Institute. EPA admits that there are
data gaps in the inventory and lists the source categories we were unable to quantify. The
-------
need for additional emissions data for a number of these sources is listed as a research
need.
11. See response to comment number 4.
12. The social cost analysis section was deleted from the report.
13. The EPA's decision not to have a formal public comment period was made in February
1994 and was based on the belief that comments from the general public would not
advance the science given the review the study received from technical experts in the this
field. In addition, the schedule did not allow for what we believed would be a 4 to 6
month effort. The emissions inventory has been published twice as well as presented at
numerous conferences and workshops. Written comments have been received (and
incorporated) from nearly every industry that is listed in the emissions inventory.
14. No change.
15. No change.
16. The executive summary will be revised to provide balance between all aspects of the
report. In particular, more information will be added regarding the emissions inventory
and the control technologies analysis.
17. As noted above, the social cost section was removed.
18. See response to comment number 6.
19. The Mercury Study seeks to provide information on whether air emissions of mercury for
U.S. sources are having an impact on public health and the environment. DOE correctly
notes that there are other emissions from global and natural sources as well as direct
discharges to water that influence mercury exposure. Unfortunately, we are unable to
quantify these other sources beyond what we have already included in the report. We
have recently contacted other mercury experts to be sure there has not been any recent
literature that estimates global or natural emissions, and there has not. However, DOE is
correct that these issues can be better highlighted in the Executive Summary.
20. These comments on control costs for the utility industry are more relevant to risk
management rather than to the technical evaluation of control costs. The Mercury Study
made no judgements as to the economic feasibility of control. The cost numbers were
presented objectively.
21. No chan ge needed.
-------
22. Through 35.
The language suggested in these comments was incorporated into the report essentially
verbatim.
36. Comment on p. 3-2, cost effectiveness vs HgCl, content of flue gas stream.
To determine the cost effectiveness sensitivity of mercury removal with carbon injection.
models were used with several values for the fraction of mercury as HgCK in the flue gas.
The table below gives the results. As shown in the table, cost effectiveness as measured
by the additional cost of producing electricity (mills/kWh) does not change as the HgCU
fraction of total mercury decreases. This lack of change occurs for two reasons: less
carbon is used for lower amounts of HgCl: and equipment costs are relatively insensitive
to small changes in particle concentration in the flue gas. However, cost effectiveness as
measured by quantity removed from the flue gas stream (S/lb of mercury), increases by a
factor of about 3.3 in going from 100 percent HgCK to 30 percent HgCK. This increase
occurs because the costs of flue gas treatment decrease only slightly while the quantity of
collectible mercury decreases by 70 percent.
Source
Utility Boilers
<\
Mercury Control
Technique
Activated carbon
injection (coal fired)
100%ofHgas
HgCU
80 % of Hg as HgCK
50%ofHgasHgCl,
30 % of Hg as HgCU
Cost Effectiveness
$/lb Hg Removed
4,970-27,700
6.190-34,600.
9,840-55,400
16,300-92,000
Cost Effectiveness
mills/kWh
0.41-2.26
0.41-2.26
0.40-2.25
0.40-2.25
Table 3-1 was revised to include the four HgCU ratios. The following text was inserted at
the end of Section B.3.2:
Table B-17 presents the results of a sensitivity analysis for mercury species in the flue
gas (elemental mercury vs mercuric chloride). As shown in the table, cost effectiveness as
measured by additional cost of producing electricity (mills/kWh) does not change significantly as
the HgCl2 fraction of total mercury decreases. This lack of change occurs for two reasons: less
carbon is used for lower amounts ofHgCl2 and equipment costs are relatively insensitive to (,
small changes in particle concentration in the flue gas. However, cost effectiveness as measured
-------
by quantity removed from the flue gas stream (S/lb of mercury), increases by a factor of about
3.3 in changing from 100 percent HgCL_ to 30 percent HgCL. This increase occurs because the
costs of flue gas treatment decrease only slightly while the quantity of collectible mercury
decreases by 70 percent.
Table B-17
Sensitivity Analysis for Carbon Injection on Coal-fired Utility Boilers with Different Ratios of
Elemental to Oxidized Mercury
Otcidi^ed Mcrcur\' in Flue Cras
100 % ofHg as HKCL
80 % of Hf> as HgCL
50 % ofHR as HgCL
30 % of Hz as HKCL
Cost Effectiveness
f>/lh HP Removed
4,970-27,700
6.190-34,600
9.840-55.400
16,300-92.000
Cost Effectiveness
mills/kWh
0.41-2.26
0.41-2.26
0.40-2.25
0.40-2 25
37. Language was added to Volume VII to emphasize that information gaps must be filled
before a thorough assessment of specific impacts on utility boilers can be determined.
38. The editorial changes were corrected. The section on exposure already states that fish
consumption dominates over other exposure pathways. In Volume EH, this dominance
can be seen quantitatively in the numerous tables provided for the results of the exposure
analysis. No other language was added.
39. The executive summary (Volume I) and the emissions inventory clearly state the
limitations of the inventory and the emission factor approach. Where data were available
they were included in the inventory. The research needs section clearly lists that
emissions data are needed for a number of source categories including those that are used
by the commenter as examples of data gaps in the inventory.
40. Emissions trading is discussed briefly in Volume VII as an alternative control. The acid
rain market is used as an example.
41. through 44. These comments were taken, and changes made as suggested.
45. See response to comment number 47 below.
46. Careful consideration was given to reorganizing the inventory as suggested, but it became
apparent that time constraints did not allow for a total reorganization of ohe report to
-------
accommodate this comment. The perspective illustrated by the commenter is utilized
somewhat in Volume VII however where various source categories are used to illustrate
management strategies such as pollution prevention.
47. The report was revised to recognized that air emissions are not the only contributors to
mercury in the environment. Direct water discharges are listed as a specific example both
in the inventory, the exposure assessment and the executive summary. While an
inventory of water discharges is beyond the scope of the study, the EPA's Mercury Task
Force, which covers all media, is addressing these other types of sources (water.
hazardous waste, etc.) In the Task Force effort, the Mercury Study serves as a resource
for estimating air emissions only. Other program offices are addressing other media so
that a multimedia approach is assured.
48. Comment on p. B-3. scaling factor of 0.6
The 0.6 scaling factor has been in use for more than 50 years (and introduced into the
literature by R. Williams in 1947). The value of 0.6 is an average over many kinds of
equipment and facilities. Contrary to the commenter's assertion, the rule can be used for
facilities (Aries and Newton), although the average exponent is nearer 0.7 than 0.6 (range
of about 0.3 to more than 1.0). Because 0.6 is an average, different values can be used for
specific types of equipment or facilities. Remer and Chai list about 200 values for the
exponent depending on type of process or equipment being considered. Their work
suggests that the "six tenths rule" should be more like the "seven tenths rule" (0.68 for
their average), but 0.6 is still commonly used. For the carbon injection modeling, a value
of 0.68 was used.
The commenter states that "Most emission control equipment scales at 0.8 or higher."
However, this average value is probably based on 10 values in Remer and Chai weighted
towards water and wastewater treatment and stacks. Particulate and venturi scrubbers are
also listed with environmental equipment, but scale at 0.6.
It must also be recognized that the costs for fixed bed absorbers in the mercury report
obtained by the 0.6 rule apply primarily to tanks and compressors, which scale at values
of about 0.3 to 0.65 (tanks) and 0.32 to 0.75 (compressors). If anything, the 0.6 value is
probably too high rather than too low.
49. Only landfill disposal costs were estimated; use of the ash as a pozolonic building
material was not considered.
50. See response to comment number 53, below.
51. Comment on p. B-19 re: 86 percent increase in capital
-------
The sentence was changed to read: A size increase from 100- to 975-MW increases the
total capital cost for the control arrangement by a factor of 7.4 and increases the
annualized cost by a factor of 6.1: the cost effectiveness value decreases by 38 percent
and the annualized mills/kWh cost impact decreases by 32 percent.
52. Comment on p. B-21 re: low baghouse costs
The commenter states that the fabric filter capital costs seem low by a factor of about 2.
We compared the costs for the two tables questioned (Tables B-14 and B-15) with costs
available for several utility baghouses (including an EPRI economic analysis). The
mercury report costs are SI 1.57/acfm and S9.99/acfm for the small and large model
plants, respectively. These values are near the bottom end of the range for other utility
baghouses, which run from less than $10 to more than S20/acfm. with one station at
$42/acfm, all updated to 1993 dollars. The latter baghouse was an early unit built very
conservatively and heavily instrumented. The costs for these baghouses include extra
ductwork and retrofit costs as needed for the individual installations. The mercury report
baghouse costs do not include extra ductwork, conservative design, or retrofit costs
(which are site specific), as cleariy stated in the text accompanying the tables. The text
states that including these extra costs could increase total costs by 50 percent or more,
which would put the $/acfm costs right in the middle of the range found for other utility
baghouses.
A footnote was added to the two tables on the fabric filter total capital cost cell that says
"See text for comment on factors that may increase costs."
53. Comment on p. B-25 re: high cost for carbon bed
The commenter cites Vatavuk as giving a rough estimate for carbon bed adsorbers at $20
to $25 per scfm in carbon steel and double for stainless steel. These are summer 1988
dollars, which would change the range to about $21 to $26 per scfm for 1993. These costs
were reviewed and as the commenter correctly noted, were high. Both the MWC and
utility boiler carbon bed costs were revised (as noted in the red-line draft). The issue was
the amount of carbon needed for these applications. The estimates were double-checked
with a vendor to make sure they were reasonable.
54. The commenter states that no mention of retrofit costs exists. This is correct for carbon
beds, but not for carbon injection. This clarification was added.
55. The highest temperatures found in the utility emission tests were 250°F. At these
temperatures the mercury removal efficiency ranged from 69 to 91 percent. It is also
important to note, as the Mercury Study does, that when the collection efficiency is
varied, only the cost-effectiveness calculation changes, not the overall cost. The national
-------
costs calculated for the executive summary was appropriately caveated and presented a
range of collection efficiencies.
v
56 and 57.
The effects of spray cooling and downstream condensation problems are discussed in
section 2.3.1 in Volume VII. The text notes that utilities firing high sulfur coals would
have problems lowering their temperatures sufficiently.
58. An extensive analysis was done to reflect different ratios of mercuric chloride in the flue
gas. This analysis was described in the response to comment number 37.
59. DOE is correct in noting that the activated carbon to mercury ratio is based on data from a
pilot scale test. This was the best information available to the EPA at the time of this
analysis. All of the cost analyses done to date by EPA and all other parties have been
based on pilot scale tests as these are the only tests available on which to base costs.
60. EPA has been asked by a number of reviewers to justify the national costs presented by
the Mercury Study as compared to the costs mentioned in a number of DOE comments.
DOE was asked to provide it's cost analysis so that a comparison could be made. EPA"s
analysis is described in detail in Volume VII of the report so that all assumptions can be
understood. DOE's cost analysis is described by DOE as simply using a different carbon
to mercury ratio than EPA, and using a 50 percent removal efficiency rather than a 90
percent removal that EPA uses.
The 90 percent removal efficiency that EPA used to calculate national costs (assuming all
boilers would spray cool, use activated carbon and have a fabric filter) was based on the
data presented in Appendix A to Volume VII. Pilot-scale data for boilers using this
configuration of controls achieved mercury removals ranging from 76 percent to 99
percent. Mean percent removal was 95 percent. No other comparisons can be made with
the DOE analysis because there was no other information provided.
61. Numerous statements regarding the uncertainties of activated carbon injection for utility
boilers were added to the Executive Summary. Most changes were taken verbatim from
DOE comments.
-------
SECTION FOUR
REVIEW OF THE MARCH 22.1996 DRAFT OF THE MERCURY STUDY REPORT TO
CONGRESS
This section contains correspondence from several Federal agencies. Subsequent to
receiving these comments, the decision was made to refer the draft report to the Science
Advisory Board. Consequently, EPA has not prepared responses to these comments or made any
additional changes to the document (with the exception of minor edits). This section contains the
following in order:
1. NIEHS comments dated March 29, 1996,
2. NOAA comments dated April 1, 1996,
3. Council of Economic Advisors comments, dated April 1, 1996,
4. Office of Science And Technology Policy (OSTP) comments dated April 1, 1996,
5. USDA comments dated April 2, 1996,
6. Centers for Disease Control comments dated April 4, 1996,
7. FDA comments dated April 5, 1996,
8. OSTP comments dated April 11, 1996,
9. National Biological Service comments dated May 9, 1996.
-------
SENT BY:OFF. of COMM. - NIEHS ; 4- 1-36 ; 10:27 '-OFF. of COMM. - NIEHS- 313 541 0840;# 2/ 2
Date: March 29, 1996 .
To: Martha Keating, US EPA
From: Robert A. Goyer, M.D.
Re: Comments on EPA Mercury Study Report To Congress, Draft
April, 1996
I have only reviewed volume IV, Health Effects and volume I, the
Executive Summary
Many of the comments previously submitted by NIEHS and others
regarding problems and appropriateness of the Iraqi study as a
basis for determining the RfD are now Included In the report.
The new comments in the report considerably enforce the previous
criticism that the Iraqi data do not provide support for reducing
the RfD for methyl mercury from 0.3 to 0.1 yg/kg bw.
Nevertheless, the calculation of the RfD remains as in previous
draft, (volume IV, page 6-29).
The Executive Summary adds uncertainty to The EPA position
regarding the current RfD, (Executive Summary, Volume I p 3-21,),
acknowledging emerging new data. Para 1 states that there will be
future review of the current RfD. Therefore, this Report to
Congress may not be based on definitive health effects data.
Copies: Dr. Ken Olden, Director NIEHS
Kevin Neyland, OMB
-------
UNITED STATES DEPARTMENT OF COMMERCE
National Qeaanic and Atmospheric Administration
NATIONAL. \iAf=iNE = S^-EPIES SERVICE
1335 Ease-West Hicnwav
Stive1" S~r^g, ^'CZ 2CS^O
THE DIRECTOR
Ms. Martha Keating
U.S. Environmental Protection Agency
North Carolina Mutual Life Building
Durham, North Carolina 27701
Dear Ms. Keating:
The National Marine Fisheries Service (NMFS) has reviewed
the revised EPA Mercury Study report to Congress. We note the
straightforward statements in the EPA Administrator's transmittal
letter (Volume I), reiterated elsewhere in the report, that the
typical U.S. fish consumer is not in danger from exposure to
methylmercury from fish and is not being advised to limit fish
consumption. However, NMFS is disappointed by EPA's disregard of
the agreements reached at the February 28, 1996, interagency
review meeting and addressed in my letter to John Seitz dated
March 8, 1996. The continued and additional inclusion of
discussion points contrary to the interagency agreement dilutes
the message regarding the safety of fish for the typical
consumer. Therefore, we cannot accept the latest revision of the
report.
One of the points agreed to by the interagency group was
that the report would mention, but not emphasize, the RfD issue.
However, in the revised Volume IV, EPA now chooses to cite
liberally, but selectively, from the New Zealand (and post-study
evaluations), Peru, and Seychelles Islands studies to support the
current RfD. By preempting the agreed-upon comprehensive review
of all information relevant to determination of the RfD, EPA has
now taken a position which places it at odds with its fellow
agencies. Further, this was done in contradiction to a statement
made earlier in Volume IV (and repeated in Volumes I and VI) that
"an interagency process with external involvement will be
undertaken for the purpose of reviewing these new data, their
evaluations, and the evaluations of existing data." We strongly
urge EPA to delete from the report the numerous elaborative
statements on the RfD and conclusions based on the various
studies cited.
Also, contrary to interagency agreement, EPA continues to
refer to a fish consumption level of 100 grams/day in its
analysis of the safety of the commercial fish supply. The report
notes that the value of 100 g/day comes from the World Health
Organization recommendation that, as a preventive measure, women
of childbearing age in subpopulations consuming 100 g/day of fish
should have their hair methylmercury levels monitored. EPA has
chosen to use this screening level as a cutoff for safe fish
©
Primed on Recycled Paper
THE ASSISTANT ADMINISTRATOR
FOR FISHERIES
-------
consumption by the typical U.S. consumer. EPA's statement, "the
commercial U.S. fish supply is safe for the U.S. population who
consume less than 100 g/day of fish and shellfish, and a wide
variety of fish types," may be readily interpreted as providing a
benchmark above which eating seafood is unsafe, and below which
it is safe. Neither EPA nor any other agency is in a position to
provide a quantitative measure of health risk to the U.S. seafood
consumer from mercury; hence the agreement to await the
deliberations of an expert work group following review of all
relevant information.
We believe it is essential to resolve these key issues before the
report is submitted to the Congress. We look forward to
discussing our concerns with the relevant parties in the
conference call on Tuesday, April 2.
Sincerely,
/
L
Rolland M/Schmittenx
cc: Kevin Neyland, OMB
-------
PPR-01-1996 19=25
CEP!
EXECUTIVE OFFICE OF THE PRESIDENT
COUNCIL OF ECONOMIC ADVISERS
WASHINGTON,TS. C. iOSOO
P.001/001
SENIOR ECONOMIST
Memorandum
TO:
FROM:
SUBJECT:
April 1, 1996
John S. Seitz, Director, office of Air Quality Planning
and Standards FAX 919-541-Q84O
Raymond Prince
Review of Revised Draft of Mercury study Report to
congress
The CEA has several concerns about the Revised Draft of the
Mercury study to Congress. These are as follows:
• The revised draft indicates that a concentrated deposition
of air-borne mercury occurs in the Northeastern and Mid-
Atlantic states but does not discuss mercury levels in fish
taken from Atlantic fisheries off the northeast coast or
from the Chesapeake Bay. This information should be
included in the final report.
• The revised draft fails to mention a DOI study which is
reported to have been submitted to EPA one year ago that
questions the amounts of mercury in wildlife from
anthropogenic sources. This information should be included
in the final report.
• The revised draft seems to underestimate the costs of
reducing mercury emissions by utilities given that current
technology captures a very low percentage of gaseous
mercury. Opportunities for substitution and new process
technology should be included in the final report.
• The revised draft has not been peer reviewed by the
Scientific Advisory Board. The final report should be held
back until Board approval is obtained.
Sincerely,
Raymond Prince
c. Alicia Hunnell, CEA
Kevin May land, OMB (FAX 5-5836)
OPTIONAL FORM 99 (7-90)
FAX TRANSMITTAL
TOO/TOO©
oa-sdovo --*• osdo/assov/vda
-------
- i-ao
EXECUTIVE OFFICE OF THE PRESIDENT
OFFICE OF SCIENCE AND TECHNOLOGY POLICY
WASHINGTON, D.C. 20502
April 1. 1996.
Memorandum to: Lynn Goldman, J_iPA
Bob Huggeit, EPA /jl
TJ. Glauthier, OMB
Sally Katzcn, OIRA
Katie McGinty, CEQ
Mary Nichols, EPA
Alicia Munnell, CEA
From Jack Gibbons and Bob Watson, OSTP
CC: Martha Keating, EPA
Brad Campbell, CKQ
Ray Prince, CEA
Re: Review of the Revised EPA Mercury Report
While OSTP recognizes that EPA has worked very hard to improve the seven volumes in a
very short time, we continue to believe the report is not ready to be released. The report will
be a very important statement on the quality of EPA's science and the Administration's
endorsement of it. This report is simply not of the caliber of many of the premier national
and international assessments. In its revisions, EPA has gone a long way toward admitting
there is inadequate data and many uncertainties and we commend them for these efforts.
However, there is still much work to be done to make even the Executive Summary a useful
readable document In the last few weeks the author of the Iraqi study and an author of the
Seychelles study have been quite vocal and spoken to several agency and Hill staff
denouncing the use of the Iraqi study at this time. EPA must thoroughly discuss the
limitations of the current data and the opportunities posed by the new data directly hi their
report. We must make sure this report is scientifically credible.
Many OSTP staff have contributed to our review of the report in the few short days allotted.
We conclude that essentially there are only two options to consider:
Option 1: Seek legal means to delay release of the report
OSTP found that EPA made many changes in the Summary of the document in response to
the very extensive agency comments submitted, but far fewer to the back volumes. However,
even the Executive summary needs considerable work. We: continue to believe the report is
not ready to be released and the optimal solution would be a multi-month delay to fully
incorporate the Seychelles and Faroe Island analyses. Such an extension would also permit
validation of the EPA models of exposure and adequate response to many remaining
-------
i i_. < i I l\.Ui V"t— < 1 U 1 » I O I U. *
concerns on all 7 volumes— expressed not only by OSTP, hut by other Federal agencies und
Congressional staff". This would ensure a credible product after the many years of work that
have gone into its preparation. Our detailed comments are attached (Attachment A). We will
be happy to provide a detailed markup of the summary volume text as well.
Option 2: Release the report on April 15, clearly marked "draft" if a set of important
changes are made and, approved by agency signoi'f.
In order to ensure signoff on the report, EPA must clearly state the strengths and weaknesses
of the Iraqi, Seychelles and Faroe Island data and unequivocally state in Carol Browner's
cover letter and in the Summary that no regulations will be based on Ibis report but will
await the analysis of the Seychelles data. In order to be suitable for release, the Summary
needs a short (5 page) Executive Summary. The cover letter must be rewritten and a minimal
number of other changes must bejnade in several areas: elaboration of the data analyses done
to date, comparison of the FDA and jgEAJlfp. description of the strengths and weakness of
the Iraqi and Seychelles arid Faroe data sets, clarification of past present and future emissions
sources, availability and costs of control technology, alternate data bases that could be used to
estimate wildlife effects, and some regional depiction of mercury concentrations in coastal \^*
areas. We amplify on these below and in our extensive accompanying comments
(Attachment A). We will be happy to provide a detailed markup of the summary volume text
as well..
Changes needed for conditions of Options 2 to be met:
Cover Letter:
The cover letter must be totally rewritten. The tone of the letter implies the current report
is based on the best available science which the backup documents refutes by .saying FPA
must analyze the better Seychelles and Faroe Island data. Further, the Seychelles data should
not be referred to as unpublished or abstracts; eleven papers have been published. The letter
implies further control is prudent but the backup report doesnt clearly explain why. The
backup report does not support that populations of loons and panthers may be declining from
Summary and Volume 4 and 6 (Health Effects and Risk Characterization)
1) Based on the amount of media coverage on the recently published studies of
methyhnercury exposure in the Seychelles Islands and comments on the utility of those
studies versus the Iraqi studies for estimating risk from exposure to mcthylmcrcury in fish, it
ia critical that the EPA report provides a qualitative comparison of the quality and
reliability of those two studies for risk assessment. A major issue is whether the Iraqi
study which was used to estimate the reference dose (RID) in the EPA report is suitable for
estimating risk in segments of the population exposed to melhylmercury from consumption of
fish. The authors have told OSTP, EPA and House Congressional Staff that it should not. A
comparison should be provided in a text and table format that identifies the strengths and
weaknesses of each of the 3 studies and their impact in estimating risk. The comparison
-------
IL.UI rwun-i , •*- i-ao , -t-iorti ,CAYIKUAMLAI uiviilOiN- a2bU51oo;? 4/22
should include at least the following characteristics:
i) study design: number of child-mother pairs, endpomt measurements and
observations, age of child when evaluated, retrospective vs prospective analysis
ii) potential confounders: nutritional status, alcohol consumption, exposure to other
developmental toxicants
iii) exposure: fish vs grain, chronic vs acute, exposure range
iv) quality and reliability of the data sets: potential errors in endpomt measurements or
observations, exposure classification errors
2) The report mentions an intcragcncy process with external involvement that will be
undertaken to review the new data from the Seychelles and Faroe Islands to determine if a
change in the RfD is warranted. Most importantly, in the transmittal letter* irom Carol
Browner to Congress as well as in the report, it must be made clear that any regulatory
decisions on mercury will be based on the RfD that includes evaluations of the data from the
mercury studies in the Seychelles and Faroe Islands.
3) An important statistical issue concerns the grouping of data for the calculation of the
RfD. Since (he manner in which the data were grouped could affect this value, the analysis
should also be performed using ungrouped data. The presentation of grouped data, is useful to
help illustrate the dose-response, however, results of the analysis with ungrouped data should
also be included in the report because that would provide a better portrayal of the entire data
set. .
4) A comparison of the FDA and EPA RfD should be presented
Volume 2 and Summary- Anthropogenic emissions of mercury. EPA did not address all of
the detailed comments submitted previously by OSTP but did make some minor revisions.
Overall this chapter suffers from the lack of detailed emission data from the sources .
considered. It also is very misleading in that the text does not specifically state for which
year, or years, the emissions tables refer to. The executive summary and text of the
chapter needs to be rewritten so that it gives 4 main messages (for which the data is in the
text but not well organiysd):
1) There has been a real success in the US in the dramatic drop in mercury
emissions from manufacturing over the past decade.
2) The largest identifiable sources of mercury emissions currently are municipal
solid waste incinerators and medical waste incinerators. Regulations now under review by
EPA will cut these emissions by over 50%.
3) After the municipal solid waste and medical waste incinerators have been dealt
with, the largest remaining identified source of mercury emissions are coal-fired utility
boilers. The next challenge will be identifying cost-effective means of limiting emissions
from these sources.
4) Much of the emissions data is of poor quality or based on very few analyses. We
need additional focused research to define emissions from other sources, such as petroleum
3
-------
01 -urrc ^n ^ *cui ru_iLi; -i- i-at>
refbnung, which may be significant.
Additionally, both the Executive Summary and the chapter need to show a range of future
emissions so the reader can understand how fast or slow emissions could grow.
Volume 3 and Summary: Focussing on Freshwater tish (less than 20% of the consumed
fish in a normal diet) does not alleviate concerns about saltwater fish. We continue to think
that data on coastal ocean sediments, bivalve tissues concentrations, or fish tissue, could be
used to characterize regional saltwater mercury concentrations (see attached map—Attachment
B). EPA's own description of the increase in marine sediment concentrations of mercury
combined with the deposition map and advection statistics wijl lead readers to conclude where
marine concentrations arc likely to be highest anyway. We anticipate immediate questions
about Chesapeake Bay loadings. We strongly disagree with averaging concentrations within
each fish type—use a weighted average for clams, tuna, etc based on percent of each
species/subspecies eaten.
Volume 5 and iSummary: Ecological Assessment. Previous comments made by DOI a year
ago were minimally addressed. We attach these for your further consideration (Attachment
C). Of primary concern is the case of a bioaccumulation factor for fish accumulation of
mercury based on water total mercury concentration. There should be a more systematic
review of wildlife data bases relative to both contaminants and population status and trends.
There are data on more than fish. Someone from the research side of DOT must review this
volume.
Volume 7 and Summary: OOE raised many significant concerns about (he control costs
cited. EPA has raised their estimates of costs somewhat, but they still do not approach DOE's
estimates. Some description of why these numbers differ must be made. Also, DOE claims
the control technologies needed are not close to commercialization. EPA does not address
this and must before the report is released.
Additional Questions EPA should resolve:
What is EPA's response to EPRI's concerns that EPA models overestimated deposition?
Error bars? Is use of Relmap (2x deposition) and COMPDEP (used for ecological impacts
even though EPA deemed COMPDEP unsuitable for its dioxin study. Why didn't EPA use
1SC3 as it did in the dioxin study?
Is it true that the US is only 5% of world mercury emissions???? Page 3-5 and Table 3.1
would lead you to believe this. Given that the US is roughly 20% of global energy use we
find it unusual that we are such a low contributor to mercury. Explain why this is (do we
have vastly better controls?)
Why didn't EPA use the Florida's ambient Mercury Study?
-------
OCA i oi-utTL, o^i ,* IC.LH rULiCY; 4- i-db , 4;i3tJM ,LNVIKU,\\it.M DIVISION- 92605155i# 6/22
Volume II: Anthropogenic Emissions of Mercury
EPA did noc address all of the detailed comments submitted previously by OSTP but did
make some minor revisions. Overall this chapter suffers from the lack of detailed emission
data from the sources considered. It also is very misleading in that the text does not
specifically state for which year, or years, the emissions tables refer to. The executive
summary and text of the chapter needs to be rewritten so that it gives 4 main messages (for
which the data is in the text but not well organized):
1) There has been a real success in the US in the dramatic drop in mercury emissions
from manufacturing over the past decade.
2) The largest identifiable sources of mercury emissions currently arc municipal solid
waste incinerators and medical waste incinerators. Regulations now under review by
EPA will cut these emissions by over 50% .
3) After the municipal solid waste and medical waste incinerators have been dealt
with, the largest remaining identified source of mercury emissions are coaJ-flred
utility boilers. The next challenge will be identifying cost-effective means of limiting
emissions from these sources.
4) Much of the emissions data is of poor quality or based on very few analyses. We
need additional focused research to define emissions from other sources, such as
petroleum refining, which may be significant.
-------
iNI BY:or~FC SCI & TECH POLICY; 4- 1-36 ; 4:13PM ;ENVIRONMEi\T D1VIS. - 32605153;* 7/22
A-2-
Additional Comments on Volume IT: Inventory of Anthropogenic Mercury Emissions
There has actually been little change to the document. Some of the limitations to the data
are more clearly stated which helps. However, with some rearranging and minor edits the
document can be much* better. The tone of the document and arrangement of various tables
makes it appear that EPA is focuses almost exclusively on the utility industry, 1 think EPA
can make this point a bit more obliquely and more effectively. One of the main messages
here should be that over the past 10 years mercury usage (and emissions) in the US has
dropped dramatically due ro taking mercury out of paint, batteries, and reducing the level in
fluorescent lights. With the proposed rulcmaking for medical waste and municipal waste
incinerators it should drop significantly again. Then EPA can make the argument mat die
last remaining major source of emissions is the utility industry (and it follows logically that
perhaps they should move to cut emissions).
Specific comments:
1) ES-2, paragraph beginning with "Moreover, a complete..." next to last sentence ending
with "mobilized by natural sources." This statement needs to be backed up by a reference
citation or removed.
2) ES-4, 1st paragraph under Anthropogenic Emissions Summary, 2nd sentence "While
/ these emission estimated for anthropogenic sources have important limitations, they..." (add
underlined word.
3) ES-4,'2nd paragraph under Anthropogenic Emissions Summary, last sentence. The
, sources should be listed in decreasing order from the largest source to smallest source, i.e. 1.
medical waste incineration, 2. municipal waste combustion, 3. utility boilers, 4
commercial/industrial boilers. This change needs to be made throughout the document in
both the text and tables. While it is true that proposed rulemaking will hopefully change the
order, since this is current estimated emissions, the order should be followed. By putting
utility boilers first in all these lists it appears that EPA is targeting them at the expense of the
, larger emitters. The report should state in this paragraph or the next that EPA is currently
«/ proposing a rulemaking to cut emissions from the two highest emitters.
4) ES-4, 3rd paragraph under Anthropogenic Emissions Summary, 1st sentence, should be:
"All four of these sources represent high temperature waste combustion or fossil fuel
processes." (again put currently largest first).
5) ES-5, Accuracy of Inventory. The text suggests that the emissions inventory could
understate emissions by as much as 20%. What is not mentioned is that the inventory could
overstate the emissions by probably a.similar amount. Based on the newer 1995 data it is
obvious that the inventory is overstating a number of categories. What did the review panel
say concerning the possibility of overstatement of emissions? Need to be fair here and give a
range (plus and minus) not just a number and a higher estimate.
6) Table ES-3 For what year is this lable? Is it an estimate of 1990 emission rates? It
\f
-------
,? a/22
A -3
clearly is not current because you have numerous notes stating lower 1995 values. Please
live, the reader some indication of what this data represents. Perhaps average 1990-1993?
The Mg/yr needs to be explained in a footnote that Mg is Megagrams - note a normal unit of
measurement. In the point sources section of the table you need to rearrange so that MWI
and MWC are the first two entries under combustion sources. Again list sources within both
Combustion and Manufacturing sources by order of abundance.
7) Table ES-4. For what year is this table? Is it an estimate of 1990 emission rates? It
clearly is not current because you have numerous notes stating lower 3995 values. Please
give the reader some indication of what this data represents. Perhaps average 1990-1993?
8) Table ES-5. For what year is this table? Ts it an estimate of 1990 emission rates? It
J clearly is not current because you have numerous notes stating lower 1995 values. Please
give the reader some indication of what this data represents. Perhaps average 1990-1993?
~ 9) Table ES-6. For what year is this table? Is it an estimate of 1990 emission rates? It
clearly is not current because you have numerous notes stating lower 1995 values. Please
give the reader some indication of what this data represents. Perhaps average 1990-1993?
Again, sources should be listed from largest to lowest, i.e. starting with MWI and MWC and
then utility boilers.
In the basis column for utility boilers the statement is made "industry estimates are . . " In
/footnotes please say what industry estimates, National Coal Assoc.?, Utilities groups?
There are also interesting questions regarding the column degree of uncertainty. For two
categories that EPA says it has only medium uncertainty (MWC and MWI) EPA now says in
oasis for emission estimates or footnotes that the estimated emissions are too high, in the
case of MWC it appears that the estimate is too high by almost 50% . This certainly seems
like a high degree of uncertainty. Also, it makes one wonder if the medium uncertainty of
the utility boilers means that EPA may have also grossly overestimated them as emissions
sources. I understand that EPA put medium uncertainty for any group that had test data, but
your data suggest that even in these cases the true emissions amounts are still highly
uncertain.
10) Table ES-7. For what year is this table? Is it an estimate of 1990 emission rates? It
clearly is not current because you have numerous notes stating lower 1995 values. Please
give the reader some indication of what this data represents. Perhaps average 1990-1993?
Put sources in table in order of emissions.
/ Need a brief note or reference to section in report to explain why the EPA emission rate of
6.5 t/yr. is nearly twice that of the industry estimate of 3.3 t/yr. Same argument with lead
smelting and copper smelting. What looks bad is the amount of overestimation by EPA of
these emission values. If they are due to increasing success of industry in arneHorxting
mercury usage and escape this should be acknowledged.
/' -
11) Table ES-8. For what year is this table? Is it an estimate of 1990 emission rates? It
clearly is not current because you have numerous notes stating lower 1995 values. Please
J give the reader some indication of what this data represents. Perhaps average 1990-1993?
Perhaps EPA should also prepare a table like Table ES-3 that is limited to best data for 1995
-------
A-t
to illustrate what the current status of emissions is.
NOTE: These tables are scattered throughout the report. Please make changes to them
similar to the suggested for the executive summary.
12) ES-12, top paragraph. What are the years for the estimates of mercury per person for
die European countries? 1980? 19907 1995? It is hkely that emissions have dropped in
several of the European countries over the past few years.
13) ES-12, Trends in Mercury Emissions. The first sentence does not appear to be true
based on the data presented in the report. Suggest tagging on a phrase at the end so the
sentence reads, "It is difficult to predict with certainty the temporal trends in mercury
emissions for the U.S. though there appears to be a trend toward decreasing total mercury
emissions from 1990 to 199S."
14) ES-12, Trends in Mercury Emissions, 4th paragraph. There should be a sentence here
praising the success of industry in recycling and thereby (hopeful) reducing the total amount
of anthropogenic mercury available in the environment.
15) ES-13, top paragraph, next to last line. Talking about some state regulations - need a
footnote to put in more specific information - which states? date of regulation? effectiveness
of regulations?
16) ES-13. Many of the conclusions deal with land areas having greatest mercury
concentrations. 1 would suggest that a paragraph be put in before the conclusions that deals
with this aspect of the report, abstracting section 5 of volume 2.
17) ES-13, Conclusions, 1st point, modify to: "Numerous industrial and manufacturing
processes emit mercury to the atmosphere. Recent decreases and phasing out of mercury in
a number of processes is lowering emissions."
18) ES-13, Conclusions, 6th point, "Anthropogenic sources in the U.S. emit approximately •
220 Mg (243 tons) of mercury annually into the atmosphere. This estimate is believed to be
accurate within 30 percent." Need to state that the 243 tons is for some year previous to
1995, perhaps use the 243 ton number for 1990-93 and then put in the new calculation
(suggested above) for a 1995 number.
19) 2-2, 1st paragraph. Sentence beginning with "For example, analysis of sediments from
Swedish lakes ..." this statement needs to be supported by a reference: Sentence beginning
with "In Minnesota and Wisconsin, an investigation ..." This sentence needs to have a
reference to document the statement.
20) 2-4, 1st complete paragraph. This paragraph is confusing and somewhat misleading.
Mercury emissions are declining in US (and probably other industrialized countries).
Concentrations are probably certainly increasing due to deposition. The data then talk about
fish. But is the problem here emissions or deposited, and re-cycled mercury? Perhaps this a
-------
,c.\v
place to put in a paragraph or two about the importance of concentration and re-cycling or
mobilization of previously deposited mercury.
21) 3-1, section 3-1, 1st paragraph, last sentence. Modify to: "It is unknown exactly what
—*• effect these measures might have on mercury emissions, but it is suspected that they should
at least slightly reduce mercury emissions."
_ 22) 3-2, Table 3-1, see comments for this table from ES
/ 23) 4-1, Combustion Sources, 1st, 2nd, 3rd paragraph. Put combustion sources in order of
emissions. Then reorder sections so that 4.1.1 is Medical waste incinerators, etc.
, 24) 4-2, 4.1.1 Utility boilers, 2nd paragraph, 4th sentence, rewrite: "Almost all of the coal
J burned in the US(95 percent) is bituminous and sub-bituminous, while only 4% being
lignite.^
— 25) 4-3, Table 4-2, see comments on table in ES
26) 4-7, Mercury Concentrations in Oils and Natural Gas. This paragraph does not track
/ with data on 4-67. There you say petroleum has 0.023 to 30 ppm mercury from the 1990
EPA report - is this data from USGS? So is petroleum data from analyzing oils for from
emissions tests? You should state that there is extreme variability in the Hg contents of
natural oil with some (such as Cymric Field in S. CA) having very high Hg contents
(mercury beads are visible in it) and most others having low contents.
You need in this section to make 2 important points:
1) Hg concentrations in oils can be high (30 ppm is an order of magnitude higher than any
of the analyses you present for coal, though I'm sure there are coals with 30 ppm Hg).
x ^ Therefore petroleum is a significant potential source of emissions.
sl 2) Emissions from petroleum are most likely at the refining stage when oil is heated. After
fy this process most of Hg will have been driven off so that oil and natural gas fired plants
f should have relatively low Hg emissions.
27) 4-8, 1st paragraph, 1st sentence. Give reference for USGS data. Also present the
"""* range of mercury values in a footnote.
28) 4-8, 3rd paragraph. "Since approximately 77 percent of the eastern and midwestcrn
/ bituminous coal shipments are cleaned ..." State who conducted the analyses discussed in
this sentence (reference) and how many analyses were done.
y 29) 4-14, Table 4-4. Coal - Comments on industry estimate. Who? How?
30) 4-18, last paragraph. "Industry estimates of mercury emissions from this source
,, category [referring here to MWCs] for 1990 are 40 Mg/yr." This is the same figure used on
V Table 4-4 for the coal column. Has something got mixed up here?
31) 4-20, 4.1.5,2, 1st paragraph, last sentence add phrase "with concentrations ranging
-------
oi
from 1 to 6 ppm by weight and a typical value being 4 ppm by weight, or four times the
amount of the most mercury-rich raw coal listed on Table 4-3."
32) 4-35, Table 4-10. see suggestions in ES.
33) 4-41, Portland cement box. No "ores" are used in the cement process. Mercury
content of limestones is exceedingly low - an order of magnitude below coal or petroleum.
, The mercury involved in the Portland cement process comes from the fuel used in the
J thermal treatment. You should change second paragraph of box to read: "The initial step in
the production of Portland cement is acquisition of raw materials - limestone (calcium
carbonate) and other minerals such as sand (silica). These minerals have virtually no
w» contained mercury. "
34) 4-59, last paragraph. Why doesn't this paragraph track the data in table 4-3? I think
v you are right that coal can have up to 8ppm Hg (probably higher) - what is the reference for
this? -"but you should cross-reference this number with the number for raw coal analyses in
table 4-3. Note that the argument here is that coking facilities can emit as much Hg as utility
plants - should they have the same regulation? ~ ~~
J 35) 4-62, Primary Lead Smelting, 1st paragraph. Rewrite: "Primary lead smelters recover
lead from sulfide ores which may contain mercury. The smelters emitted an estimated.:."
J 36) 4-62, Table 4-21 . Is it 1990 Production, Mg of lead? or mercury? Specify.
/ 37) 4-64, 4.2.12, last sentence "Date pertaining to mercury contents of the ore are not
available." Yet you show them in table 4-23. Delete sentence?
J 38) 4-64, Table 4-22. MAGMA Copper Co. no longer exists, it is now BHP Copper Co.
(applies also to table 4-23).
30) 4-67, Table 4-23. I would be very suspicious of the numbers given by Phelps Dodge-v.^
_^ Hidalgo of 5,768 Ibs. Hg in ore concentrate/yr. and only 0.09 Ibs./yr. mercury emissions J^1
when compared to all the other data provided. Is the Engineering Judgment from the d?*
company or EPA? In this case if from the company I would so state.
31) 4-68, 4th sentence. Need to put in a sentence stating that "Petroleum refining may be a
significant contributor to mercury emissions. More analyses of oils and refinery emissions
v are needed to evaluate this source."
*«, 32) 4-69, 1st paragraph, add sentence: "Mercury contents of limestone are extremely low.
"""" Emissions from ..."
/ 33) 4-72, 1st sentence after table. This sentence is meaningless. Just because steam comes
through the earth does not mean it will contain mercury. Cut the first sentence.
34) For section 5 & 6 see comments in the executive summary and apply.
-------
DJ -urfL, i^J ,* ICLH rULJOl. -i- I -do . -±-20:^ .cAVlKUtVUCM UmilUV - 320U5155 - S12/22
7-1, 4th bullet, rewrite: "More data are needed on the mercury concent of various coals
and petroleum and 'the trends in the mercury content of coals burned at. utilities and
.petroleum refined in the U.S."
36) 7-1, 6th bullet, Rewrite: "The importance (quantitatively) of re-cm ission of mercury
from mercury-bearing mining waste and previously deposited emissions ..."
-------
ICLH rui_iLi . -i- i-ao . t••i.zr.M -CAV IKUIYMLM uiviiioA- 32ouaioo,;?i;j/22
A-?-
Additional Comments on Volumes n and IV: Health Effects and Risk Characterization
1) The transmittal letter from Carol Browner claims that "the health assessment portion of
the report is based on i.he best scientific data available at. the present time." Based on the
fact that there have been 11 recent publications from the studies of mercury exposure in the
Seychelles Islands that are not evaluated in this report, the above statement is misleading.
The transmittal letter should also indicate that external experts will participate in the
evaluation of the RfD based on review of the new studies of mercury exposure hi the
Seychelles and Faroes Islands.
2) The report should acknowledge that results from the studies of the Seychelles islands have
been reported in more than just abstract form. Eleven papers from these studies were
published in Neurotoxicolgy in 1995.
TI
3) A fuller description of the hair sampling procedures for the Iraqi study would be helpful.
Did the segments of hair analyzed for mercury correspond to the full period of gestation for
the mother-child pairs'?
4) The data on mercury concentrations in fish species used by EPA (Vol. VI, Table 4-8)
should include information on variability and range of these estimates.
-------
O^tJUQiOO i rri-t/ <
Volume HI: Exposure from Anthropogenic Mercury and Summary:
Overall, Volume III has been improved significantly. There is substantially less emphasis on
the derivation of the Reference Dose and on the potential hazards of fish consumption. The '
entire report appears to be more focused on mercury emissions, transport, and deposition. In
addition to clearer text, several useful fiigures and tables have been added. Some or the
supporting data is presented in a more relevant and understandable form.
*
Focussing on Freshwater fish (less than 20% of the consumed fish in a normal diet) does not
alleviate concerns about saltwater fish. We continue to think that data on coastal ocean
sediments, bivalve tissues concentrations, or fish tissue, could be used to characterize
regional saltwater mercury concentrations. The Coastal Ocean Database—COSED— (13,500
samples from around the country—some from the Estuarine component of EPA's
Environmental Monitoring and Assessment Program, and NOAA's Mussel Watch Project
(originally an EPA program) with 154 coastal sites, and the National Status and Trends Fish
liver assessment all give some insight as to regional differences (see attached map—
Attachement B). At least show a map of mercury concentrations in one or all of these data
sets. EPA's own description of the increase in marine sediment concentrations of mercury
combined with the deposition map and adveption statistics will lead readers to conclude
where marine concentrations are likely to be highest anyway. So summarize the available
information for the audience.
It is quite clear that people will use information in the Executive Summary to wonder about
the Chesapeake Bay loadings. We think the clam number hi the Executive Summary is in
error—does not agree with Volume 3. A number closer to 0.12 would also be more
representative of the Mussel Watch numbers. We strongly disagree with-averaging
concentrations within each fish type—use a weighted average for clams, tuna, etc based on
percent of each species/subspecies eaten.
Appendix II
Data and default values for fish consumption appear to be better substantiated and
consumption patterns among Alaskans, Hawaiians, and other native Americans are now more
thoroughly documented. Much of the material is now presented in helpful tables and
summaries. The data regarding fish consumption by recreational anglers and subsistence
fishers has been expanded, and data regarding mercury concentrations in fish has been
updated and refined. The methodology used to estimate fish consumption is explained in a
manner that will be more understandable to all readers and areas of uncertainty are presented
in a more direct and unambiguous fashion.
-------
Volumes IV and VI: Health Effects and Risk Characterization
1) Based on the amount of media coverage on the recently published studies of
methylmercury exposure in the Seychelles Islands and comments by the authors-on the utility
of those studies versus the Iraqi studies for estimating risk from exposure to mechylmercury
in fish, it is critical that the EPA report provides a qualitative comparison of the quality and
reliability of those two studies for risk assessment. A major issue is whether the Iraqi study
which was used to estimate the reference dose (RfD) in the EPA report is suitable for
estimating risk in segments of the population exposed to mcthylmcrcury from consumption of
fish. This comparison should be provided in a text and table format that identifies the
strengths and weaknesses of each of these studies and their impact in estimating risk. jne
comparison should include at least the following characteristics: •
i) study design: number of child-mother pairs, endpoint measurements and
observations, age of child when evaluated, retrospective vs prospective
analysis
ii) potential conftmnders: nutritional status and general health, alcohol consumption,
exposure to other developmental toxicants
iii) exposure: fish vs grain, chronic vs acute, exposure range
iv) quality and reliability of the data sets: potential errors in endpoint measurements
or observations, exposure classification errors
2) The report mentions an interagency process with external involvement that will be
undertaken to review the new data from the Seychelles and Faroes Islands to determine if a
change hi the RfD is warranted. Most importantly, in die transmittal letter from Carol
Browner to Congress, as well as in the report, it must be made clear that any regulatory
dccifnons °n mercury will be based on the RfTJ thai; includes eygfoiarions of the data from the
mercury studies in the Seychelles and Faroes
3) An important statistical issue concerns the grouping of data for the calculation of the
RfD. Since the manner in'which the data were grouped could affect this value, the analysis
should also be performed using ungrouped data. The presentation of grouped data is useful
to help illustrate the dose-response, however, results of the analysis with ungrouped data
should also be included-in the report because that would provide a better portrayal of the
entire data set.
-------
rY-
Chapter 7: TECHNOLOGIES
In general, the section on technologies has improved and the piece in the executive summary
provides a good overview. As far as we can determine, most of the comments received were
incorporated.
However, there still is a discrepancy between the report and the DOR comments concerning
the cost-effectiveness of mercury control for coal-fired utility boilers. DOE figures remain
higher than the numbers used in the report thought the report surveys a broader range of
emerging options.
In terms of balance, the report still tilts towards control technologies over pollution
prevention. Since, as the report admits, most of the control technologies are still in the
prototype stage, there is a real option to go down a pollution prevention path; A significant
portion (over 70 percent) of the emissions are associated with the production of products so -
that process changes and material substitution strategies are possible. Rven in the case of
utility boilers, all estimates arc derived from retrofits whose cost could be 50 percent higher
than new installation cost (integrated systems will be cheaper in the long-run than end-of-
pipe). In terms of their mandate, they were supposed to look at "Control Technologies and
Costs" but, hi reality, you cannot do that in isolation.
The report does not mention the need to develop better monitoring technologies. Our ability
to optimized control technologies is dependent on our ability to track the amount and
chemical composition of Hg-containing compounds in flue gases, process streams, etc. This
needs to be listed as a major research need because of the relationship between monitoring
precision and control/avoidance success.
-------
National Status and Trends Program
MARINE ENVIRONMENTAL QUALITY
Benthic Surveillance -
Mussel Watch - l
Bioeffects Surveys
Cl
c
-------
-t- i-ao.
,c.NviNu>v>iC.\i
Distrthuttnn tif chemical «>ncciitrtiiii>n\ in < titniui
Fig. 2.
£5
Percentages of '5 x high' sites in COSKD. Mercury and Kidrnium an: by far the
most commonly found contaminate u( '5 x high' levels.
example, found thai the concentration of PAH increases with proximity to
urban areas. Cantillo & O'Connor (1992) showed by factor analysts that
chemical concentrations at NS & T sites are strongly influenced by proxi-
mity to population centers.
Area! extent of chemical concentration?)
i
Converting COSED to an estimate of the arcal extent of chemical concen-
trations is, in principle, possible. The procedure would begin with creation
of a concentration contour map of the US coast. The coastal area with Cu
concentrations below SO /Jg/g, fur example, would be the sum of all areas
minus those inside the 50 /ig/g contour. However, that would be not only
tedious but also subjective given the clustered distribution of COSED sites.
Sampling schemes have been designed .specifically to provide statisti-
cally rigorous estimates of environmental conditions on a spatial basis
(Holland, 1990; Paul et al,, 1992). The EMAP/EC Programme sampling
design is based on a grid pattern of 270 km2 hexagons mapped onto the
US prior to data collection, and sampling sites arc chosen by randomly
selecting hexagons. While the single sample in a hexagon reveab nothing
about that particular hexagon, the resulting data set is a random sample
of a large region. In particular, the proportion of total urea with sediment
a
-------
r/.v <>/ (Itcnin tit < ntrt i'ntr(tnetn\ tn tnn\\t'i\ ttti<{ f
TABLKs
Annual Geometric Mean (.'ninji-niraiions of Chemicals in Mollu^kv Tor Sites Sampled in ;u 1 e;ni S>*
Years Iroin 19X6 ta 199). t'oiicentnitmii "Units: ;,y jf ; Hry Weight lor blcmc-nts .mil nu, g ' I3iv
Weight tor Organic Compounds With tBT in ["cruis of 119 of Sn i>~ ' Ory Welkin I ,i\l ('iilunin is the
Spcarrruin Rank Corrthition Cocfliuicm of Conccntrnuon Versus Ye,ir With an Astcriik (*)
licnotiiic Slanslioil Signidciiiicc >.!
2.1
H.7
no
0.11
2.3
1.6
0.5
2.5
130
7000
7.3
4.4
26
86
ISO
47
S.5
2.5
8.1
P.O
0.12
1.7
1.7
059
2.4
130
1900V
7.2
3.7
25
82
ino
32
0 780'
-0.857'
-0810'
0.294
I!.42S)
-0.333
-0.57 1
0.667
0.381
-0.291
0.133
O.»8h"
0.810'
-0.878°
-n.j»33'
-0.4SA
-1.000*
Annual data have been aggregated to national geometric means in Table 5, where
because of species difTerences, concentrations in mussels have been separated from those
in oysters for Cu, Pb and 7.n. Figure 3 visually displays some of the decreases identified in
Table 5. Decreasing trends at the 95% confidence level and at this national level ol
aggregation eiist for As, Cd, Cu (in mussels), all the chJurinalcd organics, and tBT. This
list is approximately the same as that based on simply counting site-by-site trends.
86
. 3- Data I'mm Table 5 for live chemicals with decreasing annual geometric means.
yjgsir?*"'-
»#:£'•;*
-------
• —
jy-yo uq. to ru
C i
Sumnwry o£ Commonto o£ Te«y A. Halnw fas ifa« Inlengeacy Ravie* of tho U.S. EPA;s "Meicuiy Study
Report to Coogreu"
1. Of the wvea volumes cf tic report,' I have noi rev»wod the following volumw, JOT reuoni stated:
L EneutivB Sumnuiy: Dot received.
H Inventory of Alahiopogenic Mnrcury BtaiHum in the United Suto«: no ezpeitue in thia u«t.
TV. Health Effects of M*roary ud Mnettry Campoiindt; no expoctbe in aarnia l»»lth.
m An BvmUutiott of Moreuty Canccnl -n»chnologie«, Cotu end JUgukwry IWBM: na ezputu>
intbi* me.
2. Vohiaw EL Aa AuMnueac of Exponuie Fiom AnUaopogBnic Mucnxy BnuMioM in the United Slates:
a. Th» bMkgiDand »&nn«doB on nwscaty i* rtuanabiy compete «ed wtcnrxte. and i» A good
. •mmnarjrof the lituatioa.
b. The long raage tnuwport tn*ly«l« prodncH reocnablo estinuuM of total «nau»l Kg wet
depOiiticmfijrlhB trr-ttt in thff nnrfh"«* *»4 n^honatral US for wMeli I Juve Man deposition
BUuatBxaoatB. l.o^ xottl depoaidon in the iu£o of 5- 10 pgta2. Bewover, the model piedicti that
most of the deposition i> in the form of Hg, uid •intoit none U particoUte Hft, wnerea* all of ti«
awuamnBflE ^lta 1 us swrnre of fi"H« matt Kg la the putkulat* fbonin w«t depoaitioa, aod
vimuily none it duaolwd Hftj. Thcio «n vtiy fijv«a«npt»uj qoudfy 017 depodtioa of Hg, «od
Futlw,
thit •—• Titr-*^ ignoxcB the methyl mercany eooiponeaf hi wet dapraitioa, wUch nuy be 1-Sft of
total «vt depoaftioa. In wme location! (»,g., Swvdwn and Maine) dw atmoapheric depoaWoa of
mtthyl IfMMCWy y«* ff^^ttf1 ^ F"?""1^ fnr "n "«^«Mry aemnnilifari in flik llu« la llifliy JO be
avMtin&wiuat la areaa wfaara aqoatle cooditiona an aot conductv* to methylation of inorganic
c.Taere ia couridorable oYBttap betw«eu VeduaiB V(EeologteaI AJBeamneot..) xod this aectioo,
Tail ZMQiB IncouidanfalB conftialoaon my part ecaopniag Ju« wnat yon old and why. UN me
fgctuc ^"T flab *j*j~"r""i«^"" erf ousazy buad on watar total tnecBoiy
coaemtration it likely TO produce lugs czran. Fiah do not anmnnilate toorgaiac fatau of
meiemy, «ad tbere ia no dentonsontad eomlanon between water total meicuryoo¢ntioni and
wiMmetb^LcneiaiuyoaaceDtratiana. Methyl mucuycxnn&gD fiomlou than 196 of total
mmcoiy la u nmrfi u 1091 or mare. Vary high anoooxandana of Inorpnio zaeicaty naulting
feom-wcata ditpoami do notnaee«auily retail In high canccntrmjona of methyl aaxevy. The
pmdudioB of xnBfhyl mercury in the •trruoamanl naolta from bacterial action, and vaiiea
mbataadally with sach tUagc aa water chemistry (especially acidity), diuolvBd organic caxbon
t, ptupoitloa of vretlanda In m. deainage. Mopetauue, dlaaolved ozyym, and oth»r factor*- In
a, and aize o£ fiah>
ia gnctMt in large, old fuh that nMd high oa tte aqaatic food chain and h*w lived long enough to
aocumolate high body bordeaa. TheBAFwiUaotbecoiutantfiQtallfpedeaaBdagesof Sab. The
food chain vptak* cannot "bm readily aaMaaed by ue of *BAF. A bioenergctie modal approach
-------
t L-*~tl i Ul_ l v_ i . i " j. ~ Ou i 1 • £,ui .u ' I— * » li\W, \.I1L-\ 1 u I » I O l U. l
ra
P03
d. Bald eagles (in Maine at least) may consume fish of various trophic lorveU, sad not ail trcplie
level 4 aa assumed bate.
3. Volwno V. Aa Ecological AawMroent iof Aninropo jpsnic Mfircnry Emiwons in the United States:
a. Given the importance of dry depoMUaa for other pirticulate meUla, it is likely Hut dry
U •!«« •« JmpnT^nf i^^h.^^r. ftw^^r^my «• «
-------
£-3
k. The calculated mercury rwlduoi ja flub csjrwponding with the above wiMifa criterion value
are slndltriy highly uaiviliitic. Methyl mo nary caaceatnuum at 0.008 m/g fat trophic level 3
and 0.04 far trophic kvcl 4 are at lean aa aider of magnitade too low. Measured nercaiy
coaceptraiioiu (virtually
-------
'ftPR-02-1996 23=55
P.02/04
SSx Unrtea States
Department c;
Agriculture
Office of Risk Assessment
and Cost-Benefit Analysis
Washmoton, D.C.
20250-3300
Date: -April 2, 1996
To: Alwyneilc Ahl
Director, ORACBA
From: Michael D. McElvaine
Public Health Specialist
Subject: Review of EPA Mercury Study to Congress, April 1996 Draft
BACKGROUND
As public health specialist of USDA's Office of Risk Assessment and Cost Benefit Analysis
(ORACBA), I reviewed portions of the December 1995 draft of this document. Our joint
, comments were sent to Martha Keating at EPA through a facsimile dated February 16, 1996. We
commented on three specific topics: 1) reference dose of methylmercury, 2) mercury levels and
US fish consumption, and 3) an alternate fish consumption model. EPA agreed to make
revisions to the report that would incorporate the comments from USDA and other agencies. The
current draft, dated April 1996, was delivered to ORACBA on Monday March 25,1996. EPA
asked that final comments be returned by April 1. They also stated that a specific reply to
ORACBA comments would be sent at a later date (this has not been received as of this date).
Following is our evaluation of the revisions that have been made in the current draft and how
well these revisions address the concerns stated in our February 16 transmission. These
comments are presented under the same three headings that we used previously. Also included
are some specific comments about typographical errors. As before, this review covers only
Volume I: Executive Summary, Volume HI: Exposure Assessment and Appendix H: Estimation
of Human Exposure Through Consumption of Fish.
REFERENCE DOSE OF METHYLMERCURY
In our comments of February 16, we stated that the reference dose (RfD) may be too
conservative, (i.e,, too low). We based this on current research and questions about the
assumptions used for estimating the current reference dose. If the RfD is too low, then there
would be an overestimate of the population at risk.
In Volume I, page 3-21, EPA acknowledges that there is much new information available which
may have an influence on setting the reference dose. This new information comes from such
-------
fiPR-02-199S 23=5^
P.03/04
sources as the Seychelles and Faroes Islands studies as well as smaller studies done in
populations around the US Great Lakes. When all of this data has been published and rigorously
reviewed, they will reassess the RfD for methylmercury to determine if change is warranted.
We believe that this is a suitable response to our comments, depending on whai the time frame
for reassessing the RfD would be. Any change in the RfD wouid have a great impact on the
estimates of the population at risk. Because of this, we suggest that EPA keep this report in draft
form until the RfD can be reassessed and then evaluate whether changes need to be made in the
report.
MERCURY LEVELS AND US FISH CONSUMPTION
In our comments of February 16, we stated that the report gives the impression that any
consumption offish was unhealthy because there is mercury contamination at some level in
every finfish and shellfish. We emphasized that the focus should be on what the mercury
contamination levels were in the fish that were actually consumed, not on the fish that come from
contaminated lakes or streams from which the general public has already been told not to
consume the fish. We were especially concerned about the lack of information about mercury
levels of farm raised fish which, based on limited testing, may have much lower levels of
mercury than wild caught fish.
EPA has included several statements in this current draft report and in the draft cover letter
which specifically address our concerns That the public would get the wrong message about the
relative risk of eating fish. In Volume I, page 1 -1, they state "the typical US consumer of fish is
not in danger of consuming harmful levels of methylmercury and is not being advised to reduce
fish consumption." Similar statements appear at other places in the report. In Volume I, page 3-
27, they identify three factors which modify the potential exposure to methylmercury through
fish consumption. These factors were 1) local or regional advisories about eating fish from
certain sites known to have high levels of mercury contamination, 2) most freshwater fish in the
US have mercury levels which are only a fraction of the FDA action level of 1 ppm, 3) even
though some locally caught fish may contain high mercury levels, much of the" fish consumed in
these areas comes through commercial channels and will have mercury levels similar to the
national averages, levels significantly lower than the FDA action level.
We believe that the changes made in this draft respond to most of our concerns in this area. The
clear and repeated statements about the risk of mercury exposure to the average fish consumer
helps to clarify the muddled message that comes from the highly technical sections. However,
we feel that our concern about the lack of information about mercury levels in farm-raised fish
has not been dealt with adequately. In Volume I, page 3-33 there is an additional statement
about the mercury levels in catfish as measured by FDA for their Pesticide and Chemical
Contaminant Data Base. They report levels significantly lower than those used for the EPA
report. If these levels are in farm-raised fish (which we believe they are), then the previous
statement on page 3-33 should state this explicitly. We also believe that the mercury levels
found in farm-raised fish should be considered at each point in this document wherever there is
-------
ftPR-02-1996 23=56 ' . ' p_04,.Q4
discussion of the average Level of mercury in fish consumed in the US (for example, in Volume
III. Table 2-11}. Fish farming is already an important source of animal protein for the American
diet and will become increasingly important in the future.
AN ALTERNATE FISH CONSUMPTION MODEL
In our comments of February 16, we suggested an alternate fish consumption model that would
have focused more on the mercury levels in the specific type offish consumed rather than using
an estimate of the average level of mercury contamination. By suggesting this alternate model.
we wanted to point out that there is more than one way to look at a question. We also wanted to
emphasize that the exposure risk depends on the amount of mercury in the fish actually eaten by
each individual rather than the average amount eaten by the average person.
It was not our intention that EPA should respond directly to this suggestion in the report and they
did not. They did respond to our request to look at the species offish actually consumed by
noting in the Research Needs chapter of Volume I (page 5-3) that, more information was needed
about fish species-specific consumption rates. We would also suggest another research need is
better information about mercury levels of the fish in commerce, especially the mercury levels of
farm raised fish.
DISCUSSION
In general, we believe that EPA has made a good effort in responding to most of the concerns we
expressed in our February 16 response. However, we think that the additional suggestions
offered here would provide even more improvement in the report. We continue to have concerns
that some consumers may be unnecessarily scared away from fish consumption because of
concerns about mercury. This message can be clarified even more. We also feel strongly that
there needs to be more information about the levels of mercury in farm-raised fish. There is a
continuing increase in the consumption of these fish. We need to test to assure that general
contamination levels are no higher than early reports indicate. If this is true, we cannot allow
them to be tainted by association with wild-caught fish with much higher levels of
contamination.
TOTfiL P.04
-------
DEPARTMENT OF HEALTH & HLMAN SER\ ICES
Centers rcr Disease Control
ana Preven:i::" CCC.
Atlanta GA 30341-3724
(770) 488-7001
FAX (770) 488-7015
April 4, 1996
Ms. Martha Keating
USEPA
North Carolina Mutual Life Building
411 West Chapel Hill Street
Durham, North Carolina 27701
Dear Ms. Keating,
Thank you for the opportunity to provide additional comments on the EPA Mercury
Study Report to Congress. I have reviewed the latest version of the report. My
comments are limited to the materials in Volume I, the Executive Summary. Fn
general, I like the tone of the report and believe it continues to be improved.
My comments begin with Table 3-5, titled Mercury Concentrations in the Top Ten
Types of Fish Consumed by US Residents. There appear to be several
inconsistencies with the table. The table title should be changed so that it reflects
that both fish and shellfish are included in the table. I am also concerned that
providing an average may not reflect the best measure of central tendency for
mercury concentrations in fish and that you may wish to add the median or
geometric mean to the table. There are also several inconsistencies in the table,
for example, under the fish species pollack, the mean is given as 0.15 while the
comments suggest that the concentration is 0.04.
Perhaps my greatest concern occurs on page 3-34 where the report sites
recommendations from the World Health Organization. The second paragraph of
the page, last complete sentence suggests that subpopulations that consume large
amounts of fish be monitored for methyl mercury. I suggest that this sentence be
deleted from the text since it implies that subpopulations be monitored for methyl
mercury when no such recommendation actually exists in the United States and
thus would result in confusion. On the same page, the last full paragraph, the
WHO recommendation is again cited as "a screening analysis to identify
populations potentially at increased risk". Since there exists no such
recommendation in the United States, I believe this paragraph should also be
deleted.
The discussion that begins on page 3-34 seems to conflict with earlier statements
made on page 3-30. The earlier statements suggest that the best measure of
-------
human health risk occurs from following exposed populations prospectively to
observe health outcomes. The efforts to link risk based on dietary surveys that
begins on page 3-34 does not reflect the inferior nature of this method.
I would also like to comment on Table 3-6 that appears on page 3-36. I believe
the title of this table should be changed and the term "of concern" be deleted from
both the title and the column heading. A better title would be ... "The Estimated
Numbers of Persons Consuming 100 grams or more of fish per day." I do not
believe we need to be concerned about the amount of fish consumed unless we
characterize the fish as contaminated.
Last fall, EPA stated that it desired to coordinate the risk communications on the
release of this report with the Public Health Service. Unfortunately, there has
been no follow-up on the risk communication effort. I am concerned that there
may be expectations of CDC regarding the risk communications that cannot be
met. Please let me know, at your earliest convenience, if there is a need to
coordinate the risk communication efforts at this time. This is especially important
if EPA plans to release the report on April 15.
If you have any questions regarding my comments please feel free to call me at
770/488-7001.
Sincerely yours,
Thomas Sinks, Ph.D.
Associate Director for Science
National Center for Environmental Health
-------
fiPR-12-1996 14:35 FROM OFFICE OF SEflFOOD
DEPARTMENT OF HEALTH *- HUM AN SP
TO
919195410840 P. 02
Public Health Service
Fond anri Drug Arfminisrration
Washington DC 20204
-5 1936
'Ms. Martha Keating
U.S. Environmental Protection Agency
North Carolina Mutual Life Building
Durham, North Carolina 27701
Dear Ms. Keating:
We appreciate the opportunity to provide comments on EPA's redraft of its
Mercury study report. These comments refler.r a quick review on our part, in
order to meet your original time frame for comments. If we have additional
comments, we will provide them at a later date.
In general, it is apparent that EPA has made significant efforts to address FDA's
concerns about Uiis report. However, the draft continues to include text that
does not appear to reflect agreements made among EPA, FDA and the
Department of Commerce in February. Areas of concern in this regard involve
(a) the emphasis to be placed on the RfD; (b) the acknowledgement of new
studies without the drawing of conclusions about them; (c) the emphasis of the
report on freshwater species rather than commercial species. All of these will be
addressed below. We also have comments on other aspects of the report. A
number of our comments recommend additions and deletions to the report. All
of our suggested additions appear in bold.
VOLUME 1-EXECUTIVE SUMMARY FOR ENTIRE REPORT
Transmittal memo, p2, last par, 1. 3:
In order to convey uncertainties associated with data used in the health
assessment portion of the report, we suggest deleting the period after time and
adding as follows:
"the present time but that data have several limitations."
In the absence of qualifying language, the reader could be lead to believe that the
health effects assessment of methylmercury exposure among fish consumers is
based on conclusive evidence.
-------
fiPR-12-1996 14:35 FROM OFFICE OF SEAFOOD TO 91919541084Q P.03
Page 2 - Ms. Keating
Vol 1, p2-6, last par., 1. 2:
In order to properly characterize FDA's current action level, we recommend the
following modification:
"The current action level is 1 ppm methylmercury "
Vol 1, p2-6, last par, 1. 4:
Further clarification concerning the safety of commercial seafood would be
appropriate here in keeping with the Browner trausmitxal letter to the Congress,
as follows:
"levels of methylmercury in seafood and is not being advised, directly
or indirectly, to reduce fish consumption."
Without such statements, the average consumer may well wonder whether he/she
needs to reduce their commerical fish consumption in accordance with this
report.
Vol 1, p3-21, par 1,1. 7:
Since smdies hi the Great Lakes are still ongoing and nothing has been published
or peer reviewed regarding findings from these smdies, we suggest that EPA
delete the following statement:
"which describe effects in populations"
This language could lead the reader to conclude incorrectly that adverse health
effects have been observed in methylmercury exposed individuals and that these
studies have been completed, peer-reviewed, and published.
Vol 1, p3-22, last par., 3rd line from bottom:
In order to properly reflect FDA's plans regarding review of its action level, we
suggest deleting the current text and substituting the following:
The U.S. FDA has indicated that it will review these new studies and
make adjustments to its action level should this analysis indicate that
change is warranted.
-------
fiPR-12-1996 14:35 FROM OFFICE OF SEAFOOD . " TO 919155410840 P.04
Page 3 - Ms. Keating
This statement accurately reflects FDA's plans regarding its action level.
Vol 1, p3-23, par 1, 1. 4:
FDA's ADI and EPA's RiD differ by a factor of more than four. This is a
significant difference. In order to convey this difference we suggest the
following text:
"The existing advice and action level of the U.S. FDA falls within the
order of magnitude uncertainty surrounding the U.S. EPA RfD and the
assessment..........w
Without such modifying language, we believe that the existing sentence would be
inaccurate.
Voi 1, p3-23, par 3, 1. 6:
The hair mercury levels from WIIO reflect mean levels from around the world.
Since populations elsewhere around the world may consume more seafood than
in the U.S., the levels of methylmercury observed may be well be different from
the U.S. population. In fact, available information seems to confirm mis.
Therefore, we suggest the following changes:
"The world-wide mean of reported concentrations of...Wide variation
occurs about these values ....In contrast, available information for U.S.
populations suggest that mean hair levels are 1 ug/g or less.
Without such modifying text, the reader might well conclude that hair mercury
levels from elsewhere in the world reflect averages in the U.S.
Vol 1, P3-24, par 3, 1. 1:
The risks associated with methylmercury depend upon amount (e.g. more than
100 grams per day) as well as frequency of consumption. Therefore, we suggest
the following:
"WHO/IPCS recommends that as a preventive measure, in a
subpopulation that routinely consumes...."
Without such modifying language, the reader might incorrectly conclude that
-------
PPR-12-1996 14:36 FROM OFFICE OF -SEAFOOD TO 919195410840 =.05
Page 4 - Ms. Keating
consuming a single meal of more than 100 grams in a day is cause for concern.
Vol 1, p3-30, par 2, 1. 10:
Persons who might be more highly exposed not only consume large amounts of
fish "regularly," they do it frequently. In order to convey this other important
factor, we suggest the following text:
"are sufficiently high to have potential for increased risk if consumed
regularly and frequently."
Without this modifying language, a reader might incorrectly conclude that
consumption with any regularity (e.g., once a year or once a month) might result
in increased risk.
Voi 1, p3-30, 4th line from bottom:
The comments made above regarding frequency arc relevant here as well. We
suggest the following text:
"...are those who routinely and frequently eat freshwater fish...."
Vol l,p3-32, Table 3-5:
The mercury levels presented in this table are mean values. To avoid confusion,
we suggest changing the title and column 2 headings to read as follows:
"MEAN MERCURY CONCENTRATIONS
Vol, p3-38, 1. 3:
EPA has used two different data sets to describe the size of the population
consuming more than 100 grams of fish per day. Conclusions from both data
sets should be provided. We suggest the following text:
"89/91 are ill, 133, and 175 grams , respectively. In contrast,
data from the NPD survey indicates that less than 1% of respondents
,, ate more than 100 grams per day."
-------
fiPR-12-1996 14:36 FROM OFFICE OF SEfiFOOD TO 919195410840 P.06
Page 5 - Ms. Keating
If conclusions from the NPD survey are not discussed here, the reader might
conclude that the size of the population eating more than 100 grams per day is
larger than it probably is. We continue to believe that the use of food
consumption data from a 3-day survey can lead 10 erroneous conclusions about
seafood consumption. While 3-day survey data can be useful depending upon
the circumstances, it has significant limitations when applied to infrequently
consumed foods such as seafood. The longitudinal data from the NPD survey
provide a more realistic perspective on the fraction of the population who are
seafood consumers as well as the size of the population eating more than 100
grams of seafood per day.
Vol 1, p3-38, par 4:
The current draft could mislead a reader into concluding that a consumption rate
of 100 or more grams per day is a dividing line between safety and danger. As
we know, risk depends on how much methylmercury is in the fish, and how
much and how often it is consumed. For the most frequently consumed
cuumiercial species., consumers can regularly consume more than 100 grams of
fish per day without appreciable risk from metnylmercury. We therefore,
suggest the following text:
"In summary, the U.S. EPA analysis indicates that the U.S.
population is not expected to be at risk from consuming typical
commercial species or freshwater species from diverse sites, even if
they regularly consume on average 100 grams of fish per day. Those
persons who routinely eat more than 100 grams of fish per day with
greater than average levels of mercury (e.g.t large predator marine
species such as shark or swordfish or fish from contaminated sites)
may incur increased risk as reflected in FDA's advice concerning
consumption of these species."
Vol 1, p3-39, par 3, 1. 8:
We believe that the unpublished hair mercury data set supplied by FDA provides
valuable information on U.S. exposures. The reasonable inferences that can be
extracted from the data should be included in EPA's analysis. We suggest the
following text change beginning on line 8:
"one meal per week based on data shown in Table 3-3. The magnitude of
the increase in hair mercury concentrations shown in these unpublished
-------
OPR-12-1996 14:37 FROM OFFICE OF SEAFOOD TO 3191354106413
Page 6 - Ms. Keating
data for those women reporting seafood consumption suggests that most
of the women surveyed are typical seafood consumers, eating seafood
no more than once a week. Mean hair levels found in this and in the
study of Creason et al. suggest that the mean hair mercury levels in
the U.S. average around 1 ppm or less. Given that conclusions about
upper-percentile hair mercury levels cannot, be reached for those who
are most highly exposed, survey data from a representative sample of
the U.S. population should be developed to better characterize
exposures among high end consumers."
If reasonable inferences from this data set are not presented, readers might
incorrectly conclude that the extent of exposure in the general U.S. population is
essentially unknown.
Vol 1, p3-39: Summary of Risk Characterization
This section should address the entire US population, not just those eating more
than 100 grains of fish per day. Further, important qualifiers regarding high
methylmercury levels hi the fish consumed and the frequency of consumption
.should he integrated into the discussion so that readers get a better understanding
about what factors contribute to elevated or reduced hsk from memyknercury.
We suggest deletion of the present text and inclusion of the following:
The typical consumer eating fish in moderation from a variety of
sources and eating a variety of species is not believed to be at
increased risk. This consumer is not being advised to limit fish
consumption. Those who may be at increased risk are women of
child-bearing age and children who regularly consume over 100 grams
of fish per day if those fish contain elevated levels of mercury that
might be encountered at contaminated sites or in large marine
predator fish like shark and swordfish. Groups such as some Native
Americans or subsistence fishers may also be at risk because they do
consume fish in large quantities for cultural or economic reasons.
Additionally,, EPA has estimated that between 1 and 5% of the general
population consumes more than 100 grams of fish per day. The actual
risk to persons consuming more than 100 grams of fish per day will
depend, however, on the methylmercury concentration in the fish
being consumed. Individuals consuming over 100 grams of fish per
day should limit their consumption of shark and swordfish and pay
particular attention to mercury consumption advisories for freshwater
-------
fiPR-12-199S 14:38 FROM OFFICE OF SEflFOOD TO 9191*35410840 P.08
Page 7 - Ms. Keating
fish." ,
Without mention of the many factors that affect consumer risk, readers will
reach erroneous conclusions about eating fish that are in conflict with text
elsewhere in this report.
Volume VI.
Vol VI, Executive Summary, p xi, par 3., I. 7:
EPA has previously agreed to eliminate comparisons of exposure to the RfD in
both graphs and in text. We urge. EPA to remove" this material for essentially
the same reason that certain graphs have been removed: the public should not be
mislead to believe that exceeding the RfD by any amount poses an immediate
risk of adverse health effects, and that the RfD is a dividing line between safe
and unsafe. Beginning on line 7. we recommend deletion of all text starting at
"This section presents plots of...." continuing on line 9 through ".....the
reference, dose (RfD) and benchmark doses." Following the entire sentence
beginning on line 9 with "Literature " the rest of the paragraph should be
eliminated.
Vol VI, Executive Summary, p xiv, par 2, 1. 1O-12:
The entire text on these lines should be removed since it refers to discussion of
consumption and the RfD that would lead the reader to incorrectly conclude that
there is linkage between the RfD and risks of adverse health effects from
methylmercury and that the RfD is a dividing line between safe and unsafe
exposures.
Vol VI, Executive Summary, p xiv, last par:
The discussion here and associated tables (Tables 4-19, 4-20, and 4-21) and
discussion elsewhere provides direct comparison of the RfD with fish
consumption/exposure. Like the graphs that have already been removed, the
discussion and tables will lead readers to erroneous conclusions about the safety
of" consuming fish. Again, the public should not be mislead to believe that there
is a linkage between exceeding the RfD and risk of adverse health effects, and
that the RfD is a dividing line between safe and unsafe. These tables, charts,
and discussion should remove comparisons related to human fish consumption
-------
RPR-12-199S 14=38 FROM OFFICE OF SEflFOOD TO ' 919195410840 P. 09
Page 8 - Ms. Keating
and the RfD. As an additional matter, we are concerned that this material could
appear to be in conflict with EPA's acknowledgement elsewhere in the report of
the validity of FDA's risk management and consumer advice.
Vol VI, Executive Summary, p xv, fifth bullet:
This paragraph presents conclusions that some U.S. consumers are exposed to .
methylmercury at 10 times the RfD. Like the graphs that have been deleted,
without considerable caveats such text can mislead the reader to believe that
there is a linkage between the RfD and risk of adverse health effects and that the
RfD is the dividing Hue between safe and unsafe. We recommend that this
discussion be removed.
As previously pointed out, the use of food consumption data from a 3-day survey
can lead to erroneous conclusions about seafood consumption. While 3-day
survey data can be useful, it has significant Limitations when applied to seafood
consumption. As an example, EPA's discussion on women of child-bearing age
(see p4-40, par 3) concludes that the highest exposures among this group (over
ten times the RfD) occur among women who consume much higher quantities of
shark, swordfish, and barracuda. The use of 3-day consumption data to describe
long term consumption patterns is misleading for species such as shark and
swordfish that are consumed relatively infrequently. There are at least two
reasons for not drawing conclusions about exposures from these species using
this data base. First, the data for consumers of these species is very limited
(only seven women of child-bearing age reported eating any one of these
species). Second, since shark, swordfish, and barracuda are consumed relatively
infrequently, reliable conclusions about long term exposures cannot be derived
from three-day survey data for those individuals who happened to consume- these
fish during the survey period. Anyone who happens to be a consumer during the
three-day survey period would, given typical methylmercury levels for these
species (about 1 ppm) and typical serving sizes (about 160 g), automatically
exceed—for the 3-day period—an exposure of 1 ug/kg/day. This 3-day average is
not to be confused with exposures that are likely from typical patterns of
consumpDon.
Vol VI, p4-34, par 3,1. 9:
EPA's conclusion about the size of the population who are seafood consumers
needs to be related to the duration of the survey from which the conclusions are
drawn. We suggest the following:
-------
OPR-12-1996 14:39 FROM OFFICE OF SEAFOOD TO 919195412840 P.10
Page 9 - Ms. Keating
"..in the U.S. consumes fish once every three days."
As noted earlier, readers need to know that these conclusions are based on data
from a 3-day survey aud uxay not necessarily reflect long term consumption
patterns. Limitations of such data for seafood have been described elsewhere.
Vol VI, p4-40, par 2, last line:
We recommend deleting the entire phrase "present, and because these data were
not designed to be representative -of the United States population." This text is
inaccurate regarding the design of the NPD survey. It was intended to provide a
nationally representative and projectable sample of seafood consumers.
p4-40, par 3:
We recommend deleting the entire paragraph since it draws inaccurate
conclusions about long term consumption patterns based on 3-day consumption
data. Sec comments elsewhere regarding our concerns about the limitations of
such data for making projections about long term consumption patterns.
Vol VI, p4-44 & 41: Tables 4-19, 4-20, and 4-21:
We recommend deleting these tables. The information in these tables will lead
to misleading conclusions about the safety of consuming particular commercial
species that are in conflict with statements elsewhere in this report.
Additionally, the analysis and conclusions are based on data from a 3-day
consumption survey that has associated limitations as discussed elsewhere.
Vol VI, p5-10, Table 5-6:
This table should be to modified by deleting all entries for human fish
consumers. The entries provides a direct comparison of consumption and the
RfD that could lead readers to the erroneous conclusion that there is a linkage
between the RfD and the risk of adverse effects, and that the RfD is a dividing
line between safe and unsafe. Further, this table could lead readers to
conclusions that are in conflict with EPA's acknowledgement elsewhere in the
report of the validity of FDA's risk management and consumer advice and call
into question the safety of consuming typical commercial seafood species that
EPA has said elsewhere are safe to eat.
-------
fiPR-12-1996 14=39 ' FROM OFFICE OF SEAFOOD
TO
919195410840 P.11
Page 10 - Ms. Keatlag
Vol VI, p5-14, 6th par:
As noted in our comments on Vol 1, p3-23, distinctions need to be made
regarding hair mercury data from around the world and hair mercury dala for the
U.S. population. Therefore, we recommend the same additions noted earlier, as
follows:
"The world-wide mean of reported concentrations of...Wide variation
occurs about these values ... .In contrast, available information tor U.S.
populations suggest that mean hair levels are 1 ug/g or less.
We appreciate this opportunity to provide comments on this revision of the
Report to Congress.
•Sincerely yours
\
L. Rot
•^^^^^
Director, Office of
Policy, Planning
and Strategic Initiatives
Center for Food Safety
and Applied Nutrition
cc: TJ. Glauthier
-------
'SENT BY:OFFC SCI & TECH POLICY; 4-11-36 ; 5:38PM ;EIMVIRONMEVT DIVISION- 313 541 0840;# 2
EXECUTIVE OFFICE OF THE PRESIDENT
OFFICE OF SCIENCE AND TECHNOLOGY POLICY
WASHINGTON, D.C,
April 11. 1996
Memorandum to Karen Ixvy
From: Bob Watson, Rosina Biertiaiihi, and Ron Melnick, OSTP
Re: Mercury Report Comments
Enclosed is a detailed markup of the 50 page EPA summary, a draft 7-page executive
summary we compiled, and the sediment/bivalve maps generated by NOAA at our request
for coastal areas.
Please send copies of this information to Martha Keating and the other Federal agency
representatives who were on the Conference call for their review as well. (We do no( have
list of the participants.)
We are still waiting for FDA to supply information on their fish data bases so that we can
write a few paragraphs on needed research.
Thanks very much. Feel free to call if you need more information at 456-6202.
-------
SENT BY:OFFC sc, . TECH POUCY: 4-n-» : 3:3« :ENV,RO»,ENT D,V,S,ON- M3 stl mo:s 3
Preliminary Draft 4/11/96 for discussion purposes only L
EXECUTIVE SUMMARY
Ingestiou of high levels of mercury have been demonstrated to cause severe damage to die
nervous systems of both humans and wildlife. This report provides a provisional assessment
of the magnitude of anthropogenic mercury emissions by source, the health and
environmental implications of these emissions, and the availability and cost of control
technologies.
The most important conclusions of this assessment include:
* U.S. sources of metuylroccury (the most toxic form of mercury) are in general
decreasing or expected to decrease;
* the primary exposure pathway for both humans and wildlife is consumption of fish;
* the mcthyunercury levels in most freshwater and marine fish are only a few tenths of
a ppm or less, meaning that the general population is not exposed to dangerous levels
of mercury;
* vigilance should be maintained to limit exposure t fish contaminated with mercury
from a small number of sites.
* data from recent health studies in the Scychclle Islands and Faroe Islands will
significantly improve our quantitative understanding of the adverse health effects of
mercury on the fetus.
Background
Ingestion of mercury has been shown to cause damage to the nervous systems of both
humans and wildlife. This report provides a provisional assessment of the magnitude of
anthropogenic mercury emissions by source, the health and environmental implications of
those emissions, and the availability and cost of control technologies.
Mercury can cycle in the environment as a result of natural and human (anthropogenic)
activities. The amount of mercury mobilized and released into the biosphere has increased
since the beginning of the industrial age. Most of the mercury in the atmosphere is elemental
mercury vapor, which has a long atmospheric residence time, and hence can be widely
dispersed and transported thousands of miles from likely sources of emission. Most of the
mercury in water, soil, sediments, or biota is in the form of inorganic mercury salts and
organic forms of mercury (e.g., methyJmercury). The mercuric form of mercury when
either bound to airborne particles or in a gaseous form, is readily removed from the
atmosphere by precipitation and is also dry deposited. Wet deposition is the primary
mechanism for transporting mercury from the atmosphere to surface waters and land.
Mercury accumulates up aquatic food chains so that organisms in higher trophic levels
generally have higher mercury concentrations; of the total mercury body burden,
methylmercury generally constitutes the largest percent. Inorganic mercury, which is less
efficiently absorbed and more readily eliminated than methylmercury, docs not tend to
hioaccumulate. Nearly all of the mercury that accumulates in fish tissue is medrylatcd, and
-------
BY:OEFC SCI * TECH POLICY: 4-11-36 : 5:33PM :E\V I RasMEST DIV
319 541 U84U;s 4
Preliminary Draft 4 / 71 / 96 for discussion purposes only 2
fish consumption dominates the pathway for human exposure to this compound.
Consequently, concerns for potential human risk of mercury toxicity are greatest fur those
who consume species of fish that are contaminated with mercury. Advisories have been
issued in a majority of states, warning against consumption uf certain species of fish
contaminated with mercury. Mcthylmercury levels in most freshwater and marine fash are
only a few tenths of a ppm or less meaning that the general population is not exposed to
dangerous levels of mercury; however, vigilance should be maintained to limit exposure to
fish contaminated with mercury from a small number of lakes and rivers.
The FDA has authority for establishing administrative guidelines for mercury levels in food.
A daily tolerable intake (TDI) for methylmercury of 0.47 /zg/kg of body weight was
established, based largely on an epidemic of methylmcrcury poisoning in Japan. The TDI is \
the amount of mcthylmercury that can be consumed daily over a long period of time with a
reasonable certainty of no harm. The action level by the FDA for methylmercury in fish in
commerce is 1 ppm. Observations from a poisoning event in Iraq, led FDA to recognize that
the fetus may be more sensitive than adults to the effects of mercury. Hence, FDA advises
pregnant women to limit their consumption of fish known to have high concentrations of
mercury.
New data are emerging from large studies in the Seychelles and Faroes Islands that were
designed to evaluate childhood development and ncurotoxicity in relation to fetal exposure to
methylmercury in fish-consuming populations. Because nf various limitations and
uncertainties in the Iraqi data set, an interagcncy and external review of all the human data
on methylmercury, including tbe more comprehensive studies from the Seychelles and Faroes
Islands, will be undertaken to provide a consensus estimate of exposure to methylmercury
that may be used by EPA and FDA in future regulatory decisions.
Elevations in concentrations of methylmercury in fish may be due in part to anthropogenic
releases'of mercury from industrial and combustion sources; however, it is worth noting that
there has been a dramatic drop in mercury emissions from manufacturing in the US over the
past decade. Industrial demand for mercury has declined by about two-thirds between 1988
and 1993, due largely to tbe elimination of mercury additives in paints and pesticides and the
reduction of mercury in batteries. Mercury emissions from municipal waste combustors and
medical waste incinerators, the largest identifiable sources of mercury emissions, are
expected to decline significantly (>50%) by the year 2000 due to regulatory action by EPA
under the statutory authority of section 129 of the Clean Air Act. After the municipal solid
waste and medical waste incinerators, the largest remaining identified source of mercury
emissions are coal-fired utility boilers. The next challenge will be identifying cost-effective
means of limiting emissions from these sources. The Clean Air Act Amendments of 1990
required the USEPA to transmit to Congress a study of mercury emissions from electric
utility steam generating units, municipal waste combustion units, and other sources, and to
address the rate, and mass of such emissions, the health and environmental effects of such
emissions, technologies available to control such emissions, and the costs of such
technologies.
-------
SENT BY:OFFC SCI i TECH POLICY; 4-11-36 ; 5 = 39PM '.ENVIRONMENT DIVISION- 319 541 US40;; 5
Preliminary Draft 4/11/96 for discussion purposes only 3
The major findings of this report include the following:
Anthropogenic Mercury Emissions and Deposition
• Annual anthropogenic US emissions of mercury in 1990 were about 220 Mg (242.5
tons), compared to global anthropogenic emissions of about 4000 Mg. Thus, US
anthropogenic emissions are about 5% of the global emissions.
• The three major anthropogenic sources of mercury emissions are from municipal
waste combustors (22.7%), medical waste incinerators (26.7%), and utility
combustors (21.2%). In addition, there are a variety of minor manufacturing sources
including chioro-alkali (2.7%), Portland cement (2.7%), primary lead (3.7%) and
secondary mercury production (3.1%).
• Future emissions from utility combustors, primarily coal combustion, depends upon
the nation's energy needs and chosen fuel.
• The chemical form of emitted mercury is a critical factor in its fate, transport and
toxicity in the environment. Mercury is emitted in three forms (elemental mercury:
(Hg": 41%); divalent vapor-phase mercury: Hg2+: (41%); and particulate-bound
mercury: (Kg,,: 18%).
• Long-range transport models suggest that only 35% (77.9 Mg) of U.S. anthropogenic
emisions are deposited, through wet and dry deposition, within the lower 48 States
(2% of Hg°; 68% of Hg2*; and 36% ot"Hgp). In addition, another 33 Mg of mercury
from natural sources is deposited.
• Model results suggest that while deposition within 10km of a facility is normally
dominated by emissions from the local source rather than from emissions from
regional sources, in most instances at least 75% of the mercury emitted from each
facility is transported more than 50km.
• The highest deposition rates for mercury occur in the Eastern part of the U.S., and in
particular the Northeast.
Assessment of Exposure from Anthropogenic Mercury Emissions
• Although, there are no data that conclusively demonstrate a relationship between
anthropogenic sources and increased mercury concentrations in environmental media
or biota, emissions of mercury from anthropogenic combustion and industrial s sources
can plausibly be linked with incremental increases in methymcrcury levels in
freshwater fish.
-------
SENT BY:OFFC SCI & TECH POLICY: 4-11-36 ; 5:40PM ;ENV[RONMENT DWSIO.V
319 341 0840;;= 6
Preliminary Draft 4/11/96 for discussion purposes nnlv 4
• The major exposure route for both humans and wildlife is consumption of fish. The
magnitude of the risk to human health from eating fish is dependent upon the species
offish consumed, the concentration of methylmercury in the fish, the quantity of
fish consumed, and the frequency of consumption.
• Mercury concentrations in the top ten types of marine fish and shellfish that make up
80% of the seafood consumed within the US range from 0.01 to 0.21 ppm. Two
types of marine (Ish, shark and swordfish, have mercury levels that approach 1 ppm.
The national average for mercury levels in freshwater fish has been estimated, based
on two different surveys, to be 0.11 or 0.26 ppm.
• While mercury residues in fish have been found in 92% of the 370 surface bodies
tested in the U.S., mercury levels above 1 ppm were only found in 2% of the sites
tested, and above 0.5 ppm at ]5% of the sites.
Health Effects of Methylmercury
• Epidemics of'methylmercury poisoning in Japan and Iraq demonstrate that
neurotoxicity is the effect of greatest concern when exposure occurs to the developing
embryo/fetus during pregnancy and when adults and children are exposed to
methylmercury from consumption of contaminated fish (Japan) or bread (Iraq).
• Dietary methylmercury is almost completely absorbed into the blood and distributed to
all tissues including the brain; it also readily passes through the placenta to the fetus
and fetal brain.
• Measurements of scalp hair concentrations of mercury can serve as an excellent stable
biomarker of previous exposures to methylmercury.. The incorporation of methyl-
mercury into scalp hair is proportional to the mercury concentration in blood; the
hair-to-blood ratio La humans was estimated to be 250:1 (expressed as //g Hg/g hair to
mg Hg/L blood).
• Based on data of the combined incidence of neurological effects recorded for Iraqi
children exposed to methylmercury in uiero, a benchmark dose of 1.1 fig Hg/kg body
weight/day (11 ppm hair concentration of mercury) and a reference dose (RfD) of 0.1
/tg Hg/kg body weight/day were estimated.
• The benchmark dose for methylmercury. which was determined by modeling
neurological effects in 81 children in relation to their mothers' hair concentrations of
mercury during the period of gestation when exposure occurred, represents the lower
95% confidence limit on the dose corresponding to a 10% level of risk above
background. A dose conversion equation with reported parameter values for
absorption and elimination of methyknercury was used to estimate a daily iutate of
-------
SEYT 8Y:OFFC SCI 4 TECH POLICY: 4-11-36 : 3.40PM ENVIRONMENT DIVISION- 313 541 0840;,
Preliminary Draft 4/11/96 for discussion purposes only 5
t methylmercury that would maintain a blood concentration or hair concentration of
mercury at that level of risk. The RfD is the benchmark dose divided by 10 to
account for variability in how individuals process ingested methylmercury.
• The RfD is a daily ingcstion level of methylmercury that is anticipated to be without
adverse health effect to persons, including sensitive subpopulations, over a lifetime nt
exposure. Critical elements in estimating methylmercury exposure and risk from fish
consumption include the species of fish consumed, the concentration of
raethylmercury in the fish, and the quantity of fish consumed.
• Based on data from national surveys of fisli consumption and databases of
methylmercury in marine fish and freshwater fish from across the US, the mean
mercury concentration in fish when weighted by the frequency of consumption is
approximately 0.134 jig/g wet weight (0.134 ppm). An estimated 1 to 5 percent of
women of child-bearing age in the US consume 100 grams or greater of fish and
shellfish per day. Thus, a 60 kg individual who eats 100 grams (approximately 3.5
ounces) of fish each day at this concentration of mercury (0.134ppm), would have a
daily mercury exposure of 0.22 ng/kg body weight. Consumption of greater
quantities of fish at this mercury concentration or of more highly contaminated fish
would lead to higher daily doses, while consumption of less contaminated fish would
result in lower daily doses of mercury.
• Humans most at risk arc those who cat large quantities of freshwater fish from a
single location where the fish arc known to contain high concentrations of mercury.
Hence, consumers arc urged to heed the advice of state and local health departments
concerning local conditions.
• Because of limitations in the Iraqi study; it is expected that evaluations of all of the
human data on methylmercury, including data from the more comprehensive studies
of the Seychelles and Faroes Islands, will lead to a more reliable estimate of human
risk from consumption of low levels of methylmercury in fish.
Ecological Assessment of Anthropogenic Mercury Emissions
• Plants, animals and humans can he exposed to mercury by direct contact with
contaminated environmental media of ingestion of mercury-contaminated water and
food. Mercury exposure and contamination is of particular concern for animals at the
highest trophic levels in aquatic food webs and for animals and humans that feed on
these organisms.
• Effects of mercury on fish include death, reduced reproductive success, unpaired
growth and development and behavioral abnormalities. Exposure to mercury can also
cause adverse effects in plants, birds, and mammals. Reproductive effects are the
-------
SENT BY'OFFC SC! & TECH POLICY; 4-11-36 : 5 = 40PM ENVIRONMENT DIVISION- 913 541 0840;;; 8
Preliminary Draft 4/11/96 for discussion purpose": only 6
primary concern for avian mercury poisoning and can occur at dietary concentrations
well below those which cause overt toxicity. Sublethal effects of mercury on birds
include liver damage, kidney damage, and neurobeaaviorai effects. Effects of
mercury on plants include death and sublethal effects.
• Massive poisonings or birds and wildlife from mcthylmercury treated seed grains were
identified during the decades preceding the 1970s. In Minamata, Japan from about
1950-1952 (prior to recognition of human poisonings) severe difficulties with flying
and other grossly abnormal behavior was observed among birds. Signs of
neurological disease including convulsions, fits, and highly erratic movements were
observed among domestic animals, especially cats that consumed seafood.
• Laboratory studies under controlled conditions can be used to assess the effects of
methylmercury from fish on mink, otter, and several avian species. Effects can occur
at a dose of 0.25ug/g bw/day or 1.1 ug/g methylmercury in the diet. Death may
occur in species at 0.1-0.5 ug/g bw/day or 1.0-5.0 ug/g in the diet. Smaller animals
(for example, minks, monkeys) are generally more susceptible to mercury poisoning
than are larger animals (for example, mule deer, harp seals). Smaller mammals eat
more per unit body weight than larger mammals. Thus, smaller mammals may be
exposed to larger amounts of methylmercury on a body weight basis.
• Whole body residues of mercury in acutely poisoned birds usually exceed 20ug/g
fresh weight. Although sublethal effects include a number of different organ systems.
reproductive effects are the primary concern. These occur at concentrations far lower
than those that cause overt tnxicity.
• Mercury contamination has been documented in endangered species such as the
Florida panther and the wood stork, as well as populations of loons, eagles, and
furbearers such as mink and otters. These specks are at high risk of mercury
exposure and effects because they either are piscivores or eat piscivores. :
• The broad ecosystem effects of methylmercury arc not completely understood. No
applicable studies on the effects of mercury on intact ecosystems were found.
Consequently, characterization of risk for non-human species did not attempt to
quantify effects of mercury on ecosystems, communities, or species diversity. The
characterization focused on quantities of mercury that adversely affects the health of
sensitive subpopulations of wildlife species and on the collocation of these populations
with areas of elevated mercury secondary to ambient, anthropogenic emissions of
mcthylmercury.
• Wildlife criteria (WC), i.e., the mercury level in water which is expected to be
without harm for the species, were calculated for piscivorous birds and mammals:
The criterion value protective of piscivorous avian wildlife is 405 pg/L.
-------
SENT BY'OFFC SCI & TECH POLICY: 4-11-36 ; 5:41PM :ENVIROMMENT DIVISION- 913 541 0840;# 3
Preliminary Draft 4/11/96 far discussion purposes only 1
The criterion value protective of piscivorous mammalian wildlife is 346 pg/L.
Mercury Control Technologies and Costs
• Mass balance analyses indicate that a significant, amount of mercury is embedded in
products which eventually make their way into municipal and medical waste
incinerators. Opportunities tu deal with Hg during the product life-cycle, rather than
just at the point of disposal, need to be pursued. A balanced strategy which integrates
end-of-pipe control technologies with material substitution and separation, design-tbr-
environment, and fundamental process change approaches is needed.
• Though mercury control technologies exist, many are still in the prototype stages
making it difficult to predict final cost-effectiveness as well as the time required to
scale-up and commercialize the technologies. Because the chemical species of
mercury emitted from boilers varies from plant to plant, there is no single control
technology that removes all forms of mercury.
• There remain a wide variation in the end costs of control measures for utilities and
the possible impact of such costs on utilities.
• An R&D and demonstration strategy is needed which addresses the performance,
time, and cost uncertainties of many of the existing approaches to mercury control
and the trade-offs, opportunity costs, and interdependencies of these various
approaches.
-------
SENT BY:OFFC SC! & TECH POLICY: 4-11-36 ; 5:41PM ;ENVIRONMENT DIVISION- 313 541 0840:?11
-------
SENT BY:OFFC SCI & TECH POLICY: 4-11-36 ; 5:41PM 'ENVIRONMENT DIVIS10.V
319 541 U840;?12
„_/
no -*• —*• o o o
a -ex —k 03 o> o
a a ^ o o ^
p p p p p p
co Ka ^ ^ o o
O -Ck —*• OD C35
CO
o
Z
W
B"
H
q^
<
<•
(D
CO
pT
r-*1
CO
-------
30
O
r-H
*v
in
(TJ
•••.
o
s
oJ
CD
en
u
§
U
National Status and Trends Program
Hg
84-88 NS&T Fish Means
0.00 to 0.12
0.12 to 0.21
0.21 to 0.32
0.32 to 0.46
0.46 lo O.B6
0.86 lo 43.19
>•
ca
-------
SENT BY'OFFC SCI $ TECH POLICY; 4-11-36 ; 5:42PM :LNVIRO.NMEVT DIVISION
319 541 0840:^14
Naticinul Status and frcnos Proqram
a
a
a
•
a
•
86-UO
a. 03
0.12
0.20
0.29
0.38
0.47
Ig (ppm)
NS&T Hivaiwn Means
to
to
to
to
to
to
0
0
0
0
0
0
12
.20
.29 '
.28
.47
.56
-------
SENT BY'OFFC SCI & TECH POLICY; 4-11-36 ; 5:42PM ;L\VIROiVjEM DIVISION-
541 ua40;=15
National Status .=ind Tiv.nds Prcgrurr;
Hg (ppm)
84-91 Combined Sediment Means
l 0-00 to 0.77
0.77 10 1-54
r.54 to 2.31
2.31 to 3.08
3.08 to 3.85
3.85 to 4.62
r*- -*«"..
-------
SENT BY:OFFC SCI 4 TECH POLICY; 4-11-96 : 5-43PM ;ENVIRONMENT DIVISIO.V
313 541 Ub4U.-lb
National Status and Tiendr, Progiam
Hg (ppm)
86-80 NS&T Bivalve Means
D 0.08 to 0.1?
• 0.12 to 0.17
M 0.17 lu 0.21
• 0.21 to 0.26
D 0.26 to 0.31
8 0.31 to 0.3S
-------
SENT BY:OFFC SCI & TECH POLICY; 4-11-36 ; 5:43PM :E.NVIRO.NME.NT DIVISION-
313 541 U840:=17
National Status and Trends Program
Hg (ppm)
84-91 Combined Sediment Means
D 0.00 to 0.31
• 0.31 to 0.61
B 0.61 to 0.92
• 0.92 to 1.23
D 1.23 to 1.54
• 1.54 to 1.84
-------
SENT BY:OFFC SC! i TECH POLICY; 4-11-96 ; 5:44PM ENVIRONMENT DIVISION-
313 541 0840:^18
National Siaius and Trends Program
Hg (ppmj
86-90 NS&T Bivalve Moans
0.05 to 0.07
0.07 lo O.OG
a.oa io 0.10
0.10 Io 0.12
0.12 to 0.13
0.13 to 0.15
-------
SENT BY:OFFC SCI & TECH POLICY: 4-11-36 ; 5:44PM ;ENVIRO,\ME.NT DIVIS10.V
319 541 0840;?13
National Status and Trends Program
Hg (ppm)
84-91 Combined Sediment Means
S 0.00 to 0.08
• 0.08 to 0.16
M 0.16 to 0.23
3 0.23 to 0.31
D 0.31 (o 0,39
• 0.39 to 0.47
-
2?»T:fcJ«c;>*— ^—* «*•
5yi--vffijrif5fc L»r->.ryT t^\n\
^gfejlet^^^.^'-v
-------
O
OJ
II
oo
in
r
s^
O
C/l
lElioiuil Sta'.us and Irenes Program
Hg (ppm)
35-93 NS&T Bivalve Means
D 0 34 to 0 17
H 0 17 lo 0 30
S G 30 lo 0 44
8 0.44 lo 0.57
D 057 lo 0.71
H 0.71 (o 0.64
O
tti
s
0-
10
CO
en
O
s
U
V)
U
CO
-------
SENT BY:OFFC SCI & TECH POLICY: 4-11-36 : 5:46PM ;ENVIRONMENT DIVISION-
541 0840;
-------
SENT BY'OFFC SCI & TECH POLICY; 4-11-36 ; 5:46PM ;ENVIRONME.NT D1VISIO.V
919 541 Ob40;;?22
Nalionnl bttiUis and Iro-ids Pioqram
Hg (ppmj
86-90 NS&T Bivalve Weans
D 0.00 10 0.10
M 0 10 to 0.19
3 0.19 to 0.28
HI 0.28 lo 0.3S
D 0.38 to 0.4/
• 0.47 to O.G7 |
-------
SENT BY'OFFC SCl 4 TECH POLICY; 4-11-36 ; 5:47PM ; ENVI RO.NMENT DIVISION-
319 541 (J840;=23
H : \^J- ' •-i.'-vT'.. * •«.•*$?• ;r • •';•/-•-•..
-^VVj.-.-. - ;. •-.''ci.iSi^5"'>''V'^4_>'^1*s- -• T-'iT'^J',-.- I
;V*WM~-" • : . • -/ •s*j5^-Aii»&i5fe?3:1>--«.'-' •••-•rt-'-s,- - i
c si,^" T.'/TBi-v< •-• T i-T3a%v^"-'.?l/l':'J--'*r^-1'. •••'.-ijvj* - - • I
^^5^p^'^-:«-^2^i^v*-s-l^ w9*° n^aSpfe? i
W'v ^-3^^^§^::fefrt':':'\
•J-al.— • ..." ->.-«^aXS^>rrii;,,--. •' •-->JJ. •>,. .. X^
.me I rcna."
H«,*»-
-------
SENT BY'OFFC SCI & TECH POLICY: 4-11-36 ; 5=48PM :ENVIRONMENT DIVISION-
United States
Environmental Protection
Aaencv
319 541 0840:^24
E>A-452;'R-96-001d
cury Study
Report to Congress
Volume I:
Executive Summary
t« /ini* Ira n«*.tUI>J
Office of Air Quality Planning & Standards
and
Office of Research and Development
efi3035-1.ecir
-------
SENT BY'OFFC SCI & TECH POLICY: 4-11-36 : 5:48PM 'ENVIRONMENT DIVISION- 919 541 0840:^25
MERCURY STUDY REPORT TO CONGRESS
VOLUME I: .
EXECUTIVE SUMMARY
March 1996
Office of Air Quality Planning and Standards
and
Office of Research and Development
U.S. Environmental Protection Agency
-------
SENT BY'OFFC SCI & TECH POLICY; 4-11-96 : 5 = 49PM . EWI ROiNMEVT DIVISION- 919 541 Ub4U;?2S
• Volume 1: Executive Summary
• Volume II: Inventory nt" Anthropogenic Mercury Emissions in the United States
• Volume III: An Assessment of Exposure from Anthropogenic Mercury Emissions in
the United States
• Volume IV: Health Effects of Mercury and Mercury Compound
• Volume V: An Ecological Assessment of Anthropogenic Mercury Emissions in fhc
United States
• Volume VI: Characterization of Human Health and Wildlife Risks from
Anthropogenic Mercury Emissions in the United States
• Volume VTT: An Evaluation of Mercury Control Technologies and Costs
The various analyses documented in this Report were designed and conducted in accordance
with accepted guidelines and procedures. For example, ihe human health risk assessment perfornjed
for this Report follows published. Guidelines for Risk Assessment (including guidelines on Exposure
Assessment, Developmental Toxicity, Carcinogenicity and Germ Cell Mutagenidry) and uses
estahlished methodologies for quantitative assessment of general systemic toxicity (e.g.. in the
calculation of reference doses (Rl'Ds) and reference concentrations (RfCs)). Moreover, the assessment
of ecological effects, presented in Volume V. follow!, U.S. EPA's Framework for Ecological Rink
Assessment. Criteria values for protection of piscivorous wildlife were developed using the
meuhodology developed for the Great Lakes Water Quality Initiative.
In 1994, the National Research Council of the National Academy of Sciences, in Science and
Judgment in Risk Assessment, recommended several areas in which U.S. EPA could improve its risk
assessment and risk characterization practices. These recommendations are listed below along with a
description of how they were implemented in this Report.
• Provide an understanding of the type and magnitude of an adverse effect That a
specific chemical or emission could cause under particular circumstances. The Report
characterizes both the type and magnitude of health and ecological effects associated
with airborne emissions of mercury from anthropogenic sources.
• Validate methods and models. All models used for die Report were critiqued hy
scientific experts and model predictions were compared 10 measured mercury levels
using the most appropriate data available.
• Describe the basis for default options. All assumptions arc described and justified
based on available data. Where appropriate, exposure models were modified to
improve assumptions and to focus on areas of prediction where use of model
assumption* is most justified.
• Articulate and prioritize data needs. The Report includes a section on Research
Needs in each volume.
• Distinguish between variability and uncertainty. The Report provides discussions that
attempt to make these distinctions for the risk results.
• Perform formal uncertainty analyses. Uncertainty analyses were formally conducted
for the dose-response and exposure assessment steps of the study, and were implicit in
weight-of-evidcnce processes used in the hazard identification step of the human health
risk assessment and the problem formulation phase of the ecological risk assessment.
1-2
-------
SENT SY'OFFC SC! & TECH POLICY: 4-11-96 : 5 = 50PM :ENVIRONMENT DIVISION- 919 541 U840;z30
was^atfalyzejlx^antiuuivpi^ in ot£y«- companeiits of t^t^study, such is
e qzfculatinoxft bioac*rfjniulatiotv£Cctors jirta ihe jiflj for meih^fTinercury
Q JU. : ^ /
t-3
-------
SENT BY:OFFC SCI 4 TECH POLICY; 4-11-36 ; 5:30PM :E.NVIRO.NME.M DIVISIO.V
313 541 U840;?31
r.oi foDdr.iJU^iarion.exp€^T&-io>",^^rneritaJ Truaemy- -is Tafgdv
y
~--
tesiiaataipiaaffiaaiifflti
^•^^'^y^\'^~^--^-ii:~-:'i^.Vi^'f--~^^
lilisT&ffi^^
'oKT'-'-- ^*.v^^»'*«HfiC'S;^C'K>^i^s!:^;S'*^«^C;-%:k^^ .-^:^-^:,^^.^r*•^*t?^-:¥-'^:?*5'**^:?*-* ><•!>>• j???t"^;-'.'5v!'.!iffTj*:Vr*•!•*'^ " i-. ""sli'^*--^t-'Jti;i:;'^-j'''" '""
Current levels of mercury in freshwater fish in the United States are such that advisories have
been issued in 35 states that warn against the consumption of certain amounts and species of fish that
are contaminated with mercury. Six states (based on 1994 data) have statewide advisories (i.e.,
advisories posted on every freshwater body in that state).
sms^3^»ssss^aS^^tsiK^ssss^^&rf.f^^-^is;-"'
In some areas, freshwater fish can have mercury levels which exceed the U.S.
FDA action limit of 1 ppm. The concentration ofmethylmercury in commercially important marine
species is. on the average, lower than thgs^^^^^| level.
2-6
-------
SENT BY:OFFC SCI & TECH POLICY; 4-11-36 ; 5:51PM : ILNVIRUNMEM DIVIS1U.V
313 541 UH4U;=32
^ improved
j|ii|u
S:^;4,.Ji>..U^-£iSr^L'^^_:^^^ ....... •
tfi9ttgi!i^ii&
' J[<: 'J^j't..»J.H- \*-'»'''J)>'*J'f »>•>*• •*< •*>• •-:>*•*• ..... ' "" • '..-f-ji-ii-L-in,.- ;• »»\V*'"J',A''f T "1 '•**!•* 8'*** 5 Vrt '•''•• ••>'->J'1 »*•*'**••*<* >*>^*. .^ JJ^.i^ 'O ' ' '"• " •^••V ^ . '.,..kfc.... <.v^-'" • • <>• >** ' ^ ^^ . *, *C^ : T "**
..-f-ji-ii-L-in,.- ;• »»\, •• * Vrt ••• ••>'->'1 ^ ^ .^ O ' . .,..kfc.... .-" . , *
>;^ff-^r;y^tf
-
i ..
pk|^^^fflp|i^^§ii^p||^l^^|i^
^^:>••v'JssriTfi"'its*-'^'^1'*' *-"^'iii':*?*vS* .^trr;:"i.;-?*M.i'M'S<''"i',i'w!^»^d^S3^5S2SSSU><*>.>.: >J
'!T?S7r-aMiYritrt-"'-'f''-''-'*l^^l^-*"u
-------
SENT BY'.OFFC SCI & TECH POLICY; 4-11-36 ; 6:USPM :ENVIRONMENT DiVISlOV
919 541 U840;- 2
In 1965. ar. additional methylmercury poisonmc outbreak occurred in the nrea nf Niigata.
Japan. As in Minamata. multiple chemical plant sources QT the chemical were ccnsideri-d. Scientific
detective work identified the source again to be a chemical faciory releasing raethylinercury into [he
Agano River. The signs and symptoms of disease in Niigata were those of methylmercury poisoning
and strongly similar to the disease in Minamata.
The abnormalities (or pathology.) in the human brain thai result from mcthylmercury
poisoning are well described. There is an extremely high level of scientific certainty that
methylmercury causes these changes. Similar pathology has been identified in nther countries where
mediylmercury poisonings have occurred. Methylmercury contamination of other food products
(including grains and pork products) has resulted in severe methylmercury poisoning with pathological
changes in the nervous system and clinical disease virtually identical to Minamata disease.
Methylmercury poisoning occurred in Iraq following consumption of seed grain iliat had been
treated with a fungicide containing methylmercury. The first ouibrealc occurred prior to i960 and
resulted in severe human poisonings. The second outbreak of methylniercury poisoning from gram
consumption occurred in the early 1970s. Imported mercury-treated seed grains arrived after the
planting season and were subsequently used as grain to make into flour that was baked into bread.
Unlike the long-term exposures in Japan, the epidemic of methylmercury poisoning in Iraq was short
in duration, but the magnitude of the exposure was high. Because many of the people exposed to
methylmercury in mis way lived in small villages in very rural areas (and some were nomads), ihe
number of people exposed to these mercury-contaminated seed grains is not known. The number of
people admitted to the hospital with symptoms of poisoning has been estimated to be approximately
6,500, with 459 fatalities reported.
As in the Japanese poisoning epidemics, the signs and symptoms of disease were
predominantly those of the nervous system: difficulty with peripheral vision or blindness, sensory
disturbances, incoordinarion, impairment of walking, slurred speech and in some coses, death.
Children were affected, as well as adults. Of great concern was the observation that infants, bom of
mothers who had consumed the mcthylmcrcury-contaminated grain (particularly during the second
trimester of pregnancy) could show nervous system damage even chough the mother was only slightly
affected herself.
In both the Iraqi and Japanese epidemics, the levels of methyimercury consumed were much.
levels currently reorted in the U.S. food supply.
Health endpoims other than neurotoxicity were evaluated by U.S. EPA using established risk
assessment Guidelines.
Methylmercury has been shown to cause tumors in mice at high doses that produce severe non-cancer
toxicity. I,ow-dose exposures to methylmercury are not likely to cause cancer in humans. Data oa
effects related to mutation formation (changes in DNA) indicate that methylmercury could increase
frequencies of mutation in human eggs and sperm. These data were noi sufficient, however, to
permit estimation of the amount of methylmercury that would cause a measurable mutagenic effect in
a human population.
3-19
-------
iCl & TECH POLICY; 4-11-36 ; tr.uoPtt :L\VIRONME.NT DIVISION-
313 541 0840;* 3
What Is A Reference Dose?
A reference dose or RfD/s definec m the foilowins way by U.S. EPA: oil estimate twuh
unccnainty spanning perhaps anf order of magnitude) at u daily exposure co the human population
(including sensitive subgroups! that is likely to be without aii appreciable risk of deleterious effects during
a lifetime. RfDs are reviewed by a standing workgroup nf Agency scientists tor accuracy, appropriate
use uf risk assessment methodology,.appropriate use of data and other .scientific jssucs. When consensus
has been reached by the workgroup, information on the RfD is made available to the public through a
U.S. EPA database: namely, the Integrated Risk Information System (IRIS).
The RfD is based on die best available data that indicate a so-called critical effect: this is
generally the first indicator or most subtle indicator of an adverse effect in the species under study. In
calculating RfDs U.S. EPA generally uses a no observed adverse effect level (NOAEL). This ts found
from either inspection of or modeling of do.se-response data on the critical effect. U is a means ol
estimating the threshold for effect in the reported study. The NOAEL is most useful when ii is from a
study in which a determination of the lowest observed adverse effect level (LOAEL) can also be done.
The LOAEL is the lowest tested dose at which the critical effect was seen in the species under study.
In calculating the RfD the U.S. EPA divides the NOAEL or LOACL by a series of uncertainty
and modifying factors. The uncertainty factors (which may be as much as 10 each) are for the following
areas: extrapolation of data to sensitive human suupupulations; extrapolation from animal data to
conclusions tor humans; lack of chronic exposure data: lack uf certain other critical data; and use of a
LOAEL in the absence of a NOAEL.
The RfD is used for risk assessment judgments dealing with evaluations of general systemic
toxicity. It is intended to protect sensitive (but not hypersensitive) members of the human population; the
rationale is dtat if exposure to the RfD is likely to be without appreciable risk for sensitive members of
Che population, chen it is without appreciable risk for all members of the population. The RfD is
generally applicable to men and women and to adults, to children and to the aged, n"|M« data support ihe
calculation of separate RfDs for these groups.
The RID is a quantitative risk estimate. Risk management decisions may consider the RfD but
will take into account other risk estimates and other non-risk factors as well.
i*fl~T
How Much Methylmercury is Harmful to Humans?
Information on the amount of mcthylxnercury exposure producing particular combinations of
signs and symptoms in people has been analyzed to yield whac are called quantitative dose-response
assessments. Both the Japanese and Iraqi epidemics are important to understanding how
methylmcrcury from food produces neurological disease in humans. In the epidemics in Minamata
and Niigata, the exposures were long-term, and the tissues of fish and shellfish were the sources of
methylmcrcury exposure. This establishes with highest scientific confidence that methylmcrcury in
fish can produce human disease. A limitation to these data is that many patients were severely
affected. The extent of methylmercury poisoning was so severe that finding subtle indications of
disease is difficult. Subtle indicators of poisoning are important in identifying levels of exposure
which will not cause any adverse effects. The U.S. EPA calculates one such measure, called a
reference dose or RfD (see the box above).
3-20
-------
SENT BT-OFFC SCI
s
6:06PM ; ENV1 ROMvlENT_ DIV1S10V
313 541 0840;? 4
w^"?'-«»^lv^1'BtT^"'i^'V '"*V^^- ': ST".'-'-" • ••••-..' ^ft^'"' .*•••"••»-?- r
^
'!'• \
rsr
SI Iraqi children who had been exposed in utcro: their mothers had eaten ___
contaminated bread during pregnancy. The data were collected by interviewing the mothers of the
children and by clinical examination by pcdiatric neurologists approximately 30 months after the
poisoning episode. The incidence of several endpoints (including late walking, late talking, seizures 4 ^ }
or delayed mental development, and scores on clinical tests of nervous system function) were niTj^jK-.
mathematically modeled to determine a mercury levei in hair (measured in ail the mothers in die ^* ^.
study) which was associated with no adverse effects. These effects were delays in motor and
language development defined by the following:
• Inability to walk two steps without support by 2 years of age;
• inability to respond to simple verbal communication by age 2 yeans among children
with good hearing;
• scares on physical examination by a neurologist that assessed cranial nerve signs.
speech, involuntary movements, limb tone, strength, deep tendon reflexes, plantar
responses, coordination, dexterity, primitive reflexes, sensation, posture, and ability to
sit, stand, walk, and run; and
• assessment of mental development or the presence of seizures based on interviews
with the child's mother.
In calculating die mercury levei in hair which was associated with no adverse effects, the U.S.
EPA chose a benchmark dose (in this instance the lower bound fur 10 percent risk of neurological
changes) based on modeling of all effects in children. This lower bound was 11 ppm hair
concentration for methylmercury. A dose-conversion equation was used to estimate a daily intake of
1.1 /zg methylmercury/kg body weight/day that when ingested by a 60 kg individual will maintain a
blood concentration of approximately 44 ^g/1, of blood or a hair concentration of 11 jig
mercury/gram hair.
Dua on the behavior of mercury in the human body were used to estimate the amount of
mercury ingested per day at this no adverse effect level. Due to variability in die way individuals
process methylmercury in the body and the lack of data on observed adult male and female
reproductive effects, an uncertainty factor of 10 was used to derive the RfD from the benchmark
3-21
-------
BY'Ot-FC SCI & TECH POLICY; 4-ll-9b : b:u7PM ;E.NVIROiYME\r DIVISION 313 541 US40;* 5
dose. The RfD for methyimercury was determined ro he IxiO J mg/kg-duy. thai is a person could
consume 0.1 ^g methylmcrcury for every kg of his/her body weight even,' day widnout .anticipation of
fl^f^5^wse;:e£feci. The RfD is a daily ingesiion levei anticipated to be wuhouc adverse effect to
persons, including sensitive subpopulations. over a lifetime. Judgments as to a "sate" dose and
exposure represent decisions that involve risk management components. The RfD majf;J3e> considered
the midpoint in an estimated range of about an order of magnitude. This range reflects variability and
uncertainly in the estimate.
The range of uncertainty in the RfD and the factors contributing to this range were evaluated
in qualitative and quantitative uncertainty analyses. The uncertainty analyses indicated that paresthesia
(numbness or tingling) in the hands and feet, and occasionally around the mouth, in adults is not the
most reliable endpoint for dose-response assessment because ic is subject (o the patient's recognition
of the effect. Paresthesia in adults is no longer the basis for U.S. EPA's methyimercury RfD. There
are, however, uncertainties remaining on the current RfD based on developmental effects from
methyimercury in children exposed in utero. There arc difficulties with reliability in recording and
classifying events like lace walking in children, especially as the data were collected approximately 30
months artcr the child's birth. It should be noted, however, that the
.•. . ••. . • ... .. !,.>.„.,., ...•...-.,••*
substantial developmental delays; for example, a child being unable to walk two .steps without support
at two years of age, inability to talk based on use of two or three meaningful words by 24 months, or
presence of generalized convulsive seizures. There is uncertainty in the physiologic factors which
were used in estimating the ingested mercury dose. There is also a degree of uncertainty introduced
by the size of the study population (81 mother-child pairs). Nevertheless, the RfD for methyimercury i\
is a reasonable estimate and is very likely to be protective of human health. — f*»o uwiJ. an ,/»»«» t ™° i .
/X«.m0/><^ _ . •u./ft •SiJk\*tJ*-
"This dose-response estimate is supported by additional studies in children exposed in utero.
These include investigations among Cree Indians in Canada and New Zealanders consuming large
amounts of fish. In these studies the hair concentration of mercury is used to monitor mercury
exposure over time. Conclusions by the investigators in their official reports cite developmental
delays among che children born of mothers whose hair mercury concentrations during pregnancy were
6 to 18 ppm, consistent with the benchmark dose of 1 1 ppm.
Currently a number of research studies are underway that further address the question of what
exposures to methyimercury in fish are associated with neurological disease. These studies include •""
more subjects than did the Iraqi study, are prospective in design, and utilize endpoints that are
anticipated to be more sensitive than the clinical signs and symptoms of methyimercury poisoning
observed in Iraq. These studies of fish consumption, rather than poisoning, are conducted in the
Seychelles Islands in the Indian Ocean, the Faroe Islands in the North Atlantic Ocean (sponsored by
the Department of Health and Human Services), and in the United States: this last study is sponsored
hy the Agency for Toxic Substances and Disease Registry (ATSDR). Data from these studies, when
available, should be useful in decreasing the uncertainty surrounding both the benchmark dose and the
RfD. The U.S. FDA has determined that revisions of its action level for mercury concentrations of
fish in interstate commerce should wait until the new studies have reduced the level of uncertainly
surrounding the RID. The availability nf results from the above studies will likewise enable U.S.
EPA to re-examine and adjust its RfD as needed.
3-22
-------
SENT BY'-of-FC SCI & TECH POLICY; 4-11-36
; E.MV I ROiSMEVT DIVISIO.V
319 541 1)840;? 6
Levels of Mcthylmcrcury Exposure Addressed by the U.S. Food and Drug Administration.
World Health Organization and State Recommendations
The U.S. EPA RID is a daily intake level and is a risk assessment tool: the use of the RfD
is not limited to fish. The discussion that follows covers risk assessment and risk management
activities concerning fish. These consider fish consumption patterns and risk management
policy factors. The existing advice and action level of the U.S. FDA is compatible with the U.S.
EPA RfD and the assessment information presented in this Mercury Study.
There are numerous local and state warnings in the U.S. to limit intake offish because of
chemical contamination.' Warnings arc issued because of a number of contaminants,
Methylmercury is most often included as one of the contaminants that form the basis for the
warning. Often these warnings are issued based on local conditions.
iJ^
^^
^,—.fXXKZ.^. ^^-
ItUiiiJt-.VJ-I
f total i
A number of different estimates exist for hair mercury levels that are associated with low
risks of neurological cndpoints such as parcsthesia. These estimates are sensitive to variables
such as the half-life of mercury in the body (time to eliminate half the dose of mercury). Half
life is usually estimated as an average of 70 days, with extremes of about 3 5 to just over 200
days reported for different individuals. The half-life of mercury in pregnant women has not
been directly measured. The half-life of mercury in women during lactation is shorter, possibly
due to excretion of mercury into milk produced during lactation.
Cross-comparison of
World Health Organization
(WHO) recommendations
regarding risk associated
with hair mercury
concentrations is
facilitated by data reported
by the WHO on mercury
concentrations in 559
samples of human head
hair from 32 locations in
13 countries. The WHO
report found that mercury
3
Table 3-$
WHO Data on Mercury in Hair
Fish Consumption Frequency Avenge Mercury Concentration in
1 Hair (pg mercury per g of hair)
No unusual mercury exposure
Less than one fish meal per month
Fish meals twice a month .
One fish meal a week !
One fish meal each day
2
1.4 (range 0.1 to
1.9 (range 0.2 to
2.5 (range 0.2 to
1 1-6 (range 3 .6 to
6.2)
9.2)
16.2)
24.0)
3-23
-------
SENT BY-'OFFC SCI & TECH POLICY: 4-11-96 ; b = UbPM :E-NVIRO.NME.NT DIVISION- 313 541 ua4U:? 7
III
y I
concentrations in hair increased with increasing frequency of fish consumption i see 1 able 3-3).
The World Health Organization's International Programme tor Chemical Safety
(WHO/IPCS) concluded thai the general population of adults (males and non-pregnant females)
docs not face a significant health risk from meihylmcrcury when hair mercury concentrations arc
under 50 ug mercury/gram hair. In recent evaluations of the Niigata epidemic of Minamata
disease, study authors reported louer thresholds with mean values in the ranue ol' 25 lu
approximately 50 j.ig mercury/gram hair.
Clinical observations in Iraq suggest that women during pregnancy are more sensitive to
the effects of methylmcrcury with fetuses at particularly increased risk. The WTTO/IPCS (1990)
analyzed the Iraqi data and identified a 30 percent risk lo the infant of abnormal neurological
iigns when maternal hair mercury concentrations were over 70 ug/g. Using an additional
statistical analysis, WHO/IPCS estimated a 5 percent risk of neurological disorder in the infant
when the maternal hair concentration was 10 to 20 |ig mercury/gram of hair. The U.S. EPA RfD
is within an order of magnitude of the dose described by WHO. — _ cLcb--^. * Q *»!•*•
*
WHO/IPCS recommended that as a preventive measure, in a subpopulation that
consumes large amounts of fish (for example, one serving or 100 grams per day), hair levels for
women of child-bearing age should be monitored for methyimercury.
The WHO/IPCS estimated (1990) that a daily methyimercury intake of 0.48 ug
mercury/kg body weight will not cause any adverse effects to adults and that a methyimercury
intake of 3 to 7 ug/kg body weight/day would result in a <5 percent increase in the incidence of
paresthesia in adults. Risk to this extent would be associated with hair mercury concentration of
approximately 50 to 125 ug |^tttr^^§ gram hair. By comparison, the U.S. EPA's reference
dose, or the amount of methyimercury any person (including children and pregnant women) can
ingest every day without harm is 0.1 fig/kg body weight per day. This was based on a
benchmark dose equal to 11 ppm (ug/g) hair. Children anKexpeiiKl to have a higher exposure to
methyimercury (on a per kg body weight basis) than do adults.
In 1969, in response to the poisonings in Minamata Bay and Niigata, Japan, the U.S.
FDA proposed an administrative guideline of 0.5 ppm for mercury in fish and shellfish moving
in interstate commerce. This limit was converted to an action level in 1974 (Federal Register 3JL
42738, December 6, 1974) and increased to 1.0 ppm in 1979 (Federal Register 44_. 3990, January
19, 1979) in recognition that exposure to mercury was less than originally considered. In 1984.
the 1.0 ppm action level was converted from a mercury standard to one based on methyimercury
(Federal Register 42, November 19.1984).
The action level takes into consideration the tolerable daily intake (TDI) for
methyimercury, as well as information on seafood consumption and associated exposure to
methyimercury. The TDI is ihe amount of methylmcrcury that can be consumed daily over a
long period of time with a reasonable certainty of no harm. U.S. FDA (and WHO) established a
TDI based on a weekly tolerance of 0.3 mg of total mercury per person, of which no more than
0.2 mg should be present as methylmercury. These amounts arc equivalent tn 5 and 3.3 ug.
respectively, per kilogram of body weight. Using the values for methyimercury. this tolerable
3-24
-------
BT-UrrC iCI 4 it.Cn rULlCY; 4-il-Sto ; O'UtSPM iLNVIROiVMtNf DIVI510.N-
313 541 U840;? a
level would'Correspond to approximately 230 uu/wcek for a /O kg person or .11 ui^pcrson/dav.
The TD1 was calculated from data developed in pan by S\\edi.sh studies of Japanese individuals
poisoned in the episode of Niigata which resulted from the consumption of contaminated rish
and shellfish and the consideration or other studies 01 rlsh-catuig populations.
Based on observations from the poisoning event later in Iraq. U.S. FDA has
acknowledged that the fetus may be more sensitive than adults to the effects nr' merr.ury (Federal
Register 44j. 3990, January 19, 1979; Cordlc and Tollefson. 1984. U.S. FDA Consumer.
September, 1994). In recognition of these concerns. U.S. FDA lias provided advice to pregnant
women and women of child-bearing age to limit their consumption of fish known to have high
leveis of mercury (U.S. FDA Consumer, 1994). U.S. FDA believes, however, that given existing
patterns offish consumption, few women (less than 1%) eating such high mercury fish will
experience slight reductions in the margin of safety. However, due to the uncertainties associated
with the Iraqi study. U.S. FDA has chosen not to use the Iraqi study as a basis for revising its
action level. Instead, the U.S. FDA has chosen to wait for findings of prospective studies of
fish-eating populations in the Seychelles Islands and in the Faroes Islands.
^d^
ffiipftii^atg^
K5IW««*><*«»*»>w*>«»*'**ii>t*t»»*t'»'>l'li7j»VA *<<«'»»
mmamiKssfSi
B^ij-r^^VAh &wffSMB^ --a.
3-25
-------
BY:C*K: aci & TECH POLICY; 4-n-ao ;
;C.NVIROINMEM
aia 541 u«4u;. a
'l-fifekexiSSue.adiirBSsed.ib,the tisic.ciiaraoterizaftnii-was-vrfT.GLhcrraixrfarbpogei7ic. mercury
oiK?feTO^S=^au^e^av£.^epotOTiat;t^
i^tl§nKOTC^quss5;$^S^ die
e^;pw£iO;H
'^
s^^ •" •
»?s^^ga»a^a>^^^«g^fgjjjj!s •'
3-26
-------
SENT BY=OfFC SCI 4 TECH POLICY; 4-11-36 , b=HPM ENVIRONMENT DIVISION
919 541 0840^10
-.:;> Tabled- 4
ts and Fist Consumption Values Used in Exposure-Modeling
-A"jj:sQnied-Body.-.Weight
cimg^analysisrshow that:hiiman^i^nsuiumgcS^v/3th. predicted
1
>A><»<»«.
Pf'
3-27
-------
SENT BY'OFFC SCI 4 TECH POLICY: 4-11-36 : 6:13PM ENVIRONMENT DIVISION- 313 541 0840;sll
Distribution of Mercury Concenrrati
^ Sampied Fish Tissue Throughouc the TJ.«
*P«re«nt of SJMS in catagety cumulative
Total Situ: 374
RI»t Qnty sa«a: 123
Maximum was Whola Body: 19-"
,
we
3-28
-------
SarfMX Water and pH £ SJ tad Anthropogenic Mercury Depusillon
00
b
-n
Tl
n
n
NORTHEAST
WEST
MID-ATI ANTIC
COASTAL PLAIN
IN ERIOR SOUTHEAST
r ""'
| . . j 6-MH pH < " 6.6. fepotMan 1-10 ug/nl
\-"'\
[ . J 6-SOH pH < - 66. QcpaiHan > » uglmt
fjj] > »OH pH < - B.fc 0*|x»Hloo « ug/tn]
|; I > 20* fH < - B.i, (tepiMMant-IO ug.'ina
liJjIJL' > 20M FH < » 6.1. topwidaa >
rn
s
o
n
CD
CT)
Co
15
3
2
o
CD
CD
DO
O
II
I—'
ro
-------
SENT BY'OFFC iCI & TECH POLICY; 4-11-96 ; C>:13PM iLNVIRO.NMEM DIVISION
319 541 Uti4U;?13
fcxi^a^.iSSetayijn'^ui^^^
*"' '' M**W-::***!*:-' ''" • ^J' "^''*"'" " "
' • .^"•*VJ^^i^l> '•;•'•• •'• -••!'•..--*—•^.-t'-Jv *W-^«1 • • '•*>.
fSS^S»5*w_
itilBSik
^s^»as»«^:;s^a^^^^saE*is;^s:i;«?:s;;^^K;;^^ ; » -:' ' '
-. .. .; .-«<«.»... f'"*1-;'
-.... .
i&ll^tiiiElppai^ilii^
:-Sf^¥SS?^SS*:S^^:-^S;ViS^SJ^^SS^';'S;^KS&i:2M^
3-30
-------
SENT BY-'OFFC SCI & TECH POLICY; 4-11-36 ; 6:15PM ;L\VIRO\M£NT DIVISlOiV
919 541 US4U;?14
Table J-S
Mercury Cunceutrations iu the Top Ten Types ol' Fish
Consumed hv U.S. Residents "
Fish
Tuna
Shrimp
Pollack
Salmon
Cod
Mercury x
Concentration
(jig/g, wet
weighO
Comments
0.2U6
0.047
0.15
0.035
The mercury content for tuna is the average of the mean
concentrations measured in 3 types of tuna: albacorc tmia
(0.264 ng/g), skipjack luna(0.136 (ig/g) and yellowfin tuna
(0.218 — - - —-•-
The mercury content lor shrimp is the average of the
mean concentrations measured in seven types of shrimp:
royal red shrimp (0.074 ng/g), white shrimp (0.054
fig/g), brown shrimp (0.048 |ig/g), ocean shrimp (0.053
M-S/S). pi^ shrimp (0.031 (Jg/g), pink northern shrimp
(0.024 }ig/g) and Alaska (sidestripe) shrimp (0.042
mercury content for salmon is the average of the
mean concentrations measured hi .five types of
Salmon: pink (0.019 /ig/g), chum (0.030 ngjg), coho
(0.038 Mg/g), sockeye (0.027 Mg/g), and chinook
(0.063 pg/g).
0.121 The mercury content for cod is the average of the
mean concentrations in Atlantic Cod (0.114 //.g/g) and
the Pacific Cod (0.127
3-32
-------
SENT BY'OFFC SCI & TECH POLICY; 4-11-96 ; 6:16PM ;E.NVIROiNMLNF DIVISIU.V
319 .541 UB4U;?.l3
Mcrcur-y
Concentration
Fish
weight)
Comments
Catfish
0.088
0.16
Clam
0.023
Flatfish
(Flounder)
0.092
Crab
0.117
Scallop
The sources of mercury content in catfish are Bahnick
ct al., 1994 and Lowe et aL, 1985. Both data sets
were collected from U.S. freshwater sources. The
Bahnick data (mean = 0.088) include channel,
largcmoudi, rock, striped and white catfish, and the
Lowe data (mean = 0.16) include channel and
flathead catfish. It should be noted that neither
survey included farm-raised catfish, which Is the type
of catfish predominantly consumed in the U.S. The
mercury content of farm-raised catfish may be
significantly different than freshwater sources.
The mercury content for clam is the average of the mean
concentrations measured in four types of clam: hard (or
quahog) clam (0.034 Hg/g), Pacific litdcncck clam (0
ug/g), soft clam (0.027 ug/g), and gcoduck clam (0.032
The mercury content for flounder is the average of the
mean concentrations measured in nine types of
flounder Gulf (0.147 ug/g). summer (0.127 ug/g),
southern (0.078 ug/g), four-spot (0.090 ug/g),
windowpane (0.151 ug/g), arrowtooth (0.020 ug/g),
witch (0.083 u.g/g), yellowtail (0.067 M-g/g), and winter
(0.066 ue/e). _ . _
The mercury content for crab is the average of the mean
concentrations measured in live types of crab: blue crab
(0.140 ug/g), dungencss crab (0.183 ug/g), king crab
(0.070 ug/g), tanner crab (C. opilio) (0.088 M-g/g), and
tanner crab (Cbairdt) (0.102 ltg/g).
0.042 i The mercury content for scallop is the average of the
mean concentrations measured in four types of scallop:
sea (smooth) scallop (0.101 yg/g), Atlantic Bay scallop
(0.038 ug/g), calico scallop (0.026 ug/g), and pink
scallop (0.004 uc/g).
3-33
-------
SENT BY'OFFC SCI & TECH POLICY; 4-11-36 ; 6= 16PM ENVIRONMENT DIVISION-
313 541 0840;?16
Swp^li^
?fP^lMg$|ag$gj$^^
,'.«*i,«.,. 4 :*:•»?**•*»'..,• • -..--. .»,. *«-2,x-.;-v«>,,-(-;*:"-)«*.< * r« .j,:, '.->•*•*"**•*• '*'.*»*»*••.*.• •• - *x<-*•••.>. •"•;• --•• *«..*••• rS-S-'V, «* »>J55.A: Sy-'i:;: »J'S*V**«'*v.. -»x-> ^
™~**<>»*~*^*»^>
Ti^T^CTvTiy^If^T-trHt'-'lIlUiGilS
-t~k.i-.. .-- !'.>... . ^i.jn-«>,, »*••'••• •_ •* v-«- ••>•-««.« s - » * .•-• ,v. -.•• * » •••-.' . .» ....
H£!&BJ vffy^SMVP.J?S*3<**?.*?S**;*ft^.^' ^.^*>^'*;'<*^'^*«-riJ-^r-,-j-«-Tiimii»»>£fcM^^-^'^—^?^'^?^'^^*i^*'^'>^''•'•*• •J"^-.<--****^*j*a*«n
yi-**'gr*rvt^-;-y • »i.. 11. .<< r»* N^»,Mi;Mt7^,T<^a7i^;B&^i3SSSj!JSJ;.r.CiS:SS
-------
SENT BY:OFFC SCI & TECH POLICY; 4-11-36 : 6:18PM : E.MV 1ROMMENT DIVISION-
919 341 0840;?17
om&s^^buSary^ v?iihirr't&£rpqp:u!£tian of.im
t33E|i:^iy'3^^
,;!«3&£».*x«^ .:-ix***w«*r-«j v'-"""^"*rT!"JTt">"rs*^>s:^'**:i^-'Vi'''';;*'::t-i;:.^--<;--=' ~*-<»»»»«x:* :«.»--».>%•'» v*x*"**M<*'>-*^w**:™*;***™«-Y'^"'*7"».P«'-««- •••%:-*
**>r^'*' -'• -"*-'' ^x-*->»>-' ^-y*1 tfc .._•*. J^ t pB , g r»>t /t-aJMr'T- im* I mm ' "* " ' »"*i^i i» ** J-_Vf**iJ^__>tLt _' _i_^T«^ ^J*H^^wV^-. j*f j^vf^'B-g* i^l^'in 4TJ"r^ T''yMj*>'i^f*
]**'**SiT^±L>*S£fi!*A_ t i-'"^***'**
jgBmaaftaBa
'•-^ij:« .-VLjrvriTuci'-'T-'-jUtfi'W' --•,-••-•-•-'-•-•-^••-i-'-1 -•-•---•-•-•-•; Ay/-v/^-[v/ty'Tvy//.>y^.v^.w^.:^^vj;:^^:r-4TLiyj>^j^^si:j^L;A^v^..
SazSaaraSsieojtHa^
jj£^s"aaasffi^;.aas^^^>rii;s;rM«itg5^^t^?gsasgBgt^ •
:• .•MuiK^.'i; * ?. •5<***^|H|W«^*.'* > *y*tA-*^*^T*^^-->tf •*••• '•'-..iii'-* * •^•^.* * - **Ki*-* •»•»••• ij il j
&OT
3-35
-------
SENT BY'OFFC SC! & TECH POLICY; 4-11-36 : 6:20PM 'ENVIRONMENT DIVISION- 313 541 0840:^18
Grams or More of Fishier JL)ay
pir.esentSr 1 lq--;5.;'jJBrcent 'of tota 1
-------
SENT BY:OFFC SCI & TECH POLICY; 4-11-96 : 6 = 22PM :E.NVIROiN'MEM DIVISION- 313 541 0840;?13
BfftffifttffiffinTJ
S3«aaiR5^3ig8g^»««i8iaS53£ig«.l2;jtiS5i;^iS
3-38
-------
SENT BY':OFFC SCI '& TECH POLICY: 4-11-36 ; 6 = 24PM :E\V 1ROXMEST DIVISIO.V , 319 541 (J84(J;=20
are more
with
aSsSed&iiSgs^^
':,^^--^?*t;V^v5';,^:;^-*^:;^^*^^ '*:"T
•M<£B<.^*M}.£fKt};„.> ...,.v
:«**«««•*•*•* •s-***""**'*-:**:-'^'*"!- »»•:•- ^: :
-^^v^w*^riv^:;*K5«^^¥»^^^tSSl^^
aasa^SSEaja»a45^«£aa»ciyas;r i&aSa
Kj%?s=sr?sss;s=:i:;
;.!g-ffri^.i-y^.-^x'^£.£&y>?siyx«itw«.t-<• '^y•'*^<*'c^''s.viif^S'tf'•*{* f*^**j*^%.**w*"»y!.^'>i'^>• o-' ^* >"«*' "
^^CjKC^^|?;^|^^m:;Ja^4^^i^^
^sillijJ|g^^^^ff5S55a^?^y^M'«iP-" ^
y^i^^^^^S^^^^^att^stm&m
ataggiSSSESm! ¥*r*w*i*K«iSia*
j^a^g^Kgpgfl
£522 ^ay szwr^^WrSSswMwS
^^^ 'ffi^S^3a
7 77 pofpS^S
, i/y
><»H«H •«« « y I »»^..^.; ^^ >; V.»>rt^ <^*^< j' " *"'"' •»**••*• N**NW«.
3-39
-------
SENT'BY:OFFC sci & TECH POLICY; 4-11-35 ; e;26PM ;ENVIRONMENT DIVISION- • sia 541
Ptt^n^Qfewoincny wpmen^child-bearing ageandchiWren consuming-above. 1 00 grams -(about
3^QUB^};of;Ss^
$$|i?&|u&^
r?^&^^
The primary purpose of the Mercury Study Report to Congress was to assess the impact of
U.S. anthropogenic emissions on mercury exposure to humans and wildlife. The size of some
populations of concern have been estimated; namely women of child-bearing age and children who
eat fish. In the general population, people typically obtain their fish from many sources. The
question on whether or not the impact of mercury from anthropogenic ambient emissions can be
proportioned to the overall impact of methyimercury on wildlife is a much more difficult issue. i ,
^V
As with environmental monitoring data, information on body burden of mercury in U^rf**J
populations of concern (blood and/or hair mercury concentrations) are not available for the general "u ^-^
U.S. population. Data on higher-risk groups arc currently too limited to discern a. pattern morc_
predictive of methyimercury exposure than information on quantities of fish consumed. ~~ ^~
content of certain foods has been suggestive as a basis for modifying estimates of the quantities~oT
methyimercury that produce adverse effects. Currently, data on this mercury/selenium association
form an inadequate basis to modify quantitative estimates of human response to a particular exposure
to mercury.
Available data for human health risk assessment have limitations as described in the Report
and in this summary. Studies of human fish-consuming populations in the Seychelles and Faroes
Islands address some of these limitations; *j"-y ar" "Tecfed to *** published within a year of release of
this Report. Additional studies on U.S. populations who consume fish from the Great Lakes arc in
progress. Public health agencies of the U.S. government as well as the U.S. EPA will evaluate these
new d?** when they aze available. Risk management decisions beyond the ongoing activities specified
in the Clean Air Act Amendments of 1990 will be based on consideration of all human data includin
results of these new studies.
The benchmark dose methodology used in estimating the RfD required mat data be clustered
'into dcsc groups. Most dan on neurologically based developmental endpolnis are continuous; that is,
not assigned to dose groups. For example, scoring on scales of IQ involves points rather than a
"yes/no" type of categorization. Measurements on the degree of constriction of the visual field
involve a scaling rather than a "constricted/unconsiricted" type of variable. Although jirtjitrajry scales
can be constructed, these groupings have generally not been done^g current systems
*' ~'
&iaaaaawiiK^^SSSSSW;
LUJTJtTfli jyThVtP^rW ™uXTVtvvi?jCTvra* "^CvTOKKTTClW'ISUI.f.CHr'tC^^ 1- rTITtTtfftl'J*'***'! !•• ~ iif I i ii : • I i 1 I ^^ ITiT 11 I
An additional difficulty occurs in estimation of benchmark dose for multiple
SaThavc been measured. Further research on appropriate methods for mathematical
modeling is needed. For some situations such information is known, but for methyimercury exposure
and multiple endpoints assessing the same system (i.e., developmentally sensitive neurological,
neuromotor and neuropsychological effects) the time-course/dose-rresponsc of such changes have not
3-40
-------
b>Y:orFC 5C1 & FECH POLICY; 4-11-36 : &:27FM ; E.NV I RO.NME.YT DIVISION- 319 541
been clearly estanlished. Development ol" the inmheinauciu mudeb needs re be accompanied by
understanding the physiological/pathological procesici, of mctliylmcrcury iiuoxicaLion.
How Much Methylmercury Exposure is Harmful to Wildlife and What Are the Effects?
Massive poisonings of birds and wildlife from mechylmercury-trcated seed grains were
identified during the decades preceding ihe 1970s. These findings resulted in substantial limitation on
use of raethylmercury-treated seed grains. However, mcthylmercury contamination of the aquatic
foodchaiii from many sources continues to adversely affect wildlife and domestic mammals and wild
birds. In Minamata, Japan from about 1950-1952 (prior to recognition of human poisonings) severe
difficulties with flying and other grossly abnormal behavior was observed among birds. Signs of
neurological disease including convulsions, fits, highly erratic movements (mad running, sudden
jumping, bumping into objects) were observed among domestic animals, especially cats that consumed
seafood.
Generally the place of wildlife in the aquatic foodchain of the ecosystem and their feeding
habits determine the degree to which the species is exposed to methylmercury. Fish-eating
(piscivorous) animals and those which prey on other fish-eaters accumulate more mercury than if they
consumed food from terrestrial food chains. In a study of fur-bearing animals in Wisconsin, the
species with the highest tissue leveis of mercury were otter and mink, which are top mammalian
predators in the aquatic food chain. Top avian predators of aquatic food chains include raptors such
as the osprey and bald eagle. Smaller birds feeding at lower levels in the aquatic food chains also
may be exposed to substantial amounts of mercury because of their high food consumption rate
(consumption/day/gram of body weight) relative to larger birds.
Laboratory studies under controlled conditions can be used to assess the effects of
methylmercury from fish on mink, otter and several avian species. Effects can occur at a dose of
0.25 ng/g bw/day or 1,1 jig/g methylmercury in diet. Death may occur in species at 0.1-0.5 Mg^g
body weight/day or 1.0-5.0 /zg/g in the diet. Smaller animals (for example, minks, monkeys) are
generally more susceptible to mercury poisoning than are larger animals (for example, mule deer,
harp seals). Smaller mammals eat more per unit body weight than larger mammals. Thus, smaller
mammals may be exposed to larger amounts of methylmercury on a body weight basis.
Whole body residues of mercury in acutely poisoned birds usually exceed 20 /ig/g fresh
weight. Although sublcthal effects include a number of different organ systems, reproductive effects
are the primary concern. These occur at concentrations far lower than those that cause overt coxtcity.
The broad ecosystem effects of mercury are not completely understood. No applicable studies
of the effects of mercury on intact ecosystems were found. Consequently, characterization of risk for
non-human species did not attempt to quantify effects of mercury on ecosystems, communities, or
species diversity. The characterization focused on quantities of mercury thai adversely affects the
health of sensitive subpopulations of wildlife species and on the co-location of these populations with
areas of elevated mercury exposure secondary to ambient, anthropogenic emissions of mcthylmercury.
To this end wildlife criteria (WC) were calculated for three piscivorous birds and two mammals (see
Table 3-7). The WC is a mercury level in water which is
3-41
-------
SENT BY'OFFC SCI 4 TECH POLICY; 4-11-36 ;' tj:27PM ENVIRONMENT DIVISION-
313 541 0840 :?23
4. MANAGEMENT ALTERNATIVES
Possible Control Strategies
Effective control of mercury emissions may require a mix of strategies including pollution
prevention, materials separation and conventional regulatory strategies to control mercury emissions a.t~Tf4J~ CT
the stack. Pollution prevention would he suitable for those processes or industries where a mercury /t
substitute is demonstrated and available (e.g., mercury cell chlor-alkali plants). Material separation
an appropriate approach for processes where mercury-containing products arc disposed of by
incineration, or where mercury can he reduced in the fuel prior to the fuel being combusted (e.g.,
medical waste incineration). The third approach, conventional regulatory strategies, nji||8i5
a$|rtip8Te when mercury is emitted to the environment as a result of trace contamination in fossil fuel
or otncr"esscntial feedstock in an industrial process (e.g., smelting). Other non-traditional approaches
such as emissions trading or application of a use tax, or other market-based approaches may also
prove feasible for mercury control.
Si|;3fl^
' W
^''fi^i^
••* * ••.•.-•.• .. -. • ••••• ~~^ .
if
The analyses of control technologies and 'costs presented in this Report are not intended to
replace a thorough regulatory analysis, as would be performed for a rulemaking. The i
presented is intended to present the range of available options and provide a relative sense of the f j&p.
extent of mercury reductions achievable and the general magnitude of the cost of such reductions
m
mi^^^^^^^^^^^^^^— -~»..
Ig^i^i^a^M^aE^^^^g^^^^i^^
'. :•". • y^»"wBwg»«j"^«iig«».Mg^t«.!f.i ,-\" ^_-T-=~
^^^gsgg^iiLisrMS^g^^^^S^
4-1
-------
OJ'urrL id i ItCH fULlCY; 4-11-atj : c-2bPM ;t.\VIRO;NML\T DiViSIOiV
313 541 0840:^24
i^^j^iMtion-mcasuri:,.fiiduding.,any changes, in the., quatitj-of the-xnd^
!a^^^am^S^^B^^e^M^lmu^acf
v -™-—.-,;J,,..iT,,..,.,.......<-..,.^gngjg^
Sjjlcmjijiei
HmrfDB
MMii
|p|l|]^f^^^^^^^
atfiKfisi-|^3rjS
iitffiaefliii
,«*»**M*;^-.-.-^->^K*.w..-h^^. ^£*£.-l^.±^£^^^i^:?#•*&***>''*••*••••.-•*• > »-^<-«-«v<%--.--s-....-. V.,«.*«:A;::.^.^;\,\ .,; liX^'-r,"-;^?."^ ; .v,"v«»-«*-~-.-*.*»-„„.*.»,, «;.v,„_,..........^.; .;
jj^ajppq^mliii^^l^ia^^
--
ifflaas^.«a^jiii.iHisaiSiij
casasd^^i^tJKtttlRKrar^nDraigm^
5TiZC5SJ33^JII2ISS^IS5M*«1^k-«»-!«»'r'!^..~... .,v; ~
4-5
-------
SENT BY'OFFC iCI & TECH POLICY; 4-11-36 : &:31PM ; ESVI ROiNMEVT DIVISION-
541 0840:£25
jfo>^||^
Ki~™«ia»wM'«m
7.fiftttti?&ttftf$#l
•Mffyitli.!****!***** «•'».••'«•&£
M^^^ffJjIjjjSsSs^
^^S^SiR^^i^gV^M^i^^^^^^l^igiMga
-------
£5
OJ
(b
X)
Z>
-f
in
CTJ
r—i
cn
a
u
rt
U
u_
o
)-
CQ
s
Ml!!
ill
I!
-------
-I • (-It-It. ,
alb 041 ua4u;s=27
sndppinis .including peuratosiciiy and-reproductive
assSiSfflj
jp$j|$ir^
iip^^fiai^
%<^:iHaB%i&^^
•' •'«•"". v "***"*•** 'W;,';'^;--x-><«..T>^;- •'.',', ;••"•""'."', •;•••?•*••*" • "•"«.*«*«•—•!•«.".. *C ."*•*••' '' •»'%v.".>>%v.%^.1 ,v..'.. ;-V."1''*»*"»|Sl*wi'>J -,'••-• • ^H"t--;
5-7
-------
SENT BY'-OFFC SCI £ TECH POLICY; 4-11-36 : 5:48PM ; ENV1 R
-------
SENT BY'OFFC SCI £ TEOt'POLICY; 4-11-36 : 5'43PM :E.NVIROMMENT DIVISION 313 541 U840;?27
mercury tixpcoute lo Oi. par r.icjj.dr .-ancarn icr ^.-.ima^a -n'_ r._-.e ~op
of the food web, including humans . Human and wildlife exposure 'jo
methylmercury oc:c:ura primarily by ingesting fish. The amount of
fiah consumed and Lhe ^.evel of methyJ-itiercury in the species o~
fish that are eauen determine che level of exposure.
Methylmercury primarily affects the nervous system in humans and
causes reproductive damaqe in wildlife, particularly birds.
Although direct cause and effect relationships between . ~rL - /\
rw»— ^ jiff.
mercury contamination and poulation dec] ines in various wildlife / J
species have net been established, mercury may be a contributing sc*»r»^
" factor to population declines of the endangered Florida panther
and the common loon.
For human, health impacts, it is emphasized that ill: current
levels in the environment the typical U.S. consumer of fish is
riot in danger of consuming harmful levels of methylmercury and is
not being advised, directly or implicitly, to reduce fish
consumption. Existing Food and Drug Administration management
and consumer advice remains protective of the food in commerce
and the health of individuals consuming it. However, a jSma 1 .1
percentage of the U..S. population whose diet consists, largely of «»
fish (e.g., subsistence fishers) may be at risk if_ those fish are
obtained from mercury-contaminated waters and _if the f inn contain
relatively high levels of mercury.
•>
The EPA is bringing to public notice, through this report,
Lhe potential risks to human and wildlife species of continued or
increasing mercury emissions. <~ •• 5 --^ %
It should also be noted that the health assessment portion
of the report is based on the^es£) scientific data available at
the present time. New studies of potential human risks from'
ingesting mercury through fish consumption, have been completed
-------
SENT BY:OFFC SCI & TECH POLICY; 4-11-36 ; 5:43PM ;E.NVIRO.VMENT DIVTSIOiV
313 541
rf«~* P»p
and arc cx.ccczcd "o be p.ibii^hud "r -be ,:OTrnr.a v^ar . Cnce '..hs -%* aS^lJL.
complete resull..1-; ot those stud: e.5 are axmilab" e, the EFA will f"
collaborate with other Federal ^qencics in a scient i ir; r review Q-
A
Lhe total body of health eifecus information. If, following t,h: :
review, a revision to the risk assessment portion of .this report
is warranted, the EPA will i.ssue a revised .risk assessment.
Although it is important to continue developing the best
possible scientific evidence relating to mercury exposure and its
effect:?/ it appears that reducing additional mercury loadings to*
the environment is desirable and prudent, if feasible. The hiPA,
many individual States, arid several industries have already !-.aken
a number u± steps, including regulatory actions and voluntary
programs, to reduce such releases to the environment. The SPA
intends-to furt.her refine its approach for reducing mercury
releases. The fcli'A' s focus will be on common sense solutions and
pollution prevention programs. A wide range of options will be
examined within a multi-media framework. Further research and
science will continue and be coordinated within EPA as well as
with other Federal atgenci es and institutions to better understand
Eiercury Issues.
Sincerely yours,
Carol M. Browner
-------
United States Department of the Interior
NATIONAL BIOLOGICAL SERVICE
Washington, DC 20240
MAY -9 19
Dr. Martha Keating
U.S. Environmental Protection Agency
Risk and' Exposure Assessment Group (MD-15)
Research Triangle Park, North Carolina 27711
Dear Dr. Keating,
The National Biological Service (NBS) has completed a review of the draft report titled
Volume V: An ecological assessment of anthropogenic mercury emissions in the United States.
This review was done over 3 weeks and focussed on content issues such as data sources and
interpretation of results rather than on a line-by-line evaluation. Comments of the reviewers
from the Midwest Science Center and Patuxent Wildlife Science Center are enclosed. The
important issues raised by these reviewers and those received from technical staff at
headquarters are summarized below.
Reviewers generally agreed that the draft report was balanced, scholarly, and comprehensive.
The authors offer a thorough discussion of limitations and uncertainties concerning the models
and the derivation of wildlife criteria for mercury levels in water. Because of these issues,
conclusions in the draft report should be viewed as provisional pending additional research and
monitoring to address these limitations and gaps in our knowledge and understanding of the
effects of mercury on biological resources.
Some specific concerns raised by NBS reviewers were:
• The simple pH-based model of mercury availability may not be adequate to characterize
aquatic ecosystems. This model in conjunction with predictions on atmospheric
deposition did not identify south Florida as an "at-risk" region, an area where mercury
is accumulating in food chains.
• Bioaccumulation factors (BAFs) were computed from the ratio of mercury levels in
biota to those in water. This method of estimating BAFs has a great deal of
uncertainty. Toxic thresholds based on tissue concentrations of fish and wildlife or on
concentrations in their food would have much less uncertainty.
• On page 2-16, the authors state that only one nationwide fish collection effort provides
suitable data for estimating average concentrations of mercury in freshwater fish. For
this effort (Bahnick et al. 1994) fish were collected at sites chosen because of their
proximity to point and non-point sources of pollution. Although 35 "remote" sites
were included, the majority of samples were collected from locations likely to be more
-------
"polluted" than sites not sampled. This effort is likely to provide a biased estimate of
average contaminant levels in fish tissues.
• A data set more appropriate for estimating average concentrations of contaminants in
fish tissues is provided by the sampling conducted through the National Contaminant
Biomonitoring Program (NCBP). -Mercury has been measured regularly since 1969 in
whole fish composite samples collected at more than 100 stations situated at key points
in major river systems and the open waters of the Great Lakes. The samples were
collected according to consistent field protocols. Cold-vapor atomic absorption was
used to measure concentrations in fish tissue in a reliable, quality-assured manner. The
last full set of data is from 1986, however 34 stations in the Mississippi River drainage
were sampled in late 1995. The chemical analysis of samples collected in late 1995 is
now beginning. Although mercury concentrations in many of these waterways have
been influenced by point sources, the sites were not selected to reflect those sources.
Consequently, the NCBP data set provides unbiased estimates of average concentrations
in freshwater fish. Findings through 1986 have been summarized (Schmitt and Bunck
1995 and Schmitt and Brumbaugh 1990) and the raw data are included in reports on
individual collections (e.g., Schmitt and Brumbaugh 1990 for data from 1986). Data
from 1970-1986 are available through STORET, both on-line and in CD-ROM format.
• Overlap between the range of a species and "regions of concern" is used to assess the
potential risk. While this approach is a reasonable beginning, such an analysis must be
taken further. All parts of a species range are not equally important to sustaining that
species and organisms are not equally distributed throughout their range. For
migratory species, breeding range should be considered separately from wintering
range, because consequences of exposure may be different during these seasons.
• The interactions of mercury and selenium are especially important to interpreting
mercury residues from field samples. It is not uncommon to find high levels of
mercury and selenium in healthy organisms. Apparently selenium can protect against
mercury poisoning and, although less frequently discussed in the literature, mercury
can protect against selenium poisoning. The authors should take this issue into account
when discussing field levels.
Our knowledge and understanding of environmental exposures and the effects of mercury on
biological resources can only be improved through further research and monitoring. NBS
scientists identified areas for future research that were similar to those listed in Section 6 of the
draft report. These topics included study of ecosystem cycling of mercury and factors that
affect its bunavailability, controlled dosing studies with free-ranging avian species, controlled
laboratory studies on the toxicity of chemical forms of mercury, study of factors that influence
the bioaccumulation of mercury in lower trophic levels and in aquatic food chains, and
development of improved analytical tools for sampling and measuring various forms of
mercury.
-------
Research is also needed to understand the interactions of mercury and selenium. Some initial
work at the Patuxent Wildlife Research Center (described in the enclosed reviewers' comments)
has highlighted the complexities of these interactions and the importance of understanding
these interactions for the interpretation of field data.
Both research and monitoring are needed to evaluate the assumptions of the models and
assessments made in this report. The authors provide a sound and comprehensive beginning to
objectively define the ecological risks associated with mercury emissions. A critical next step
would be to design survey and monitoring efforts to evaluate the broad geographic predictions
made in' this assessment, and to design research to study assumptions behind the models used
in this assessment. * \
We appreciate the opportunity to review and comment on this report. If NBS can be of further
assistance, please give me a call at 202-482-3555.
Sincerely,
Dr. 0.J. Reichman
Assistant Director for Research
Enclosures
References
Bahnick, D., C. Sauer, B. Butterworth, and D. Kuehl. 1994. Chemosphere 29:537-546.
Schmitt, C.J., and W.G. Brumbaugh. 1990. Archi. Environ. Contam. Toxicol. 19:731-747.
Schmitt, C.J., and C.M. Bunck. 1995. Pp. 413-416 in Our Living Resources. A report to
the Nation on the distribution, abundance, and health of U.S. plants, animals, and ecosystems.
U.S. Department of the Interior, National Biological Service. Washington D.C.
-------
April 23, 1996
Memorandum
To: Chris Bunck
From: Gary Heinz
Subject: Review of EPA mercury document.
e>
The draft of "Mercury Study: Report to Congress, Volume V is : very impressive
document. It is apparent that a great deal of careful work went into it." I feel the best help I
can give is to suggest some ways to improve section 6, Research Needs, specifically section
6.2 Oi. vVildlife Toxicity Data. Here are some suggestions.
1. Species Differences
The report recognized the need to get more toxicity data for a broader range of
wildlife. However, this is very difficult because so few wild species can be bred easily in
captivity, and results from field studies are much more difficult to interpret. At least with
birds, I believe the way to go is to conduct field studies in which relatively clean colonies of
birds are made to lay eggs with varying levels of mercury in them. This would be
accomplished by capturing adult females prior to the reproductive season or catching females
on the nest and destroying their eggs (making them later renest), then injecting or implanting
them with a device to release methylmercury. Different groups would get different doses of
methylmercury, based on a randomized scheme.
If the specie: were a colonially nesting fish-eating bird such as the herring gi.ll or
double-crested cormorant, one could conduct the entire experiment on one nesting island.
Perhaps 20 females would be injected or implanted with only the carrier (these would be the
controls). Another 20 randomly selected females would be injected with a dose that one
would guess to produce eggs containing 0.2 ppm mercury, another 20 with enough
methylmercury to produce eggs containing 0.5 ppm mercury, another group 2 ppm mercury,
and perhaps a final, group with eggs containing about 5 ppm mercury. Of course, owing to
variations in individual absorption of methylmercury and the timing of subsequent egg
laying, actual mercury levels in eggs would encompass a broad range around the projected
level. But this would be okay. As long as the study generated a range of mercury levels in
eggs going from a control level up to 5 or 10 ppm, the residue data could be grouped and
analyzed statistically. One would, for example run a correlation between ppm mercury the
sample egg taken from each nest versus hatching success of the remaining eggs in the nest.
Or one would run an analysis of variance for hatching success, wherein eggs residues were
grouped as 0 - 0.2 ppm mercury, >0.2 - 0.6 ppm, >0.6 - 1.5 ppm, > 1.5 3 ppm, >3 - 6
ppm, and > 6 ppm, or some such arrangement.
The beauty of this controlled field study approach is that all the birds would be
exposed to the same weather, food supply, and other conditions, just like in a lab study. For
-------
birds such as loons, that do not nest in colonies, the same basic approach could be used, but
it would take a lot more work to get adequate sample sizes. For bald eagles, this approach
could not be used.
Taere are two obstacles to conducting controlled field studies. The biggest is the
resistance of State Departments of Natural Resources and the public to intentionally dosing
wild birds with toxic substances. The second is the methodological difficulties in injecting
and implanting slow i Mease doses of methyl mercury. The state and public concern must be
overcome with education about the benefits of learning how toxic mercury is. The dosing
problems can be overcome; we have people here at Patuxent working on th issue.
2. Chemical forms of mercury in adult tissues and eggs."
The report addresses the need to measure methyl mercury as wen as total mercury, but
the needs do not stop there. Methylmercury is known to be more toxic than other commonly
known forms of mercury, but what is really known about the toxicity of the metabolites of
methylmercury or the combined treatment of a bird with methylmercury plus inorganic
mercury. In other words, are we ready to simplify the issue of mercury poisoning by
thinking that only methylmercury in nature is responsible for wildlife problems?
This area of research will be best approached through controlled laboratory studies
using birds and mammals fed various chemical forms of mercury, at first singly, and then in
combinations. A lot is know about mercuric chloride toxicity to animals, but there are a lot
of metabolites of mercury in animals.
If only methylmercury turns out to be important in tissues and eggs, then the entire
question of water based criteria must be revisited, because even if we know that most of the
merc>:ry in fish, for example, is methylmercury, it is really the amount of methylmercury in
the mammal or bird eating the fish that counts. Animals can eat food containing
methylmercury and then convert it to metabolites. And it appears that there may be
mechanisms that regulate the amount of methylmercury that accumulates in animals; it some
species, the greater the amount of mercury in the body, the smaller the percentage that is
methylmercury. Marine species in particular seem to be able to regulate methylmercury.
Do water based mercury criteria apply to these marine species that seem to be more immune
to mercury poisoning?
3. Mercury/selenium interactions.
The draft document recognizes that other factors can affect mercury toxicity, but the
interactions of mercury and selenium are especially important and relevant to interpreting
mercury residues in birds and mammals, and this problem needs to be addressed in the
section on Research Needs. It is not uncommon to find high levels of mercury and selenium
going hand in hand. The existing literature on these interactions tells us that selenium can
protect against mercury poisoning, and, although it is less frequently discussed, mercury can
protect against selenium poisoning. But there have been exceptions, where the combination
of mercury and selenium was worse than the single dosage by only one.
-------
We did an interesting study here a couple of years ago that really opened our e\es to
the complexities of mercury/selenium interactions. We fed mallards a control diet, a diet
containing 10 ppm mercury (as methylmercury), a diet containing 10 ppm selenium (as
selenomethionine), or a diet containing both 10 ppm mercury and 10 ppm selenium. Adult
males fed the 10 ppm mercury diet suffered from severe methylmercury poisoning. Males
fed 10 ppm selenium plus 10 ppm mercury were protected against mercury poisoning. So
far, this fits in with what might have been predicted. The surprising result was that
reproductive effects were most severe in the group fed mercury plus selenium. Both
chemicals by themselves reduced reproductive success, but together they were worse. So,
within the same experiment, adult males were protected from mercury poisoning by the
addition of selenium to the diet, but when the adult females incorporated both selenium and
mercury into their eggs, there was no protection for the embryos; there was at least an
additive effect.
We know the chemical forms of mercury and selenium we fed to the mallards, but we
only analyzed for total mercury and total selenium in eggs and adult tissues. So we do not
know what metabolized forms the adults and embryos might have been exposed to. We do
know that the presence of mercury in the diet greatly enhanced the storage of selenium.
especially in the livers of the adult males. Selenium, however, did not enhance storage of
mercury.
After all the years that publications have been coming out on the toxic interactions of
mercury and selenium, we still cannot advise people on what the significance is of high
residues of both elements simultaneously in mammals and birds. This might be considered
an academic question, where it not for the frequency with which both elements are found at
elevated levels in the same animal.
Additional research must look at various concentrations of each element in the diet
and what the chemical forms end up being in tissues and eggs.
I hope these suggestions on research needs will help the Environmental Protection
Agency in their revisions of this draft document.
Sincerely,
Gary Heinz
-------
United States Department of the Interior
NATIONAL BIOLOGICAL SERVICE
Midwest Science Center
4200 New Haven Road
Columbia, MO 65201
MEMORANDUM
Date: April 26, 1996
To: Wilbur Mauck, Assistant Center Director, MSC
«.
From: John M. Besser, Research Fisheries Biologist
Jimmie D, Petty, Senior Research Chemis
Subject: Review of EPA Mercury Study
We have reviewed "Mercury Study Report to Congress, Volume V: An
Ecological Assessment of Anthropogenic Mercury Emissions in the United States."
The authors have addressed a difficult task and their report represents a balanced
view of the mercury problem. We agree with tt\e fundamental approach used in
this report and with its general conclusions regarding the behavior of mercury in
aquatic food chains and potential risks of mercury toxicity to fish-eating wildlife.
The report does not understate the limitations of the models and data used to
identify ecosystems at risk from mercury contamination and to generate Wildlife
Criteria for mercury, As a result, the conclusions of the report should be
considered provisional, pending additional research to address these limitations.
Chapter 6 (Research Needs) adequately summarizes most of the items that
must be considered in order to adequately assess risks to piscivorous wildlife from
mercury contamination. Although we do not disagree with the research needs
identified in the report, we would like to emphasize the following concerns about
the report's methodology and the research needed to improved the scientific
foundation for these methods:
*
Identification of Ecosystems at Risk (Chapter 3). The simple pH-based model of
mercury availability used In this report may not be adequate to characterize
aquatic ecosystems with widely differing characteristics. Although the influence
of pH on mercury bioavailability has been well documented, the study
acknowledges that "factors other than pH" may also affect bioavailability. For
example, the report notes the strong evidence for mercury risks to wildlife in the
Florida Everglades, yet this ecosystem is not identified as "at-risk" by the pH-
based model.
-------
Research Need: Ecosystem characteristics affecting Hg bioavailability. Examine
differences in mercury bioavailability among geographic regions and
ecosystem types, including associations with factors other than pH which
may affect bioavailability (e.g. dissolved organic matter, productivity, and
hydrology).
Research Need: Patterns of mercury bioaccumulation in aquatic food chains.
Characterize patterns of mercury bioaccumulation in aquatic food chains
potentially affected by mercury inputs, considering of factors su~h as length
and complexity of food webs and differences in species composition and
feeding habits.
\
Derivation of water-based Wildlife Criteria {Chapter 4). The use biota:water
bioaccumulation factors (BAFs) to calr"!ate wildlife criteria based on total aqueous
mercury is difficult, since bioavailable mercury species (monomethyl mercury)
make up a small and variable fraction of total mercury concentrations in water.
The resulting high variation of BAFs derived from total aqueous mercury
concentrations (e.g. range of two orders of magnitude for BAFs for 3rd-order
predators) reduces'confidence in these values and may contribute to unrealistically
low Wildlife Criteria values. There is much less uncertainty about toxicity
thresholds based on mercury concentrations in tissues of wildlife or their food
items, making these values a more reliable basis for risk assessment.
Research Need: improved methods for sampling and analysis of bioavailable
mercury species. Adequate analytical facilities are currently limited and
extreme care is necessary to avoid sample contamination. Innovative
sampling techniques, such as selective passive samplers, could facilitate
sampling and analysts of low concentrations of methylmercury from water.
Research Need: Bioaccumulation of mercury in lower trophic levels. Better data
on mercury uptake by organisms at the base of aquatic food webs
(macrophytes, phytoplankton, zooplankton, and benthic invertebrates),
combined with data on bioavailable mercury species, should reduce variation
in BAFs and allow better linkages of Hg bioavailability to ecosystem
characteristics.
These problems will not be solved without a long-term commitment to
multidisciplinary research. We recommend that EPA convene workgroups, with
representation from other government agencies, universities, and the private
sector, with the goal of formulating a research approach that will both address
fundamental research needs and generate data needed for regulatory decisions.
-------
Supplemental
Scientific
Materials
June 1996 Supplemental Materials SAB REVIEW DRAFT
-------
GUIDE TO FIGURES
The following text and figures constituted pan of the Risk Characterization (Volume VI) in the
December 1995 version of Volume VI of the Mercury Study Report to Congress. U.S. EPA was advised
to remove the graphs. These are provided to the Science Advisory Board to facilitate a complete and
thorough scientific review. Numbers in brackets [ ], refer to the graphs in this supplemental material.
Human Intake of Methvlmercurv Estimated Through Dietary Surveys and Mercury Residue Data
Food intake primarily from the ingestion of contaminated fish is the only significant source of
methylmercury exposure to the general human population (Stem, 1993; Swedish EPA, 1991; WHO, 1990).
Total mercury concentrations in meats and cereals often measure hundreds of times less than in fish
(Swedish EPA, 1991). In most non-fish foodstuffs mercury concentrations are typically near detection
limits and are comprised mainly of inorganic species (WHO, 1990). In contrast, most of the mercury in
fish is methylated and fish methylmercury concentrations are typically above analytic detection limits.
Available techniques to estimate fish consumption include long-term dietary histories and
questionnaires to identify typical food intake or short-term dietary recall techniques. Temporal variation
in dietary patterns is an issue to consider in evaluation of short-term recall/record data. For
epidemiological studies that seek to understand the relationship of long-term dietary patterns to chronic
disease, typical food intake is the relevant measure to evaluate (Willett, 1990). Because methylmercury
is a developmental toxin that may produce adverse effects following a comparatively brief exposure period
(i.e., a few mouths rather than decades), comparatively short-term dietary patterns can have importance.
Human methylmercury intake from fish for the general U.S. population was estimated in this
Report by combining data on mercury concentrations in fish species (expressed as micrograms of mercury
per gram fresh-weight of fish tissue) with the reported quantities and types of fish species consumed by
fish eaters or "users" in the USDA's Continuing Surveys of Food Intake by Individuals (CSFII 89/91).
The dietary survey methodology consisted of an assessment of three consecutive days of food intake and
selection of interviewees from probability samples for non-institutionalized United States households.
Survey respondents numbered 11,706 individuals who were surveyed across all four seasons of the year
and all seven days of the week. Respondents also reported their body weights, and these data were
utilized in Volume in to estimate fish consumption on a per body weight basis. Use of these survey data
provides a nationally based estimate of fish intake by the general population of the United States.
The CSFII 89/91 survey design reflected known sources of variability in estimating dietary intakes
in general. The extent to which comparatively short-term assessments of dietary intake predict long-term
fish consumption patterns remains an uncertainty. Nutritional epidemiologists (among others see Willet,
1990) have observed that these surveys provide a cross-sectional view of dietary intake that better predicts
central tendency than the extremes of the range of typical fish consumption behavior. In Volume III
comparisons were made between quantities consumed and the upper quartile of the fish-consuming
subpopulation of the general United States population and estimates of quantities of fish consumed by
subpopulations of high fish-consuming Native American Tribes and anglers. Fish consumption rates
reported by several tribes and by high fish-consuming anglers, to some extent, corroborate the daily
consumption rates of the extreme end of the distribution of the CSFII 89/91. Since these individuals are
June 1996 Supplemental Materials SAB REVIEW DRAFT
-------
part of the U.S. population, their consumption rates may be reflected in the CSFII 89/91 It should be
noted that the angler and Native American fish consumption surveys utilize different types of survey
methods; this further corroborates the high-end estimates of CSFII 89/91.
Among nationally representative weighted samples of individuals, 50.9 percent reported
consumption of fish and/or combinations of fish, shellfish, or seafood with starches in a 3-day period.
Of individuals reporting fish consumption, approximately 98 percent consumed fish only once, and about
2 percent consumed fish in two or more meals during the 3-day survey period. For foods consumed by
only a minority of the population, estimates of per capita consumption rates overestimate the consumption
rate for the general population, but underestimate the consumption rate among the portion of the
population which actually consumes the food item. As a consequence, in this risk characterization fish
consumption estimates are based on a "per consumer" basis.
Respondents in CSFII 89/91 who reported eating fish indicated the fish species and quantity of
fish consumed. Fish consumption has been reported to be recalled with greater accuracy than other food
groups (Karvetti and Knuts, 1985). Nevertheless, an uncertainty in these data is the ability of consumers
to identify the species of fish consumed. The species of fish identified~by the respondents were recorded
as part of the dietary records of the survey. These fish species were identified and used to estimate dietary
intake of methylmercury. The survey and results are described in Appendix H to Volume III.'
Selection of a database for mercury residues in fish was determined by several factors. Because
the dietary consumption data were nationally based, preference was given to fish residue data that included
fish from widely diverse geographic areas. This choice aimed at avoiding data more appropriate for site-
specific assessments that would result from the sampling of a very limited geographic area (e.g., a state-
wide survey). In addition, the preferred database should include as many individual fish to represent the
consumed species as possible; this is considered a better approximation of the estimate of central tendency
due to individual samples. A third preference was for data collected over a time period that approximated
the years of the dietary survey. The third criteria was judged the least important in selection of the
database because residues of mercury in soils and sediments continue to contribute mercury to the aquatic
food-web and are considered to minimize year-to-year variability in mercury concentrations in fish tissues.
Data describing methylmercury concentrations in marine fish were largely based on the National Marine
Fisheries Service (NMFS) database, the largest publicly available database on mercury concentrations in
marine fish. This NMFS database has been collected over the past two decades. Comparison of the
values for central tendency (e.g., 50th percentile) in mercury concentrations between the NMFS database
and FDA's compliance data on selected species (Carrington et al., 1995) indicated close agreement in
mercury concentrations.
For freshwater fish species two publications reporting mercury concentrations in multiple species
offish were chosen. Data reported by Bahnick et al. (1994) and Lowe et al. (1985) were used to estimate
average mercury concentrations in fresh-water finfish from across the United States. Both Lowe et al.
(1985) and Bahnick et al. (1994) used nationwide surveys of freshwater fish. Both databases suffer from
a limited number of samples at any one site; however, a strength of each data set is that many sites were
sampled. These data are described in detail in Appendix H of Volume III. If average mercury
concentration in fresh-water fish are compared, the results from Lowe et al. (1985) and Bahnick et ai.
(1994) differ by a factor of approximately two. Consequently separate analyses of methyl mercury intake
June 1996 Supplemental Materials SAB REVIEW DRAFT
2
-------
from fish were prepared to assess the impact of the database chosen for mercury resides.
Findinas of this evaluation based on dietary patterns for the general U.S. population and mercury
residue data in fish were used to estimate mercury consumption from fish. Figures [1] through [25]
present methylmercury intake from marine and/or freshwater fish and shellfish for the general U.S.
population and subpopuiations of either women of child-bearing age or children aged 14 years or younger.
Plotted on each of these figures are lines labelled for three estimates of hazard assessment and dose-
response: the RfD for methylmercury based on combined developmental effects, the benchmark dose for
combined developmental effects among Iraqi children exposed to methylmercury in utero (rounded to one
figure) and an estimate of the Lowest Observed Adverse Effect Level (LOAEL) (estimated from Table
5-7 of this Volume). The fraction of the U.S. population or subpopuiations that exceed these values can
be estimated by comparing the percentile of fish consumers that exceed these lines. The number of
persons in age-gender categories come from the US Census (1990) data. These are enumerated and not
projected data. Consequently the number of individuals who exceed a particular percentile can be directly
estimated by combining estimated percentile of fish-consumer and the number of persons of the
appropriate age-gender category who exceed a particular fish and/or methylmercury intake.
Consumption of methylmercury (expressed per "kilogram self-reported body weight) from marine
and freshwater fmfish and shellfish for the general U.S. fish-consuming population were estimated using
the Bahnick et al. (1994) data shown in Figures [1] and [2]. For comparison, methylmercury intake from
total fish and shellfish for the general U.S. population estimated from the data of Lowe et al. (1985) are
shown in Figures [3]and [4]. Use of the data of Bahnick et al. (1994) which reported overall higher in
mercury concentrations than those of Lowe et al. (1985) provided estimates of methylmercury intake for
the general U.S. fish-consuming population. Overall, the difference hi concentration of mercury is
approximately two-fold between these reports of mercury concentrations hi fish. Methylmercury intake
from only freshwater are shown in Figures [5] through [8]. The most commonly consumed marine
finfish for the general population is tuna. Estimated consumption of methyhnercury from tuna for the
general U.S. fish-consuming population is shown in Figures [9] and [10].
Methylmercury intakes calculated in Appendix H of Volume HI have been developed for a
nationally based rather than a site-specific assessments. The CSFH 89/91 from USD A was designed to
represent the U.S. population. The concentrations of methylmercury in marine fish and shellfish were
derived from a database that is national hi scope and the data on fresh-water finish were from two large
studies that sampled fish at a number of sites throughout the United States. The applicability of these data
to site-specific or region-specific assessments must be judged on a case-by-case basis.
In selection of sensitive subpopuiations of humans, sensitivity may reflect an inherent
responsiveness to die hazard (i.e., toxicity based) or reflect elevated exposures to the agent of concern.
With respect to risks posed by methyhnercury from fish, two subpopuiations of humans are of particular
interest hi this risk characterization: women of child-bearing age and children. Women of child-bearing
age are of concern because developmental effects following in utero exposures are the basis for the RfD
and because of the developing nervous system would be expected to be most sensitive to mercury toxicity.
Because 9.5 percent of women ages 15 through 44 years are pregnant hi a given year and the half-life
of mercury averages 70 days, the entire population of women of child-bearing age is judged to be of
concern. Figures [11] and [12] present estimated methylmercury intake from finfish and shellfish of both
June 1996 Supplemental Materials SAB REVIEW DRAFT
3
-------
marine and freshwater origin for women of child-bearing age. The data of Bahmck et al. (1994) were
used to calculate the estimates shown in Figure [11]. These data are generally higher in methylmercury
concentration than are those of Lowe et al. (1985) which were used to present the estimates shown in
Figure [12]. Because of the quantities of freshwater fish consumed are a small part of the total fish intake
for the general population, differences in average mercury concentrations between the data of Lowe et
al. (1985) and Bahnick et al. (1994) do not result in marked differences in average mercury intake from
fish. Figures [13] and(14] give estimated methylmercury intakes from freshwater fish for U.S. women
of child-bearing age. Figure [15] presents methylmercury intake from tuna only.
The second subpopuiation identified in this risk characterization to be of concern is children aged
14 years and younger. The basis for this concern is that intake of methylmercury from fish is greater
(if expressed on a per kilogram body weight basis) than for adults. When the methylmercury intake is
expressed on a per kilogram body weight basis, the exposure for children aged 14 years and younger is
approximately two-to-three times that of the adult. These data for children are presented in Figures [16]
through [25]. The higher estimated exposure to methyl-mercury is the result of the higher intake of food
on a per weight basis among children. A major uncertainty identified in this risk characterization is the
absence of data to assess health hazards of methylmercury for children who have low methylmercury
exposures in uiero.
One strength of the data from CSFn 89/91 is that individual body weight data were available and
were utilized in calculation of dietary intakes on a per kilogram body weight basis. Consequently actual
body weights rather than default values were utilized to estimate methylmercury exposure per kilogram
body weight. The methylmercury intakes of adult males and females are comparable on a body weight
basis. The maximum intakes on a per kilogram body weight basis are also provided for each group
considered. The intake for the high-end fish consumer (the mariTmim respondent hi each group of adults)
is at least four times that of the intake for the individual at the 95th percenrile.
Methylmercury intake from fish estimated hi this way does not permit attribution to anthropogenic
or "natural" sources. Because of the magnitude of anthropogenic, ambient mercury contamination, the
estimates of methylmercury from fish do not provide a "background" value. "Background" values imply
an exposure against which the increments of anthropogenic activity could be added. This is not the
situation due to release of substantial quantities into the environment.
The following text accompanied selected figures in the 10/31/95 version of Volume I, Executive Summary
of the Mercury Study Report to Congress.
There are two ways to estimate the size of the population at risk. The first evaluates the quantity of
methylmercury exposure on a body weight basis and compares this exposure to the U.S. EPA's reference
dose. The second looks only at the quantity of fish consumed without malting any assumptions about the
concentration of mercury in the fish.
The first method requires information on the amount of fish consumed and the concentration of
mercury in various species of fish. Data on fish consumption for a general population of women in the
United States were developed from the United States Department of Agriculture's Continuing Surveys
June 1996 Supplemental Materials SAB REVIEW DRAFT
4
-------
of Individual Food Consumption for the period 1989-1991 (CSFII 89/91). These surveys assessed fish
consumption over a three-day period. Data on mercury concentrations in fish were available from :he
National Marine Fisheries Service and from published information on mercury in freshwater fish. These
sources have been used in this Report to estimate methylmercury intake from fish.
The RfD for methylmercury is 1.0 xlO" mg/kg body weight/day or 0.1 ,ug/kg body weight/day.
Data on methyimercury consumption from fish for women of child-bearing age (ages 15 through 44
years) show the 50th percentile for adult females to be 0.08 fig/kg body weight/day with 75th percentile
exposure to be 0.16 jig mercury/kg body weight/day. Consequently approximately one-half of the
general U.S. population of women of child-bearing age who consume fish at least once within a 3-day
period are predicted to ingest methylmercury at the RfD. The RfD is expected to be a level of exposure
to be without adverse effect.
These data are shown in Figures 3-1 [16], 3-2 [17], and 3-3 [11], which describe the. total
quantity of fish and shellfish consumed by chifdren (ages 14 years or younger) or adult women of child-
bearing age (15 through 44 years) who eat fish often enough to be identified as fish consumers in the
three-day dietary survey, CSFn 89/91. Because freshwater fish may be of particular interest around
anthropogenic sources, Figures 3-4 [20], 3-5 [21], and 3-6 [13]show the quantity of freshwater fish
consumed by children (aged 14 years or younger) and adult women of child-bearing age (15 through 44
years) who ate fish one or more times during die three-day dietary survey period.
General U.S. Population
Figure 1. General U.S. Male Population, concentration data provided by National Marine
Fisheries Service and by Bahnick et al. for freshwater fish, all fish and shellfish.
Figure 2. General U.S. Female Population, concentration data provided by National Marine
Fisheries Service and by Bahnick et al. for freshwater fish, ail fish and shellfish.
Figure 3. General U.S. Male Population, concentration data provided by National Marine
Fisheries Service and by .Lowe et al. for freshwater fish, all fish and shellfish.
Figure 4. General U.S. Female Population, concentration data provided by National Marine
Fisheries Service and by Lowe et al. for freshwater fish, all fish and shellfish.
Figure 5. General U.S. Male Population, concentration data provided by National Marine
Fisheries Service and by Bahnick et al. for freshwater, freshwater fish.
Figure 6. General U.S. Female Population, concentration data provided by National Marine
Fisheries Service and by Bahnick et al. for freshwater fish, freshwater fish.
Figure 7. . General U.S. Male Population, concentration data for freshwater fish from Lowe
et al., freshwater fish.
June 1996 Supplemental Materials SAB REVIEW DRAFT
5
-------
Figure 8. General U.S. Female population, concentration data from Lowe et al., freshwater
fish.
Figure 9. General U.S. Male Population, concentration data provided by National Marine
Fisheries Service, tuna fish.
Figure 10. General U.S. Female Population, concentration data provided by National Marine
Fisheries Service, tuna fish.
Women of Child-Bearing Age
Figure 11. U.S. Women of Child-Bearing Age, concentration data provided by National
Marine Fisheries Service and by Bahnick et ai. for freshwater fish, all fish and
shellfish.
Figure 12. ' U.S. Women of Child-Bearing Age, concentration data provided by National
Marine Fisheries Service and by Lowe et al. for freshwater fish, all fish and
shellfish.
Figure 13. U.S. Women of Child-Bearing Age, concentration data from Bahnick et al.,
freshwater fish.
Figure 14. U.S. Women of Child-Bearing Age, concentration data from Lowe et al.,
freshwater fish.
Figure 15. U.S. Women of Child-Bearing Age, concentration data provided by National
Marine Fisheries Service, tuna fish.
US. Children
Figure 16. U.S. Male Children, concentration data provided by National Marine Fisheries
Service and by Bahnick et al. for freshwater fish, all fish and shellfish.
Figure 17. U.S. Female Children, concentration data provided by National Marine Fisheries
Service and by Bahnick et al. for freshwater fish, all fish and shellfish.
Figure 18. U.S. Male Children, concentration data provided by National Marine Fisheries
Service and by Lowe et al. for freshwater fish, all fish and shellfish.
Figure 19. U.S. Female Children, concentration data provided by National Marine Fisheries
Service and by Lowe et al. for freshwater fish, all fish and shellfish.
Figure 20. U.S. Male Children, concentration data from Bahnick et al., freshwater fish.
Figure 21. U.S. Female Children, concentration data from Bahnick et al., freshwater fish.
June 1996 Supplemental Materials SAB REVIEW DRAFT
-------
Figure 22. U.S. Male Children, concentration data from Lowe et al., freshwater,fish.
Figure 23. U.S. Female Children, concentration data from Lowe et al., freshwater fish.
Figure 24. U.S. Male Children, concentration data provided by National Marine Fisheries
Service, tuna fish.
Figure 25. U.S. Female Children, concentration data from National Marine Fisheries Service,
tuna fish.
Data for Freshwater Fish Only
Figure 5. General U.S. Male Population, concentration data provided by National Marine
Fisheries Service and by Bahnick et al. for freshwater, freshwater fish.
Figure 6. General U.S. Female Population, concentration data provided by National Marine
Fisheries Service and by Bahnick et al. for freshwater fish, freshwater fish.
Figure 7. General U.S. Male Population, concentration data for freshwater fish from Lowe
et al., freshwater fish.
Figure 8. General U.S. Female population, concentration data from Lowe et al., freshwater
fish.
Figure 13. U.S. Women of Child-Bearing Age, concentration data from Bahnick et al.,
freshwater fish.
Figure 14. U.S. Women of Child-Bearing Age, concentration data from Lowe et al.,
freshwater fish.
Figure 20. U.S. Male Children, concentration data from Bahnick et al., fres'hwater fish.
Figure 21. U.S. Female Children, concentration data from Bahnick et al., freshwater fish.
Figure 22. U.S. Male Children, concentration data from Lowe et al., freshwater fish.
Figure 23. U.S. Female Children, concentration data from Lowe et al., freshwater fish.
Data for Tuna Fish Only
Figure 9. General U.S. Male Population, concentration data provided by National Marine
Fisheries Service, tuna fish.
June 1996 Supplemental Materials SAB REVIEW DRAFT
7
-------
Figure 10. General U.S. Female Population, concentration data provided by National Marine
Fisheries Service, tuna fish.
Figure 15. U.S. Women of Child-Bearing Age, concentration data provided by National
Marine Fisheries Service, tuna fish.
Figure 24. U.S. Male Children, concentration data provided by National Marine Fisheries
Service, tuna fish.
Figure 25. . U.S. Female Children, concentration data from National Marine Fisheries Service,
tuna fish.
June 1996 Supplemental Materials SAB REVIEW DRAFT
8
-------
5.5
5.3
I
Figure 1
Estimated Exposure to Methylmercury from Fish and Shellfish: General U.S. Male Population Who
Reported Consuming Fish and Shellfish in a Cross-Sectional Survey
With a Three-Day Sampling Period
Enlargement of upper 5 percent
Human LOAEL
RfD Range (0.07 - 0.3)
96lh 97lh ' 98lh 99ih Max
Human NOAEL
RfD Range (0.07 - 0.3
0.1
Min 5!h 25lh 50ih . • 75»h
Percentile of U.S. Fish Consuming Population*
* Estimated from fish consumption ilula in CSFII 89/91, and mercury conccnlruliuns provided by National Marine Fisheries Service and Balmick ct ul.
Note: The percenlilc indicates the fraction of the surveyed population with an estimated melhylmercury exposure < the value shown on the Y-axis.
For example, at the 25lh percentile, 25 percent ol'the surveyed population had an estimated melhylmercury exposure < 0 04 ug/kg BW/day.
Max
Page 9
-------
Mill
Figure 2
Estimated Exposure to Methylmercury from Fish and Shellfish: General U.S. Female Population
Who Reported Consuming Fish and Shellfish in a Cross-Sectional Survey
With a Three-Day Sampling Period
Enlargement of upper 5 percent
Human LOABL
RED Range (0.07 - 0.3)
95lh 96ih 97lh 98th Wih
Human NOAEL
RED Range (0.07 - 0.3)
25lh
50lh
Percentile of U.S. Fish Consuming Population*
75th
'J5lh Max
* Hstimaied horn fish consumption data in CSIJII 89/91, and mercury concentrations provided by National Marine Fisheries Service and Bahnick ct al.
Note: The percentile indicates the fraction of the surveyed population with an estimated melhylnjercury exposure < the value shown on the Y-axis.
For example, al (he 25lh perceniile, 25 percent of the surveyed population hud an estimated melhylmcrcury exposuie < 0.04 jig/kg BW/day.
Page 10
-------
0.1
Mm
Figure 3
Estimated Exposure to Methyimercury from Fish and Shellfish: General U.S. Male Population Who
Reported Consuming Fish and Shellfish in a Cross-Sectional Survey
With a Three-Day Sampling Period
Enlargement of upper S percent
Human LOAEL
RtD Range (0.07 - 0.3)
Human NOAEL
RfD Range (0.07 - 0.3
25ih
50ih
Percentile of U.S. Fish Consuming Population*
75ih
95ih Max
* lislimaled fmin fish consumption data in CSHI 89/91, and mercury concentrations provided by National Marine l;ishenes Service and Lowe el ul.
Note: The pcrcenlilc indicates the fraction of the surveyed population with an estimated mclhylmercury exposure < the value shown on the Y-axis.
For example, at the 25th percenlile, 25 percent of the surveyed population had an estimated mclhylmercury exposure < 0 04 pg/kg BW/day.
Page
-------
Figure 4
Estimated Exposure to Methylmercury from Fish and Shellfish: General U.S. Female Population
Who Reported Consuming Fish and Shellfish in a Cross-Sectional Survey
With a Three-Day Sampling Period
Enlargement of upper 5 percent
Human LOAEL
RED Range (0.07 - 0.3)
Human NOAEL
RfD Range (0.07 - 0.3)
Min 5lh 25lh 50th 75lli
Percentile of U.S. Fish Consuming Population*
* Estimated from 1'ish consumption cliila in CSFII 89/91, and mercury concentrations provided by Nulionul Marine Fisheries Service and Lowe ct al.
Nole: The pea-entile indicates the fraction of the surveyed population with an estimated ineihylmereury exposure < the value shown on the Y-axis.
For example, at the 25ih perecnlilc. 25 percent of the surveyed population had an estimated nicthylmercury exposure < 0 04 pg/kg BW/day.
95lli Max
Page 12
-------
5.5
5.3
f
o>
1.0
3
M
O
J1
I
0.5
*w
I
0.1
Figure 5
Estimated Exposure to Methyimercury from Freshwater Fish: General U.S. Male Population Who
Reported Consuming Freshwater Fish in a Cross-Sectional Survey
With a Three-Day Sampling Period
Enlargement of upper 5 percent
Human LOAEL
Human NOAEL
RfD Range (0.07 - 0.3)
RfD Range (0.07 - 0.3)
Min Slh
25lh
50lh
Percentile of U.S. Fish Consuming Population*
75lh
95th MM
* Estimated from fish consumption duia in CSPII 89/91, and mercury conccnlrulions provided by National Marine Fisheries Service and Buhnick el al.
Note: The perccniile indicates I he fraction of the surveyed population with an estimated mclhylmcrcury exposure < the value shown on the Y-axis.
For example, al the 25th perccniile, 25 percent of the surveyed population had an estimated mcthylniercury exposure < 0.1 pg/kg BW/day.
Page 13
-------
5.5
5.3
I
ffi
D)
1.0
S
I
a
I
">.
0.5
Figure 6
Estimated Exposure to Methylmercury from Freshwater Fish: General U.S. Female Population Who
Reported Consuming Freshwater Fish in a Cross-Sectional Survey
With a Three-Day Sampling Period
o.i
Enlargement of upper 5 percent
Human LOAEL
Human NOAEL
RtD Range (0.07 - 0.3)
95lh %lh 97th 98lh l>Jth Max
RfD Range (0.07
Min 5lh 25lh ' 50th 75th
Percentile of U.S. Fish Consuming Population*
* Lsiiinaiol I'rom 1'ish consumption data in CSFII 89/91, and mercury concentrations provided by National Marine Fisheries Service and Bahnick et al.
Note: The percenlile indicates Ihe traction or Ihe surveyed population with an estimated mclhyhnercury exposure < the value shown on the Y-axis.
For example, at Ihe 25lh percenlile, 25 percent of Ihe surveyed population had an estimated njelhylmercury exposure < 0.08 p.g/kg BW/day.
Max
Page 14
-------
5.5
Figure 7
Estimated Exposure to Methylmercury from Freshwater Fish: General U.S. Male Population Who
Reported Consuming Freshwater Fish in a Cross-Sectional Survey
With a Three-Day Sampling Period
Min 5th
25lh
Enlargement of upper 5 percent
R£D Range (0.07 - 0.3)
50th
Percentile of U.S. Fish Consuming Population*
75th
'J5lh Max
* Estimated from fish consumption data in CSF1I 89/91, and mercury concentrations provided by National Marine Fisheries Service and Lowe el al.
Noic: The pcrccniilc indicates the fraction of the surveyed population with an estimated methyl mercury exposure < the value shown on the Y-axis.
Por example, ill the 25ih pen-entile, 25 percent of the surveyed population had an estimated incihylmercury exposure < 0.06 Jig/kg BW/day.
Page 15
-------
Figure 8
Estimated Exposure to Methylmercury from Freshwater Fish: General U.S. Female Population Who
Reported Consuming Freshwater Fish in a Cross-Sectional Survey
With a Three-Day Sampling Period
Enlargement of upper 5 percent
Human LOAEL
RCD Range (0.07 - 0.3)
Human NOAEL
RfD Range (0.07 - 0.3)
25lh
50lh
75lh
Min 5th
Percentile of U.S. Fish Consuming Population*
* Estimated from fish consumption data in CSFII 89/91, and mercury concentrations provided by National Marine Fisheries Service and Lowe el al.
Note: The percenlile indicates the fraction of Ihe surveyed population with an estimated methylmercury exposure < ihc value shown on the Y-axis.
For example, at the 25ih percenlile, 25 percent of the surveyed population had an estimated melhylmercury exposuie < 0.05 Jig/kg BW/day.
95th Max
Page 16
-------
5.5
I
Figure 9
Estimated Exposure to Methylmercury from Tuna Fish: General U.S. Male Population Who Reported
Consuming Tuna Fish in a Cross-Sectional Survey With a Three-Day Sampling Period
Enlargement of (he upper 5 percent
Min 5lh
25lh
50lh
Percentile of U.S. Fish Consuming Population*
75th
* Estimated from I'ish consumption data in CSFII 89/91 survey.
Note: The percentile indicates the fraction of the surveyed population with an estimated inelhylmercury exposure < the value shown on the Y-axis.
For example, at the 2.5th pcrcenlile, 25 percent of the surveyed population had an estimated mclhylmcreury exposure < 0.06 ug/kg BW/day
RfD Range (0.07 - 0.3)
95111 Max
Page 17
-------
Figure 10
Estimated Exposure to Methylmercury from Tuna Fish: General U.S. Female Population Who
Reported Consuming Tuna Fish in a Cross-Sectional Survey With a Three-Day Sampling Period
Enlargement of upper 5 percent
Human LOAEL
RED Range (0.07 - 0.3)
Human NOAEL
RfD Range (0.07 - 0.3)
Min 5lh
25lh
5()lh
Percentile of U.S. Fish Consuming Population*
Max
* Ustimaied IIODI fish consumption data in CSI-'ll 89/91 survey.
Note: The pcrceniilc indicates the fr.it I ion of the surveyed population with an estimated meihylmercury exposure < the value shown on Hie Y-axis.
For example, al the 25ih pcrccnlilc. 25 peiceni of the surveyed population had an estimated meihylmercury exposure < 0 06 fig/kg BW/day.
Page 18
-------
Figure 11*
Estimated Exposure to Methyimercury from Fish and Shellfish: U.S. Women of Child-Bearing Age
Who Reported Consuming Fish and Shellfish in a Cross-Sectional Survey
With a Three-Day Sampling Period
Enlargement of upper 5 percent
Human LOAEL
Rfrf Range (0.07 - 0.3)
Human NOAEL
RfD Range (0.07 - 0.3)
Min
Max
Percentile of U.S. Fish Consuming Population*
* Estimated from lish consumption Jula in CSFII 89/91, and mercury concenlralions- provided by National Marine Fisheries Service and Hahnick el. al.
Nole: The pcrcenlile indicate!) the traction of the surveyed population with an estimated melhylmercury exposure < the value shown an the Y-axis.
For example, al the 25ih pcrcenlile, 25 percent of the surveyed population had an estimated melhylmercury exposure < 0 04 jig/kg BW/day.
Page 19
-------
Figure 12
Estimated Exposure to Methylmercury from Fish and Shellfish: U.S. Women of Child-Bearing Age
Who Reported Consuming Fish and Shellfish in a Cross-Sectional Survey
With a Three-Day Sampling Period
5.5
5.3
Enlargement ol upper 5 percent
R£D Range "Ttf.
RfD Range (0.07 - 0.3)
Min
25lh
50th
Percentile of U.S. Fish Consuming Population*
75lh
95lh Max
* Estimated Iron) fish consumption dala in CSFII 89/91, and mercury concentrations provided by National Marine Fisheries Service and Lowe el al.
Note: The perceniilc indicates the fraction of ihe surveyed population with an estimated melhylmercury exposure < the value shown on the Y-axis.
For example, al the 25lh percentile, 25 percent of the surveyed population had an estimated melhylmercury exposure < 0.04 Mg/kg BW/day.
Page 20
-------
5.5
5.3
1.0
0.5
0.1
Figure 13
Estimated Exposure to Methylmercury from Freshwater Fish: U.S. Women of Child-Bearing Age
Who Reported Consuming Freshwater Fish in a Cross-Sectional Survey
With a Three-Day Sampling Period
Enlargement tig upper 5 percent
RfD Range (0.07 - 0.3)
Min 5ih 25th 50lh 75lh
Percentile of U.S. Fish Consuming Population*
* lisiimaled Iron) li.sh consumption data in CSFH 89/91, and mercury concentrations provided by National Marine Fisheries Service and Hahnick el. al.
Note: The percentile indicates the fraction of the surveyed population with an estimated mclhylrncrcury exposure < the value shown on (he Y-axis.
I-'or example, al the 2Sih perccmile, 25 percenl of (he surveyed population had an estimated inclhylmercury exposure < 0 OH |4g/kg BW/day.
Max
Page 21
-------
5.5
Min
Figure 14
Estimated Exposure to Methylmercury from Freshwater Fish: U.S. Women of Child-Bearing Age
Who Reported Consuming Freshwater Fish in a Cross-Sectional Survey
With a Three-Day Sampling Period
Enlargement of upper 5 percent
Human LOAEL
Human NOAEL
RfD Range (0.07 - 0.3)
RfD Range (0.07 - 0.3)
25lh
50lh
Percentile of U.S. Fish Consuming Population*
75th
«)5lli Max
* Estimated from fish consumption data in CSFII 89/91, and mercury concentrations provided by National Marine Fisheries Service and Lowe el al.
Note: The percentile indicates the fraction of the surveyed population with an estimated melhylmcrcury exposure < the value shown on the Y-axh>.
For example, al the 25lh percenlile, 25 percent of the surveyed population had an eslimaled melhylmcrcury exposure < 0.04 ug/kg BW/day,
Page 22
-------
5.5
5.3
SK
CD
5
ffi
o»
i.o
0.5
a
'•g
UJ
Figure 15
Estimated Exposure to Methylmercury from Tuna Fish: U.S. Women
of Child-Bearing Age Who Reported Consuming Tuna Fish in a Cross-Sectional Survey
With a Three-Day Sampling Period
Human LOABL
Human NOAEL
Enlargement of upper 5 percent
RfD Range (0.07 - 0.3)
95ih 96lh
97lh
98th
99th
Max
0.1
A
Min 5lh
RfD Range (0.07 - 0.3)
25lh
5()th
Percentile of U.S. Fish Consuming Population*
Max
* Estimated from fish consumption data in CSFII 89/91 survey.
Note: The percenlilc indicates the fraction of the surveyed population with an estimated inethylincrcury exposure < the value shown on the Y-axis.
For example, ai the 25th percentile, 25 percent of the surveyed population had an estimated melhylmercury exposure < 0.06 MB&g BW/day.
Page 23
-------
5.5
5.3
UJ
0.1
Figure 16
Estimated Exposure to Methylmercury from Fish and Shellfish: U.S. Male Children Who Reported
Consuming Fish and Shellfish in a Cross-Sectional Survey With a Three-Day Sampling Period
Human LOAEL
Enlargement of upper 5 percent
RfD Range (0.07 - 0.3)
RfD Range (0.07 - 0.3)
Min
Percentile of U.S. Fish Consuming Population*
* lislimaled from hsh consumption ihila in CSFII 89/91, and mercury conccnlralions provided by National Marine Fisheries Service and Bahnick el. al.
Note: The pcrcentilc indicates the fraction of the surveyed population with an estimated meihylmercury exposuie < the value shown on the Y-axis.
For example, al the 25ih percenlile, 25 percent of the surveyed population had an estimated meihylmercury exposuie < 0 09 ug/kg BW/day.
Max
Page 24
-------
0.1
Figure 17
Estimated Exposure to Methylmercury from Fish and Shellfish: U.S. Female Children Who Reported
Consuming Fish and Shellfish in a Cross-Sectional Survey With a Three-Day Sampling Period
Enlargement of upper 5 percent
Human LOAEL
RfD HaiKje (0.07 - 0.3)
Human NOAEL
RfD Range (0.07 - 0.3)
Min 5lh
25th
50lh
Percentile of U.S. Fish Consuming Population*
75ih
*)5lh Max
* IJMimaleJ from fish consumption data in CSFII 89/91, and mercury concentrations provided hy National Marine Fisheries Service and Bahnick el. al.
Note: The pereenlile indicates the fraction of the surveyed population with an estimated melhylmercury exposure < the value shown on the Y-axis.
I'or example, at the 25ih perecnlile. 25 percent of the surveyed population had an estimated melhylmcrcury exposure ^ 0 07 MB^B BW/day.
25
-------
Figure 18
Estimated Exposure to Methylmercury from Fish and Shellfish: U.S. Male Children Who Reported
Consuming Fish and Shellfish in a Cross-Sectional Survey With a Three-Day Sampling Period
Enlargement of upper 5 percent
RfD Range (0.07 - 0.3)
RfD Range (0.07 - 0.3)
Min 5th
25lh
50lh
Percentile of U.S. Fish Consuming Population*
75th
95lh M.ix
* Estimated I'rum lish consumption data in CSFII 89/91, and mercury concentrations provided by Nulionul Marine l-'ishcrics Service and Lowe ct al.
Note: The percenlile indicates the fraction of the surveyed population with an estimated inclhylmcrcury exposure < the value shown*on the Y-axis.
For example, at Ihe 25th percenlile, 25 percent of the surveyed population had an estimated inclhylmcrcury exposure < 0 08 HB/^8 BW/day.
Page 26
-------
0.1
Figure 19
Estimated Exposure to Methylmercury from Fish and Shellfish: U.S. Female Children Who Reported
Consuming Fish and Shellfish in a Cross-Sectional Survey With a Three-Day Sampling Period
Enlargement of upper.5 percent
Human LOAEL
RtD Range (0.07 - 0.3)
95lh 96th
Human NOAEL
RfD Range (0.07 - 0.3)
Min 5lh
25th
50th
Percentile of U.S. Fish Consuming Population*
75lh
* lislnnalcd rioiii I'i.sh consumption dala in CSFII 89/91, and mercury concentrations provided by National Marine Fisheries .Service and Lowe el al.
Nolc: The percentile indicates the fraction of the surveyed population with an estimated mclhylmcicury exposure < (lie value shown on the Y-axis.
For example, at the 25th pen-entile, 25 percent of the surveyed population had an estimated nielhylmercury exposure < 0 07 MU^g UW/day.
95th Max
Page 2 7
-------
Figure 20
Estimated Exposure to Methylmercury from Freshwater Fish: U.S. Male Children Who Reported
Consuming Freshwater Fish in a Cross-Sectional Survey With a Three-Day Sampling Period
5.5
Enlargement of upper 5 percent
Human LOAEL
Human NOAEL
RED Range (0.07 - 0.3)
RfD Range (0.07 - 0.3)
Min 5lh
25lh
50(h
Percentile of U.S. Fish Consuming Population*
75th
95lli Max
* Intimated from fish consumption dulii in CSPII 89/91, and mercury concentrations provided by National Marine I-'ishenes Service and Bahnick el. al.
Note: The perceulile indicates the fraction of the surveyed population with an estimated mclhylmcrcury exposure < the value shown on the Y-axis.
Tor example, al the 25th percenlile. 25 percent of the surveyed population had an estimated inelhylmercury exposure < 0.15 MB/kg BW/day.
Page 28
-------
Figure 21
Estimated Exposure to Methylmercury from Freshwater Fish: U.S. Female Children Who Reported
Consuming Freshwater Fish in a Cross-Sectional Survey With a Three-Day Sampling Period
5.5
5.3
ffi
1.0
S
a
•o
S
a
3
ui
o.i
Enlargement of upper 5 percent
RfO Range (0.07 - 0.3)
Min 5lh 25lh 50lh 75lh
Percentile of U.S. Fish Consuming Population*
* Estimated from fish consumption data in CSFII 89/91, and mercury concentrations provided by National Marine Fisheries Service and Buhnick cl. al.'
Note: The pcrccmilc indicates the fraction of the surveyed population with an estimated mclhylmercury exposure < the value shown on the Y-axis.
For example, at the 25th perccntile, 25 percent of the surveyed population had an estimated melhylmerctiry exposure < 0.12 \\g/kg, BW/day.
i)5th Max
I'age 29
-------
5.5
5.3
m
o>
$...
(0
o
3
>.
a
E
. 5
o.i
Figure 22
Estimated Exposure to Methyimercury from Freshwater Fish: U.S. Male Children Who Reported
Consuming Freshwater Fish in a Cross-Sectional Survey With a Three-Day Sampling Period
Human LOAEL
Human NOAEL
RfD Range (0.07 - 0.3)
V
Enlargement of upper 5 percent
RfD Range (0.07 - 0.3)
95lh
96lh
97lh
98lh
Max
Mm 5lh
25lh
50lh
Percentile of U.S. Fish Consuming Population*
75th
95lh Max
* Estimated Irom fish consumption dula in CSPII 89/91, and mercury conccnlrulions provided by National Marine Fisheries Service and Lowe el ul.
Note: The pcrcentile indicates the fraction of (he surveyed population with an estimated melhylmercury exposure < the value shown on the Y-axis.
For example, at the 2Sth percentilc, 25 percent of the surveyed population had an estimated methyl mercury exposure < 0.08 |Jg/kg BW/day.
ae 30
-------
5.5
0.1
Figure 23
Estimated Exposure to Methyimercury from Freshwater Fish: U.S. Female Children Who Reported
Consuming Freshwater Fish in a Cross-Sectional Survey With a Three-Day Sampling Period
Enlargement of upper 5 percent
Human LOAEL
RED Range (0.07 - 0.3)
Human NOAEL
RfD Range (0.07 - 0.3)
Min 5lh
25lh
50th
Percentile of U.S. Fish Consuming Population*
75th
Max
* lislimaled from fish consumption data in CSh'll 89/91, and mercury concentrations provided by National Marine Fisheries Service and Lowe el al.
Note: The percemile indicates (lie fraction of the surveyed population with an estimated niclhylmeicury exposuic < the value shown on the Y-axis.
l;or example, at the 25lh percenlilc, 25 percent of the surveyed population had an estimated melhylmercury exposure < 0.06 ug/kg BW/day.
Page 3 I
-------
5.5
5.3
Figure 24
Estimated Exposure to Methyimercury from Tuna Fish: U.S. Male Children Who Reported
Consuming Tuna Fish in a Cross-Sectional Survey With a Three-Day Sampling Period
Enlargement of upper 5 percent
RfD Range (0.07 - 0.3)
Min
5lh
25lh
50th
Percentile of U.S. Fish Consuming Population*
75lh
95th Max
* Estimated from fish consumption data in CSFII 89/91 survey.
Note: The percenlilc indicates the fraction of the surveyed population with an estimated mclhylmercury exposure < the value shown on the Y-axis.
For example, at the 25th pcrccntile, 25 percent of the surveyed population had an estimated mclhylmercury exposure < 0.09 im/kg BW/day.
''"ge 32
-------
5.5
Figure 25
Estimated Exposure to Methyimercury from Tuna Fish: U.S. Female Children Who Reported
Consuming Tuna Fish in a Cross-Sectional Survey With a Three-Day Sampling Period
Enlargement of upper 5 percent
RfD Range (0.07 - 0.3)
RfD Range (0.07 - 0.3)
Min
Percentile of U.S. Fish Consuming Population*
\
* Estimated Iron) fish consumption data in CSi-U 89/91 survey.
Nole: The pcrccniilc indicates the fraction of the surveyed population with an estimated meihylmercuiy cxposuie < the value shown on the Y-axis.
(•'or example, al the 25th pcrccniile, 25 percent of ihe surveyed population had an estimated melhylmcrcury exposure < 0 11 |Jg/kg HW/day.
lJ5lh Max
Page 34
-------
Dietary Exposures to Selected Metals and Pesticides
David L Macintosh, John D. Spongier, Haluk Ozkaynak, Ling-hui Tsai, and P. Barry Ryan
Environmental Science and Engineering Program, Department of Environmental Health, Harvard School of Public Health,
Boston, MA 02115 USA
il
estimated for ayytv
120«000
£ romluningdata on annual diet, as measured by a food frequency questionnaire.
^with contaminant readne data for table^ready foods due were collected as pare of die annual
t US. Food'and Drag Adminisiratioa Total Diet Study. The contamina
i 'rarfminni. lead, mercury), duee «
.). ?~f f^'^rM^i*. pm^Mfc. (JiMri^ JL/-DPE, Eft.
•; dane, heptadiioa,eDaxidc). Dietary expoeores to
i included in dw anaiy*
opaospfaaie pesticide*
Pearson sooc-
. For some of
~1.~~. o~ffix-^. ^-g^i fi~- ft T«
; to bare dietary intakes in
risk aHTMiiM-iic or epide
be evaluated bjr
biological
enablished by die EPA. Before use for
, however, die validity of die exposure estimates most
tdtcaton of chronic exposure. Because of their low
detection rate in table^ready foods,, die '~-ri"im<~< dutributions of exposures for dieldrin, p,f'-
, DDE, heptachlor epoxide; K^»*^i J?«»i«>m; and. cbJorpyrifos were found to be semitire to
assumed falnes foVnondetect samples,. Reliable rtrimatrs of die population distribution of
, &•***$'tsfamm^ m'm«»Mli^yf^Mmin«i«. ^.imor K. m«ti» qnmeqdy, An* ta dwCT low rate of
'- detection uu tahW fnyrgjfeods.Mongnrinf; proframs dial use more sensiaTe itndy designs and
ud. IM «i pcuMicy, nc^nttnR roGucD*
•ire, haffmccadaS
Ui Ptnp* 104002-205 (19%) _, ^
'•-•-''• --^
Hie consumption of foods containing envi-
ronmental concuninancs is 2 poocnoally siff»
nificant source of human exposure to
numerous metals and pesticides (1-3). The
U.S. Environmental Protection Agency
(EPA), Food and Drug Administration
(FDA), and Department of Agriculture
(USDA) are responsible for ensuring the
safety of the U.S. food supply. These agen-
cies, in conjunction with state and local pub-
• lie health departments, meet this obligation
by conducting various food-contaminant
monitoring programs and assessment activi-
ties (2,4,5). Nevertheless, current assessments
of dietary exposures co food contaminants
for the U.S. population are limited by
incomplete information on diets of individu-
als and residue levels in foods (7,6). Little is
known about die variability of dietary expo-
sures among individuals, and hence health
risks, that arises from different diets or differ-
ent contaminant levels in foods. The degree
of uncertainty about the dietary intake esti-
mates published by the EPA and FDA is also
poorly understood. Consideration of vari-
ability and uncertainty is a fundamental
component of effective environmental health
management strategics (7} and is a key ele-
ment of a federal bill designed to standardize
exposure and risk assessments (
-------
Articles • Exposure to environmental pollutants in food
within each of four geographic regions and
sent co an FDA laboratory in Kansas Citv,
where the three samples of each food item
from each region are composited, prepared
for consumption, and subsequently ana-
lyzed for contaminant and nutrient levels
(75). This design yields four residue values
(one for each region) for each of the 234
foods each year. The FDA reports residue
concentrations in one of three ways: those
that exceed the limit of quantification
(LOQ) are quantitatively reported, chose
below the LOQ but above the limit of
detection (LOD) are quantitatively report-
ed and identified as "trace,'' and those that
are below the LOO are reported as "not
detected" (16).
We assumed that the mean concentra-
tion for a given contaminant in each of
the 131 foods listed on the food frequency
questionnaire represented the average con-
centration of the contaminant that a per-
son would be exposed to after repeated
consumption of portions of that food over
a year. Residue data for individual samples
•collected from the 1986-1991 Total Diet
Studies were obtained from Technical
Assessment Systems, Inc., an EPA con-
tractor working on dietary exposure issues
(17). Mean residue levels in each of the
234 foods were estimated by contaminant
from all of the concentrations measured
over the entire 6-year period and were
subsequently matched to the foods on the
diet questionnaire. The standard error of
the mean was also computed and was
assumed to be a measure of the uncertain-
ty about the true but unknown mean
residue level. A statistical analysts of the
foods for which all the samples (n = 24)
were above the LOD indicated that the
contaminant levels were approximately
lognormally distributed. Therefore, the
contaminant concentrations in all foods
were assumed to follow a lognormal distri-
bution for purposes of estimating the
mean and standard error of the residue
levels. One of three methods was used to
characterize the uncertainty about the
mean residue concentrations based on the
fraction of food samples that contained
detectable residue concentrations.
For foods for which all the sample con-
centrations were greater than the LOD, the
mean concentration was computed direct-
ly, and the variance of the estimated mean
residue concentration (the square of the
standard error) was computed as described
by Gilbert (18) for lognormally distributed
random variables. This method takes
account of the fact that the distribution of
sample means obtained from repeated sam-
ples of small size from a skewed distribu-
tion will be asymmetric
For foods for which at least 50% but
not all the sample concentrations were
greater than the LOD (12
-------
Articles • Macintosh et at.
x recision of Estimated Contaminant
Exposures
The precision of die estimated contami-
nant exposures was investigated by quanti-
fying die uncertainty about die residue and
food consumption components of dietary
exposure and using analytical methods to
propagate die uncertainty about die inputs
through to the estimated exposures.
Uncertainty about food consumption rates
was estimated by using data collected dur-
ing a validation study of die 131-item diet
questionnaire (11,22). In 1986,127 partic-
ipants in the HPFS completed food fre-
quency questionnaires 1 year apart and
completed two 1-week diet records 6
months apart during the intervening year.
The mean daily consumption of <-q«-h food
was computed from the diet records for
each participant and was used as the true
measure of average daily food-specific con-
sumption over the year. The daily average
consumption rate reported on the second
questionnaire for each food, which is
designed to represent consumption over
the preceding year, was regressed against
die corresponding mean consumption rate
determined from die diet records:
(1)
where i • number of subjects; 1.....127;
DR.{ * mean consumption rate estimated
from die diet records of person /' (servings/
day); FPQ .» mean consumption rate
reported on die food frequency question-
naire by person i (servings/day); e - a nor-
mally distributed random variable with
mean zero and standard deviation of the
regression residuals, die root mean square
error (RMSE; servings/day)'. The RMSE
for each food was used to characterize the
measurement error or uncertainty about
die true daily consumption rate of the food
1« SO 84 90 3U 99.99
«rty(%]
Figure 1. Distributions of dietary exposures to 4
metals estimated for 78,882 adult females in 199O
Circles indicate intakes estimated for a typical
adult female by previous researchers (23).
relative to that reported on the diet ques-
tionnaire.
The uncertainty about the dietary
intake of die 11 contaminants for a hypo-
thetical individual who consumed the
mean amount of each of die 131 foods was
calculated by decomposing che dietary
exposure variance into the absolute and rel-
ative contributions of die variances of die
food-specific consumption and residue val-
ues (22); the covariance between the con-
sumption and residue values was assumed
to be negligible. For example, die variance
(Le., uncertainty) of die mean arsenic expo-
sure (s2^) was computed as follows:
J ,-{J ~"
-ic,
U)
where, u 2;, = square of che mean daily
consumption (/) of food j (servings/day)2;
s~£ = variance of die mean As concentra-
tion (C) in food j ((ig/serving)2. |a '2C =
square of die mean As concentration (C)
in food j (ug/serving)2; s2r = square of
die RMSE of die consumption estimate (/)
from the regression equation for food j
(serving/day)2.
Tabto 2. Summary of estimated dietary exposure tug/day) distributions for 78,882 adult females and 38,075
males in 1990*
Gender Chemical
Mean
GM
GSO
Min
Max
Rt(r')
Female
Mahi
Arsenic
Cadmium
Lead
Mercury
Chlorpynfos
Oiazinon
Malathion
AP'-DDE
Dieldrin
Heptachlor epoxida
Undane
Arsenic
Cadmium
Lead
Mercury
ChJorpynfos
Oiazinon
Malathion
AP'-DOE
Oieldrin
Heptachlor epoxida
Undane
50.6
18.5
14.9
3.2
0.8
IS
5J
U
0.5
-------
Articles • Exposure to environmental pollutants in food
Results
The distributions of dietary exposures ro
11 food contaminants estimated for NHS
participants who completed valid diet ques-
tionnaires in 1990 are summarized in
Figures 1-3, and summary statistics for
both the 1990 NHS and HPFS partici-
pants (1990 responses) are presented in
Table 2. All of the distributions were
approximately lognormaily distributed; the
r2 between the natural log of the estimated
exposures and the corresponding z-scores
was greater than 0.95 for all (£0.99 for
most) cohort-chemical combinations.
Differences between intakes estimated for
the male and female cohorts were small.
Summary statistics of the estimated expo-
sure distributions based on the 1986 diet
questionnaires from both males and
females were within 3% of the values
developed from the 1990 food consump-
tion data. Individual dietary exposures to
each compound were estimated to range
over two to three orders of magnitude.
Pearson correlation coefficients
between predicted exposures to the 11 con-
taminants were estimated from the 1990
results for the NHS and HPFS participants
(Table 3). Because of the large sample size
for each cohort, all of the estimated correla-
tion coefficients were highly statistically
significant. There was little difference
between the correlation coefficients esti-
mated for males and females. Pairwise cor-
relation coefficients among all of the metals
were at least 0.4 and reached as high as
0.83 between arsenic and mercury, most
likely because the highest concentrations of
both chemicals are found in fish. Pairwise
correlation coefficients between contami-
nants in the respective pesticide groups
ranged from 0.3 to 0.72, indicating that an
individual highly exposed to one of these
compounds is likely to also be highly
exposed to others.
Using rhe error propagation technique
described previously, che uncertainty about
the mean daily dietary exposure to arsenic,
cadmium, p,p'-DDE, lead, malathion, and
mercury was estimated for a hypothetical
member of the 1986 HPFS cohort who
consumes the mean amount of each food
per day reported, on the diet questionnaires.
The other five chemicals (chlorpyrifos,
diazinon, dieldrin, lindane and heptachior
epoxide) were excluded because, as
described later, the population exposures
for these chemicals were dominated by
food items for which the majority of sam-
ples contained residue concentrations
below the LOD. Therefore, we believe that
uncertainty about dietary exposures to
those chemicals is dominated by a lack of
knowledge about their true average concen-
trations in food.
The coefficient of variation (CV), com-
puted as the square root of the estimated
variance (Eq. 2) for a contaminant divided
by the mean exposure for the contaminant,
ranged from 21% for cadmium to 49% for
malathion, indicating that the exposures to
these chemicals estimated for a given indi-
vidual may be accurate to within approxi-
mately a factor of 2 (Table 4). Lack of data
about the am»gl amount of food consumed
accounted for at least 80% of the total
uncertainty for arsenic, cadmium, mercury,
and malathion. Individual food items con-
tributing most to uncertainty for these
chemicals were, for arsenic, fish, canned
tuna, and shrimp; for cadmium, spinach,
coffee, lettuce, nuts, potatoes, and assorted
beverages; for mercury, canned tuna and
other fish; and for malathion, white and
dark bread. The source of uncertainty
about exposures to p,p'-DDE and lead was
approximately equally split between a lack
of data about consumption rates and
residue levels. The foods contributing most
to uncertainty about the p,p'-DDE esti-
mate were whole milk, spinach, and beef.
Table 3. Estimated Pearson correlation coefficients Between dietary intakes of 11 food contaminants for
78.882 adult females (upper right) and 38,075 adult males (lower left)*
As
Cd
Pb
Ch
Oz Mai
DOE
Dial HCH Hep
As
Cd
Pb
Hg
Ch
Oz
Mai
ODE
Diel
HCH
Hep
036
0.71
OM
0.22
0.31
0.06
0.14
0.27
0.04
0.19
039
157
037
055
0.70
0.41
0.50
0.68
0.34
0.60
069
06«
U4
0.40
0.50
0.17
0.32
0.45
0.25
0.46
OS3
(X40 :
0*? \
0.24
0.30
0.09
0.17
0.25
0.06
0.21
074
054
0.41
026
061
036
030
OS4
078
03
0.31
069
052
0.32
069
047
0.46
063
0.43
077
0.08
039
070
0.12
038
0.48
073
030
072
0.36
073
055
0.40
0.27
035
0.51
075
052
039
063
033
064
048
030
051
062
0.27
052
035
072
0.05
031
076
0.07
0.27
0.40
071
035
028
051
074
059
052
078
060
079
0.36
062
O69
045
Abbreviations: Ch, chlorpyrifos; Oz, diazinon; Mai, malathion; DOE, p.p'-OOE; Dial, dieldrin; HCH, lindane
(hexachlorocyclohexane); Hep, heptachior epoxide.
'Correlation coefficients between the contaminants comprising the metal, organophosphate and
organochlorine groups are shaded and all coefficients greater than 0.5 are in bold type.
while those for lead were canned tuna,
skim milk, peaches, coffee, and white wine.
To investigate the uncertainty about
the estimated mean daily exposure to these
chemicals among the study population,
uncertainty about food consumption was
assumed to be negligible (reflecting the
large sample from which the average con-
sumption rate of each food item was
obtained). In this scenario, the coefficient
of variation ranged from 7% for arsenic to
30% for lead (Table 4) and varied inversely
with respect to the percentage of the popu-
lation exposure composed of foods for
which at least half of the Total Diet Study
samples were above the LOO.
Discussion
We assessed average daily dietary exposures
to 11 food contaminants for approximately
120,000 U.S. adult males and females.
Because of the large sample size and geo-
graphic diversity of the study population,
we believe the results are generaiizable to
die majority of die U.S. adult population.
However, because of the age and occupa-
tional restrictions on admission to the
NHS and HPFS cohorts and the potential
for correlations between demographics and
diet, the results may not be representative
of dietary exposures to members of certain
age, ethnic, sodoeconomic, and other sub-
populations.
The residue levels used to estimate the
dietary exposures presented here were based
solely on the results of the 1986-1991
Total Diet Studies. While information
from more recent Total Diet Studies were
not used due to our inability to easily access
the data, summary reports indicate that
contaminant levels in food have been rela-
tively constant in recent years (5). The
Total Diet Study data represent the best
available measures of contaminant levels in
table-ready foods; however, they are subject
Table 4. Summary of uncertainty about estimated
dietary exposure to six contaminants in food for a
person who consumes the mean amount of each
food reported on the 1986 HPFS diet question-
naires*^
* *
CV{%) Residue Consumption CV*(%)
Arsenic
Cadmium
AP'-OOE
Lead
Malathion
Mercury
41
21
35
43
49
44
3
19
34
50
4
5
97
81
66
50
96
95
7
9
20
30
10
9
'Estimated coefficient of variation (CV) about
average daily intake (ug/day): the relative contri-
bution of uncertainty about mean residue concen-
trations and food consumption patterns*, and the
CV assuming uncertainty about food consumption
is negligible (CV*).
a ffld Unmhar *) PohniatV 7.436
205
-------
Articles • Macintosh at al.
i the limitations of relatively small sample
sizes (n = -i tor each food item each year,
based on a composite of three items per
sample). In the Total Diet Study, samples
are obtained from grocery stores and tius
may not be representative of homegrown
toods or those available at specialty retail
outlets such as farmers' markets and organic
grocers. We assumed that any systematic
regional or seasonal effects on contaminants
levels in food and individual diets were neg-
ligible. In reality, seasonal and regional vari-
ability of food consumption rates, possibly
due to availability and prices, and residue
levels, from nonuniform pesticide applica-
tion rates, for example, may be important
contributors to interindrvidual variability of
dietary exposures. Currently, there are
insufficient data to investigate die veracity
of the simplifying assumptions made in our
analysis. However, our estimates likely
underestimate the actual variability of
dietary exposures to these chemicals among
U.S. citizens, as well as die degree of uncer-
tainty about the true exposure distributions.
Nevertheless, this work is believed to be
the first attempt to characterize the
interindividual variability of exposures to
contaminants in food. Numerous other
•esearchers, primarily affiliated with the
FDA, have estimated typical intakes of the
11 contaminants considered here. The
dietary exposures estimated for a typical
female adult by previous researchers are
indicated in Figures 1-3. The typical
intake for arsenic estimated elsewhere is
well within the distribution estimated from
die diet questionnaire (23). The ra<4mhim
and lead intakes estimated by Gunderson
(23) lie in die upper tail of the NHS distri-
bution, whereas the mercury estimate is
equivalent to the 9th percentile of the
NHS distribution. In general, typical pesti-
cide exposures estimated by the FDA fall
well into the lower tail of the correspond-
ing NHS distribution, except for the
malathion and p,p'-DDE estimates, which
correspond to the 37th and 21st percenriles
of die estimated distribution, respectively.
Differences between the typical esti-
mates made previously and those based on
me questionnaire are most likely due to me
treatment of residue values that were not
detected; the FDA estimates were based
solely on foods for which trace or quantifi-
able residue concentrations were measured
(i.e., nondetection samples were set to
zero), while in our analysis, residue levels
cor food samples that contained nonde-
tectable concentrations were set to 1/2
LOD. Comparing the typical pesticide
intakes estimated by the FDA to the maxi-
mum values obtained from die distribution
estimated for the 1990 NHS cohort shows
that the combination of using average food
consumption patterns ror a subpopuianon
and setting nondetection values to zero may
underestimate exposures for some members
of die population by a factor of 10-60.
Sensitivity of Results
The sensitivity of the predicted exposure
distributions to treatment of residue sam-
ples below the LOD was investigated by
estimating the distributions for the 1990
NHS respondents under cwo cases in addi-
tion to die current estimates: 1) nondetect
samples set to zero and 2) nondetect sam-
ples set to the LOD (Fig. 4). The EPA
Integrated Risk Information System (IRIS)
was queried on 8 November 1994 for
human health-based exposure standards for
die 11 chemicals included in our analysis.
We retrieved cancer potency values f^,"
(mg/kg/day)"1] for ingesrion of compounds
treated by the EPA as carcinogenic in
humans and reference doses (RfD;
mg/kg/day) for ingestion of the noncar-
cinogens (Table 5). Exposures exceeding
die RfD and die level estimated to produce
an excess lifetime cancer risk (ELCR) of
10"* are compared to the predicted dietary
exposure predictions in Figure 4. The orig-
inal exposure estimates (ug/day) were con-
vened to units of uz/kg/day by assuming a
uniform body weight of 65 kg for adult
The estimated exposure distributions
for cadmium, lead, arsenic, mercury, and
malathion were relatively insensitive to dif-
ferent assumptions about the true concen-
tration in foods with residues below the
LOD. Note the divergence of the lower
end of die predicted mercury distributions
due to diets primarily composed of foods
containing mercury levels below the LOD.
Assuming inorganic arsenic accounts for
10% of all arsenic in foods (24), a substan-
tial fraction of die population was estimat-
ed to have dietary exposures to inorganic
arsenic that exceed the RfD (13% of the
population) and a 10"4 ELCR (80%). The
validity of this model result must be evalu-
ated by future research of arsenic speciation
in different foods.
The predicted dietary exposure distrib-
utions for die remaining six contaminants
were sensitive to die treatment of nonde-
tect residue samples (Fig. 4). For example,
median exposure estimates ranged by a fac-
tor of 2 for diazinon to a factor of 10 for
heptachlor epoxide. This degree of uncer-
tainty is of apparent little consequence in
some cases, such as chlorpyrifbs, where all
predicted exposures are well below the
health-based RfD. In contrast, the fraction
of die population predicted to be exposed
to dieldrin at levels estimated to produce
an ELCR greater than 10"1 raneed from
approximately 10% to 85% over the three
cases, while the traction of estimated hep-
tachlor epoxide exposures above the RfD
and equivalent to an ELCR greater than
10"4 ranged from 0 to approximately 20%.
These results indicate that the number of
individuals predicted to bear health risks
above what may be considered a tolerable
level can change by tens of millions
depending on assumptions made about die
contaminant concentrations in foods with
residues below die LOD. This uncertainty
can only be resolved by additional moni-
toring efforts that use a more sensitive
design, which should be a priority for
future research.
The sensitivity of the error analysis
results to assumptions made regarding\the
uncertainty about mean residue levels in
food items for which less than half of the
samples were above the LOD was also
investigated. The error analysis was repeat-
ed after estimating the variance of the
mean residue concentration in- food items
for which less dian half of die samples were
above the LOD based on die mean GSD
rather than the 90th percentile GSD for
metals and pesticides, respectively. The
results were found to be virtually identical
to diose presented above.
Excluding water from the analysis is
not expected to have a substantial impact
on our results because of die generally low
levels of contaminants in drinking water.
For example, arsenic is typically present in
drinking water at approximately 2 ug/1
(25), which at a consumption rate of 2
I/day would increase the mean arsenic
exposure of approximately 55 ug/day by
about 7%. Similar results were estimated
for cadmium, mercury, and die pesticides
based on typical tap water concentrations
published in the literature (26-29). The
EPA (3ff) estimated that drinking water
supplies used by 99% of die U.S. popula-
tion contained lead levels less than 5 ug/1.
Assuming a typical lead concentration of 3
ug/1, water consumption may account for
an additional 6 ug/day of dietary exposure,
nearly 40% of the 15 ug/day estimated
from food alone.
Health risk assessments of environmen-
tal contaminants are typically based on
exposure or dose rates expressed on a body
weight basis (e.g., ug contaminant/kg body
weight/day). The exposure rates presented
here were expressed on a mass per day basis
(ug/day), which may limit their utility for
use in risk assessments due to concerns
about correlations between food consump-
tion and body weight. However, Wtllett et
al. (31) found that neither height nor body
weight were significantly correlated with
206
Volume 104, Number 2. February 1996 • Environmental Health Perspectives
-------
Articles • Exposure to environmental pollutants in food
total caloric intake among a group of 194
adult women who completed four 1-weck
diet records over a year. Physical activity is
a major determinant of energy intake, and
metabolic efficiency may be a minor deter-
minant (32). Investigations of potential
' correlations among these factors, body
weight, and the types of foods consumed
by individuals are likely co yield results char
would be more useful for environmental
risk assessments.
Determinants of Dietary Exposure
To better understand the sources of dietary
exposures to the contaminants considered
in our analysis, we determined die relative
contributions of the 131 food items on the
food frequency questionnaire to the mean
exposure estimated for individuals compos-
ing the upper and lower deciles of the
exposure distributions estimated for the
NHS participants who -completed valid
diet questionnaires in 1990.
For individuals in the first decile of the
arsenic exposure distribution, canned tuna
(28%), chicken (11%), and white rice
(10%) were estimated to be the principal
contributors to exposure, while arsenic
exposures for individuals at the upper end of
die distribution were due to frequent con-
sumption of fish (92%). Osmium expo-
sures for individuals at cither end of the esti-
mated exposure distribution were due to
consumption of liver (10%), potatoes (8%),
spinach (8%), iceberg lettuce (7%), and
pasta (5%), with no single food making a
particularly large contribution. For lead,
canned tuna was estimated to be the princi-
pal contributor to the average exposure
among individuals in the first (11%) and
tenth (34%) deciles; exposures in die upper
decile were marked by more frequent con-
sumption of canned tuna and other fish.
Dietary exposures to mercury at die upper
end of die estimated distribution were dom-
inated by consumption of fish products
(87%), principally canned tuna (65%).
Table & IRIS exposure standards for the 11 chem-
icals analyzed
Chemical
RfD
(lig/kg/day) 9,
Arsenic (inorganic) 03
Cadmium
Mercury
Lead
Chlorpyrtfos
Oiazinon
Malathion
ftp'-OOE
Oieldnn
1.0
(under review)
_
3.0
(under revew)
20.0
_
0.05
Heptachlor epoxide 0.013
lindane
OJ
•tug/kg/dayi"
0.00175
__
—
_^
—
—
_
0.00034
0.016
0.00091
—
"— Zero
— U2LOO
— LCD
— ELCH !=!»-»>
— RFO
3E-2
IE-2
3E4
IE-2
364
16-4
3E-S
JE-2
2E-2
1E-2
SE-3
2E-3
IE-3
5E-4
2E-4
3E-1
lE-t
3t-2
1E-2
3E-3
IE-3
S-l
2E-I
IE-!
5E-Z
lE+t
IW
1E-1
1E-2
IE-3
OieUfw
5E-2
2E-2
IE-2
5E-3
2E-3
IE-3
5E-4
2E-4
3E-2
IE-2
3E-3
IE-2
3E-4
IE-4
3E-5
5E-2
2E-2
IE-2
5E-3 ?
2E, |
IE-3 o.
5E-« 5=
2E-4 S
3T
m
3E-1 ^
It, |
3E 2 *
1E-2
3E-3
IE-3
3£-t
1E»0
SE-I
2E-1
1E-I
5E-2
001
31 9U9 0.01 2
P*rcmt>9* oi population
SO
99.99
tE-1
IE-2
IE-3
Abbreviations: IRIS, EPA Integrated Risk
Information System; RfD, reference dose.
figure 4. Estimated average daily contaminant intakes from food among 78,882 adult females in 1990
under 3 different assumptions about the true mean contaminant concentration in foods containing
amounts below the Total Diet Study limit of detection (LOD). Average daily intakes that correspond to an
excess lifetime cancer risk (ELCR) of 10"* and the reference dose (RfD) for noncancer effects established
by ttia EPA are shown for purposes of comparison.
environments/ Hfalth Persoectives • Volume 104. Number 2. February 1996
207
-------
Articles-Macintosh eta/.
Apples were found to be the largest con-
tributor to chiorpyrifos exposures, account-
ing for 9% and 36% of the mean exposure
for die tendi decile, respectively, of die esti-
mated distribution. Diazinon exposures
were estimated to be due to consumption of
wheat-based products, such as english
muffins (11% of tenth decile) and pasta
(9% of mean for first and tenth decile).
Estimated exposures to malathion were also
found to be dominated by consumption of
wheat-based products, although die specific
items contributing most to exposure at bodi
ends of the distribution were white and
dark breads (24% of 1st decile mean and
65% of 10th decile mean).
Beef-related foods (15%) were estimat-
ed to be die principal contributors to />,/>'-
DDE exposures at the lower end of me dis-
tribution, while consumption of whole milk
(42%) dominated exposures in the tenth
decile. High levels of dietary exposures to
dieldrin were estimated to be primarily due
to frequent consumption of summer and
winter squash (38%), while diose at the low
end were dominated by foods that con-
tained residue levels below the LOD.
Residue concentrations in all but one of the
principal contributors to both high and low
dietary exposures to heptachlor epoxide
were set to 1/2 LOD, indicating that little is
actually known about the magnitude of
exposure to diis pesticide. High exposures
to lindane were estimated to be due to fre-
quent consumption of chocolate (48% of
mean for 10th decile).
Correlations among Contaminant
Exposures
Exposures to selected contaminants were
estimated to be positively correlated within
individuals, which has implications for
understanding che full public health
impacts of exposures to contaminants that
have the same toxicological effect (e.g.,
cholinesterase inhibitors and carcinogens).
It seems reasonable to sum the exposures of
toxicologically identical compounds in
environmental health assessments. In a
population-based assessment, simply sum-
ming die correlated exposure distributions
of toxicologically identical compounds
would underestimate die true variability of
total exposure among the population, as
shown in the following example.
The distribution of total exposure to
die seven contaminants (Cd, Pb, Hg, />,/>'-
DDE, dieldrin, lindane, heptachlor epox-
ide) included in our analysis diat have been
identified as hormonal agonists (33) was
simulated for the 1990 NHS cohort using
the summary statistics of the lognormal
distributions, shown in Table 2, and the
correlation coefficients, presented in Table
3. Simulations consisting of 10,00,0 trials
were conducted with and without correla-
tions among contaminant-specific expo-
sures for individuals. When correlations
were not considered, the standard deviation
(10.9 ug/day) of total exposure among
individuals underestimated the standard
deviation (16.0 fig/day) from die simula-
tion that did consider correlated exposures
by nearly 50%, and die 95th percentile of
total exposure was underestimated by 15%.
The true number of food contaminants
with identical or similar toxicological
effects is likely to be much greater than
seven. For example, the EPA lists 34 com-
pounds in the class of cholinesterase-
inhibirihg pesticides (4), and Colborn et aL
(33) identified 45 ubiquitous endocrine-
disrupting substances. Because of the
potential for intraindividual correlation
among dietary exposures to all of die mem-
bers of these two classes of contaminants,
simply summing contaminant-specific
exposure distributions to estimate the total
exposure to these contaminant classes is
likely to produce greater underestimates of
exposure for individuals in die upper end
of the joint exposure distribution than diat
observed in this simple example.
Conclusions
Food consumption data collected as part of
two large prospective epidemiologic studies
and contaminant residue data collected as
part of the FDA Total Diet Study were
combined to estimate the distribution of
average daily dietary exposures to 11 food
contaminants for a large population of U.S.
adult males and females. The estimated dis-
tributions of dietary exposures were shown
to be comparable to point estimates made
by other researchers, indicating that the
food frequency questionnaire-based
approach produces reasonable results.
Exposures were estimated to be highly vari-
able among individuals, spanning two to
three orders of magnitude, indicating that
it is important to examine the range of
dietary exposures when considering the
public health risks of food contaminants.
Intraindividual exposures to the 11 conta-
minants were estimated to be strongly cor-
related, which has implications for assess-
ing the full public health impacts of expo-
sures to contaminants that have the same
toxicological effect.
For all of die chemicals included in our
analysis, except heptachlor epoxide, expo-
sures estimated at the upper end of the
respective distributions were due to con-
sumption of food items that contained
measurable levels of the contaminant; diat
is, relatively reliable residue concentrations.
This finding suggests diat it may be possi-
ble to use dietary exposures estimated from
diet questionnaires in epidemiologicai stud-
ies. The correlation coefficients presented
earlier suggest that a principal components
or factor analysis may reveal types or
groups of food items that together deter-
mine die approximate level of an individ-
ual's exposure to compounds with similar
toxicological action, which could also be
used in epidemiologicai studies. Prior to
such analyses, we recommend that a statis-
tically designed study be conducted to vali-
date these estimates by comparing biologi-
cal indicators of exposure to die estimated
exposures.
The estimated exposure distributions
for some of the compounds were shown to
be sensitive to valuation of the nondetect
residue samples because of- the low detec-
tion rate observed in the Total Diet Study
data for diese chemicals. We selected die 11
contaminants considered in this analysis
based on dieir relatively high detection rates
in die Total Diet Studies conducted from
1986 to 1991; dius, most odier contami-
nants were detected less frequendy. It can
be inferred that estimates of dietary expo-
sures to many odier food contaminants will
also be sensitive to the treatment of nonde-
tect samples and the determinants of dietary
exposure to these contaminants may not be
readily identifiable. Therefore, we recom-
mend that new monitoring studies be con-
ducted that use a study design more sensi-
tive than that employed by the Total Diet
Study. In addition, we recommend that
population-based estimates of exposures to
contaminants in food be conducted for
odier subgroups of the United States, such
as children and ethnic populations.
REFERENCES '
1. National Research Council. Pesticides in che
diets of infants and children. Washington,
DCNarional Academy Press. 1993.
2. Berry MR. Strategy for a dietary exposure
research program. J Exp Anal Environ
Epidemiol (Suppl) 1:97-110(1992).
3. Whitmore RW, Immerman FW, Camann DA.
Bond AE, Lewis RG, Schaum JI_ Non-occupa-
tional exposures to pesticides for residents of
two* cities. Arch Environ Contam Toxicol
26:47-59 (1994).
4. Environmental Protection Agency. Code of
Federal Regulations, 40, Pan 180. Washington.
DOOffice of the Federal Register National
Archives and Records Administration, 1994.
5. U.S. Food and Drug Administration. Food and
Drug Administration pesticide monitoring pro-
gram—1993. J AOAC Int 77(5):163A-185A
(1994).
6. Blair D. Uncertainties in pesticide risk estima-
tion and consumer concern. Nutr Today
November/December 13-20 (1989).
7. Sexton K. Callahan MA. Bryan EF, Saint CG,
Wood WP. Informed decisions about protect-
208
Volume 104, Number 2, February 1996 • Environmental Health Perspectives
-------
Articles • Exposure to environmental pollutants in food
ing and promoting public health: rationale for a
national human exposure assessment survey. J
Exp Anal Environ Epidemiol 5:233-256
(1995).
8. Job Creation and Wage Enhancement Act of
1995. Title III—Risk assessment and cost/ben-
efit analysis for new regulations. Report 104-33
Part I. HR 9. Washington, DC:104th
Congress, 1st Session. 1995.
9 Coidirz GA. The Nurses Health Study: find-
ings during 10 years of follow-up of a cohort of
U.S. women. Curr Prob Obiter Gvnecol Feral
13(4):131-174 (1990).
10. Rimm EB, Stampfer MJ, Ascherio A,
Giovannucci E, Colditz GA. Willett WC.
Vitamin E consumption and the risk of coro-
nary heart disease in men. N Engl J Med
328:1450-1456(1993).
11. Rimm EB, Giovannucc! EL, Stampfiar MJ,
Colditi GA. Utin LB, Willett WC.
Reproducibility and validity of an expanded
self-administered food frequency questionnaire
among male health professionals. Am J
Epidemiol 135:1114-1126(1992).
12. Willett WC. Nutritional epidemiology.
London:Oxfbrd University Press, 1990.
13. Sampson L Food frequency questionnaires as a
research instrument. Clin Nutr 4:171-178
(1985).
14. Pennmgton J, Gundenon EL. History of the
Food and Drug Administration's Total Diet
Study—1961 to 1987. J AOAC Int
70(5):772-782(1987).
15. Pennington J. The 1990 revision of the FDA
Total Diet Study. J Nutr Educ 24(4):173-178
. (1992).
16. Gartrell MJ, Craun JC, Podrebarac DS,
Gundenon EL Pesticides, selected elements,
and other chemicals in adult total diet samples.
October 1980-March 1982. J AOAC Int
69:146-161 (1986).
17. Tomeriin JR, Kidwell JL, Tucker KD. Chew S.
Lee KH. Dietary exposure assessment. Phase il
report, vol 1. Development of summary residue
database for dietary exposure assessment.
Washington, DQTechmcal Assessment
Systems, Int, 1993.
18. Gilbert RO. Statistical methods for environ-
mental pollution monitoring. New YoricVan
Nosoand Reinhold. 1987:167-168.
19. Haas CN, Scheff PA. Estimation of averages in
truncated samples. Environ Sci Technol
24(6):912-919 (1990).
20. Hunter OJ, Sampson L Stampfer MJ, Colditz
GA, Rosner B, Willett WC Variability in por-
tion sizes of commonly consumed foods among
a population of women in the United States.
Am J Epidemiol 127(6):1240-1249 (1988).
21. Feskanich O, Rimm EB, Giovannucci EL,
Colditz GA, Stampfer MJ. Latin LB, Willett
WC. Reproducibility and validity of food
jfifalry measurements from a semiouanocative
food frequency questionnaire. J Am Diet Assoc
93(75:790-796 (1993).
22. Bevington PR. Data reduction and error analy-
sis for the physical sciences. New York:
McGraw-Hill, 1969.
23. Gundenon EL. FDA Total Diet Study, April
1982-Apnl 1984, dietary intakes of pesticides,
selected elements and other chemicals. J AOAC
Int71(6):1200-1209 (1988).
24. Lawrence JF, Michalik P, Tarn G, Conacher
HBS. Identification of arsenobetaine and
i ^id A»jlfi«li by
liquid cuiuiuaujgnpny witn
atomic absorption detection and conflnnat
by fast atom bombardment mass speorome
J Agric Food Chem 34(2)315-319 (1986).
25. ATSDR. Toxicological profile for arsenic.
Adanta. GAiAgency for Toxic Substances and
Disease Registry, 1991.
26. ATSDR. Toxicological profile for aldnn/dieid-
rm. Atlanta. GA:Agency for Toxic Substances
and Disease Registry, 1992.
27. ATSDR. Toxicologicai profile for mercury.
Atlanta, GAiAgency for Toxic Substances and
Disease Registry, 1992.
28. ATSDR. Toxicological profile for DDT, DDE.
ODD. Atlanta, GA:Agency for Toxic
Substances and Disease Registry, 1992.
29. ATSDR. Toxicological profile for cadmium.
Adanta. GAiAgency for Toxic Substances and
Disease Registry, 1991.
30. U.S. Environmental Protection Agency.
Drinking water regulations: maximum contam-
inant IdveJ goals and national primary drinking
water reguiaaons for lead and copper. Proposed
rule. 40 CFR Parti 141 and 142. Fed Reg
53(160):31515-31578 (1988).
31. Willett WC Sampson L, Stampfer MJ, Rosner
B, Bain C Witschi J, Henneken CH, Spexzer
FE. Reproducibility of a semiquanotanve food
frequency questionnaire. Am J Epidemioi
122(1):51-65(1985).
32. Wilier WC, Stampfer MJ. Total energy intake:
implications for epidemiologic analyses. Am J
Epidemiol 124(l):17-27 (1986).
33. Colborn T, von Saal FS, Soto AM.
Developmental effects of endocrine-disrupting
chemicals in wildlife and humans. Environ
Health Perspect 101(5):378-384 (1993).
34. U.S. Food and Drug Adminstrarion. Food and
Drug Administration pesticide program-residue
monitoring 1992. J AOAC Int 76(5):
127A-148A(1993).
0IV1ROHEALTH@NIEHS.NIH.
Environmental Health Perspectives • Volume 104, Number 2. February 1996
209
-------
May 10, 1996
NOTE
TO:
Jeanette Wiltse, Ph.D.
William H. Farland, Ph.D.
FROM: Kathryn R. Mahaffey, Ph.D.
SUBJECT: Support for Mercury Exposure through Fish
This is in response to your request for information on support for mercury
exposure through fish.
Introduction
Fish consumption was one of the issues which held up the Mercury Study Report
to Congress. In the February 1996 issue of Environmental Health Perspectives.
Macintosh et al. published data on mercury exposures through food to a cohort of United
States adult women and men. Results from this paper are shown below. A brief
summary compares these data to our findings in the Mercury Study Report to Congress.
Macintosh et al. (1996) used the food consumption patterns of a different cohort
of United States adults than was analyzed in CSFII 89/91. Also Macintosh et al. used
the FDA Total Diet Study data for mercury concentrations in food; whereas data on fish
mercury concentrations from the National Marine Fisheries Service were used in the EPA
Mercury Study Report to Congress. Despite use of these different sources the estimated
dietary exposures to mercury reported by Macintosh et al. (1996) are highly comparable
to findings in the Mercury Study Report to Congress.
Mercury Exposure
/jg/kg body weight/day
Percentile
50th
Maximum
Cross-Sectional Data from
CSFII 89/91 (Cited in the
Mercury Study Report to
Congress).
Females
0.07- 0.08*
1.67-2.76*
Males
0.07-0.08*
1.57-1.98*
Long-term Food Patterns
Data from Adult Health
Professionals Study, 1990
Data Base
(Macintosh et al., 1996)
Females
0.12
3.08
Ma/es
0.10
2.02
' Range reflects different age groups of persons 15 years of age and older.
-------
2 May 1996
Note from K. Mahaffey
In the Mercury Study Report to Congress, we compared estimates of quantities
offish ingested using cross-sectional data from the 1989/1991 (CSFII 89/91) period and
longitudinal data from the National Purchase Diary (NPD 73/74) obtained in 1973/1974.
Among adults ages 18 through 98 years, the 99th percentile fish consumption from NPD
73/74 was 112 grams per day. The CSFII 89/91 data indicated that 99th percentile
consumption for adult female subjects was 182 grams of fish per day and 99th percentile
consumption among adult male subjects was estimated at 190 grams per day. Both
sources of data indicate that approximately 1% of the overall population and
approximately 5% of the population who consume fish often enough to be identified
during a three-day period will consume 100 grams or more of fish per day.
Although fish consumption for the overall US population has. increased
approximately 25% during the two decades between the early 1970s and 1990s, we
have not attempted to adjust the NPD 73/74's 99th percentile estimates because we are
uncertain that the increase would be reflected at the highest percentiles.
Detailed Summary:
Conclusions for Macintosh et al. (1996)
In the February 1996 issue of Environmental Health. Perspectives, Macintosh et
al. published analyses of dietary exposure to pesticides and metal contaminants
measured by FDA in their Total Diet Study. The contaminants included mercury. To
estimate exposure of adults to mercury through diet, Macintosh et al. combined the FDA
data on mercury concentrations in food with estimates of food intake from the adult
women who participated in the Nurses' Health Study (NHS) and the adult men who
participated in the Health Professionals' Follow-up Study (HPFS).
The NHS and HPHS are prospective epidemiological studies that originally
included 121,700 female registered nurses who were 30-55 years of age in 1976 and
51,529 male health professionals who were 40-75 years of age in 1986. Beginning in
1986 a food frequency questionnaire evaluating the previous year's diet was included
in the study. These questionnaires are designed to measure long-term average food
intake rather than provide a precise estimate of short-term food consumption habits.
The analyses reported by Macintosh et al. were based on the NHS and HPFS
cohorts who returned diet questionnaires in 1986 and/or 1990. The total number of food
consumption records analyzed by Macintosh et al. were 75,542 and 78,882 from 1986
and 1990, respectively for the NHS cohort and 49,934 and 38,075 from the 1986 and
1990, respectively, for the HPFS cohort.
-------
2 May 1996
Note from K. Mahaffey
Based on these records, dietary intake of mercury was estimated and reported as
micrograms/day for the 78,892 adult females and 38,075 males in the 1990 cohort.
Mercury intake was largely dependent on fish and shellfish consumption especially at the
upper end of the estimated distributions. Eight-seven percent of consumption of mercury
came from fish products, with 65% from canned tuna.
Comparison with CSFIi 89/91 Data Used in EPA 's Mercury Study
The data reported by Macintosh et al. (1996) were as total micrograms of mercury
ingested per day. Although the two surveys are of a large number of adult men and
women, these individuals are representative of well-educated health professionals and
are not representative of the United States population overall. By contrast CSFII 89/91
is designed to be representative of the United States population when properly weighted.
The analyses we reported in the Mercury Study Report to Congress were weighted to
reflect the United States population.
Macintosh et al. had log-transformed their data to conduct their statistical analyses
making it difficult to compare directly with the CSFII 89/91 data until more detailed
information from Macintosh et al. is obtained. In the Environmental Health Perspectives
report, arithmetic means, geometric means, geometric standard deviations, minimums,
and maximums were reported. There is a graph plotting mercury intake against percent
of the population is provided; however, the scale is too small to identify the highest
percentiles with accuracy.
To compare the total daily intakes of mercury from food reported by Macintosh
et al. with the CSFII data, I used the average body weights for adults males (82 kg) and
adult females (66 kg) reported in CSFII. This comparison is shown in Table 1 (above).
When expressed as //g/kg body weight/day, mercury exposures estimated in these two
studies based on two entirely separate data bases for both mercury concentration in food
and food intake are within less than a factor of two.
Comparison with EPA's Reference Dose
Macintosh et al. compared the results of their analyses for the 78,882 adult female
subjects and 38,075 adult male subjects who provided the 1990 dietary data with EPA's
references doses for inorganic arsenic, cadmium, chlorpyrifos, malathion, dieldrin,
heptachlor epoxide, and lindane) and with cancer potency values for inorganic arsenic,
p,p'-DDE, dieldrin, and heptachlor epoxide. Since the mercury RfD was under review
at the time this paper was written comparison was not made for mercury.
-------
2 May 1996
Note from K. Mahaffey
In May, 1995 the revised RfD for mercury was published on IRIS as 0.1
body weight/day. Macintosh et al. (1996) reported the 50th percentile intake of mercury
from diet for the 78,882 female subjects was 0.12 j/g/kg body weight/day and 0.1 0 //g/kg
body weight/day for the 38,075 male subjects. The estimate in the Mercury Study was
0.07-0.08 //g/kg body weight/day for both females and males.
Comparison of Estimates of Total Fish Ingested Based on CSFII 89/91 and the National
Purchase Diary Data from 1973/74
The quantity of fish consumed by women of child-bearing age was estimated in
the Mercury Study Report to Congress based on CSFII 89/91 and NPD 73/74 data. NPD
73/74 recorded fish consumption for a one-month period in a large sample of households
during the period 1973/74. Since these data were based on frequency of consumption
and not on estimates of quantities offish ingested, average portion sizes were estimated
from another USDA source. Within the NPD 73/74 sample, 94% of people reported
consuming fish or shellfish at lease once in a one-month period. Within this sample, the
99th percentile consumers reported an average fish/shellfish intake of 112 grams/day.
Within the CSFII 89/91 cross-sectional data that identified 30.5% of females of child-
bearing age consuming fish/shellfish in the 3-day sampling window. Within this group
the 95th percentile consumption was greater than 100 grams/day.
Extrapolation of these data to the overall United States population of women of
child-bearing age estimated comparable numbers of pregnant women consuming 100
grams or more offish per day; i.e., 84,300 based on the 1989/91 cross-sectional data
and 51,900 based on the 1973/74 longitudinal data.
•
Conclusions
Comparison of the results of Macintosh et al. (1996) with findings of EPA's
Mercury Study report to Congress shows close agreement of mean mercury intake
between these two analyses. The highest levels of mercury consumption were within
a factor of two.
REFERENCES:
Macintosh DL, Spengler JD, Ozkaynak H, Tsai L-h, and Ryan PB (1996) Dietary
exposures to selected metal and pesticides. Environmental Health Perspectives 104:202-
209.
U.S. EPA (1996) Mercury Study Report to Congress. Volume III, Appendix H. Draft.
Saved as: c:\HgNote2.fsh
-------
Evolution of Our Understanding of
Methylmercury as a Health Threat
Chiho Watanabe and Hiroshi Satoh
Department of Environmental Health Sciences, Tohoku University
School of Medicine, Sendai, Japan
Methylmercury (MeHg) is recognized as one of the most hazardous environmental pollutants.
pnmanly due to endemic disasters that have occurred repeatedly. A review of the earlier literature
on the Minamata outbreak shows how large-scale poisoning occurred and why it could not be
prevented. With the repeated occurrences of MeHg poisoning, it gradually became dear that the
fetus is much more susceptible to the toxicity of this compound than the adult. Thus, recent
epidemiologic studies m several fish-eating populations have focused on the effects of in utero
exposure to MeHg. Also, there have been many studies on neurobehavioral effects of in utero
exposure to methylmercury in rodents and nonhuman pnmates. The results of these studies
revealed that the effects encompass a wide range of behavioral categories without clear
identification of the functional categories distinctively susceptible to MeHg. The overall
neurotoxicity of MeHg in humans, nonhuman pnmates, and rodents appears to have similarities.
However, several gaps exist between the human and animal studies. By using the large body
of neurotoxicologic data obtained in human populations and filling in such gaps, we can use
MeHg as a model agent for developing a specific battery of tests of animal behavior to predict
human risks resulting from in utero exposure to other chemicals with unknown neurotoxicity.
Approaches developing such a battery are also discussed. — Environ Health Perspect 104
-------
WATANABE AND SATOH
system. "Incoordinauon was noticed, espe-
cially in rabbits, and motor paralysis was
observed in dogs and cats. Tremors, blind-
ness, loss of the sense of smell, deafness, and
attacks of wrath on the slightest provocation
were observed in many of the dogs" (2).
The toxicity of alkylmercury compounds
had therefore already been recognized in the
19th century.
Although there had been accidental
cases of mercury poisoning and related
findings in experimental studies as men-
tioned above, organic mercury compounds
such as aryi and alkyi derivatives continued
to be used for seed dressing. "In 1940,
Hunter reported four cases of merhyimer-
cury poisoning in a factory where fungici-
dal dusts were manufactured without an
enclosed apparatus... [The symptoms were]
severe generalized ataxia, dysarthria and
constriction of the visual field" (2). The
characteristic symptoms of mercury vapor
poisoning, with the exception of tremors,
were not observed. One of die victims suf-
fered from symptoms (mainly ataxia) for 1 5
years after exposure had ceased.
"Ac [patient] necropsy, generalized »«*"
was referable to cerebfllar cortical atrophy,
selectively involving die granule-cell layer
of the neocerebelluin. The concentric con-
striction of die visual fields was correlated
with bilateral cortical atrophy around the
calcarine fissures" (2). This was originally
reported in 1954 and later methyimercury
poisoning was referred to as Hunter-Russell
syndrome. The emergence of a methyimer-
cury poisoning epidemic, Minamata dis-
ease, coincided with these yean.
Thereafter, cases of organic mercury
poisoning were reported in the United
Kingdom, die United States, Canada, and
Sweden. Since organic mercury com-
pounds were used mainly for seed dressing,
most victims were workers in chemical
manufacturing plants and farmers and
members of their families who aeriAmntty
ingested dressed seeds.
From these cases of accidental human
exposure, die health hazard of methyimer-
cury and other organic mercury deriva-
tives had been well recognized by the
1950s. Despite this awareness, however,
Minamata disease occurred in the same
decade, and later methyimercury poison-
ing surfaced in Iraq.
Minamata disease was defined as methyl-
mercury poisoning that occurred among
die people living along Minamata Bay in
Kyushu, Japan (3). The way in which die
victims became exposed to metnyimercury
was uncommon; they consumed substan-
tial amounts of contaminated fish and
shellfish. The source of mediyimercury was
effluent from a chemical company where
mercury was used as a catalyst to produce
acetaldehyde. Although methylmercury
concentration in die seawater was not high,
it was concentrated as it ascended die food
chain and thus was in die fish and shellfish
that were the staple diet of die villagers.
The concentrations of mediyimercury in
the fish were high enough to cause
methyimercury poisoning. Minamata dis-
ease is evidently unique in its origin as it
involved die bay's ecosystem.
Minamata disease was first officially
reported on 1 May 1956 to the public
healdi authority of Minamata, Kumambto
prefecture (4). During die preceding 10
days, Dr. Hosokawa, die head of die hospi-
tal that was affiliated with Chisso (the
responsible company), and his colleague
experienced two infantile cases of an
unknown .disease that resulted in death.
Since die two infants were sisters and so
severe a disease occurred in one family at
die same rime, me doctors felt due the sit-
uation required serious attention and
reported it to die public healdi audiority.
Moreover, before these two infantile cases,
diey had dealt wim sporadic occurrences of
a similar disease (5).
Abnormal gait, dysardiria. ataxia, deaf-
ness, and me constriction of me visual field
were die main symptoms (6). It was also
common to find emotional lability in die
form of euphoria or depression. Serious
cases displayed states of mental confusion,
drowsiness, and stupor. Sometimes, how-
ever, die victims were resdess and prone to
shouting, which often led into coma.
After Dr. Hosokawa's official report, a
committee to study diis serious disease con-
sisting of representatives from Minamata
City, die affiliated hospital, die municipal
hospital, and the Minamata Medical
Association was formed. This was called
die Kibyou Taisaku Committee, which lit-
erally translates as die anrimysterious dis-
ease committee. The committee found 30
cases widiin several months. The epidemic's
first case was reported in December 1953.
Then 10 and 11 more patients were
confirmed whose onsets began in 1954 and
1955, respectively (Figure 1). The progno-
sis was poor, and more than 30% of the
cases-were fatal.
In August 1956, Kumamoto University
School of Medicine was asked to join
the committee and the Study Group of
§
S
M
5
5-
-50
-30 I
-10
II I 4 I I 10 n 2 4 t I 10 IJ I 4 t I 10 12
19S3 1954 19S 1956
Rgurt 1. Onset of Minamata disease in early stage by
month. Arrow indicates the case by which Hosokawa
decided to report the mysterious disease.
Kumamoto University was organized. The
initial epidemiologic study revealed an
entire range of characteristics related to
"the mysterious disease" (7}. The disease
occurred regardless of age. Although family
clustering was observed, there was no proof
of infectious transfer. Most of the popula-
tion used wells, but wells were not involved.
Ninety percent of the victims' households
were related to fishery, whereas less than
30% were in the control group formed from
die households in die neighborhood. More
dian two diirds of die households consumed
fish caught in die bay every day in substan-
tial amounts (sometimes several hundred
grams and even up to 1 kg per person in one
meal); among die control group, only 6%
ate local fish daily and in lesser amounts.
The deadi rate of domestic cats in die vic-
tims' households was also higher than that
of die controls; during die period of 1953 to
1956, 50 cats out of 61 died in the victims'
houses whereas only 24 out of 60 died in
die control houses.
These epidemiologic findings clearly
indicated that substantial fish consumption
was the cause of the mysterious disease.
These findings also demonstrated that a
toxic agent, not a biologic one, in fish was
responsible. The study group suggested a
ban on the catching and selling of fish
from die bay although die local authorities
were opposed to this policy.
Having occurred by a unique route of
exposure, various extraordinary phenom-
ena and ecological changes preceded che
outbreak of Minamata disease; floating
dead fish and empty shellfish had been
observed a few years before. In one area of
che bay, observations of floating fish date
back to 1949 (4). Crows were also affected.
Cats that were housed in the villagers'
homes showed symptoms similar to those
manifested in human victims; the cats
showed ataxic gait, slowness, and unsteady
368
Environmental Health Perspectives • Vol 104. Supplement 2 • April 1996
-------
METHYLMERCURY AS A HEALTH THREAT
movement. Sometimes cney dashed around
in a circle and ran hysterically, che latter
causing some of them to jump into the sea
and drown. There was no evidence that
these events had been seriously recognized
by the local authorities. Witnesses were
later collected in che epidemiologic studies
conducted by the study group of
Kumamoto University School of Medicine.
Early pathologic examinations (6) of
victims suggested that the disease was
encepalopathia toxica (toxic encephalopa-
thy). Similar pathologic changes were
found in affected cats, birds, and even fish.
It is noteworthy that characteristics of
the disease were revealed and the study
group concluded by the end of 1956 that
an unidentified toxic agent in fish was
responsible for the disease.
A long period remained, however, before
the causal agents could be specified. First,
manganese was suspected in November
1956, followed by selenium in April 1957
and thallium in 1958; however, no link
between these agents and the disease could
be found in feeding experiments.
Moreover, clinical and pathologic findings
did not support any of these substances as
the causal agent (6').
Finally in 1957, organic mercury was
first suspected. In pathologic examinations
of four cases, lesions in the granular cell
layers of the cerebellum were noted (6").
Professor Takeuchi, at the Department of
Pathology in Kumamoto University con-
sulted pathologic textbooks and found that
carbon monoxide and mercury poisonings
caused such lesions. A volume that fol-
lowed the pathologic texts was published in
1958, and a chapter of this volume intro-
duced the study by Hunter and Russell
(8). The pathological changes described in
the book resembled the findings in the
cases of Minamata disease. Since die study
group was unable to chemically analyze
mercury at that time, Professor Takeuchi
and his colleagues tried and succeeded in
identifying mercury histoiogically.
Clinical observations of constriction of
the visual field and ataxia also indicated
organic mercury poisoning (9). Professor
Tokuomi of the Department of Internal
Medicine came across a clinical toxicology
book (10) in April 1957 that classified
symptoms and listed possible agents. Under
the item "ataxia." alkylmercury was named
widi odicr agents such as atropine, barbitu-
rates, etc. Alkylmercury was also listed as an
agent under "restriction of visual fields." At
this point, he almost determined the toxic
agent responsible, but the conventional
wisdom of chemistry did not support tnis
idea; because mercury, especially alkvlmer-
cury, was expensive, it did not seem logical
that such materials would be discharged.
Moreover, it was not understood how
alkylmercury could have been synthesized.
Thus, Professor Tokuomi abandoned the
idea that alkylmercury was a possible agent.
Later, when Professor Takeuchi almost
concluded that alkylmercury was the causal
agent. Professor Tokuomi once again
strongly suspected alkylmercury and
decided to reexamine die patients. In addi-
tion to making clinical observations, he
noted an increase in urinary excretion of-
mercury, while the administration of
British antilewtsite (BAL) further revealed
increased excretion of mercury in urine (9),
After the establishment of a chemical
analysis for mercury, environmental inves-
tigations directed by Professor Kitamura of
the Department of Public Health also
showed elevated mercury concentrations in
sediments near the factory waste-water out-
let (7). Mqreover, organ samples from die
victims and affected cats contained high
concentrations of mercury.
These results were reported by the
Study Group of Kumamoto University in
July 1959. Their conclusion was that
"Minamata disease occurred by eating con-
taminated fish and shellfish and organic
mercury is most suspected as the causal
agent" (6}. Though their findings seemed
conclusive, arguments about the causal
agent continued for several years, partly
because the indicated causal agent linked
legal responsibility to the company. In
addition, the methylmercury synthesis
mechanism was not clarified until 1964.
A careful review of the literature
resulted in the discovery of a paper pub-
lished in 1930 that described a type of
mercury poisoning different from typical
(metallic) mercury vapor poisoning (11).
The different type of mercury poisoning
was observed among workers in acetalde-
hyde plants who handled mercury-contain-
ing sludge. They did not have stomatitis,
which is commonly observed in mercury
vapor poisoning. The author had even sus-
pected that mercury in its organic form
could have been the cause of the poisoning.
Therefore, this different type of mercury
poisoning had already been described
before the observation of Hunter (12).
Moreover, the formation of organic mer-
cury as a by-product in the production of
acetaldehyde using mercury was also sus-
pected. Somehow, this literature was not
found by the study group.
It took a long time to reach the conclu-
sion that the organic mercurv ingested bv
fish and shellfish was the cause ot Minamata
disease. However, it was rapidly concluded
from the epidemiologic study that the dis-
ease was caused bv an unidentified toxic
agent and that fish and shellfish were
involved. Here lie the strength and limits
of epidemiology: it is not difficult to rec-
ognize a risk factor but it is difficult to
specify the causal agent. Considering the
state of analytical chemistry at that time,
it was more difficult to identify the toxic
agent than it is now. It is regrettable that
the local authorities did not prohibit
fishing in the bay in the early stages of the
epidemic of Minamata disease. The con-
clusion to be drawn after reviewing the
events that occurred at the onset of
Minamata disease is that epidemiology is
able to provide enough evidence to pre-
vent the spread of an unknown disease,
even though the specific agent involved
has not been determined.
Fetal Minamara Di^-a^y
Fetal Minamata disease was first detected in
1958 by Professor Kitamura and his col-
leagues in the Minamata Bay area (13).
They found nine infants who manifested a
severe disease resembling cerebral palsy dur-
ing dieir epidemiologic investigation. The
incidence of the cerebral-palsy-like-disease
was extremely high among infants who were
bom in and after 1955. Of 188 births in die
area during 1955 to 1958, 13 cases were
found. The incidence rate was calculated at
6.9%. Later, three more cases surfaced
involving mental retardation and minimal
neurologic symptoms. By 1974, 40 cases
were confirmed as fetal Minamata disease.
Examination of these children revealed
the following signs and symptoms in high
incidence: mental retardation, cerebeilar
ataxia, primitive reflex, and dysarthria in
all children (17/17), seizure in 82%, and
pyramidal signs in 75%. Sensory distur-
bance, constriction of the visual fields, and
hearing impairment could not be exam-
ined because of the serious conditions of
the patients.
It was a tradition in Japan to preserve a
pan of the umbilical cord that remained on
a baby after birth which later fell off.
Methylmercury concentrations in the cords
of the victims were high, and exposure to
mercury was thus confirmed.
The mothers of these children had
seemed healthy at the time their children
were confirmed to have fetal Minamata
disease. However, 11 mothers out of 15
Environmental Health Perspectives • Vol 104. Supplement 2 • April 1996
369
-------
WATANABE AND SATOH
showed slight symptoms of Minamata
disease in 1962. Later the mothers devel-
oped further symptoms, and in 1974, 57%
of these mothers experienced constriction
of the visual field, one of die typical symp-
toms of Minamata disease.
From die experience in Minamata, birth
control (actually abortion) was advised by
the local government to women of child-
bearing age who lived in die polluted area
and who had hair mercury concentrations
of 50 ppm or higher. Only one case of fetal
Minamata disease was confirmed in
Niigata by 1974 (4).
Metfayimexcurp Poisoning in Iraq
Since organic mercury compounds have
been used as seed dressings, poisoning by
eating dressed seeds (mainly wheat) have
occurred repeatedly (14). In Iraq, three
epidemic poisonings were reported: one in
1955 to 1956, another in 1959 to 1960,
and die third and largest outbreak in 1971
to 1972 (15). These outbreaks were caused
by the distribution of seed grain 'treated
with alkyimercury compouds. Rural people
consumed the grain to make homemade
bread. The total number of official victims
was 6530 including 459 4*9****-. Symptoms
were paresthesia or malaise followed by
ataxia, visual field constriction, and hear-
ing impairment.
In the investigation of the tragedy,
dose—effect and dose—response relation-
ships were established. Since there was
possibly a background incidence, a hockey-
stick model, which is composed of a hori-
zontal line and a sloped line, fitted well. In
addition, a relationship between mercury
concentrations in the hair and blood was
also established. Since mercury concentra-
tion in hair strands recapitulates die history
of metnvimercury exposure, analysis of hair
mercury provided abundant information
about the course of exposure.
Fetal Exposure to
Methyimercusy in Iraq
In die Iraqi outbreak (15-17), babies with
in Men exposure to methylmercury were
investigated for physical and mental devel-
opment. The mothers were interviewed as
well. Exposure was estimated by die peak
mercury concentration in a single hair
strand from each modicr.
A scoring system of examination results
was adopted in the investigation. Although
individual scores exhibited variability, a
dose-response relationship was found.
Statistical analysis suggested greater effects
in boys than in girls.
The data were statistically analyzed in
detail to establish a dose-response relation-
ship between the effect and the hair mer-
cury concentration (18). Both logit and
hockey-stick models were fined to die data.
From these analyses, the estimated lowest
effect level (ELEL) was proposed as a
threshold for human populations.
Recent Epidemiologkai Studies
Since fish-eating populations are exposed
to die threat of methylmercury, effects of
in utero exposure to methylmercury have
been studied (Table 1). In New Zealand,
a group with high fish consumption
(more than 3 times per week during preg-
nancy) was identified and the risk of in
utero methylmercury exposure was evalu-
ated (23}. When the children were 4 years
of age, they were tested with the Denver
Developmental Screening Test. Children
born to mothers with hair mercury levels
higher than 6 ppm had a higher preva-
lence of abnormal results. More compre-
hensive examinations were given at 6
years of age. At this age, children with
Tabto 1. Behavioral functions/domains examined in recent epidemiologie studies.
Study
Functions/domains examined
References
New Zealand
1983 Survey subjects: 4 years old; /j»31 matched pairs
(mean pregnancy maternal hair Hg of the exposed group > 6 mg/kg)
1985 Survey subjects: S years old: n»46 matched pairs
lhair Hg >10 mg/kg n-15: &-10 mg/kg n»31)
Canada
Subjects: 12-30 months old, n»234
(maximum gestational hair Hg; mean*6 mg/kg)
Faroe Islands
Subjects: 7 years old
(Pilot study showed maternal hair Hg median 4.5 mg/kg; n» 1023)
Seychelles Islands
Subjects: 6.19.26. and 66 months old
Gross motor, fine motor, language, personal-social functions with DOST,
abnormal or Questionable score increased
Visual discrimination
Sensory tests (touch, tactile, ttiermosensirjvityl
Poorer performance in W1SC-R or in Test of Language Development
Neurologic examinations
Abnormal tendon reflex (only in boys)
OOST—no effect
Vigilance-attention (frontal lobes)'
Manual motor coordination (cerebral motor system)
Mood
Tactile processing and memory (parietal lobes)
(Bender Gestalt test) Nonspecific brain damage
Short-term memory (left temporal lobe)
Attention and tracking (frontal lobes)
WISC-R Block Designs (cortcally right frontal and parietal:
subcorncally basal ganglia and white matter)
Naming (left temporal)
Reasoning and cognitive flexibility (frontal lobe)
Global-cognitive (human, monkey)"
Visual-perceptual (monkey)
Speech-language (human)
Visual memory (monkey)
Visual attention (monkey)
Neuromotor/neurologic (human)
Social-emotional {monkey)
Learning-achievement (rat)
1/9)
(20)
(21}
(22)
Abbreviations: DOST. Denver Developmental Screening Test WISC-R. Wechsler Intelligence Scale for Children-Revised. 'Brain area whose function is associated with the
performance of the behavioral task. *Data from the species provided the rationale for including the behavioral task.
370
Environmental Health Perspectives • Vol 104. Supplement 2 • April 1996
-------
METHYLMERCURY AS A HEALTH THREAT
/Imcrcury exposure performed worse
man children with less exposure, but the
/anance explained by methyimercury expo-
ure was small.
Currendy possible neurobchavioral out-
-omcs of prenatal merhylmercury exposure
:rc being evaluated in large-scale prospec-
:ive studies on human fish-eating popula-
aons. In the Seychelles (22), children up
:o 5 years old are being studied in terms of
development of cognitive functions and
more specific effects. Test items were
selected based on the preceding reports on
behavioral consequences of prenatal low-
'evei methyimercury exposure in human as
well as in nonhuman primates. In the
Faroe islands (21), a cohort of 7-year-old
children being studied. Test items were
chosen so as to cover a wide variety of
behaviors, but at the same time, maximize
:hc specificity of the evaluated functions.
Results of these studies are yet to come but
ire expected to reveal possible neurobehav-
lorai consequences of perinatal methyimer-
cury exposure.
Neurobehavioral Profile
of Prenatal Methyimercury
Toxicrty in Experimental
Animals
Past experiences have shown that fetuses are
much more vulnerable to methyimercury
exposure than adults. In the Minamara dis-
ease epidemic and in the methyimercury
poisoning in Iraq, infants were affected by
in utero exposure. Methyimercury readily
crosses the placental barrier and is trans-
ported to the developing nervous system.
Embryos and fetuses have been considered
much more susceptible to methyimercury
than adults. In Minamata, mothers with
minimal symptoms, such as numbness of
the extremities and perioral region, gave
birth to severely affected infants. Moreover,
pathologic changes observed in the patients
of fetal Minamata disease were much more
destructive, presumably because die archi-
tecture of die nervous system in die fetuses
was under development during die in utero
exposure to methyimercury. In the Iraqi
tragedies, perinatal exposure cases were
observed. How in utero exposure to
methyimercury affects postpartum life is or
interest and importance in terms of suscepti-
bility. In this section, therefore, die focus is
on the neurobehavioral consequences of in
uiera exposure to methyimercury. Because
most of the regulatory agencies required
behavioral tests in rodents, rodent and pri-
mate studies are described separately.
Studies in Rodents
Since die pioneering study of Spyker et al.
(1), a considerable number of investiga-
tions on die effects of in utero methyimer-
cury exposure have been reported. Tables 2
through 11 summarize the results of these
studies according to the test procedures
employed and according to functional cate-
gories [as proposed by Rees et ai. (49}].
These tables are an expansion of the com-
pilation by Shimai and Satoh (50).
Motor Development And Functions
(Tables 2 and 3). The reflexes of rats were
not affected except in die results of Olson
Tabl« 2. Effects of prenatal methylmercury exposure on tfie development of reflexive behavior.
•1 (strain)
MiceUCl-ICBI
Vtice(CFW)
lats(SO)
=ats (Charles River)
3ats (Hotaman)
?ats
Dosels).
nxjHg/kQ {routs)
0.4 or 4 daily (gi)
6.0x1 (sc)
4.3x4 |sc)
0.08. 0.4. or
2.0x10(gi)
2 ppm daily (diet)
0.2.or6x4
(gavage)
Period of
ouumiiauauum
60 15 to weaning
GO 9
GO 0.3. 7. and 11
GO 6-1 5
GO 0 throughout
experimental period
GO 6-9
Behaviors examined
Reflexes, cliff avoidance
.,. .— _• a j „ •
Observation of reflexes
and behavior
Righting reflexes
Righting reflexes by
dropping
Negative geotaxis
Agate) at
examination, days
4-21
1.3. and 8
Birth to 21. dairy
Birth to 28. dairy
7-17
7-10
findings
No change
Tendency of retardation but
not statistically significant
Coadrninistration of
Se counteracted
No change
No change
Retardation
No change
References
(24)
(25)
(26)
(27]
(29)
(23]
Abbreviations: gi. gastric intubation; sc. subcutaneous injection; diet food containing methyimercury compounds; GO, davls! of gestation.
Tabta 3. Effects of prenatal methyimercury exposure on swimming ability.
Animal (strain)
Mica(129/SvSI)
Vlice(JCL-ICR)
Jats(SO)
7ats (Holtzmani
H
q
Doseis).
mgHgAg (route)
5.4x1(ip)
0.4 or 4 daily (gi)
4.3x4 (sc)
2 ppm daily (diet)
0.025. 0.05. 0.5. or
5.0x4 (gavage)
Period of
ouini nau duui Kt
GO 7 or 9
G0 15 to weaning
GO 0.3. 7. and 11
GO 0 throughout
experimental period
GO 6-9
Agelslat
examination, days
31
7. 10. and 14
21
7-17
14
Findings
Impairment
Retardation in 10-day-old females
No change
Retardation
Impaired in 0.5 and 5.0 mg/kg groups
References
W
(24)
(26)
(28)
(30)'
Abbreviations: ip, intrapentoneal injection; sc. subcutaneous injection: gi, gastric intubation; diet, food containing metnylmercury compounds; GO, dav(s| of gestation.
Environmental Health Perspectives • Vol 104. Supplement 2 • April 1996
371
-------
WATANABE AND SATOH
Table 4. Effects of prenatal metnylmercurv exoosure on avoidance learning
Animal (strain!
Mice(CFW)
Mice(CFW) '
MicelCFW)
Mice(JCL-ICR)
Rats ISO)
Rats (Long-Evans)
Rats
Rats ISO)
Cosets),
mgHg/kg (route)
5x1 (gi)
2.3.or5x1fgi)
4.6oM3.5x1(sc)
0.4 or 4 dairy (gi)
8.0x1(iv)
4or6.4x1(gi)
0.0.025,0.05. 0.5. or
5x4(gavage)
6.4xl(gi| -
Penod of
administrations
GO 3
GO 8
GO 9
G0 15 to weaning
' G04
G08or15
GD&-9
G0 15
Behaviors examined
Passive avoidance
Two-way avoidance
Taste aversion
Two-way avoidance
Lever press avoidance
Two-way avoidance
Passive avoidance
'Step-down' passive avoidance
Age(s) at
examination, days
56
55
24
224-252
110-140
63
90
60
Findings
Rapid extinction
Impaired learning m
3 and 5 groups
Rapid extinction
Impaired learning in
all groups
Impaired learning
Impaired learning
No change
Rapid extinction
References
(31)
(31}
{32}
(24)
133)
(34}
(30}
(35)
Abbreviations: iv. intravenous injection; sc. subcutaneous injection: gi. gastric intubation; GO. dayts) gf gestation.
Table 5. Effects of prenatal methylmercury exposure on maze and water escape learning.
Animal (strain)
MicelCFW)
MJceUCl-lCR)
Rats(Holtzman)
Rats(Holtzman)
Oosels).
moHgAg (route)
1.2. 3. 5, or
10x1 (gi)
0.4,or4/day(gi)
2.5 dairy
(drinking water)
2 ppm dairy (diet)
Period of
administrations
60S
G0 15 to weaning
GOO to weaning
GOO throughout
expennwiiuu penoo
Apparatus or
behaviors examined
Water escape
Water T-maze
Water T-maze
Symmetrical maze
Agels) at
examination, days
56
42 and
140-147
30
60
Findings
No change
Shorter latency in
4 mgHg/kg group
Impaired teaming
Deficit in martin
meat-fed group
References
(31}
(24)
(36)
(23)
Abbreviaions: gi, gastric intubation; diet food containing metnylmercury compounds: GO. dayls) of gestation.
and Bousch (28). In mice, no change or
slight retardation was observed. The latter
was partly counteracted by co-administra-
tion of selenium. Retardation of develop-
ment of swimming ability was an important
result shown by Spyker et al. (1) and was
observed in most of the studies.
Cognitive Functions (Tables 4—6).
Most investigations showed impairment in
mice and in rats in learning a maze or
water escape.
Sensory Functions (Table 7). Most of
the behavioral studies in this functional
domain have been done on visual functions
of primates and a few have been done on
rodents. Eisner (40) trained rats to press a
lever with predetermined ranges of force
and time. The impaired performance of
methyimercury-exposed rats was consid-
ered to be a result of deficit in tactile-kines-
thetic systems.
Motivation anil Arousal Behavior
(Tables 8-10). In mice, spontaneous activ-
ities were decreased; the results were incon-
sistent in rats. Selenium supplement partly
counteracted the hypoactive effects of
methylmercury (25). In the open-field
tests, two investigations employing an iden-
tical strain of mice showed comparable
results: longer latency, decreased urination,
and increased backing. In rats, however, no
change was observed, although increased
locomotion was found when challenged
with amphetamine. Increased susceptibility
was observed in two studies although
inducing methods were different.
Social Function* (Table 11). While
three studies found slight or no effects on
ultrasonic vocalization, Eisner et al. (48),
using highly sophisticated devices, observed
significant differences between die treated
and control animals. Rats exposed co
methylmercury were found to be more
aggressive than vehicle control in dyadic
encounters (51).
Studies in Rodents: Methyimercury
as a. Model Agent
In the last decade, several attempts have
been made to evaluate various behavioral
tests as end points of prenatal neurotoxic
insults (29,30,52,53). Because methylmer-
cury was known as a typical behavioral ter-
atogen, it was included in these attempts as
a model agent. Because various aspects of
behavioral functions were examined in
each of these studies, results of these stud-
ies will be compared in terms of the toxic-
icy profile of methylmercury.
In the Collaborative Behavioral
Teratology Study (CBTS) involving six
laboratories (29), 0, 2, or 6 mg/kg body
weight (bw) of methylmercury were given
daily to pregnant rats 'at gestation days 6 to
9 (Tables 1,6,8,10). The offspring were
evaluated with a test battery covering wide
aspects of behavioral functions. Auditory
startle habituation was most consistently
affected among the laboratories. Vorhees
(52) employed the same dose regimen used
in the CBTS and evaluated their own test
battery, the Cincinnati test battery, which
emphasized reflex ontogeny. Generally, the
result coincided with those of the CBTS.
Vorhees (52) also recommended that
372
Environmental Health Perspectives • Vol 104. Supplement 2 • April 1996
-------
METHYLMERCURY AS A HEALTH THREAT
Table 6. Effects OT crenatai retnvimercurv exoosure on ooerant learning
Animal (strain)
Mice (CFWI
Rats ISO)
flats (Wister)
Rats IWister)
Rats ISO)
Rats (Long-Evans)
Rats
Rats
Rats
Oosasi.
mgHg/kg (route)
SOxllgi)
1 2. 2.S. or 4 7 x 3 (gi)
0.04 or 1.6x4 (gi|
0.004. 0.008. or
0.04 x 4 (gi)
IQ.Oxl(iv)
6.4x1(gi)
0. 2. or 6 x 4 (gavage)
0.0.025. 0.05. 0.5. or
5.0x4 (gavagel
0. 0.025. 0.05. 0.5. or
5.0x4 (gavage)
Period or
administrations
GO 8
GO 0.7, ana 14
GO 5-9
606-9
604
3D 8 or 15
GO 6-9
GO 6-9
GO 6-9
Behaviors examined
Lever press conditioneo
suppression
Lever cress
Amphetamine challenge
to Rl performance
Lever press multi DRH-TO
Lever press multi-ORH-TO
Lever press CRF
Lever press
Amphetamine challenge to
ORL performance
Discrimination and reversal
Visual discrimination and
reversal
Spatial alternation
Agaslat
examination, days
55
"80
90
120
110-140
150
7
75-145
Findings
No cnange
Low drug sensitivity in
2.6 and 4 7 groups
Deficits in both groups
Deficits in 0.008 and
0.04 groups
Slower learning and
extinction
Low drug sensitivity
6
In 6 mg/kg group, correct
response decreased
Affected 5 mg/kg group
Affected 0.05 mg/kg and
larger doses •
References
(31)
(37}
[38]
(39)
(33)
(34)
(29)
(30)
(30)
Rats
Rats
0.0.025. 0.05. 0.5. or
5.0x4 (gavagei
0. 0.025. 0.05. 0.5. or
5.0x4 (gavage)
GO 6-9
GO 6-9
Visual discrimination and
reversal
Spatial alternation
Abbreviations: iv, intravenous injection; gi, gastnc intubation; GO, dav(sl of gestation; Rl. random interval
continuous reinforcement ORL, differential reinforcement of low rates.
7 Affected 5 mg/kg group
75-1 45 Affected 0.05 mg/kg and
larger doses •
(30)
(30)
; ORH-TO. differential reinforcement of high rates-time out CRF,
Table 7. Effects of prenatal metrrylmercury exposure on sensory functions.
Animal (strain)
Rats
Dttato).
mgHg/kg Iroute)
0.1.5. or 5.0 mg/l
(in water)
Period of
administrations
2 weeks before
contention to lactation
Behaviors examined
Lever press with required
force
Ageislat
BJUMnination, days Findings
300 Increased failure in
successful performance
Reference
(40)
«
Table 1 Effects of prenatal methylmercury exposure on spontaneous activities.
Animal (strain)
Mice(129/SvSI|
Mice(C3H/HeN|
Rats ISO)
Rats (Long-Evans)
Rats (SO)
Rats
Rats
Rats ISO)
Oosetsl.
mgHg/kg (route)
Period of
administrations
5.1.6.8.or10.2x1(sc) GO 10
16.0x1|gi)
3x1 (iv)
4or6.4x1(gi)
6.4x1(gi)
0.2. or 6x4 (gavagel
0, 0.025. 0.05. 0.5. or
5.0x4 (gavage)
Q.2.or6x3(gi)
Agels) at
examination, days
24. 44, and 64
GO 13.14,15.16. or 17 3-8 weeks
(every week)
GO 4
GO 8 or 15
GO 8
GO 6-9
GO 6-9
GO 6-9
110-140
4. 8. 15. and 22
15. 22. 40. and 60
21-120
70-190
60
Findings
Decrease in 6.8 and 10.2 treated groups
Decreased
Decrease in a plain enclosure
No change in enriched environment
Increase in both groups
No change
Only males affected
5 mg/kg group affected in two types
of activity monitor
Hypoactivity. Se-supplemented diet
(1.3 ppml partly antagonized
References
(4i)
(42)
(33)
(43)
(44)
(29)
(30)
(45)
Abbreviations: sc. subcutaneous miection; iv, intravenous injection; gi, gastnc intubation: diet food containing methylmercury compounds: GO. day(s) of gestation.
Environmental Heaitn Perspectives • Voi 104, Supplement 2 • April 1996
373
-------
WATANABE AND SATOH
Table 9. Effects of prenatal mettivlmercurv exoosure on the ooen field test results
Ooselsl. Penodof Apparatus and penoa Ageulat
Animal (strain! mgHg/kg (route! administrations of observation examination
Mice(129/SvSI) 5.4x1 dp) GO 7 or 9 50x50 cm2, 2 mm 30 and 31
MicelCFW) 1,2. 3.5, or 10x1 (gi) GO 8 75x75 cm2, 5 mm 56
Mice|129/SvSI) 5.1, 6.8. or 10.2x1 (sc) G0 10 eSxeScm^min 23
3.4x3 Iscl G0 10-12 33
Mice(JCUCR) 0.4 or 4.0 daily (gi) G0 15 to weaning 50x50crm,2mm 21
70
224-252
flats (SO) 6.4x1(gi) GO 8 15 and 22
40 and 60
, days Findings References
Longer latency ( 1)
Increased backing
Decreased defecation
and urination
No change (3D
Longer latency (41)
Decreased grooming
Decreased urination
Longer latency
Decreased rearing
Increased backing
Increased activity and (24)
rearing in 4.0 group
Increased rearing In 4.0 group
No change
Apomorphme-induced stereotyped (44)
sniffing elicited only in the
treated group
Apomorphme-mduced stereotyped
sniffing potentiated
Rats (SO) 6.4x1 (gi) G0 15 60x60 cm2, 15 mm 14, 21. and 60 ' No change [35]
Amphetamine-induced
locomotion increased
Abbreviations: ip. mtraperitoneal injection; sc. subcutaneous injection; gi, gastnc intubation; GO, dayts) of gestation.
Tiote 10, Effects of prenatal metiTylmercury exposure on susceptibility to convulsions and seizures.
Oosels). Penodof Ageislat
Animal (strain) mgHg/kg (route) administrations examination, days
Mice(129/SvS!) 5.1. 6.8. or 10.2 x 1 (sc) 6010 70
Mice(HUSSabra) S.4x1(sc) G0 12 28-31
Rats 0.2. or6x4(gavage) GO 6-9 18-19.57-58
Rats 0.0.025. 0.05. 0.5. or 5.0x4 GO 6-9 60, 120. 180. and 210
(gavage)
Abbreviations: iv. Intravenous injection; gi; gastric intubation; GO. dayts) of gestation.
Tibte 11. Effects of prenatal methylmercury exposure on ultrasonic vocalization.
Oosels). Penodof Age(s)at
Animal (strain) mgHgAg (route) administrations examination, days
Rats (SO) 0.1.6.3Zor GO 7 5. 7. 9. and 11
4.8x1(gi)
Rats (SO) 6.4x1(gi) G0 15 4. 8. and 12
Rats 0.0.025.0.05,0.5, UV during sexual behavior 240
or 5.0 (gavage)
Rats(Wistar) 0, 0.18-0.27. or Two weeks before 5.7,9.11,
0.59-0.86 pairing and continued 1 3, and 1 5
(mgHg/kg/dav) during experiment
drinking water
Findings References
Increased susceptibility to (41)
fturothyl-induced convulsion
Increased susceptibility to (46)
audiogenic seizure
Startle facilitation in auditory 129)
startle habrtuaoon
Startle facilitation observed in (JO)
males of 0.5 mg/kg and more in
auditory startle habitation
Findings • • References
Slight effects: stronger effects by [47]
ages and order stimuli
No change (35}
No change (30)
Developmental delay (43) .
Reduction in number of calls
Shortening of basal-interval and -
call duration
Flattening and shift of frequency distribution '
Abbreviations: iv. intravenous injection; gi; gastric intubation: GO, dayis) of gestation; UV. ultrasonic vocalization.
374
Environmental Health Perspectives • Vol 104. Supplement 2 • April 1996
-------
METHYLMERCURY AS Al-lt/U-in inite/ii
evaluation of swimming ontogeny and Biei
maze learning should be included because of
their sensitivity to methylmercury exposure,
Collaborative studies were also done by
Eisner and his colleagues (30,53}. In the
first trial (53), female rats were given
mcdiylmercury in drinking water at concen-
trations of 0, 1.5, or 5 mg/l from 2 weeks
before pairing until weaning. Among the
various test items examined, a discrete trial
spatial alternation task was shown to be the
most sensitive, both in terms of effective
dose 50% (ED;o) and of no toxic effect level
(NTEL). In the second trial (30), a wider
range of doses was employed to include
lower exposure levels. Thus, rat dams were
administered 0.025, 0.05, 0.5, or 5.0
mg/kg/day of methylmercury during gesta-
tional days 6 to 9. Among the behavioral
tests, the discrete trial spatial alternation task
was found, as it was in the first trial, to be
the most sensitive, with effects detectable in
the 0.05 mg/kg/day group. It should be
noced that differences in performance in a
visual discrimination task, another rather
demanding operant task, could only be
detected at doses of 5.0 mg/kg/day, die
largest dose employed.
Studies in Nirmlinin-m
Motor Function*. Contrary to the
rodent studies, litde work has been done in
this category in primates (Table 12).
Cognitive Function*. Gunderson et aL
(57) reported that exposed infant mon-
keys paid less visual atcencion to novel
stimuli. The result was interpreted as a
deficit in visual recognition memorv
Object permanence development in
infants (55) and delayed alternation in
adult monkeys (56], both assumed to be
tests of spatial memory, were examined
with one cohort of monkeys. Of these two
test paradigms, only the object perma-
nence development was impaired by
methylmercury. Thus monkey studies so
far did not show any persistent cognitive
deficits caused by in utero methylmercury
exposure.
Sentory Functions. Taking advantage
of the similarity between the visual system
of monkeys and humans. Rice and Gilbert
(59) examined visual effects of prenatal and
postnatal exposure to methylmercury.
Spatial vision was affected in boch die stud-
ies, but temporal vision was impaired only
by exposures that had started before birth.
Social Functions. By observation and
coding of elements of behavior in monkeys,
Burbacher et al. (60) found less frequent
social behaviors among the exposed groups.
What Do These Results Tell
as a Whole?
The experimental studies have shown that
some of the test items detected some
behavioral alterations caused by prenatal
exposure to methylmercury, at least when
high doses (but not high enough to cause
severe maternal toxicicy or fetotoxiciry)
were given to the animals. In this sense, a
proper combination of these tests would
have been successful in detecting some
effects of prenatal methylmercury exposure,
ilthough a given single irem might not
have produced a positive resuir.
Some behavioral tests were shown to be
particularly sensitive to prenatal methyl-
mercury exposure. Among others, the
spatial alternation test (30], the tactde-
kinesthetic test (40), and the DRH task
(39) showed deviations from control at very
low dose levels. At these doses, other simple
tests such as those included in a functional
observation battery, would fail to show any
changes. It is unknown, however, whether
such differential sensitivities among die tests
reflected the nature of the behavioral tasks
per se or reflected the nature of the effect of
methylmercury. Evaluation of rhcse tasks
against other agents may show that the latter
was the case. On the other hand, more
mechanistic analyses of these behaviors
might reveal the inherent sensitivities of
these tests, which would support the former
explanation. It should also be noted diat die
reproducibility of these test results must be
demonstrated; e.g., Eisner (40) could not
reproduce the deviation in the spatial alter-
nation task (30) obtained in their first trial,
and no laboratory has published rodent
behavioral studies that showed effects from
the same low level of exposure as was
demonstrated in the DRH studies (39).
Thus, it is not clear from these tables,
which cover a broad spectrum of behav-
ioral functions, whether there are any func-
tional categories particularly vulnerable to
prenatal methylmercury exposure. It may
be that these results simply indicate that
the behavioral consequences of prenatal
Tablt 12. Behavioral effects of in utero exposure to methvlmercury in nonhuman pnmates.
Behavior examined
Species
Ooseis)
Period of
administrations
Ageislat
examination, days
Maior findings
References
Cognitive function
Spatial memory
Object permanence
development'
Spatial delayed alternation'
Visual recognition memory*
Macaca
fasdcularis
Macaca
fasoculans
Macaca
fasciculans
0,50.70.90 MgAg/day
0. SO. 70.90 ug/kg/day
(50.70 ug/kg/day)
150-750 days before
and during pregnancy
150-750 days before
and during pregnancy
Infants
7-9 years
52 days
Retardation in
development
No effect
Reduced attention to
novel stimuli
(551
156)
(57)
Discrimination reversal
performance and fixed
interval
Sensory function
Spatial and temporal visual
function
Social function
Social behavior
Macaca
fasciculans
Macaca
fasciculans
Macaca
fasciculans
10, 25. 50 pg/kg/day
10.25.50mgHg/kg/day
(SOyg/kg)
Before conception.
throughout the
experiment
From pregnancy to
4-4.5 years
150-750 days before
and during pregnancy
Infants and juveniles
From 2 weeks
to 8 months
Noeffect
Both spatial and temporal
vision impaired
Social behavior reduced.
passive nonsocial
behavior increased
153)
(55)
(60)
'Same cohort of monkeys was used.
ErMronmental Health Perspectives • Vol 104. Supplement 2 • Apnl 1996
375
-------
WATANABE AND SATOH
methylmcrcury exposure are widespread
among various behavioral functions. Also,
the fact that most of the tests (except for
some tests with primates) were apical
rather than specific to one functional cate-
gory might obscure any profile that might
be present. For example, the effects
observed in some learning evaluations,
such as Bid maze (52) or die ORH oper-
ant task (39), might result from motor
incapacity rather than from learning.
Likewise the change in audiogenic startle
habituation might result from either oto-
toxic effect or a Teaming deficit (61,62).
These facts may not be problematic if
one does not intend to characterize the
neurotoxicologic profile but only intends
to detect some adverse effects of methyl-
mercury. It is clear, however,'that for such
a characterization, one must seek another
set of test items that is more specific to
each functional category.
Comparison of Human
and Animal Data on
Neurobehavioral Effects of
Prenatal Methyimercury
Exposure
GflDS J***"»**' Animal* ami Humam
Burbacher et al. (63) thoroughly reviewed
the literature dealing with neuropathologic
and/or neurobehavioral effects of prenatal
methylmercury exposure in humans,
nonhuman primates, and rodents. They
concluded that neurotoxicity of methyl-
mercury in terms of behavioral and patho-
logic effects had remarkable similarities
among humans, (nonhuman) primates, and
small mammals at high levels of exposure
(i.e., brain mercury levels of 12-20 ppm)
and that at moderate or low levels of expo-
sure, neurobehavioral effects were regarded
as similar when functional categories (e.g.,
motor, sensory, cognitive, etc.) rather than
specific end points were compared. To be
exact, they observed that at least two of
three species shared such effects as "early
reflex behaviors, motor coordination, visual
functioning, and complex performance"
(63) as a result of prenatal methylmercury
exposure. It should be noted that these
shared responses covered a broad range of
behavioral categories.
Despite the conclusion reached by
Burbacher et al. (63), there seem to be some
gaps between human and animal studies
dealing with neurobehavioral consequences
of prenatal methylmercury exposures. The
first gap is the level of specificity. As
discussed above, epidemiologic studies of
fish-earing human populations have looked,
and are continuing to look, into behavioral
end points in a more specific way than
experimental studies in animals have. For
example, the New Zealand study (19) sug-
gested some functional domains such as
fine-motor functions and language skills
were vulnerable to methylmercury, although
no clear-cut profile had been delineated.
The Seychelles study (22) has adopted test
items that are more or less connected with
specific functional domains. The Faroe
Islands study (21) a the most explicit in
this regard because it systematically chose
behavioral end points that are related to
focal brain pathology. In comparison, most
of die rodent studies usually adopted rather
apical tests. On the other hand, what the
behavioral changes demonstrated in rodents
imply regarding human behavior might not
always be clear-cut. For example, what kind
of behavioral deficit in humans does the
impaired ORH performance in rats predict?
How well does an effect on spatial memory %
in rodents predict the effect on spatial
memory in humans? These are issues that
need to be answered to establish a test bat-
tery that aims at predicting die neurotoxi-
cologic profile of an agent, which, in mm,
can be extrapolated to human behavior.
The second gap relates to die periods of
testing. In the above-mentioned human
studies, developmental profiles of infants
and children, including such higher
functions as cognition or learning, were
examined. Such emphasis on behavioral
evaluation in the developmental period is
partly because any follow-up studies
extended into adulthood will be extremely
difficult to conduct in natural human pop-
ulations where socioeconomic factors exert
significant impacts on behavioral develop-
ment and where relocations of the subjects
are not uncommon. On the other hand, in
rodent experiments behaviors were usually
examined in the adults; this is especially
true for examining complex behaviors
including schedule-controlled operant
behaviors. Thus, developmental profiles of
such complex behaviors have been rarely
obtained except for die quantitative analyses
of the development of ultrasonic vocaliza-
tion (48) or of auditory startle habituation
(29). In monkeys as described above,
effects of in utero exposure on spatial mem-
ory were apparent in infants but not in
adults (55,56), suggesting a reversible
nature of the effect on this function. It
should be noted, however, that the test
techniques employed in the two studies
were not identical or even similar, and, as
the authors have acknowledged (56), they
might evaluate different (unctions.
The third gap refers to the difference in
die types and periods of exposures. In most
of the rodent studies, methylmercury was
administered several times between 5 and
15 days of gestation. Since the concern
regarding human populations is related to
exposure derived from fish consumption.
. lower level exposures with longer durations
(including both preconception as well as
neonatal periods) should be evaluated,
although neonatal treatment might con-
found the results by affecting the dams'
behavior. Also, it may be important to exam-
ine differential susceptibility to methylmer-
cury among different stages of both the
gestanonal as well as neonatal periods.
Development of
Specific Test Batteries
Methylmercury has acquired a unique status
among hazardous chemicals in our environ-
ment in that a) existence of developmental
neurotoxicity is apparent in both humans
and animals, b) a relatively large body of
neurologic and behavioral data is available
in both humans and animals when com-
pared to most other chemicals, and c) the
ongoing large-scale epidemiologic studies
are evaluating behavioral functions in more
specific ways than routine neurologic or
psychologic batteries. Thus, by taking into
account current and expected future
findings, as well as the human-animal gaps
described above, methylmercury may now
serve as a model agent for developing a
more specific test battery of animal behavior
that could be used to predict possible
human hazards resulting from prenatal
exposure to other chemicals.
To develop such a battery, it is essential
to have a choice of behavioral domains or
categories and a choice of specific behav-
ioral items for each domain. For the
domains or categories, the choices adopted
by Rees et al. (49), thc.Narional Center for
Toxicological Research (64), or the Faroe
Islands study (21) are useful as guidelines.
In the remaining pan of this paper, we will
focus on the second step of the procedure.
To choose specific behavioral items for
a given domain, there are two possible
approaches. The first is an approach in
which a behavior of an animal that is func-
tionally or operationally analogous to
human behavior of concern will be chosen
as the test item. For example, the results of
the discrete trial spatial alternation task
used by Eisner et al. (30) were compared
376
Environmental Healffi Perspectives • Vol! 04. Suoolement 2 • Aonl 199A
-------
METHYLMERCURY AS A HEALTH THREAT
.ie attention disorder (and minimal brain
dysfunction) seen in human children.
Testing the tacnle-(anesthetic system of rats
140) was suggested by human studies that
showed a relationship between attention
deficit disorder and poor development of
die tactile-kinesthetic system in children. If
such a functional analogy could be validated
with some appropriate experiments, this
type of approach could provide a means to
directly predict human behavior on the
basis of rodent behavior. A problem with
dais approach, however, is that an apparent
similarity between behaviors exhibited by
different species does not always guarantee
the same underlying neurologic mecha-
nisms. Thus, extensive validation is required
in this regard.
As an alternative approach, one can
examine a particular behavior with a known
neurological mechanism that is related to
human behavior of concern. Stanton and
Spear (61) argue for this approach, suggest-
ing that such neural comparability became
known not only for sensory functions but
also for a number of behavioral functions
that might be examined in a psychologic
evaluation. If so, this can be a powerful
roach, especially when a prediction of
, . neurotoxicologic characteristics of a
given agent in a human population is
needed. Although it seems diat lesion stud-
ies (65) or pharmacologic studies may pro-
vide valuable information in this regard,
few attempts at developing test batteries for
neurobehavioral toxiciry seem to have fully
used such information.
Recently, a behavioral test battery that
may be used for evaluating several aspects
of central nervous system function in pri-
mates has been developed (64}. The battery
includes several test items (Table 13), some
of which were chosen by the functional
Table 13. aenaviorai runcrc-s ana items examined in the National Center for Toxicoiogicai ^esearcn Ooerant
Test Battery.
Behavioral function
Behavioral item
Motivation (to respond for reinforcement!
Learning
Position ana color aiscnmmation
Time estimation
Short-term memory and attention
Progressive ratio task
Incremental repeated acquisition tasK ihiopocampusjj
Conditioned position resoonamg task icrerrontal cortex)
Temporal response differentiation task
Delayed matching-to-sampie task (hippocampus)
'Brain region that is thought to be related with the task is in parentheses.
analogy of certain behaviors between
humans and monkeys and others because
of suspected correlation between the task
and the integrity of certain brain areas (66).
Thus, me choice of die test items used bodi
of* the above approaches. The battery was
tested against acute effects of several refer-
ence compounds to demonstrate differential
sensitivity of each task to different com-
pounds (64). Although much work has to
be done to validate each item and demon-
strate the sensitivity of this battery, espe-
cially for chronic effects, this approach
seems worthy of extensive pursuit. It is
desirable to have a similar specific battery
applicable to rodents, because it is much
easier to run parallel experiments of neuro-
chemical and neuropharmacologic exami-
nations with rodents than it is with
primates. Although some human behav-
ioral functions exist that cannot be assessed
in rodents, e.g., color vision or language
skills, rodents can accomplish highly com-
plex behaviors such as 24-arm radial mazes
(67), or repeated acquisition of correct
sequences (68). These complex behaviors
may be used for examining specific
functional domains, such as those evalu-
ated in die monkey battery.
Finally it should be pointed out that
diere are some basic items that have been
dropped in most of the behavioral studies.
The first one is the determination of the
internal dose. Lack of an appropriate mea-
sure of the internal dose, e.g., brain Hg
concentration, makes the significance of
certain behavioral findings (regardless of
whether they are positive or negative)
somewhat ambiguous. In -the case of in
utero exposure, the dose should be deter-
mined not only at the time of testing but
also during the prenatal period (69). The
second is potential influences of the sub-
jects' genetic background. Although the
influence of genetic background on kinet-
ics (excretion and distribution) of methyl-
mercury has been evaluated, influences on
behavioral effects seem to have scarcely
been examined. In Iraq, individual differ-
ences were recognized in terms of die neu-
rologic susceptibility of infants to prenatal
methylmercury exposure as previously
described. Individual differences were also
a focus of consideration in choosing test
items in the Faroe Islands study (21).
Genetic background must be one of the
determinants of such individual differ-
ences, and thus, requires further considera-
tion. In general, systematic study of genetic
influences on behavior is best conducted
with rodents. In this respect again, a
specific battery with rodents, if properly
developed, would be of great value.
2.
REFERENCES
Spyker JM, Sparber SB, Goldberg AM. Subde consequences of
methylmercury exposure behavioral deviations in offspring of
treated mothers. Science 177:621-623 (1972).
Hunter, D. Mercury. In: The Disease of Occupations. 5d» ed.
London: English Universities Press, 1969:288-332.
Arima S, ed. Minamata Disease—Studies during These 20
Years and Problems Remaining for Today [in Japanese].
Tokyo:Seirinsha, 1979.
Harada M. Introduction—1: History of the study on
Minamata disease and problems remained for today [in
Japanese]. In: Minamata Disease—Studies during These 20
Years and Problems Remained for Today (Arima S, ed).
Tokyo:Seirinsha, 1979:3-26.
Hosokawa H, Noda K, Misumi H, Kakita T, Kojima T.
Investigation on the mysterious disease in Minamata [in
Japanese]. In: Minamata Disease—Studies during These 20
Years and Problems Remained for Today (Arima S, ed).
Tokvo:Seirinsha,1979;253-258.
6. Takeuchi T. Introduction—2: Advance of the pathological
study on Minamata disease—investigation of unknown H'"T»"
and revelation of causal agent, organic mercury [in Japanese].
In: Minamata Disease-Studies during These 20 Years and
Problems Remained for Today (Arima S, ed). Tokyo:Seirinsha.
1979:27-48.
7. Kitamura S. "Epidemiology" of Minamata disease [in
Japanese]. In: Minamata Disease—Studies during These 20
Years and Problems Remained for Today (Arima S, ed)
Tokyo:Seirinsha, 1979;8l-94.
8. Hunter D, Russell DS. Focal cerebral and cerebellar atrophy in
a human subject due to organic mercury compounds. J Neural
Neurosurg Psychiatry 17:235- 241(1954).
9. Tokuomi H. Clinical investigation of Minamaca disease [in
Environmental Health Perspectives • Vol 104. Supplement 2 • /\pn( 1996
377
-------
WATANABE AND SATOH
Japanese]. In: Minamata Disease—Studies during These 20
Years and Problems Remained for Today (Anma S, ed). 27.
Tokyo.-Seirmsha, 1979:273-288.
10. Von Oemngen WF. Poisoning; A Guide to Clinical Diagnosis
and Treatment. New YoricPad B. Hocbner, 1952. 28.
11. Zangger H. Erfahrungen liber Qiicksilvergiftungen (in
German]. Archiv fur Gewerbepathologie und Gewerbehveiene.
1:593-560(1930). 29.
12. Hunter D. Poisoning by methyl mercurv compounds. Quart j
Med 9:193-213 (1940):
13. Harada M. Congenital (or fetal) Minamata disease [in
Japanese]. In: Minamata Disease—Studies during These 20
Years and Problems Remained for Today (Anma S, ed). 30.
Tokyo:Seirinsha, 1979:345-363.
14. Doi R. Introduction—3: Review on the studies on organic
mercury poisoning and discussions from the view point of
social medicine [in Japanese]. In: Minamata Disease—Studies
during These 20 Yean and Problems Remained for Today 31.
(Arima S, ed). TokyorSeirinsha. 1979:49-77.
15. WHO. Effects on man. In: Environmental Health Criteria
101: Methylmercury. Geneva:World Health Organization, 32.
1990:68-99.
. 16. Marsh DO, Myers GJ, Ckrkson TW Amin-Zaki L, Tikriri S,
Majeed MA, Dabbagh AR. Dose-response relationship for 33.
human fetal exposure to methylmercury. Clin Toxicol
18:1311-1318(1981).
17. Marsh DO, Clarkson TW, Cox C Myers GJ, Amin-Zaki L, 34.
Tikriri S. Fetal methylmercury poisoning: relationship- between
concentration in single strands of maternal hair and child
effects. Arch Neurol 44:1017-1022 (1987).
18. Cox C, Clarkson TW, Marsh DO, Amin-Zaki L. Tikriti S, 35.
Myers GJ. Dose-response analysis of infants prenatally exposed
to methyl mercury: an application of a single compartment
model to single-strand hair analysis. Environ Res 49:318-332
(1989).
19. Kjeustrom TP, Kennedy P, Willis P, Mantell C Physical and 36.
Mental Development of Children with Prenatal Exposure to
Mercury from Fish. Stage. 1. Preliminary Tests at Age 4.
SolnaiNarional Swedish Environmental Board. 1986. 37.
20. McKeown-Eyssen GE, Ruedy J, Neims A. Methyl mercury
exposure in northern Quebec. II. Neurologic findings in chil-
dren. Am J Epidemiol 118:470-479 (1983).
21. White RF, Debes F, Dahl R. Grandjean P. Development and 38.
• field testing of neuropsychological test battery to assess the
effects of methylmercury exposure in the Faroe Islands. In:
Proceedings of the International Symposium on Assessment of 39.
Environmental Pollution and Health Effects from
Methylmercury, 8-9 October 1993. Kumamoto, Japan.
Minamata, Japan:National Institute for Minamata Disease, 40.
1994:127-140.
22. Davidson PW. Measuring neurodevelopmental outcomes of
young children following prenatal dietary methylmercury expo- 41.
sure. In: Proceedings or the International Symposium on
Assessment of Environmental Pollution and Health Effects
from Methylmercury, 8-9 October 1993, Kumamoto, Japan. 42.
Minamaca, Japan.-National Institute for Minamata Disease,
1994:96-111.
23. Kjeustrdm T. Effects of methylmercury exposure in utenr, studies 43.
in New Zealand and proposal for future studies. In: Proceedings
of the International Symposium on EpidemioloKical Studies on
Environmental Pollution and Health Effects of Methylmercury, 44.
2 October 1992, Kumamoto, Japan. Minamata, JaparuNational
Institute for Minamata Disease. 1993:67-78.
24. Tanimura TE, Ema E. Kihara T. Effects of combined treat-
ment with methylmercury and polychlorinated biphenyls 45.
(PCBs) on the development of mouse offspring. In: Neural and
Behavioral Teratology (Persaud TVN, ed). Lancaster, PA:MTP
Press, 1980:163-198.
25. Satoh HN. Yasuda N, Shimai S. Development of reflexes in
neonatal mice prenatally exposed to methylmercury and selen- 46.
ite. Toxicol Lert 25:199-203 (1985).
26. Monet NK. Effects of chronic low-dose exposure of rat fetuses
to methvlmerairv hvdroxide. Teratology 10:173-189 (19741
Sobotka TJ, Cook MP, Brodie RE. Effects of perinatal expo-
sure to methylmercury on functional brain development and
neurochemistry. Biol Psychiatry 8:307-320 (1974).
Olson K, Boush GM. Decreased learning capacity in rats
exposed prenatally and posrnatally to low doses of .mercury.
Bull Environ Contam Toxicol 13:73-79 (1975).
Buelke-Sam J, Kimmel CA, Adams J, Nelson CJ, Vorhees CV
Wright DC, St Omer V, Karol BA, Butcner RE, Geyer MA.
Holson JF, Kutscher C, Wayner MJ. Collaborative Behavioral
Teratology Study: results. Neurobehav Toxicol Teratol
7:591-624(1985).
Eisner J, Hodel B, Suter KE, Oelke D, Ulbrich B, Schreiner G.
Cuomo V, Cagiano R; Rosengren LE. Karlsson JE, Haglid KG.
Detection limits of different approaches in behavioral teratol-
ogy, and correlation of effects with neurochemical parameters
Neurotoxicol Teratol 10:155-167 (1988).
Hughes JA, Annau Z. Postnatal behavioral effects in mice after
prenatal exposure to methylmercury. Pharmacol Biochem
Behav 4:385-391 (1976).
Shimai SH. Satoh H, Yasuda N. Taste aversion learning and
perinatal methylmercury exposure in mice. Ind Health
22:41^4 (J984).
Scharlock RL, Brown WJ, Kark RAP, Menon NK. Perinatal
methylmercury intoxication: behavioral effects in rats. Dev
Psychobiol 14:213-219 (1981).
Eccles CU, Annau Z, Prenatal methyl mercury exposure: II.
Alterations in learning and psychotropic drug sensitivity in
adult offspring. Neurobehav Toxicol Teratol 4:-377-382
(1982).
Cagiano R, De-Salvia MA, Renna G, TorteUa E, Braghiroli D,
Parenri C, Zanoli P, Baraldi M, Annau Z, Cuomo V. Evidence
that exposure to methyl mercurv during gestation induces
behavioral and neurochemical changes in offspring of rats.
Neurotoxicol Teratol 12:23-8 (1990).
Zenick H. Behavioral and biochemical consequences in
methylmercury chloride toxiciry. Pharmacol Biochem Behav
2:709-713 (1974).
Hughes JA, Sparber SB. ^-Amphetamine unmasks postnatal
consequences of exposure to methylmercury in utenr. methods
for studying behavioral teratogenesis. Pharmacol Biochem
Behav 8:365-375 (1978).
Miisch HR, Bornhausen M, Kriegel H, Greim H.
Merhlmercury chloride induces learning deficits in prenatallv
treated rats. Arch Toxicoi 40:103-108 (1978).
Bornhausen M, Miisch HR, Greim H. Operant behavior per-
formance changes in rats after prenatal methylmercury expo-
sure. Toxicol Appl Pharmacol 56:305-310 (1980).
Eisner J. Tactile-kinesthetic system of rats as an animal model
for minimal brain dysfunction. Arch Toxicol 65:465-473
(1991).
Su MQ, Okita GT. Behavioral effects on the progency of mice
treated with methylmercury. Toxicol Appl Pharmacol
38:195-205(1976).
Inouye MK, Murao K, Kajiwara Y. Behavioral and neuropatho-
logical effects of prenatal methylmercury exposure in mice.
Neurobehav Toxicol Teratol 7:227-232 (1985).
Eccies CU, Annau Z. Prenatal methyl mercury exposure: I.
Alterations in neonatal activity. Neurobehav Toxicol Teratol
4:371-376 (1982).
Cuomo VL, Ambrosi Z. Annau, Cagiano R, Brunello N,
Racgni G. Behavioural and neurochemical changes in offspring
of rats exposed to methyl mercury during gestation.
Neurobehav Toxicol Teratol 6:249-254 (1984).
Fredriksson A, Gardlund AT, Bergman K. Oskarsson A, Ohlin
B, Danilsson B, Archer T. Effects of maternal dietary supple-
mentation with selenite on the postnatal development ot rat
offspring exposed to methyl mercury in utero. Pharmacol
Toxicol 72:377-382 (1993).
Menashi MA, Ormoy A Yanai J. Transplacental effects of
methylmercury chloride in mice with specific emphasis on
audiogenic seizure response. Dev Neurosci 5:216-221 (1982).
378
Environmental Health Perspectives • Vol 104. Supplement 2 • April 1996
-------
METHYLMERCURY AS A HEALTH THREAT
dams J, Miller DR, Nelson CJ. Ultrasonic vocalizations as
Diagnostic tools in studies of developmental coxicity.
Ncurobehav Toxicol Teratol 5:29-37 (1983).
<8. Eisner J, D Surer, S Alder. Microanalysis of ultrasound vocal-
izations of young rats: assessment of the behavioral teratogenic-
iry of methyimercury. Neurotoxicol Tcratol 12:7-14 (1990).
*9. Recs DC. Francis EZ, Kimmel CA. Qualitative and quantita-
tive comparability of human and animal developmental neuro-
toxicants: a workshop summary. Neurotozicology 11:257-269
(1990).
50. Shimai S, Satoh H. Behavioral teratology of methyimercury. J
Toxicol Sci 10:199-216 (1985).
51. Royalty J, Taylor GT, Korol BA. The effects of prenatal expo-
sure to methyimercury on aggressive behavior in the rat.
Neurotoxicol Teratol 9:87-93 (1987).
52. Vorhees CV. Comparison of the Collaborative Behavioral
Teratology Study and Cincinnati Behavioral Teratology test
batteries. Neurobehav Toxicol Teratol 7:625-633 (1985).
53. Eisner J, Suter KE, Ulbrich B, Schreiner G. Testing strategies
in behavioral teratology: IV. Review and general conclusions.
Neurobehav Toxicol Teratol 8:585-590 (1986).
54. Buelke-Sam J. Practical considerations in establishing reliable
and sensitive neurobehavioral test methods. Zcntralbl Bakreriol
Mikrobiol Hyg B 185:4-9 (1987).
55. Burbacher TM, Grant KS, Monet NK. Retarded object perma-
nence development in methyimercury exposued Macaca fascic-
ularii infants. Dev Psychol 22:771-776. (1986).
56. Gilbert SG, Burbacher TM, Rice DC. Effects of in utero
methyimercury exposure on a spatial delayed alternation task in
monkeys. Toxicol Appl Pharmacol 123:130-136 (1993).
57. Gundenon VM, Grant-Webster KS, Burbacher TM, Mottet
NK. Visual recognition memory deficits in methylmercury-
exposed Macaca facicularis infants. Neurotoxicol Teratol.
10-373-379(1988).
Rice DC. Effects of pre-plus postnatal exposure to methyimer-
cury in the monkey on fixed interval and discrimination rever-
sal performance. Neurotoxicology 13:443-452 (1992).
59. Rke DC, Gilbert SG. Effects of developmental exposure to
65.
66.
67.
68.
69.
methvi mercury" on spatial and temporal visual function in
monkeys. Toxicoi Appl Pharmacol 102:151-163 i 1990)..
60. Burbacher TM, Sackett GP, Mot:et NK. Metnvlmcrcury
effects on the social behavior of Macaca fasciculnns infants.
Neurotoxicol Teratol 12:65-71 (1990).
61. Stanton ME, Spear LP. Workshop on the qualitative and quan-
titative comparability of human and animal developmental
neurotoxiciry, Work Group I report: comparability of measures
of developmental neurotoxicity in humans and laboratory ani-
mals. Neurotoxicol Teratol 12:261-267 (1990).
62. Adams J. Methods in behavioral teratology. In: Handbook of
Behavioral Teratology (Riley EP, Vorhees CV, eds). New
YorlcPlcnum Press, 1986:67-100..
63. Burbacher TM, Rodier PM, Weiss B. Methyimercury develop-
mental neurotoxicity: a comparison of effects in humans and
aoimalc Neurotoxicol Teratol 12:191-202 (1990).
64. PaukMG. Analysis of brain function using a battery of sched-
ule-controlled operant behaviors. In: Neurobehavioral Toxiciry:
Analysis and Interpretation (Weiss B, O'Donoghue JD, eds).
New YorlcRaven Press, 1994331-638.
Wenk GL Olton DS. Lesion analysis. In: Neurobehavioral
Toxicology (Annau Z, ed). Baltimore:Johns Hopkins
University Press, 1986:268-276.
Paule GM, Allen RR. Bailey JR. Scallet AC, Ali SF, Brown
RM, SLikker W Jr. Chronic marijuana smoke exposure in the
rhesus monkey II: Effects on progressive ratio and conditioned
position responding. J Pharmacol Exp Ther 260:210-222
(1992).
Miller-DB, Eckerman DA. Learning and memory measures. In:
Neurobehavioral Toxicology (Annau Z, ed). Baltimoretjohns
Hopkins University Press, 1986:94-152.
Conn j, Cox C. Cory-Slecha DA. The effects of lead exposure
on learning in a multiple repeated acquisition and performance
schedule. Neurotoxicology 14:329-346 (1993).
Inoue M, Kajtwara Y, Hirayama K. Dose- and sex-dependent
alterations in mercury distribution in fetal mice following
methyimercury exposure. J Toxicol Environ Health
19:425-435 (1986).
EnvironmenHl Health Perspectives • Vol 104. Supplement 2 • Apnl 1996
379
-------
Impact of Maternal Seafood Diet on Fetal Exposure to Mercury,
Selenium, and Lead
P. GRANDIEAN
Institute of Community Health
Odense University
Odense, Denmark
P. WEIHE
Landssjukrahusid
Tdrshavn, Faroe Islands
P. j. JORGENSEN
Department of Clinical Chemistry
Odense University Hospital
Odense, Denmark
T. CLARKSON
E. CERNICHIARI
Department of Biophysics
Rochester University School of Medicine
Rochester, New York
T. VIDERO
Landssjukrahusid
Tdrshavn, Faroe Islands
ABSTRACT. (Jmbiikal cord blood from 1 023 consecutive births in the Faroe Islands show-
ed a median blood-mercury concentration of 121 nmoi/l (24.2 pg/I); 250 of those samples
(25.1%) had blood-mercury concentrations that exceeded 200 nmol/l (40 yg/l). Maternal
hair mercury concentrations showed a median of 22.5 nmol/g (4.5 ng/g), and 130 samples
(12.7%) contained concentrations that exceeded 50 nmol/g (10 Mg/g). Frequent ingestion of
whale meat dinners during pregnancy and, to a much lesser degree, frequent consumption
of fish, and increased parity or age were associated with high mercury concentrations in
cord blood and hair. Blood-mercury levels were slightly lower if the mother had occa-
sionally ingested alcoholic beverages. Mercury in blood correlated moderately with blood
selenium (median, 1.40 ^mol/l). Increased selenium concentrations were associated with
intake of whale meat, alcohol abstention, delivery after term, and high parity. Lead in cord
blood was low (median, 82 nmol/l), particularly if the mothers had frequently had fish for
dinner and had abstained from smoking.
ACCUMULATION of mercury in fish, shellfish, and exposures would be the Faroe Islands, which are locat-
cetaceans results in increased human exposures to this ed in the North Atlantic between Scotland and Iceland.
metal in populations whose diets include a high intake The population of 45 000 resides on 18 islands, an area
of marine food.1 An area with possible high mercury of 1 400 km*. A recent questionnaire study of 331
May/|un«1992(Voi.47(No.3)} 185
-------
Faroese adults revealed a daily consumption of 72 g
fish, 12 g whale muscle, and 7 g of blubber: fish and
pilot whale constituted 44%' and 9.5% of Faroese din-
ner meals, respectively.2 Most of the fish consumed in
the Faroe Islands is cod, which has an average mercury
concentration of about 0.07 jig/g (0.35 nmol/g)
(Hygienic Institute, Torshavn, personal communica-
tion). Almost all of the mercury in fish is methylated.3
Muscle tissue of Faroese pilot whale contained an
average mercury concentration of 3.3 j/g/g (16 nmoi/g),
of which about one-half was methylmercury; higher
concentrations were found in the liver, mainly in the
form of inorganic mercury.4 In both muscle and liver,
selenium varied considerably, but tended to approach
equimolar concentrations with total mercury.4
Methylmercury has caused serious disease in hu-
mans, and it has a threshold for clinical toxicjty at a
blood-mercury 18-Hg) concentration of about 200-500
^g/l - (corresponding to 1-2.5 jumoi/I in SI units).3-* A
dose-response relationship has been established be-
tween peak hair-mercury (H-Hg) levels during preg-
nancy and the prevalence of psychomotor retardation
and neurological signs in children—particularly boys-
exposed to lower levels of methylmercury in utero.7 Ex-
perimental evidence3 supports the assertion that the
fetus has an increased susceptibility to methylmercury
toxicity.
A need, therefore, was perceived to assess the magni-
tude of fetal mercury exposures in the Faroes. Selenium
may counteract some of the toxic effects of mercury9
and was, therefore, included in the study. Lead was
also determined because this trace element may be
associated with adverse pregnancy outcomes.10
In a pilot study, we determined the mercury specia-
tion in blood samples from a group of adult women
who resided in a small fishing village in the Faroe
Islands. Following the detection of increased mercury
concentrations in this group, we initiated a prospective
study of children bom ar consecutive deliveries at the
Faroese hospitals during a 22-mo period. This report
describes the trace element exposures and their rela-
tionship to seafood diet and other predictors.
Populations
Pilot study. The small fishing village of Lorvik, Faroe
Islands, was selected for a pilot study of mercury ex-
posures. All women in the fertile age groups between
20 and 50 y were identified through a municipal reg-
ister; 63 women had always resided in the community.
Blood samples were obtained from 53 of these women
(84%).
Birth cohort study. Subsequently, collection of um-
bilical cord blood was initiated with the help of mid-
wives at consecutive deliveries at the three Faroese
hospitals in Torshavn, Klaksvik, and Suderoy, from
March 1, 1986. through the end of 1987. A high con-
sent rate, successful cooperation within the healthcare.
system, and minimal practical problems ensured that
blood samples were obtained from almost all births.
However, delayed delivery of blood vials prevented
sampling during a 1-mo period (summer of 1987). Dur-
ing the active sampling period, a total of 1 386 children
were born, including 12 pairs of twins. Of the smgie-
born children, blood samples and questionnaire data
were obtained from 1 023 (75.1%).
Methods
Pilot study.
A brief questionnaire was filled in with demographic
information and data on nutritional habits, smoking, and
alcohol ingestion. Blood samples were taken in 10-mi
heparinized Vacutainer* vials (Beckton-Dickinson,
Rutherford, NJ, USA) certified to be free of mercury.
Birth cohort study
Obstetrical data. In connection with each birth, the
midwife asked the mother about the course of the preg-
nancy; nutritional habits (frequency of dinners with ffsh-
or pilot whale); and use of alcohol (never, occasionally,
or frequently) and tobacco (none, less or more than 10
cigarettes per day) during the pregnancy. The midwife
then recorded this information in a questionnaire. The
midwife also supplied information on the course of the
parturition. Parity, expected term, and other relevant
information were obtained from the patient chart at the
hospital.
Sample collection. Immediately after delivery, blood
samples from the umbilical cord were taken by the
midwife in 10-ml Abbott syringes equipped with teflon-
lined pistons (Abbott Ireland Ltd, Sligo, Ireland). Hair
samples of at least 100 mg (most were more than 5 cm
in length) were cut with a pair of scissors close to the
root in the occipital area of the mother the hair was
tied with a cotton string and saved in a small, marked
plastic bag. The blood samples were kept deep-frozen
during transport and storage until analysis.
Mercury analysis. All blood samples from the Lorvik
group, one-tenth of all cord blood samples, and all hair
samples were analyzed in Rochester, New York, by the
method described by Magos and Clarkson.12 In brief,
when CdCIa and SnCU were added, mercury com-
pounds were rapidly converted into atomic mercury
that was aspired through the detection cell of a UV-
absorptiometer (Mercury Monitor 1235, Laboratory
Data Control, Milton Roy, FL, USA). Blanks and work-
ing standards were determined every day at the begin-
ning and at the end of the working period. After
satisfactory results were obtained, standard reference
materials or quality control samples were analyzed. All
measurements were performed in triplicate. Standard
reference materials from the National Institute of Stan-
dards and Technology (Caithersburg, MD, USA) in-
cluded freeze-dried urine (#2672a) and bovine liver
(#1577a). Hair samples were weighed on an Auto-
Microbalance model AD-2Z (Perkin-EImer, Norwalk,
CT, USA) prior to analysis. Segmental hair analysis was
achieved by aligning a hair sample by the root ends and
affixing it to a plastic ruler. The hair strands were then
straightened and cut at the desired interval of 1.1 cm
with a scalpel. Hair from each segment was then
weighed and analyzed. The laboratory has participated
186
Archives of Environmental Health
-------
in several mtertaboratory comparison programs, in-
cluding determination or H-Hg; tne results have
documented a high quality of analytical data.': Thirty-
four hair samples that had mercury concentrations
above 75 nmoi/g were analyzed again several months
later. These duplicate results correlated well (correla-
tion coefficient, .97), and a regression coefficient of
0.93 (intercept, 2.34 nmol/g) and an average coefficient
of variation of 5.4% were calculated.
The UV-absorptiometry method was also established
at Odense. However, in the most recent LDC Mercury
Monitor, the light source and ail electronics are con-
tained in one unit. We found that the heat from the
mercury lamp caused electronic interference, thus lim-
iting the time of useful continuous operation. The elec-
tronic noise could be decreased if only one cell of the
detector was used; this* adjustment was deemed accep-
table because blank samples never showed any detect-
able level of mercury. We used teflon tubing through-
out. The air trap was eliminated, and the reaction
chamber was cooled in an ice bath. In the reaction
chamber, a 1-ml sample of blood was mixed with 1 ml
8.3 mM L-cystein (cat. no. 2838, Merck, Darmstadt,
Germany) solution; 20 ml 171 mM sodium chloride
(cat. no. 6404, Merck); 10 ml 8 M sulfuric acid (cat. no.
30741, Riedel de Haen, Seelze, Germany); and 1 ml of
a solution containing 0.62 M cadmium chloride (cat.
no. 31147, Riedel de Haen) and 2.2 M SnCIi,2H,O (cat.
no. 31655, Riedel de Haen) in 0.8 M H2SO4. After mix-
ing, cooling, and connection of the tubing, 20 ml 11.3
M NaOH (cat. No. 6498, Merck) was added. The meth-
od of standard addition was used, and all samples were
measured twice.
The total analytical imprecision was 14.1% at 49
nmol/l (9.9 pgrl) (n - 44), 8.6% at 563 mol/l (112.6 ng/'l)
(n - 15), and 7.0% at 580 nmoi/l (116.0 Mg/D (n - 14).
The accuracy was assessed by analyzing Control Blood
for Metals (Behringwerke, Marburg, Germany). The
average obtained for material #1, lot no. 620303, was
9.9 ^g/l (assigned value, 9.9 jig/l; acceptable range,
8.6-11.2 M8/')- Measurements of material #2, lot nos.
420403, 420404, and 420402. gave the following aver-
ages (fig/I): 112.6 (assigned value, 98; acceptable range,
83-113); 116.0 (assigned value, 103; acceptable range,
88-118); and 114.5 (assigned value, 103; acceptable
range, 91-115), respectively, for the three materials.
A comparison study was earned out with cold-vapor
atomic absorption,13 in which an atomic absorption
spectrometer model 305 A (Perkin-Elmer) equipped
with an electrodeless discharge lamp for mercury, an
MHS-20 hydride system, and a recorder were used.
Prior to analysis, 500 ^ 8.3 mM L-cystein, 100 M' 1.5
mM CuSO4 (cat. no. 2791, Merck), 6 ml 1 M NaHCO3
(cat. no. 6329, Merck), 900 jil 32 mM Triton X-100 (cat.
no. 8603, Merck), and 100 jA tri-n-butylphosphate
(BDH) were added to 1 ml of blood. A solution of 0.26
M NaBH4 (BDH) and 0.1 M NaOH (BDH) was used as a
reducing agent. All tubing was teflon, nitrogen was
used as purge gas, and the mercury vapor was amalga-
mated on a gold filter. Mercury concentrations were
determined in duplicate against a blood-based stan-
dard curve with values specified by standard addition.
The total analytical imprecision was 9.5% at 10.1 jig/I
'50 nmoi/l) (n - 34) and 1.00/0 at 102.2 jig/I (510 nmoul)
(n - 6). With regard to accuracy, the, average obtained
for Control Blood for Metals lot no. "620302 was 10.1
Mg/l (assigned value, 10.0 Mg/l, range - 8.0-12.0
for lot no. 620403, the average obtained was 102.2
(assigned value, 98 Mg/l range - 83-113 Mg/l).
Blood samples were exchanged between the two lab-
oratories and showed satisfactory agreement between
the methods for mercury analysis (Fig. 1). The correla-
tion coefficients were .96 between atomic absorption
(Odense) and UV-absorptiometry (Rochester), and .98
between UV-absorptiometry (Odense versus Roches-
ter). The joint regression coefficient was 0.92, and the
intercept was -1.5 nmol/l (Rg.1). Despite the fact that
the quality control materials in Odense suggested a
slight bias toward readings higher than expected, this
comparison study indicated slightly lower (8%) results
in Odense than in Rochester.
Selenium analysis. Selenium in whole blood (B-Se)
was determined by electrothermal atomic absorption
spectrometry. The equipment used was an atomic ab-
sorption spectrometer model 5000, equipped with Zee-
man background correction, HGA-500 graphite fur-
nace, and an AS-40 autosampler (Perkin-Elmer). the
B-Se results were read in duplicate against a blood-
based standard curve, and values were specified by
standard addition. The blood sample was diluted 1:5 by
a 1.5 mmol/l Triton X-100 solution (Boehringer Mann-
100
Od«ns« (nmol/l)
""" UV ABsorptiomctry
—> Atomic oojorption
30 -
SO -
40 -
20-
:o 40 so
Rochester (nmol/l)
30
!00
Fig. 1. Comparison of three methods for blood-mercury determina-
tion: UV-absorptiometry in Odense and Rochester, and cold-vapor
atomic absorption in Odense, with the regression line for Rochester
versus Odense.
May/lune 1992 (Vol. 47 (No. 3)1
187
-------
heim, Mannheim, Germany); 10 ni of the diluted sam-
ple was placed on the platform of the graphite tube,
followed by 10 ^ of matrix-modifying solution (30 g/l of
Ni(NO3)j-6HiO (Merck, Darmstadt, Germany) and 1.5
mmol/I of Triton X-100 in deionized water). The ashing
temperature was 1 300 °C (atomization at 2 200 °O.
The total analytical imprecision was estimated to be
5.9, 3.8, and 2.7% at B-Se levels of 1.27, 1.93, and 2.61
^moi/l, respectively. The accuracy of the selenium
determinations was ensured by the use of Seronorm*
Trace Element batch 904 (Nycomed, Oslo, Norway) as
quality control material. The average selenium concen-
tration measured in 110 determinations was 1.27
umol/l (assigned value, 1.26 ^mol/l).
Lead analysis. Blood-lead (B-Pb) was determined by
electrothermal atomic absorption spectrometry.14 The
equipment used was the same as for selenium (see
above). The B-Pb results were read in duplicate against
a blood-based standard curve with standard reference
material SRM 955 from the US National Institute of
Standards and Technology (Gaithersburg, MD, USA) as
primary standard.
The total analytical imprecision was estimated to be
5.9% (n - 85) and 1.8% (n - 87) at B-Pb levels of 0.25
and 1.63 punol/l, respectively. The accuracy of the lead
determinations was ensured by using Seronorm*
Trace Element batch 901 and 902 (NycoMed) as quality
control materials. The lead concentrations found in 85
and 8"7 determinations averaged 0.25 and 1.63 ujnol/l,
respectively (assigned values, 0.3 and 1.6 ^mol/l).
Statistical analysis. All data were entered in Clinical
Report System', and statistical evaluation was carried
out with SPSS-PC version 2.0, except for the Spearman
correlation coefficient, which was determined with
Spida® software. Multiple regression analysis was car-
ried out in SPSS, and parameters that appeared to be
relevant predictors (p < 0.1) in the bivariate analyses
were used. For this purpose, most independent varia-
bles were dichotomized. Logarithmic transformation of
trace element concentrations was performed to ap-
proach Gaussian distribution "of the dependent varia-
bles. Cases for which information was missing were
skipped. Twins were treated seoaratelv and were not
included in the general analyses.
j. -
Results
Pilot study.
The women from Lorvik had a median B-Hg level of
60 nmol/l (12.1 pg/l); values ranged from 13-250 nmoi/'l
(2.6-50.1 pg/l). Particularly in the blood samples with a
high mercury content, only a small proportion was in
the form of inorganic mercury (Table 1). Also, less mer-
cury was present in the serum, thus confirming that the
majority of the mercury present in the blood was in the
form of methylmercury. Lead concentrations were very
low (Table 1), which indicated that this metal would
not be of major significance with regard to adverse ef-
fects on pregnancy outcome.
Birth cohort study
Trace dement exposure levels. Incomplete sampling
volume prevented 26 cord blood samples from being
analyzed for mercury, whereas analyses for selenium
and lead were impossible for 3 and 8 cases, respective-
ly, because of the same problem. Incomplete labeling
prevented hair analysis for three cases.
In the cord blood, the median mercury concentra-
tion was higher in this group than in the Lorvik group
(Table 2). A total of 250 samples (25.1%) had a B-Hg
level that exceeded 200 nmol/l (40 jig/I), and 20
samples (2.0%) had levels higher than 500 nmol/l (100
30
Inorganic m»rcury (nmol/g)
25
20 -
Table 1.— Total and Inor
and Serum (nmol/D, and
Whole blood
Total mercury
Inorganic mercury
Serum
Total mercury
Inorganic mercury
Whole blood
Lead
ganic Mercury in Whole Blood
Blood Lead (nmol/l) of 53 Adult
oe Islands
Number Median
53 60
47 10.5
50 10.5
52 8.5
52 96
Range
13-250
<0.5-25
<0.5-41
<0.5-!.5
39-174
'Because there was an insufficient sample volume, not all
analyses could be performed on each sample.
15 '-
10
5 -
50 100 '50
Total mercury (,nmol/g)
200
Fig. 2. Correlation between concentrations of total and inorganic
mercury in hair samples from 34 Faroese women at delivery.
188
Archives of Environmental Health
-------
1000 T-
,0 100;
"3 :
c
0)
E
i_
'5
10d
1 j
0.1
10
100
1000
10000
Blood mercury (nmol/l)
Fig. 3. Correlation between mercury concentrations in cord blood and in maternal hair obtained in
connection with consecutive births in the Faroe Islands.
Table 2.— Umbilkal Cord Blood Concentrations of Mercury (nmol/I), Selenium (nmd/I), and Lead (nmol/l), and
centration of Mercury (nmol/g). Relative to Maternal Fish Intake During Pregnancy
No. of fish
dinners/wk
None
1
2
3
4
5 or more
Total
V, - .19fp<
tr, - .18 (p<
*rs - .15 (p<
Blood mercury*
n
26
139
356
285
157
33
997
.001).
.001).
.001).
Median
35.6
93.7
112
134
146
169
121
(50% range)
(13.2-101)
(37.9-163)
(66.4-186)
(65.1-220)
(94.2-210)
(89.5-240)
(65.0-201)
n
27
140
365
295
159
33
1 020
Hair mercuryr
Median
7.0
16.8
22.6
25.2
26.0
25.1
22.4
(50% range)
(17-16.3)
(7.3-32.1)
(12.8-36.6)
(13.9-41.8)
(15.6-40.4)
(16.0-42.6)
(12.6-38.3)
n
27
141
365
295
158
33
1 020
Blood selenium*
Median
1.28
1.40
1.38
1.41
1.42
1.42
1.40
(50% range)
(1.12-1.37)
(1.24-1.53)
(1.25-1.54)
(1.30-1.58)
(1.30-1.60)
(1.36-1.56)
(1.27-1.55)
n
26
140
360
297
162
33
1 015
Maternal Hair Con-
Blood lead
Median
92
87
77
82
82
72
82
(50% range)
(63-121)
(63-121)
(58-101)
(53-106)
(58-111)
(58-109)
(58-106)
. Likewise. 130 hair samples (12.7%) had a mercury
content that exceeded 50 nmoj/g (10 ng/g), and 5 sam-
ples (0.5%) exceeded~T25~ n'rnol/g (25 Mg/g). In 34 hair
samples that had a mercury concentration above 75
nmol/g (15 Mg/g), both total mercury and methylmer-
cury were determined. Inorganic mercury represented
an average of 11.0 ± 1.85 % of the total mercury con-
tent (Rg. 2).
Mercury concentrations in cord blood and in mater-
nal hair correlated significantly, although some scatter-
ing was apparent (Rg. 3). The regression coefficient,
with mercury in hair as the dependent variable, was
0.11 (and 4.8 for mercury in blood as the dependent
variable). The median ratio between mercury in hair
versus blood was 191 (n - 994). This ratio could be af-
fected by various predictors (see below). Also, part of
the scattering, seen in Figure 3 might result from tem-
poral variations in mercury exposure. Such variations
would affect the mercury concentrations in hair along
the length of the hair strands collected, and the hair
mercury results therefore represent averages for time
periods of variable length. Six hair samples allowed for
analysis of mercury concentrations in 1.1-cm segments
of the hair, and each segment corresponded to a 4- to
5-wk period. The average coefficients of variation for 5,
6, 7, 8, 12, and 18 segments from the same woman
varied between 8.1% and 23.8%. The mercury levels,
/hen compared for time of sampling, showed no com-
.non pattern in seasonal variation. Similarly, mercury
concentrations in cord blood showed no relationship
with the month of delivery.
8-Se varied less than B-Hg (Table 2), but the correla-
tion between selenium and mercury was significant (rs
- .35; p < .001) (Rg. 4). Once again in this population
the lead concentrations were very low (Table 2). Seleni-
um and lead showed a slight correlation (r, - . 11; p <
.001), whereas lead and mercury were unrelated.
Data on twins. The multiple births consisted of 12
May/|une 1992 (Vol. 47 (No. 3)]
189
-------
J«r
O
£ 1
1 I
H 1-5"*
-a
8
S 1-1
0.5
10 100 1000
Blood mercury (rimol/l)
10000
Fig. 4. Correlation between mercury and selenium concentrations in cord blood obtained at con-
secutive births in the Faroe Islands.
sets of twins. Sufficient amounts of cord blood were ob-
tained so that lead analyses could occur in all cases,
selenium in all but 1, and mercury analysis in all but 3
complete sets of twins. The median concentrations of
mercury (120 nmol/l), selenium (1.51 ^mol/l), and lead
(118 nmol/l) were very similar to those seen in the re-
mainder of the cohort. The coefficient of variation (%)
within each set of twins averaged 7.1, 4.6, and 11.2 for
mercury, selenium, and lead, respectively.
Predictors for trace element exposures. Data on
single births only were used for the statistical analyses
of potential predictors for trace element concentra-
tions. No difference was seen in trace element concen-
trations in cord blood from male and female infants.
The median age of the mothers was 26.9 y (range,
15-45 y). As the age of the mother increased, the trace
element concentrations tended to become somewhat
higher, particularly mercury in hair (r, - .10, p - .002).
Lead levels in cord blood were particularly low (p -
.002, Mann-Whitney U-test) if the mother was a teen-
ager. Increased (prepregnancy) maternal weight index
(weight/height1) and height were slightly associated
with higher selenium and lead concentrations in cord
blood.
As parity increased (range, 0-8), augmented mercury
and selenium concentrations were seen in cord blood
(Rg. 5). Mercury concentrations in hair increased even
as parity increased (p < .0001). The validity of these
associations is difficult to determine from bivariate
analyses because intake of fish (p < .0001) and pilot
whale (p - .001) also increased significantly with pari-
ty. Lead concentrations in cord blood appeared to be
independent of parity.
Cord blood mercury increased slightly (r, - .04; p -
.2), selenium increased considerably (r, - .16; p <
.0001) (Rg. 6), and lead was unrelated to increases in
gestational age.
Frequent consumption of fish for dinner, revealed by
the questionnaire, was significantly associated with in-1
creased B-Hg and H-Hg; selenium concentrations were y
comparatively low and lead concentrations high in \
mothers who did not eat fish during pregnancy (Table,
2). A more pronounced relationship was seen between
the mercury concentrations and the frequency of din-
ners, based on pilot whale meat; again, B-Se appeared
low in mothers who did not eat this type of food during
pregnancy, whereas B-Pb was not significantly affected
(Table 3). Inasmuch as intake of pilot whale and fish
was associated, the relationship of trace element con-
centrations to fish intake was examined in women who
had not eaten pilot whale (Table 4). The tendencies
shown in Tables 2 and 4 are similar, though at different
concentration levels. The ratio between mercury con-
centrations in hair and blood was particularly low if the
mothers had eaten pilot whale for dinner at least twice
per month during the pregnancy (median, 180), com-
pared with those who had eaten this food item once
(median, 202) or not at all (median. 213) (Kruskal-
Wallis, p - .005). Also, the hair-blood mercury ratio
seemed to decrease slightly as the frequency of con-
suming fish dinners increased (p - .03), but this effect
was not apparent when the mother had not eaten pilot
whale at all.
Alcohol intake was limited in this cohort; 771 women
(75.4%) were abstainers, and all other women drank al-
coholic beverages only occasionally. B-Hg and H-Hg
were significantly lower if the mother had occasionally
ingested alcoholic beverages during pregnancy; a sim-
ilar tendency was seen with B-Se, whereas B-Pb was
significantly higher if the mother had occasionally in-
gested alcoholic beverages (Table 5). However, moth-
ers who occasionally drank alcoholic beverages tended
to eat fish less frequently than the women who ab-
stained from alcohol (Mann-Whitney U-test, p < .01),
190
Archives of Environmental Health
-------
'80
Mercury (rtmot/O
Selenium (umot/l)
Mercury Seleniur
160 -
140 -
120-
100
Parity
fig. 5. Increases in mercury (r, - 0.12. p < .0001) and selenium (r, - 0.14, p < .0001) concentra-
tions in cord blood reiative to parity. The number of observations is indicated above each bar.
180
Mercury (nmol/t)
Selenium (umol/l)
1.8
Mercury
Sannium
1601-
~. 1.6
4.1s
140 f
120 r-
100
38-39 40
Length of gestation (weeks)
41
Fig. 6. Variations in mercury (p > .5) and selenium (r, - 0.13, 'p < .0001) concentrations in cord
blood relative to length of gestation. The number of observations is indicated above each bar.
and a minimal difference in pilot whale intake was also
seen (p - .07). Alcohol intake was also associated with
a higher hair/blood-mercury ratio (median, 208, com-
pared with 187 in abstainers; Mann-Whitney test, p -
.03).
A total of 614 (60.0%) women were nonsmokers, 251
(24.6%) smoked fewer -than 10 cigarettes/d, 126
(12.3%) smoked more than that, and 31 (3.0%) smoked
other types of tobacco. The median B-Pb was 77 nmol/l
in nonsmokers and 92 nmol/l in smokers (Mann-Whit-
ney U-test, p < .0001). A dose-response relationship
was seen with respect to dairy cigarette consumption.
As expected, fewer smokers (68.4%) than nonsmokers
(80.3%) were alcohol abstainers. In a stratified analysis,
alcohol usage was an important predictor, but only for
nonsmokers (p - .0004). No significant relationship
May/June 1992 [Vol. 47 (No. 3)]
191
-------
Table 3.— Umbilical Cord Blood Concentrations of Mercurv (nmol/D, Selenium (^mol/D, and Lead (nmol/D, and Maternal Hair Con-
centration of Mercurv (nmotfg). Relative to Maternal Pilot Whale Intake During Pregnancy
No. of whale
dinners/mo
None
1
2
3
4 or more
V, - .51
-------
Maternal age that exceeded 25 v was associated with a
16% increase in H-Hg (p - .003); if maternal age was
omitted from the equation, parity of at least three be-
came a significant predictor.
The multiple regression analysis for B-Se tended to
confirm the bivariate analyses. 8-Se appeared to in-
crease with alcohol abstention, pilot whale intake, high
parity, and long gestation period, but these factors
combined'Contributed only about 5% (rj) toward ex-
plaining the variation of B-Se.
The most significant predictor for B-Pb was smoking
status; if smoking was left out, alcohol drinking status
became a significant predictor for B-Pb. Teenage moth-
ers had a 25% lower B-Pb than the average (p <
.0001). At least two fish dinners per week were associ-
ated with a 12% decrease in lead (p - .007), whereas
pifot whale dinners during pregnancy would increase
B-Pb by 11% (p - .029). Accordingly, cord blood from
the 86 nonsmoking women who had at least two fish
dinners per week, but who never ate pilot whale,
showed a very low median of 70 nmol/l; 70 of these
blood samples had lead concentrations below 100
nmol/l.
Discussion
The main finding of this study was that mercury expo-
sures in the Faroe Islands were primarily determined by
pilot whale consumption. Thus, most of the variations
in cord blood mercury and maternal hair mercury con-
centrations were related to the frequency of dinners
that included pilot whale meat during the pregnancy
period. High mercury levels in cord blood have been
reported in previous studies of northern populations
who depend on marine food. For example, Hansen et
al.'s found similar concentration levels and a maximum
of 177 ng/l (882 nmol/l) in 89 cord blood samples from
Greenland. Also, the large-scale study of Canadian In-
dian and Inuit populations showed that about 2% of
the adults examined had a B-Hg level above 100 ng/l
(500 nmol/l); of 2328 cord blood samples, 14 were
above this level."
Mercury concentrations tend to be 20-65% higher in
cord blood than in the venous blood of the mother."
This difference can probably explain the higher average
in the cord blood samples when compared with the
blood samples from the Lorvik women (Tables 1 and 2).
For comparison, B-Hg levels in Denmark average
slightly more than 5 nmol/l, which is less than one-
„ tenth of the median found in the Faroes.18
>The ratio between mercury concentrations in hair
and cord blood (Fig. 3) is below the expected 250:1,17
probably because kinetic factors and the higher hemat-
ocrit of the cord blood cause a considerable increase in
mercury concentration in cord blood, compared with
maternal blood.17 In addition, the presence of some
inorganic mercury in pilot whale meat may result in
larger increases in mercury concentrations in blood
than in hair (see Table 1 and Fig. 2). However, the data
confirm that methylmercury constitutes the major haz-
ard in this population.
Experimental studies with pigs suggest that mercury
concentrations may increase in some fetal tissues toward
the end of the gestational period, but that these changes
are not reflected in the B-Hg concentrations/9 In agree-
ment with this observation, we found ftoafmercury con-
centrations in cord blood showed no discernible pattern
of variation relative to gestational age (Fig. 6).
The significant increase in cord blood mercury as fre-
quency of marine food consumption increases accords
with previous findings from Greenland'5 and Canada.16
Most of the fish consumed on the Faroe Islands is cod
that has a relatively low mercury concentration. With
the very high mercury levels in pilot whale,4 augmented
mercury concentrations in consumers would be expect-
ed. The increase in mercury levels as maternal age and
parity increase can perhaps be explained by a tendency
of older women to consume more pilot whale than the
average mother.
The possible influence of alcohol consumption on
B-Hg concentrations in this cohort (Table 5) is note-
worthy. Experimental studies Have documented that
ethanol ingestion increases the formation of mercury
vapor (Hg°) from ionic mercury (Hg2*) and that some of
that Hg° is exhaled.20 Parallel observations of mercury-
exposed human volunteers21 confirm that ethanol in-
gestion may actually lead to a decreased concentration
of inorganic mercury in blood. In the Faroese popula-
tion, significant exposures to inorganic mercury origi-
nate from pilot whale,4 and our data could, therefore,
be interpreted as an epidemiological confirmation of
the experimental studies.
When the selenium data are interpreted, account
must be taken of the fact that selenium concentrations
in blood' decrease during pregnancy.22 If compared
with other data on cord blood, the B-Se levels in the
Faroes are higher than seen elsewhere, e.g., in Ger-
many where the average for 17 samples of cord blood
was 80 Mg/1 (1.0 nmoi/l).23 Detailed comparison would
require that the appreciable increase in B-Se after the
expected term (Fig. 5) be considered. As is the case
with mercury, an augmented selenium intake most
likely originates from marine food.-4 The influence of
smoking and alcohol ingestion on B-Se has been re-
ported previously.25
The association between B-Hg and B-Se shows con-
siderable scattering, selenium generally being in excess
on a molar basis (Rg. 4). Experimental studies suggest
that selenium may offer some protection against
methylmercury toxicity,9 although methyimercury-as-
sociated teratogenicity may not necessarily be alleviat-
ed.26 If the methyimercury originates from fish, a major
part of the total mercury content may consist of a bis-
methylmercury-selenide complex that is retained for a
longer time, but it may be less toxic.24 However, the
considerable scatter in Figure 4 suggests that variable
amounts of selenium are not bound to mercury in cord
blood.
The B-Pb concentrations are low, compared with
average lead concentrations in cord blood elsewhere,
e.g., Denmark (23 yg/l or 0.11 ^mol/l, n - 48),27
Finland (37 ^g/l or 0.18 ^moi/l, n - 19),28 Australia (62
or 0.30 fimoi/1, n - 172),29 and the United States
May/)une 1992 [Vol. 47 (No. 3)]
193
-------
(66 fig/I or 0.32 wmol/l, n - 11 837).30 The low B-Pb
may be related to the minimal alcohol and tobacco
use, which has been suggested bv previous population
studies.3' and the low lead .concentrations in marine
food. In addition, the limited extent of air pollution
from automobiles on these windswept islands must be
an important factor.
An important question is whether the high mercury
levels constitute any hazard to the fetus. Exoenmental
data suggest that methyimercury may cause teratogenic
"and other toxic effects.17 Data from poisoning incidents
in japan and Iraq indicate that the fetus may develop
neurotoxic effects, even if the mother is clinically
healthy.32-33 However, despite the occurrence of some
very high mercury concentrations in blood, true poi-
soning cases have not been encountered in fish-eating
populations that reside far from industrial pollution
sources.14 Epidemiological studies of children from fish-
ing communities have reported minor neurological ab-
normalities that possibly relate to increased exposure
to methyimercury.35'36 The presence of very high levels
of mercury in cord blood in the Faroe Islands should,
therefore, be regarded a matter of possible concern.
Support for this study was obtained from the Danish Medical
Research Council, the Danish Health Foundation, the H6jgaard Foun-
dation, the Vestnorden Foundation, and the Danish Agency for En-
vironmental Protection. The authors are grateful to the midwives and
laboratory technicians in the Faraese health care system for their con-
tribution to this project, Hanne Albaek, Brita Andersen, and Rania
Bjemng performed the analyses of trace elements in Odense. Steen
Henrik Sand*, M.O., provided assistance with the computer software.
Submitted for publication April 10, 1991; revised: accepted for
publication October 29, 1991.
Requests for reprints should be sent to Philippe Grand jean, M.D.,
Institute of Community Health. Odense University, j. 8. Winslowsvej
17, DK-5000 Odense C, Denmark.
References
1. Turner MD, Marsh DO, Smith JC, Inglis JB, Glarkson TW, Rubio
CE, Chiriboga ). Chiriboga CC. Methyimercury in populations
eating large quantities ot marine fish. Arch Environ Health 1980;
35: 367-78.
2. Vestergaard T, Zachariassen. [Dietary survey 1981-82. in
Faroes*.] Frodskapamt 1987: 33: 5-18.
3. May K, Stoeppler M. Reisinger K. Studies in the ratio total mer-
cury/methvlmercury in the aquatic food chain. Toxicol Environ
Chem 1986: 13: 153-59.
4. Juhlshamn K, Andersen A, Ringdal O, M4rk4re ). Trace elements
in the Faroe Islands I. Element levels in edible parts of pilot
whales (Clobicephalus meieanus). Sci Tot Environ 1987; 65:
53-62.
5. Bakir F, Damluji SF, Amin-Zaki L. Murtadha M, Khalidi A, Al-
Rawi NY, Tikriti S. Dhahir HI, Clarkson TW. Smith |C. Doherty
RA. Methyimercury poisoning in Iraq. Science 1973: 181:
230-41.
6. IPCS (International Programme on Chemical Safetv). Mercury
(Environmental Health Criteria 1). Geneva: World Health
Organization. 1976.
7. Marsh DO. Clarkson TW, Cox C, Myers G), Amm-Zaki L. Al-
Tikriti S. Fetal methyimercury poisoning, relationship between
concentrations in single strand of maternal hair and child effects.
Arch Neurol 1987; 44: 1017-22.
3. Spyker )M. Sparber SR. Goldberg AM. Subtle consequences of
methvimercurv exposure: benaviorai deviations m offspring or
treated mothers. Science 1972: 177 621-23.
9. Ganther HE. Goudie C, Sunde ML. Kooeckv M|, Wagner P Sang-
Hwan O, HoeKStra WG. Selenium: relation to decreased toxicitv
or methvlmercury addea to diets containing tuha. Science 1972,'
175- 1122-24.
10. Needleman HL. Rabmowitz M. Leviton A Linn S. Schoenbaum
S. The relationship Petween prenatal exposure to lead and con-
genital anomalies. ) Am Med Assoc 1984: 251: 2956-39
11. Magos L, Clarkson TW. Atomic absorption determination OT total.
inorganic and organic mercury in olood. ! Assoc Offic Anal Chem
1972:55:966-71.
12. Lind 8. Bigras L, Cernichiari E. Clarkson TW, Fnberg L, Hellman
M, Kennedy P, Kirkbride I, Kjellstrbm T, Ohlin B. Quality control
of analyses of mercury in hair. Fresenius Z Anal Chem 1988: 332:
620-22.
13. Welz 8. MelcherM, Sinemus HW, Maier D. Picotrace determina-
tion of mercury using the amalgamation technique. Atom Spec-
trosc 1984; 5: 37-42.
14. Fernandez F), Hilligoss D. An improved graphite furnace method
for the determination of lead in blood using matrix modification
and the L'vov platform. Atom Spectrosc 1982: 3: 130-31.
15. HaAsen 1C Christensen RB. Allermand H. Albige K, Rasmussen
R. Concentrations ot_ mercury, selenium and lead in 'blood
samples from mothers and their newborn babies in four
Greeniandic hunting distncts. Meddr Gronland Man & Soc 1984:
6: 1-20.
16. Methvlmercury in Canada. Exposure of Indian and Inuit residents
to methyimercury in the Canadian environment. Ottawa: Health
and Welfare Canada. 1979: Vol.2. 1984.
17. IPCS (International Programme on Chemical Safety). Methvlmer-
cury (Environmental Health Critena 101). Geneva: World Health
Organization, 1990.
18. Grandjean P, Nielsen CD, Jirgensen P)), Harder M. Reference
intervals for trace elements in blood: Significance of risk factors.
Scand ) Clin Lab Invest (in press).
19. Kelman 8), Walter BK, Sasser LB. Fetal distribution or mercury
following introduction of methyimercury into porcine maternal
circulation.) Toxicol Environ Health 1982: 10: 191-200.
20. Ounn ID. Clarkson TW, Magos L Interaction of ethanol and in-
organic mercury: generation of mercury vapor in vivo. 1 Phar-
macol Exp Ther 1981: 216: 19-23.
21. Hursh IB, Greenwood MR. Clarkson TW, Allen I. Demuth S. The
effect of ethanol on the fate of mercury vapor inhaled by man. I
Pharmacol Exp Ther 1980; 214: 520-27.
22. Bro S. Berendtsen H, Ndrgaard ), Hiist A. Jdrgensen P). Serum
selenium concentration m maternal and umbilical cord blood.
Relation to course and outcome of pregnancy, j Trace Elem Elec-
trolytes Health Dis 1988: 2: 165-69.
23. Schramel P, Hasse S. Ovcar-Pavlu J. Selenium, cadmium, lead.
• and mercury concentrations in human breast milk, in placenta.
maternal blood, and the blood of the newborn. Biol Trace Ele-
ment Res 1988: 15: 111-21.
24. Yonemoto I, Webb M, Magos L. Methyimercury stimulates the
exhalation of volatile selenium and potentiates the toxicity of
selenite. Toxicol Lett 1985: 24: 7-14.
25. Uovd 8. Lloyd RS, Clayton BE. Effect of smoking, alcohol, and
other factors on the selenium status of a healthy population. )
Epidemiol Comm Health 1983: 37: 213-17.
26. Nishikido N, Satoh Y, Naganuma A. Imura N. Effect of maternal
selenium deficiency on the teratogenictty of methyimercury. Tox-
icol Lett 1988: 40: 153-57.
27. Milman N, Christensen JM, Ibsen KK. Blood lead and ervthrocvte
zinc protoporphyrin m mothers and newborn infants. Eur I
Pediatr1988: 147:71-73.
28. Korpela H. Loueniva R. Yrjanheikki E. Kauppiia A. Lead and cad-
mium concentrations in maternal and umbilical cord blood, am-
niotic fluid, placenta, and amniotic membranes. Am I Obstet
Gynecol 1986: 155: 1086-89.
29. McMichael A), Vimpani GV, Robertson EF, Baghurst PA. Clark
PD. The Port Pine cohort study: maternal blood lead and
pregnancy outcome. ) Epidemiol Comm Health 1986: 40: 18-25.
30. Rabinowitz MB, Needleman HL. Temporal trends m the lead
concentrations of umbilical cord blood. Science 1982: 216:
1429-31.
194
Archives of Environmental Health
-------
31 Granaiean ?, Olsen N8, Hoilnagei H. Imiuence or smoking ana
alcohol consumption on dlooa leaa levels. Int Arcn OCCUD En-
viron Health 1981; 46: 391-97
32. Harada M. Congenital Mmamata disease: mtrautenne metnvi-
mercurv poisoning. Teratol 1978: 18:285-88.
33. Amin-Zaki L. Maieed MA. Elhassani SB, Clarkson TW. Green-
wood MR. Donertv RA. Prenatai methylmercury poisoning. Am )
Dis Child 1979; 133: 172-77
34. Wheatley B. Barbeau A, Clarkson TW, Upham IW. Methyimer-
:urv poisoning in Canadian Indians—the elusive diagnosis. Can |
Neuroi Sci 1979- 6 417-22.
35. cvssen GEM. Rueav J, Neims A. vtethvimercurv exposure in nor-
;nern Quebec II. Seuroiogic findings in children.?AmJ Epiaemioi
'983: 118: 470-79
36. Kiellstrom T, Kenneov P Walhs S. Mantell C. Physical ana mentai
development or cmidren with prenatal exposure to mercurv from
•Ish. Stage I: preliminary tests at age 4 (Report 3080). StocKnoir-
National Sweaisn Environmental Protection Board, 1986.
ASSISTANT/ASSOCIATE PROFESSORS
IN ENVIRONMENTAL HEALTH
The Department of Environmental Health in the Harvard
'School of Public Health seeks two assistant or
associate professors for its Environmental Epidemiology
and Occupational Health Programs.
Candidates must hold a doctoral degree in a related
field and should have training and experience in
chronic disease epidemiology. They also must have a
record of independent research as well as
demonstrated interest in teaching and in the
supervision of graduate students and research fellows.
Interests in biomarkers and other innovative
approaches to measurement of exposure and health
impact are welcome. *
«
Send curriculum vitae. including a list of publications
and the names of 3-5 individuals who would be willing
to respond to a request for evaluation to:
Environmental/Occupational Epidemiology
Search Committee
Or. Alexander M. Walker. Chairman
677 Huntington Avenue
Boston. MA 02115
EXPRESSIONS Of INTEREST ARC PARTICULARLY INVITED
FROM WOMEN AMD MINORITf CANDIDATES.
too? fVnl. a {Nn.
195
-------
approximately the minimum size of through-holes, which is in the
range 10-100 run. This means that the practical resolution of the
holographic manipulation technique is of the same order as that of
proximity manipulation. The size of the atomic source and the
velocity spread of the atoms are not the real limiting factors of
resolution, as the coherent portion of the atomic beam may be
removed either by a diaphragm or by a velocity-selective chopper,
though this reduces the intensity of the atomic beam. The number
of atoms that contribute to the imaging is limited by the theoretical
maximum diffraction efficiency (6.25%) of a two-dimensional
amplitude hologram. This is a disadvantage for an imaging
device. However, the holographic technique described here is an
example of the more general technique of neutral atom manip-
ulation. It can in principle construct an arbitrary wavefront of an
atomic de Brogiie wave retaining both intensity and phase infor-
mation. When only the intensity of the image is required, it is
possible to disperse the information of each object point over
entire hologram surface, as we did in the present experiment, This
makes the image very resistant to defects in the imaging compo-
nent Atomic holography is not limited to the generation of a fixed
pattern with a fixed diffracting component By combining- it with
other manipulation techniques, it should be possible to make a
real-time moving pattern of atoms, or a device to investigate the
three-dimensional phase and amplitude characteristics of an
atomic de Brogiie wave. O
RmMd 20 SwaniMr 1995; accwxM 13 Mwen 1996.
3. E&r. 0. M. & Semmr. E. K. Htajn M4,524-528 (1990).
2. Hmai, r. & ScftMiow. A. Oat Common. U, 69-69 (1975).
3. PMN«. W. D. & Mmar. H. Pfw. flw. Ua. 4*. 590-599 (1982).
4. Raito. E. L. Puma* »*_ Can*. A, Chu. S. 4 fnttont. 0. E. PH*. Oft. IjUt ft, 2631-2634
(1987).
5.0*1.5.4W«nwv C. («M J. opt Sac. ttpcc. OM. an MHroaaftifan* np0n •*> 118-128 (1979).
irngm nm* nf r«i»nf Thafiti nwaf •»nl
U. Shmai, F., SMrai. K. 4 Tatam. H. Qpc UK 10, 339-341 (1991).
ACmowiEDGEMECT. w> ffanK ff» two* fcr LOT Sdra at M UnMMy of Etocm-
oonmmcMm fcr uw of Mr tattta. Th« wMtwt mpporad ki put b» IM MkMBy of
Gduaeon. Same* ami CuftM.
Photodegradation of
methylmercury in lakes
P. Sellers*, C. A. Kelly*, J. W.M. Ruddt,
A. R. MacHutchont
• Department of Microbiology, Uruvetsity of Manitoba, Winnipeg. Manitoba,
R3T2N2, Canada
t Freshwater Institute, Department of fisheries and Oceans,
501 University Crescent, Winnipeg, Manitoba, R3T 2N6, Canada
MciHVLMERCunr can accumulate in fish to concentrations that
threaten human health1. Fish methylmercury concentrations are
high in many reservoirs2 and acidic lakes9, and also in many
remote lakes44—a net that may be related to increased atmo-
spheric deposition of anthropogenkaily mobilized mercury
during the past few decades'. Although sources of methyknercniy
to lakes and reservoirs an known7, in-lake destruction has not
been demonstrated to occur at the low concentrations found hi
most water bodies. Here we report in rite incubations of lake
water that show .that metfayhnercnrjr is decomposed by photo-
694
degradation in surface waters. This process is abiotic and the rate
is first-order with respect to methylmercury concentration and
the intensity of solar radiation. In our study lake, the calculated
annual rates of methylmercury photodegradation are almost
double the estimated external inputs of methylmercury from
rain, snow, streamflow and land runoff, implying the existence
of a large source of methylmercury from bottom sediments.
Photodegradation could also be an important process in the
mercury cycle of other aquatic systems. This discovery fundamen-
tally changes our understanding of aquatic mercury cycling, and
challenges the long-accepted view that microbial demethyiation
dominates metnyimercury degradation in natural fresh waters.
Recent development of highly sensitive analytical capabilities3
have enabled us to examine the stability of methylmercury in lake
water at naturally occurring, low concentrations. At the Experi-
mental Lakes Area (ELA) in northwestern Ontario, we incubated
lake water in TeSon bottles (which transmit sunlight9) and found
that methylmercury concentration decreased in the sunlight but
not in the dark (Fig. la). Filtering the lake water with 0.45-um
filters, which removed all photosynthetic organisms and most
bacteria, had no effect on the photodegradation rate (Fig. Ib).
The fact that the reaction occurred in filtered water suggested that
the photodegradation was abiotic. To confirm the abiotic nature
of the reaction, we conducted an experiment using unfiltered.
sterilized lake water and found that sterilization did not inhibit
photodegradation (Fig. Ic).
The light-dependence of the reaction was demonstrated further
in an experiment in which one set of hordes was exposed to
sunlight, a second set was dark, and a third set was moved from
dark to light during the experiment Loss of methylmercury
occurred only during the periods when bottles were exposed to
sunlight (Fig. 1^;.
We also determined that photodegradation rates were first-
order with respect to methylmercury concentrations (Fig. 2a) and
to levels of photosynthetically active radiation (PAR, Fig. 26). The
lake water used hi all of these experiments (Figs 1 and 2) had a pH
of 6.1-6.3 and a dissolved organic carbon (DOC) concentration of
1,190-1,470unroll"1. Several comparisons of photodegradiation
rates in different ELA lakes spanning a range of water chemistry
(for example DOC 250-1,210 umoir1; pH 6.0-7J; alkalinity
127-143 ueql"1) showed no differences in rates due to water
chemistry (P.S., CAJC, J.W.M.R- and A.R.M., unpublished
data).
In situ incubations showed that methyimercury photodegrada-
tion rates decreased with depth below the lake surface, and
corresponded closely to the exponential decrease in light intensity
with depth (Fig. 3). Because we have shown that methylmercury
photodegradation is consistent with standard rate laws for photo-
chemical reactions, it should be possible to add this process to
mechanistic models of mercury cycling in lakes and reservoirs10.
The end-products of methylmercury photodegradation in nat-
ural waters have not been identified. In laboratory experiments
done at relatively high methylmercury concentrations (micro-
grams per litre), a variety of mercury and carbon species are
photo-produced11. In natural waters, however, the relatively low
concentrations of methylmercury (0.(&-3.0ngl~l in ELA lakes
and streams) are bound with complexing agents which are also
present only in trace quantities12. Therefore, me chemistry of the
reaction in natural waters may be different than our laboratory
experiments.
It is important to understand the difference between the
process of photodegradation of methyimercury described here
and the process of photoproduction of elemental mercury. (Hg°;
ref. 9). Photoproduction of Hg° is thought to occur primarily by
reduction of inorganic mercury (Hg2*; refs 13,14), and flux of Hg°
to the atmosphere is an important mechanism by which mercury is
lost from lakes14. Photodegradation of methylmercury may or ma'
not produce Hg°. If it does, the estimated production rate of Hg°
in our study would be 1.5 ugm'^yr"1. This is similar to the flux of
Hg° to the atmosphere measured in two previous studies of
NATURE1- VOL 380 • 25 APRIL 1996
-------
To estimate whether the observed photodegradation rates
could have a significant effect on the methyimercury concentra-
tion in lakes and reservoirs, we calculated a methyimercury turn-
over rate for the epilimnion of a typical ELA lake (Lake 240)
during summer stratification of 1994. We used the following Lake
240 data: (1) the average epiiimnetic methyimercury concentra-
tion (0.07ngT1), (2) the average incident PAR (35 EnT2 d~'; refs
18, 19) and (3) the average extinction coefficient for PAR
(0.5m"1) in the epilimnion (0-4m depth). We also used the
relationships between photodegradation rate and methyimercury
concentration (Fig. 20), and between photodegradation rate and
light intensity? (Fig. 3). From these data we estimated that the
average photodegradation rate constant in Lake 240 was 0.043 d~'
for the ice-free stratified season. This means that the methyimer-
cury in the epilimnion of Lake 240 would be replaced about seven
times during the period of the summer stratification.
We made a preliminary comparison of the importance of
photodegradation relative to the annual masses of methyimercury
flowing into and out of Take 240 at ELA (Fig. 4). Using previously
published20 and our unpublished data, we estimated that the total
methyimercury input (direct deposition, stream flow, and runoff)
was ~370mgvr~l and outflow was ~ 140mgyr~'. Thus, if photo-
degradation is not included in the methyimercury cycle (Fig. 40),
one would conclude that there is net flux of methyimercury from
the water column (net destruction and/or storage). But when
photodegradation is included (Fig. 46), there must be an in-lake
source of methyhnercury of ~450mgyr"'. Because the entire
water column of T-afcg 240 is oxic, and because mercury methyla-
tion is primarily an ancotic process^, this in-lake source is expected
to be the sediments at a rate of ~1 ugm"2 yr'1. This mass-balance
shows that during the year of our study, the flux from sediments
must be similar in importance to external inputs as a source of
methyunercuiy to fh*« laifg
Knowledge of photodegradation may be useful in the design of
methods to mitigate methyimercury problems. For example,
where effluent water contains high concentrations of methyimer-
cury, it could be retained in shallow ponds before discharge.
Coincidentalty, this should also decrease the concentration of
inorganic mercury because of photoreduction of Hg2* to volatile
Hg° (ref. 9). Knowledge of methyhnercury photodegradation
would also be used as a criterion for site selection of reservoirs,
which commonly contain mercury-contaminated fish. For exam-
ple, peatland sites are not recommended because water in reser-
voirs that flood peatlands is stained brown by dissolved organic
Lake budget without phcxodegradauon
Lake budget with phocodegradaiioa
FIG. 4 Preliminary annual mass-balance budget for methyimercury
(mgyr1) in Lake 240 without (a) and with (o) tne inclusion of photode-
gradatxxiintfiernetr)ylrrercurxcyde.Tnenetftuxofmetfry^^
the sediments is trie net result of methyiation, demetnylation, and accu-
mulation in tne sediments.
METHODS. Direct wet deposition and terrestrial runoffwere estimated from
previously published data for the ELA20. Stream inflows were calculated from
measured volumes and concentrations (our unpublished data). Lake 240
has a surface area of 44.1 ha and for 1994 had an average epiiimnetic
methyimercury concentration of 0.07 ng r1 and a total mercury concentra-
tion of 0.8 ngr1, an average DOC concentration of 575 nmol r1, and an
average pH of 7.06. An annual photodegradation rate was calculated as
described in text
Avenge daily PAR (E or1
-------
LETTERS TO NATURE
lakes14-23, but smaller than Hg° production rates in another study9.
Of course, identification of methyhnercury photodegradation
end-products is necessary for a realistic interpretation of these
rate comparisons. Even so, the immediate significance of methyl-
mercury photodegradations that it decreases the concentration of
FIG. 1 Methyimercury concentrations in lake water incubated in light and
dark (aluminum-foil covered) Teflon bottles on the surface of the lake. Each
panel represents a separate experiment in which all bottles were incubated
under the same light and temperature conditions. The experiments were
conducted on different dates. Water treatments: a, unfittered water; b,
unfUtered water incubated in the light and the dark, and filtered water
incubated in the light; c. unfiltered, stenlized water d, filtered water in the
light and the dark, or transferred from dark to light In a and c, there were
two sets of bottles incubated for each treatment The lines represent
separate sets, with each data point representing a single analysis from a
single bottle incubated. The low initial value for one of the light treatments
in c is due to analytical error. The data points in d represent duplicate
analyses for a single botde at each point in time, and range bars indicate
where these duplicate analyses are larger than the data points.
METHODS. AH sampling, manipulation and analyses of water were
performed using ultra-dean techniques*-. Water was incubated with a
50% headspace under the following temperature (mean daily) and light
conditions: a and c, 17.5 °C and 23 E m-J d-1; o, 16.5 °C and 21E m-J d'1;
d, 20.0 °C and 25EnrId-1. For 6 and d, the concentration of methyl-
mercury was elevated by addition of a methylmercunc chloride stock
solution (made using distilled water). Fore, water and bottles were stenlized
by autoclaving, and were not opened until analyses. Lake water used for
these experiments had a pH of 6.14, a dissolved organic carbon concen-
tration of 1,410umor1 andachtorideconcentration of 0.4mgr1. Methyl-
. mercury concentration was determined by a distillation extraction, aqueous
phase etrtylation and atomic fluorescence detection1. A control experiment
ensured that methyimercury did not absorb on the Teflon bottles: after
Incubation, acid ONHjSOJ was added to sample bottles to a final
concentration of 0.09 N. No methyimercury was released back into solution
(data not shown). .
rNATURE • VOL 380 • 25 APRIL 1996
methyimercury, which is the form of mercury that is most easily
accumulated in fish and that is the most toxic to fish consumers.
Two lines of evidence show that abiotic photodegradation is a
much more important process in epiiimnetic lake water than is
biological demethylation, which can occur in the dark. First, our
experiments repeatedly showed that methyimercury concentra-
tions in unfiltered water were stable in the dark (for example, Fig.
la), indicating that biological demethylation was either absent or
was occurring at rates undetectable with our methods. Second.
biological demethylation can be detected in lake water using I4C-
methyimercuiy at very high (1,900 ngl'1) concentrations (incu-
bated in the dark), but the turnover rate measured is about 350
times slower than it is for photodegradation13. The data wa present
here therefore call into question the long-accepted view that
biological demethylation is the dominant removal mechanism of
methyhnercury in lake water. However, biological demethylation
may be important in sediments where methyimercury degradation
rates measured using 14C-methylmercury in the dark are much
higher than in the water column13-16:
Given that numerous studies of demethylation using UC-
methylmercury have been done in the past17, one may question
why photodegradation has not been observed before. The reason
is that in all previous studies the samples were incubated in the
dark17. ' \
tototf dalr fAR - «« E ra' <
r -0.92
f
-------
LETTERS TO NATURE
3. Wemer. J. G. Trans. NA Wlkt. i, Nat. Ro. Conf. S3,6*5-657 (19S91.
4. 3oaaly.S. A.. Rudd. J. W. M.. Fudge. R. J. P 4 Kelly, C. A. Can. J. Fan. aouat Sci. SO, 980-987
(1993).
5. Haines. T. A., Xomov. V T & Jagoe. C. H in Mercury Pollution: integration ana Symnesra. ieos
Waffas. C. J. 4 Hucxauee. J. W.) 397-407 (Levus , Chelsea. Ml 19941.
6. Swan. E. 3.. Engsrom, 0. R.. angpam. M. £., Hennins, T. A. i Bnsnn*, P. L Scene* 2B7.
784-787 (1992).
7. Rudd. J. W. M. Wat. Mr Sal PtMut M, 697-713 (1995).
8. Herat. M.. bang, L 4 Bloom, n. S. Aiaryt cnim. >tca 2*3. 153-168 (1993).
9. Amyoc. M.. Mi«M. G.. Lean, 0. R. S. 4 McQueen. 0. J. £nw. So. recnrx*. 2*. 2366-2371
(1994).
10. Hudson. R. J. M.. Ghenm. S. A., wains. C. J. i Pofceaa. 0. 3. in Mercury feHuOon: IntergnOon
and Svmnesa. (eds Watras. C. J. Hucxaoee. j. W.) 473-523 (Lews . Chelsea. Ml, 1994).
13. lnc*o, M. Emr. Pw/uf. 8 2,3-10 (1981).
12. Zar*, a G_ Bauaftnan. a L. Worte. N. L 4 din*. 0. M. Emir. Lett. «, 117-127 (1974).
13. Mao, Z. F_ Mum*, j.. SBOTCen 0. & Undovat, a n AMcuy MUM: ftitaoyaoun and
SynMaoi (a* Wat/am. C. J. & HucXaUa*. J. W.) 581-592 (Lawa. Cnataaa. Ml, 1994).
14. Vandal, G. M., Mam, R. P. & Recmd. W. F. Wat. AfrSoii ft***. M, 791-803 (1991).
15. Xun, L-Y., Campbell. N. E. R. & Rudd. J. W. M. Can. An aquae So. 44,750-757 (1987).
16. Ramal. P S.. Rudd. J. W. M. 4 Hecxy. R E. ADO. Emir. MtooOKX. 81. 110-114 (1986).
17. Winfrey, M.R 4 Rudd. J.W.M. £rtv»r. Taxcol. Clem. », 853-869 (1990).
IS. Fee. E. J. Can Teen. fleo. fisfl. aouaoc. Sci. 174O, (1990).
19. Shearer. J. A.. DeBryun. E. R., DeClera. 0. R . Sctiinaler. 0. W i Fee. E. J Can Teen. flea, ns/i
Aauac Sa. 1141, 11385).
20. SL Loua. V. L et*. Can. J. Frsn. aouat. So. SI. 1065-1076 (1994)
21. Gilmour, C. G.. Hemy, £. A. 4 MitcneH. R. Emv. Sd. Tecfinol. 2«, 2281-2287 (1992).
22. HultBarg, H., Ivertetdt A. 4 Lea. L H. m Mefcuiy Pollution: imegranon and Synrnesa. (eas
Watras, C. J. & Huckaoee. J. W.) 313-322 (Lewis. Chelsea. Ml 1994).
23. Watras, C. J. eta/, in Mercury AV/uowrtntegraoon and Syntnesa (eds Watras. C. J. 4 Huckaaee.
J. W.) 153-177 (Lewis. Cnebaa, Ml. 1994).
24. Hemy, E. A.. Oodss-Murany, L J.,8i0iam, G. N.. raem. S. M.4Gilmour. C. C. Wat AtrSalPoiM.
M, 509-518 (1995).
A«NOWl£DG£MENIS. Wa mar* R. Harm. R. E. Hacxy. R. H. Hodan, 0. M. RoaantMnj. 0. W.
Sctandlar. R. Scarkng and V. L St Lou» tor mar commeno on an earner varans of tna> manuscript
WeeapeoaHytnarWK. J. Scon XVII tor ner«iom on etpanmaiiial design. Ihti wore was suoporwjoya
FeMMonip (P.S.) and a Oeoartment of Rsnems and Oceans sudwnrjon grant.
Active cycling of organic carbon
in the central Arctic Ocean
Patricia A. Wheeler*, Michel Gosselinf,
Evelyn Sherr*, Oeiphine Thibaultj:,
David L Kirchman§, Ronald Benner||
& Terry E. Whitledge||
* College of Oceanic and Atmosphere Sciences, Oregon State University,
Corvallis, Oregon 97331, USA
t Department d'oceanographie, Uruveisrte du Quebec a RfmousW,
300 AHee des Ursullnes, RImousW, Quebec, Canada GS.3A1
tINRS-Oceanologie, 310 Allee des UrsuHnes, RimousW, Quebec.
Canada G5L3A1
§ College of Marine Studies, Univwaity of Delaware, Lewes,
Delaware 19958, USA
|| Marine Science Institute, University of Texas at Austin, PortAransas,
Texas 78373, USA
THE notion of a barren central Arctic Ocean has been accepted
since English's pioneering work1 on drifting ice-islands. The year-
round presence of ice, a short photosynthetic season and low
temperatures were thought to severer/ limit biological produc-
tion14, although the paucity of data was often noted. Because
primary production appeared to be low", subsequent studies
assumed that most organic carbon was either derived from river
inputs or imported from adjacent continental-shelf regions1'4.
Here we present shipboard measurements of biological produc-
tion, biomass and organic carbon standing-stocks made during a
cruise through the ice covering the central Arctic Ocean. Our
results indicate that the central Arctic region is not a biological
desert Although it is less productive than oUgotrophic ocean
regions not covered by ice, it supports an active biological
community which contributes to the cycling of organic carbon
through dissolved and particnlate pools.
From 26 July to 26 August 1994, the United States and Canada
conducted an interdisciplinary expedition aboard USCGC Polar
Sea and CCGS Louis S. St Laurent to study the role of the Arctic
Ocean in global change and the potential effects of climate change
on biological production in this region. Beginning in the Chukchi
Sea, the trans-Arctic station line ran northwards between 170° W
and 170° E crossing the Mendeleyev ridge at 80° N, and traversing
the Makarov basin. The ships reached the North Pole on 22
August Two additional stations were completed in the Nansen
basin between 84-86° N and 35-38° E before leaving through the
Greenland Sea. Here we report microbial and mesozooplankton
standing-stocks, primary production, bacterial production and
dissolved organic carbon (DOC) concentrations. Sampling was
primarily in-the upper 100m of the water column where organic
material is most abundant""7.
NATURE "•'VOL 380 • 25 APRIL 1996
Surface water entering the Arctic Ocean through the Bering
Strait had distinct Pacific Ocean characteristics, readily seen in the
initial high Si(OH}4 concentrations which decreased by a factor of
ten along the cruise track (Table 1). Phosphate concentrations
decreased by 50% from the Pacific to the Atlantic side, while NO,"
concentrations remained at low to moderate levels in all areas
(Table 1). Severe seasonal nutrient depletion commonly observed
a
~ 40
u
0
2
u
(U
140
120
100
10
«0
.«. .*..,
b
. C
70
73 SO 83 90
Luanda fN)
FIG. 1 a, Percentage of total phytoplankton production released as
dissolved organic carbon. Phytoplankton here includes any segmenting
ice-algae that are in the water column. Dashed curve, second-order
regression with r2 = 0.51. b, Mesozooplankton standing-stocks for 0-
100m depth. Zooplankton were collected in vertical net tows (200-n/n
mesh, 0.75-m opening), dned on board ship, weighed and then analysed
for carbon wrthaCHN analyser. Dashed curve, second-order regression with
i1 = 0.834. c. Total organic carbon in the surface water. Samples were
collected in duplicate, and stored frozen with no preservative. Thawed
samples were analysed as described by Benner and Strom30. Dashed
curve, second-order regression with i* =0.634.
697
-------
JUNH34-1996 13=37
US EPfVEAD Washington DC
202 260 7185 P.02/12
645
Intake of Mercury From "Fish, Lipid
Peroxidation, and the Risk of Myocardial
Infarction and Coronary, Cardiovascular, and
Any Death in Eastern Finnish Men
Jukka T. Satonan, MD, PhD. MScPR Kari Seppanen, MSe; Kristiina Nyyssdnen, MSc
HeikH Korpela, MD, PhD; IBB* Kaunaaea. MO, PhD; Marjatn Kantol*. MSc;
• Jiakko Trowilehto, MD. PhD; Hermann Esterbaaer, PhD;
Raw Taober, PhD; Rfitu Satonon, MD, PhD
U!
Even tfceugb previous studies taw
»n association between high fish tonka snd reduead ojroiury
heart disease (CHD) mortality, men in Eavtara Finland. who
ha*« • hith H* Intake, hm sw «wepcicm*]ly high CHD
morally. W« hvpothecized that this pundai ewiW be in part
explained by high mc«ury content In ibb.
MtOutO «n4 AmAi V/e studies the relation of tnt dietary
Intake of fish and mercury, as well ts bajr coflttnt and urmsry
excretion e( eternity, to t|» ri*k oracuta myocacdtal inbroie*
(AMI) and death from CHD. canfiowaraJar disease (CVD).
and any eavse te 1883 men aged 42 » obadew«sm«i«d
toeal nantetty Ibh speetai inri «i««ate4 baic laerRny vmuuu.
la Cat moduli with the nujur aNtfanaseular rift itaon at
dietary intake* of fich and onmuf t>wn
HHth lignifleanUy mcrnsed risk at AMI md dMdi bwe CHD.
A high fish intake has been associated with a
reduced mortality from corooary heart dis-
ft-»5? (CHD) ia sevefaj prospective papula*
tion studies.'-1 la at least three more recent studies in
populations with high fish intakes, no such relation
was. however, observed.*' Also* evea though men in
Eastern Finland consume a tot of fish,* their mortality
frooi CUD is one of the highest in the world.' This is.
to controversy with the concept that a high fish intake
would uniformly be healthy for the cardiovascular
system. Studies that found an association could not
establish che nutrient ia ftsh that was protective, even
though n-3 polynnsaiuroted fate were suggested to
have a key role,1.*
Because ia Some populations a high intajee of fish
did. not appear to be associated with reduced CHD
aortality, differences have to exist in the nutrient
Recetad Jua* 1, 1994; nvfeioa acocnud Au«ust 19,179*.
From the feacach Inatiojte of Pobtic Health (/.«- JCA, fclf.
JJC, BJ.) and Depa/uwiua el Connuui9 tlnUi snd Orneral
fneoem (HJC) md dteaistrr (MJL). UfrtmsHy of Xuopie,
IfnUedt die Deeatiiwiu «f EpWcnitology aad Health Vraawrion
(J.T.X Ac Natieawl publie Haalth Inrttws of Finland, Hetainkj.
KfituuJ; and d« huntute of BRoeltetnisny (Hi. P.T V Om»enny
ofGtix, Austria.
ConespgnricBe* w Prtf JukkavT. Salonen, Univcniqr eH Kiropio,
PO Box 1627.70311 Knopie. Fjfllawt
« 1995 Ameri^n tU«« Ameiatieii. Inc.
, and any death. Men m (tie highcn icrtile (%2.0 >ig/g) of
hair merauy content bad a 2.0-fald (95% oonfidcnee interval.
1.2 te 3.1: f".OOS) *g* Md CHDsidjuntcd risk o/ AMI and a
2>Wd (95% d. U to ft* P--OJ4) Mjus«0 r*k of cardio-
rsscuinr deaiA eempared »f»h those with e tawr Mir tntrmty
ooMteac In • aesied cas*«entrol *ubs*jnpl«. die 2*-bnur
urinary mercnry aeration had a .-ritirincani (P».CHJ) mdepen-
4enr association «Kth the rwk of AML Both tne nair and
ariiMqr m«taw> tsweiateJ <%niiicanny wM liters of immur*
camplexa eontaJainy raodfaxq LDU
CMwnuiMr 3Tw« data svggwt that a high fnttte of mer»
•wr Ama R«mreojr drethwater fl»i and in« e»iiM»o«itm aeco-
et* nwrmry ia the body are associuKd «itn an exeew
liskef AMI as wtll M death from CHD. CVD, and any cause
hi Eastern tfututk men and dw< increased risk may be ihm ta
ike pfefnetbn of !!eJd eeroxidarioii ny mereury.
eompotilioa of nsh, or possibly some unmeasured
harmful saiMtaiices in fofi mifhl account for these
ineoasisteaem. We hypothesized that mercury in fish
could counteract the beneficial metabolic effects of
other nutrient* in fish.
We have earlier observed a relation between sele-
nium deficiency and aa excess risk of acute myocardial
ia/orction (AMI) M well as death from CHD and
CVD ia Eastern Finland.1* The finding was subse-
quently confirmed in another prospective population
study11 sod hi x case-control study.'* We also found an
association at both low semat selenium levels13 and
lipid perosadarion in. vivo" with accelerated progres-
lioa of dtotid aiherosclcrosii in Eastern Finnish men.
In another prospective population study in men in
Eastern Finland, the) Kuupio hchacmk Heart Disease
Risk Factor Study (JOHD), we found a relation
between hij^» dietaiy innka and stored body level* of
iron, a eaiarjrn of Upid penwidaticn. with increased
risk of AMI"
Because mercury is another transition m«tal that also
caa catalyze Upid peroxidAtioa,1*-17 dad because mereury
counteracts the antfesidative elect of selenium,'* we
hypothocfeed diat s high dieraiy intake of mercury from
freshwater fish could be associated with an increased risk
of AMI aad dearJt from CHD, eazdravucular disease
(CVD], nod any cause.
-------
JUN-04-1996 13:09
US ERVEflD Uashinaton DC
202 263 7185 P..04/12
1
r
d
;d
v-
'U
*y
or
ms
00
sir
ad
?l«
ua
W
of
!«-
35
ies
an
an
tad
Jrt
? Z
1 in
A
jht
ine
M*.
V31
V,
MM
xd
ase
rp-
111
the
serum. This raiiemi *as prepired bv addina; unmcniura
sultan (35% saturation) n the scrum u *-*glabulin fractal v» dihind LlOOO. Tw nonspecific
binding o£ ail 187 scrum aanpics was measured la tbt san»« way
except thai plstci were not coated with the rebbil aaoieivni.
The ttar w»* csprcttcd. after nbcxac&ofi of nonipedfie bind-
ing at t percent of the signal of the standard scrum.
Otter Cktmical Mtosunmtntt
Toe main Upoproteia fraction* (VU3L LW, and HDD
•ere tcpantod Ion fresh serum waplcs using- otoaesnirtf*'
' i nd precipitation*) dvsetib^ prertoady in dtaa.aThe
j sad HDL, juWtaetiom wew separated during m second
viffacsnsUugil spin at 108 000? for 42'boon apia* ^ dcrsny
of l,US e/eaV. Tin cholesterol «onMau of afl Bpopcown
annum wen atcawradr catynutkaUy (CHOO-PAP flwhod.
Boehrinawr M*n»hri») aa UM day «fttr tht hit *p»a Apdi-
B *as UcmiaiimJ by an iinaiuaoturbidiacttic
mpio* for temtn
for 1 Co 5 yeats.1* ft m'tin coneeatnaaas were nuasttred with
a double antibody ndioimmunoauay (Amenta* Inteiurion-
«l> Plasma flbtinoaea conceatnrjon wu aeaamed bated on
cloning of dinned pinna whs cacna tbcombia with the
CeaguloaMter KCA (Heinrka Ancfone GaAHM* AJbumm
eoncencmtons were nscanrcej uom itiui«i} scrutn swaptcs
photornerricaOy (Kon« Oy). The bct»v«n-baica wcfioem of
variation for afeamin *BS 15% at the lewd of 33 g/L. Blood
glucou was raeannd by (hieose dahydrogenase method
(Merck) alter precipjnrioo of pronios wfaii Biihlerie ecetie
acid.
Seruai selerrhmi coneentratio* «es measured by ta ctomie
tbsorpdoo jpcctromerrie method wring graphite fane**. Zee-
nan background eorreetion. «nd pyrvfytimUy coated graphite
mb« with • platform" aad scrom copper wixh atomic tbwrp-
don (Boetrometiy with flam* ^wmaarton,54
flairaa aicorbatc eoncsacnouna w«n mcesareJ »i»h a
luCii^pvrivreitAncB WJOK» ctutittvtt&fRtpiif IHCQ^V*** wifnvy
aieorin* raenbociM w*t* aecaycd hy a eolofhnatrie ondwd in
24-noiir yrtiM «wnp)et.^ Th« niDMim
Boti was multiplied by the 2A»h«mr wim vohme to derive lite
24-Jiour exertdon.
Aaicssmcnt of Dfefiirr Ftth and Mtmiy Inttkt
The eomunpnoa of (bods »w «i«s«d at UK ome of blood
sampling with in Butractai «d u>Ui»ia» wa* cheeked by *a mtervio»«».
Rcs«a( tRtervtav* to obuin raedkal h
. TBe £un9y lr«to<7 of CHD »a» defined M po*ci*«
if oither tha Wotey'csl father, mother, jiner. or brother oC ite
Jubjetr had « libiory of CHD.
A aioject was 4«Ancd a smoker if hs hid ever aowked otk t
rf «/ Mtreory, Bth, aad Coronary Dfccan
±« piir 30 diys. T&e life-long sxposure CO smokln; ("cjjarerre
^»n1 «r«5 otfonicd as Ote product of years tanM mi th«
camber of tobacco produea smoked dally *t the time of
mninattoa. Yean smoked were 4«An*d th« rutu «f ye*n of
smokiax jejarttea of woen itnoiiins t>"d staned. whether Un
subject had (topped anofcinj. «nd *fc«t>>er smoking tud oe-
aurred eontiiMiOuiljr or during siv«ral penod5. The eomomp-
tioa of «te»hol in (tie praviout 12 monttts vac isxeued with
the. quantity-frequency method by uxing the Nordic Alcohol
CoMwnptrO* brMntory. Therefore, (he losses to
Between Macdi 196* aad Oeeemb«r 1991. a denVtiie or
peodble htri or eonftui AMI wu rcpstared ID 73 of cfce 1833
nca at risk F« & «t the** aaeav a J4-hour urinary sample had
ben eoBectsd in ia« baselioe ennuRarJon. Thus. JBihrsas
ooneentutt} tne aCTociafton of innuuy wwuuy w^ta AMI "etc
boced Oft 69 p*Bei»a aad US control subjects. Ia the can of
mnlo'ple eventi during the follow-up, the tot cm for each
subject van taken uiha end point in the present analyses. By
me ead of 1992. af 18 desda IhW had oeeurted. T« deatfts were
due to OJO (btsmadorial OatMfieaiioB of Diuisn (1CD[
41OA1*) and 2* ware doe n CVD (ICD 390-»S8>. The longat
faUow-up pctiod for AMI« for indr>KKi«l subject* w« 7.75
ycats, and the mean jullowr-ap dau we* •ppraacinaKJy 5 years.
The MkHMq> period tot deatta waa as k»g *s 3,75
MraciiM auomiiinacely 6 jean.
-------
JUN-04-1996 13=38 US EPfl/EAD Washington DC
£46 Circaiatioa Vol 91, JSfo 3 Primary 1. 1995
202 260 7185 P.03/12
f. 3' it
Methods
Subjects
The KJHI5 it a population «udy to inveiiicate
unestsbltshed rick factors for AMf and carotid aihefowtaroiii11
in man in Eanern Finland, the population with the highest
recorded CHD Incidence and mortality.' One of the m»in
purposes of die study was to investigate the rote of lipid
peftoddaiien, pro-eadative nunwals, and amiaxiduu ia Mh-
erosdenais and CHD. Hair and urine anples were cottecsed
at tlw baseline exammaoan tor mercury denxmiaadoiu, Tbe
iabOB* wer* carried out between March 1*M
and Deeember 19».
Tie study sample «u competed of 3333 ma in Eastern
Ftoland aged 4], 43.54. or CO rev* at the hue (fee exrmiaaticn.
Of those. 2682 (SZ9%)p«rtMpate«t Men with either prevalent
CHD (n-*T7) or history of uajBUmamilar ttrake (n*«X
elaxidkaew (n«10S), or cancer (••44) were esekdcd fret*
the present analyses, ac these. djseaaea could has* influenced
the dietary babia. Pravatem CHD ** defined if either a
hiiory of AMI or angina pacearki or positive angina pccicris ra
eJfon in ROM Interview* or the use of ntoogr/eeriB taMtts
onee a 'week or mere OcquetrtT- Tlw diagnoM of stroke,
diudjcaaon, and cancer was eased oa u icfervieer and exam-
iaanon by • pbytktaa, Of die reiMdniat; 188* n«, data oa
Veto flab intake and hair mercury content were a»ai)abl* {or
1933 men. For these men, data on cynolic Mood precm w«re
missing for 10 men. on eigucae-jean for 36 man, on serum
ferriua eoneeuratiaa {or 47 man, OB senna apeiipopmtUi 8
for 61 men, on serum HDL, cholesterol for 72 men. and oa
iM&mal oxygen uptake tor 183 nan. For dxte men, the raeaa
valve of the variable in question in (ha entire cohort (n»1833)
Unnirjf inticuff excretion MO* dettnthofld in i subset of
subjects eomiiiing of tboaa ^ho bad an AMI during tbe
Wtaw-up and a double munfter of coMiol subjects «fc» had no
AMI during the follow-up. TVra eaatrai f aetjecw wen matched •
n» eMft ptilcnB actordinf to aee, munidpaliqr of residence,
and dale el batdioe oearainttjoo, A btjcme 2*^neur urine
sample •» available tor o9 pstionu and 138
toulbi 207
Laboratory Methods
Blood; flair, ANO! Umw SamfGng
The examinatioa protueul and ncasufenen.B beve been
(toseribed m O«u3 prewienttly.*'*^'* Svbfutt came la fr*
OlOOd aad hair speeanens beeween ftOO and 10^)0 AM on
Tuesday, Wejdnetday, or Ikurcday. "Diey «en iwtrueted to
abstain ftom iagesting alcohol fat 3 days. (Km smoUtu] fcr li
houn, ami from eating for 12 hews. After (he subject bad
tested in the supine position far 30 minuws, blood wa* drawn
with Tcruno Vcaoject vacuum tubes. No hwnio,«w w» used.
For Hood sampling, a bait sample avenging M m| wat cut
from die scalp hair of Ae subjects farnereuiy nmsuremenu.
The sobjeett had a It-howr urine SMipie eoueeied dvrini the
24 hours preeodnn the study visit wiwn blood umpke were
DOfrmatattoH of Bar and Urine Mmary
Mcscmy ia hair and urine samples was determined between
Mav 1993 and August 19*3 by flow inioenost aoar/sis-cold
vapor atomic absorption ipcctroiMtry BM amlevmBtioa, Hair
samples were processed in a random order W in* Department
of QnnistiT «i r4« UacVeahy «f Knopio. UK chetsost who did
the measurements w« blinded wfifc regard eo «B rUk Encrer
vahtes and health otneomcs.
AJIrc«ejwiwereat«aa (te/g (Commraswn
vi tho Eurep^n Conuaaaitiu, Commugjcy Bureau of Aofer.
cnae. Bruswh, Beicjuca), aad AW firom 2 osnffol jubject. Pot
\uin% ne quaij^r conirol ma*rtal w» S*rwnorra Thtc* 8le-
meats Urine «i^ «n added inenury of 504 >tg/L (Nyeomed
PharmaAS).
Sampto oiaen&Btion was performed with a Microwave
DfeestnA S>S(«m (node! MDS-S3D, CEM Corp). Dry hair
nmpke 0 to £0 mcj and concentrated urine nmplei (LO mL)
»en treated »ith a mixture of Suonpur niizie aad and
hydnchlene acid. After mirteraJaation, sample* were «sbi«
plat, Mmdardj dlhued boat a xioeK mercury loiurioa (1000
mc/U Merck), end aaeisiy OMtraJ maumb «v detennmcd
with dw FIAS-JOO Flow JBjecrioa A«alyji» and Amalgam
Synon (Bedeaceewerk Perife.Efcaer GmbH). Mercury was
redoeed a Ae Otemifoid of RA5.200 w atomi* BMRury
vapev wirh KaBIL Aram awreuxy v»por was firsr carried by
a nreaia of 99S9M argoa (AGA) into tbe Amalgam Syitem
ind aor {••> tta* owrtt ceO in ttte Perka-ElrBcr Zecnea 5000
Spetaimmtei (Bodewee»»rk Pcridn-BmeT GmbH),
dw eoandty of mcfoary was meeswed et 253.7 pm.
Tm we« value* of meronrf oonlem in UM UPFSfS nair
peel, «B *» Hoar, and » tbe BCR material were 1 .10. 5.4*. and
1.97 fij/B» »Bd *»e eoeOeiants «t variation between sarupte
bateba were 7 Jm 3f97v, ind «.!%, nepeedvejy. Tlte vviedoa
cocrtdcnq for a subject's hair were i.6% (n=4l), 7,8%
(a-«52)« aad 7.7% (n-92) in duee periods, duriat each of
which «ec batch was used. Tne meao value at mercury ooneen-
ntntB m the. Serooorm Trace Eteeaeats Urine »« 49J5
ut/L. and dw tocifcfcac of variattoo between sample scries
(inehtdmc fyoehfluarica, diceuiea. and measurement prow
dorc) vaf 9 J*.
To «bdV the traefcinsj of hair mercury valocs o«er lime,
repeat hair tampw wan eolleeud and the mercury comems
were measured Aril subfec«t4 to 9 years (neaa, 6 yearn) after
(b« baseline etamineboa. Peaieeei oarrelalioa ee«daaM be-
tMsei the originat ud the repeat meaannm«ia wa» .91.
DttutuautinR tftmmun* Camfiaxs ConuuuMg
eewtainiai; ojodbsed L.DL were
Serum
measured to a wbtampie of 1S7 comroJ fabjeetc wing an
ELBA ueay with cappcr-oadiacd LDL as the antisj«n. Servrn
mnpla had been kepi (TOMB at - 2/TC tar 3 »o 8 j««« and had
not been thawed previously. The a*a.y> worn perforated at the
Institute of Btacfcemiear, Uar«eftity of Graz, Austria, in 1992.
To derive owfited human LDL anOeen, LDL WBS Ittonted
from healthy daaoa by densfey gredioH vkraccntrifupiwn
and ««idi»ed with oopper iaru as previweriy deseribed.*1
T» prepere raWw antiscn M oddked LDL. twa rabbits were
imrminizcd five times with oaidiaed LDL it mletvaie of 2
weeks," Andien derived troaa these two rabbits were used in
parallel m the present study.
MTcrocrw plates Qbe» MKdforp) *ere co»t«d with 250 (O.
of UOOO dnetodi flitOMrum m CO, bii«er»ami lte«4 o-em'ght
•i -H?C. After washing (Oubeeee/e phwiphate-fcosfared saune
pBSl whb UJtNaClawiiL of serum sample and Incubation
far 2 honra at +JTC After wajhinj. dw assay w« developed
with a leaf wtHniman I«O antibodjr labeled wMi pemddvM
(Medac) and terramathyibenjidine M subttraie. Ttte absorp-
IMA was measured at 492 wn. for ealmradeix, a human scram
(hat gave & very high signal in this assay wu rosd as the
standard aod was assayed on each plate in paralM with *l>
e«her mnple: The tiler ww ftprmnd M •« pvrceot of the
sigml of the standard nerom. AU uaayn were done in duplicate.
ma* the sneaa rifcer was otedin be statistical anaiysk.
The entire setiet of hnmtiM eaeapitx RKasurctteno was
repeated ustoc • T£>obalui ihedoii ol another rabbit aiti-
setum acaiosr ondizcd LDL instead of the neat rabbic anti-
I
sulfat
The
in;. '•
UilT!
of :
free
se
-------
JUN-04-1996 13:13
US EPfVEAD Uash:ngt6n DC
202 2&3 7185 P.06/12
Salonen tt el Merony, FUh. and Cowmary Disease $49
T*ate 2. Ag»* and examlnatfon Yaar<-Adiu«ted Correlation Cceficien» 8«t«M«n Hair and IMnwy M«rcufy and
I fliatt Factors
LOW
Avaflpgpratvifl
ass
urfmnr memtnr,
.120
3St
.9*8
(-,0191
.073
J7»
jae
.094
W7S)
(JM*
nww Harinogen. syt
,073
f-.1Z«
(.027)
.^3S
-.1OS
AM
.38T
•TWZ07.
Tbe curvts diverge moooioaicaUy over (be eanre range
of follow-up time, and dM Hazvd raUa w« approxt»
mately constant over line. In Jeparate Cox propordoeal
h«2»d$ modclfi td}ostin| tor age, examination /ear (fro
indicator variables) and svbefinical CUD (bc&cmk ex*
ercise ECO and maximal oxygen uptake), both th« hair
concent of mercury (P*,B37} and the Miimated intake*
of both S006) associated
staustically significaady with ch« risk of AMI CTabie 3).
On (He everage, u increraeni of I VL% K|/e of bair (6%
incrcmeot) associated *ith an increment of 9% (95%
confidence interval [CI], 1% to t9%) in ((M 5*yur risk of
AMI. An incfetneot of 1 tng/d of dianry mercury intake
was aasocjated wrdl a 3% (95% CJ, 1% to 5%) average
uioenwai in th« 5-year AMI risk (Table 3).
These associations were attenuated by chc addJtionel
staiisa'cal conuei far potentiai coafiDunding factoa (so-
cJoeoonoTmc SIXOM. place of residence (urban -rersua
roral^ aod egarette*yean) and the major coronary risk
boon (family histoty of CHD, sneaa systolic blood
pressure, diabetes, dietary iron intake, and fcnun apo*
Itpoprotein B, HDLi cholesterol, and femtin [> versus
<200 jAg/L] coacenuatioos). and remained significani
only for the intakes of Eish and mercury (Table 4). Tic
hair mercury aMoeiaitd significandy wiin chc risk of
death from CHD. CVD. and any cause, even when we
adjusted for a]) eovartates (Table .
Men in the higheai fertile (>2.0 ^g} of The bair
mercury content bad a 2.0-told (9S% CL 1.2 to 3.1;
risk of AMI compared *\th men in the t*o
Plot of w^'sia) fUnedan estimates 4«-
»«*ing the probability ofnotamperieno-
ttiefofew up (I« days) ferimn vitth ato»
(<2.0 m^9> lair mercury eontant (0] aM
far those **h a Ngh (XLO m^gj hair
mocuiy concent D-
• w
-------
JUN-04-1996 13=10
U5 EPfi/EAD Washington DC
232 260 7185
P.05/12
P."
6*9 GrataOoq Voi 91, jVb 3 FArvary 1, 199!
Distributions of moieatDrs of Boc-y Mercury SWus; notary Intakes of Rah. Mercury, and OU**/- Relevant
Coronary Risk Factor*
MMffHim
VtBlr mvcwy consent.
IMflwy nwrcujy acrimon.
Pferwy **h Mala, g/a
, mm HO
rwww.
SdrumiW.
1.18
r.s
AS
52.*
1«
134JD
SJS
i.oe
i.aa
OM
1«
32.1
2J7
1.M
MO
7.7
S&S
114
1J«
O.M
&a
027
S.C
7.2
OJS
OJ»
OjOO
1.1
D
8T.T
iw
a«
Ml
0.07
25
1S.S7
«.«
9S.3
«18.2
6! .8
UJ
so*
tSA
Statistical Analysis
Associations her**** htir mercury content an4 otewy
intakes of fish and mercury and ride facxoa for CHD were
estimated with Pemnson oanctuioa oectficieMii vdjiusrotf Simoi-
ntneously far age n»d the y««i of At baseFm* «t«mm«Tco
(J98S versiu other, i486 versqs other. 1987 vwvus other. 1988
v«nut ortiw, 19R9 vcmis oik than CHD aad CVD in mepiwl!iy of
%ift time and wift risk factor levek The result!
indieated thar tha afpKeatwn of Ae nodeb *ai aoprafrfart.
In the nc«ed eas»«nnlMi snbnmple vitU urinary mercury
data, cbe anoeiafieit with (he nsk of AMI •m* a/raJyMd by SPSS
loc'itk rajreswofl analysis. Wik faetor-adjutwd ratatjre risfa
wen tstiMMtd M antJogariThm of t eotifieixM. Thek eoaft*
deaca inlartb ww» eoaapiHed bated rm the tscumption of tfct
atymptoUe aormtlitjr of Mtimatat.
The differencas in mew hair mexeury Mmesi according w
(be use of specifie fish specie!, tba place of residence (urban
versus rural), and snoUnf (yes «««ut no) ««r« analysed to
SA5 wo- and rhtec-way AJTCOVA with linev eo»»rnrre
nrreetiont. All tens of significance were t»o*oded
lUsults
The mean daily intake of fisfa wac 4200
tag/l» The mean hair mercury concent did not differ
significantly between subgroups according to CHD fam-
ily hioery, enreisa BOG, or scrum femtin.
The unadjusted survival curves for men in. the highest
fertile (>Z.O ya/g) and in the two lowest teniles (<2.0
Mg/f) of hair mercury content an shown in the Figure.
Assc
13*1
Punr
o( :
mat
haz
InU
er<:
cor
of
sr.i
Or.
Af
WTt
im
-------
JUN-04-1996 13U1
•WwJon, »«/«
^M^fa^ 4^^M^MM.M
«"•••» *|9'r*"*Tr
Strum tarittv eZOO »«
scs.
US EPfVEflD Washington DC
202 268 7185 P.08/12
a ai Mercury, Fish, sod Coronary Dfctase 651
Tula* S. Association of 14-Hour Urinary Mercury Bcerwfcn and Other Coronary
Hiak Factors WWi the Rb* af Acute MyocartKal Infarctfon at a MuftKoriate
Logistic Mod* in 69 Patterns and 138 Central Subjects Matched for Aae, Place of
Residence, and Date o* Examination
Ah*
as*
1.001
2.80
Z.B*
&22
ua
ow
oo .
1,01 to 1&
1.00)19 1JXM
US
4.1S
B.7S
U» to 133
&M * 1M
8,17 la isa
«.sa
&o«
z.«c
1.30
1.01
ox
.ooaa
.9107
.lisa
1
hant Tesidual association between the urinary mereucy
excretion and the risk of AMI (Table 5). For each
nictognuB of mercury excreted dafy, the risk of AMI
increased by 36% (95% CJ. 1% to S2?«} /»=».&«).
In a subsuunple of the subjects (n«!S7) for whom ve
had neasuremenn of scrum immuoa complexes eeo-
Quaiag oxidized LDL, both the hair mercury concent and
die urinary mercury excretion asfrxnled *hh immune
complex tiiers measured waft a new rabbit antiserum
ayaimt oxidized LOL (liter l, Table 6) and the rgioo-
uJin tectfon of a rabbit amfecnra against oxidized LDL
(riter 2. Tabtc 6). Of aB variables tested in mulitariate
models, the hair mercury content *as the smagest
predictor of both Immune cotnpl*x diets. Al««. a h««h
dairy excretion of nicotine metabolites, high serum cop-
per, and tow serum aflnwria concentration associated
significantly with a high immune complex deer when
adjusting for age and examination year, in nsultivariatc
step-up regression models, the half mercury content •was,
however, the only statistically significant predictor of
oxidized LDL immune complex tints fTabw 6). A Wgb
scrum ferritin and a low plasma aseorbate concentration
had nonsignificant asociaa'oni with a b«gh imranae
camplax titer. Scrum LDL cholesterol concentration
had BO association with the immune complex were (age-
and examination year-adjusted r».Q3S and .012 for the
M>O titan), in favor of specificity of the assay to oxidized -
LDL
Men who eansuroed 530 #0 of any fish had 56%
higher mean hair mercury content (/"<.001 for differ-
ence) Inao Tbota who had consumed <30 g/d of fish. The
moft eommonly used fUi species were vendance (Cor*-
genut a/oute, a snail local white fish) (17%). rainbow
troal (12%), and the Northern pike (10%).
To explore the dietary sources of mercury !n mure
detail, the mean hok mercury content WBR computed in
subgroupa of men who reported the use of specific fish
species during the four food-recording day* preceding
hair samplinf, jeparately for men residing in couaoyside
|nd in towns (Table 7). A lineat covariance correction
was appilfld fcir age, soooeeonornic sranu, and OM year
of «aamJnation. The higheti mercury hair contents were
mearand anong men who had ingested either bureat,
vondace, Northern piba, or wbiteSsh. K.unl men who
had consumed Baltic herring had a higher hair mercury
content than rural men who bad nur. There was a similar
but smaller difference fer urban men. Consutnen of buy
fan species such as salmon, herring, domestic ratnbn*
(root, and tuna did not have a higher hair mercury
content than nun who bad not eaten these fish species
TABUS.
of Hair and Urinary Mercury and Other Preaxieanta and Arrnoaidanti WtBi Tthra «t Imrrtuft*
Mvltonrt*
I.1S*
.» v»
3.10
S«av
1.00*
.69<0
rr»
991»
S*wm ttbg
.'.55
.TT!
.1*1
.17*
-.10
.101
.810
013
.008
.011
.081
am
.120
.lie
.CM
.•m
an
.in
.its
•MS
.one
.ua
.oas
.109
.Off
.017
.170
9 irrt
. ana afo
, n-iS*.
-------
JUN-34-1996 13=11 US EPfVEAD Uashin3ton DC
650 Circulation VW 91, ,Vo 3 Fibtvcuy I, 1995
202 260 7185 P.07/12
Age. and Coronary Oia»>B»»-A4)o3ted R«Mrye Risks' of Aeura h*r»e*rdial Infarction and Deatft From
Coronary Heart Disease, Cardtovaccufar Oferas*. or Any Caus« Asxoetowd VWln Hair Mercury Content as Wall M
Dietary Fisn and Mercury Intake*
*»«•!••
6*0
U Cause
ISVtCt
9614 «
tw
RR
K% a
H^trnwrewy
.P-.OOS
VOOS
2.0*
1.01 to
1.19
153 »
3.13
1.001»
1.0M
VH*
1.0Z8
P-JM
1.2M
ZJS
WO*
ZJO
Pw-100
1.013
1.«
0.98 to
SJB
oaerio
MJ10
1J06
^-J3Q2
2.95
LOW
LBTto
US
k$9
in
1.BOB
031 W
1.132
P^OO'
urn
1.0$ XI
122
1.H3TO
1400 «
voot
ojfaw
1.028
P-.1ZO
NO. «f I
n
2*
78
AMI kictaM»»eU» nr/MB^ll HvCONVCHO. (
'nfoni fHuttranMii Cw prvpwtfwfl rattf^ii'
tMduOngi
|o««n tertSes when adfwting for ago. examination year.
tsehemie exercise ECG, and the maximal oxygen uptake
(Table 3). This relative risk was 1.7 (95% CI. I.rtS io 18:
r-.OSS) in a model adjuning (or all eonraderod ecn-
founden and risk faciors (Titbit 4). The relative risk «u
similar for coronary death* but not statistically signifi-
cant due to the smaller number of events. However, a
high hair mercury content wn* significantly associaibd
with the risk of dead) from CVD (relative risk (RRJ. 2fl:
93% d. 1.2 to 6.6: P-.OH) and from *ny eauae (RR,
X3; 95% a. 1.4 to 3.6; y»<.OUl).
An average daily fish intake nf fc?0 g was associated
with 11-fold »^e-. eicammauoA year-*, aid CHt>-ad-
jujted risk of AMI (95% CL U to 3.4; P».0041 cnm-
pared *riih men cotwiming Ie« fish. For each additional
10 o/d jbh intake, there t*as an increment of 5* (95%
0.2« to 8%) in the S-year risk of AMI (Table 3).
In additional Cox models (not presented in the ta-
bles), we entered a numher of other variables that had
been associated -wrth the risk of CHD in previous
studies. Thexe included dietary intakes tit* energy, satu-
rated fatty acids, dietary poiyun5ahirarad fatty acids, and
carotene; The consumption of alcohol and coffee; l
lime physical activity; body nao index (kg'ir): Dlo°d
leukocyxe count: plasma ascorbaw and a-
T*M« 4. Biak Faetor-Adjusted R«lrt^ duila* of *cutt MyoeardM mfwetfan and Peatn from Corenwy H»an
Dtsns*, CardiovMeurar OiMac«. «r Any Cause Associated WWi Hair Mereuiy Content •* Well «• Dietary Rah and
Mercury lntak««
CMO
AHCaus*
im
onto
HtKrwrwury, |
1.0CB
l.6»
0V,
II
»»-
1*p
i.en
riTS
1.03 W
Z.78
1.00110
1JX1 »
1.404
1.1ST
M4BD
^-iT«
una
4.7«
1AIO
an to
5JT
i.ejo
P-.OOT
1.003 TO
1.T«
1.004
1.13W
3.0*
1.0002 M
ur
1MO
1JB7
J*-.iC«
ill
0.9MIO
veer
1.W
/•*.««
1.BD8
P-.S33
1.033
NO. •frnen
wftrttvtm
T3
1*
1*
•P^/B mwlnvitf* CM pnpenlenKI hguida rr««ais ^du*n91«, MarmMilan ywpb»*
fwmty >«steiy ot CJHO, clgme»>r«in. rrww syttoiie Wo«l pr««*«, dAim. Jwita«iw«iin
ionhB. and svum aeor^onvM
n*18».
riM EDO, mMlmai wygco upeefc*,
»or KMIUIU* (urban vs. fur*), giiary "*i
-------
JUTH34-1996 13=12
652 Grculatfma Vol 91 ffo 3
US EPfVEfiD Washington DC
I. 1993
202 260 7185
P.09/12'
II
TABU ?. Ag8- and Soooaconomic Stafes-Adjuated Mean' Hair Mercury Content in Men Who Reported Us« of
Various Piah Spodea During 4-Day Food ReeortBng and HI Thos» Who Wo1 Not, Accardtag to Place
Hair Mofcny O»w«t,
Usws
NOTMMI*
Wore.
urn*
Urban
Bvxtoi
Van<
SWRMi
2J
10.3
IS
u
8.S
34
3.93
3.1«
lie
4^3
tia
2.0«
Me
IT»
1JS
ISO
2.40
a.1*
2.01
t»
3.11
141
2J2
2.09
£35
ZiZ
2.0S
3.11
1.87
1JH
1J9
1JO
tJl
tJO
uz
'•91
<.001
JHO
Ma
.7T7
JW
• .an
jni
.591
.013
.COS
oot
J507
,eso
,107
.011
477
.089
JOS
-3K
.015
J81
UJM
J1J
431
.14*
jaa
•««*aiM far to* w, •nrrmtian ywr (fet eumnw va
Vdifta
i MO nomiMn m anoiMrs v«p te
daring the four food-recording days. Ail of these forty
fiib species (yjuxpt some ot the niinbow noui) are
caught from the set rather than the local lakes. In a
separan set of thr««way oovarianea analyses, smoking
Cra versus no) was entered, in addition to each fob
species and the place of residence. The associations of
th« us* of burbot (P--002), vendace (f*<0»I>. Northern
pike (Js<-001), and whinfeh r/».Q2S) remained ftath.
tieaUy stgnifieant Then were 90 significant mtcnutfens
«iih ±e use of any fish spceiei wiih imeking.
Dbcossfoo
Tbe human body (70 kg> contains on the average 13
n| Hfe.* No meubotie fanctfotw in the hrnnan body are
known for which mercury is required. At high concen-
trations, mercury is kncwn to cause IKur and kidney
damage at well as neurological symptom*." Abes inter.
esl has grown tn the possible beaitk iffeeu of mercury
liberated frwn dental amalgam nllings, which nay be the
principal source of exposure to mercury for a targe
segment of the US population.* Clarkson" has pro*
posed that given the increasing consumption of. fish in
the United States and iu important nutmtonal role, it is
vita) That the potential for prenatal damage from m»thy»
roercury be assessed. There is. however, praea^ally no
previous research concerning the pocsiale harmful ef-
fects of a usual intake of mercury with regard to the
cardiovascular system.
There *n at least three otecnarusms through which
mercary cao promote Qpid peraddaliaa. Firtt merenzy
>i a transition metal, and it can thus act as a cataljsi ia
Peotomtvpe reactions, reanlttng in the ionnaiioc of free
ruficsli. The noo'on that mereurypromoies free radical
genenaoa was first presented by Gaarher" based on the
obcervetioB that vitwnin E and the amioaidani DPPD
provided prolection against nrcehy) mcrcnry pouonin j in
rats.2* In a recent in vitro stecty. Hg(Jl) ions in rnicro-
raoltr conccatrations increased the production of super-
oxide anions in human neutrophils.** Tn another in «itro
study, mercuric ions (1 to 6 iueol/1.) caused a eoncen-
tration-dapendeDt increase (op ta fivefold) in mitochoo-
drial HjO, produeUon.4* In addition to its direct catalyoe
eSffL, mercury hu been fimaU to enhance iron-stimu-
lated lipid peraxidatton i
An in vr*o study revealed a significant concearntioa-
reiated depoltrauion of the inner mitochondrlal mem-
brane. increased HiO) rormattou, glutathione depiction,
and formation of thiooarbiturie acid reacave subxances
after the addition o< Hg(H) lo mitochondria isolated
Cram kidneys of uncreated rats.41 Thus, although cata-
lytic Fe(H) princrpany cataryzM the oxidation of HA to
the more reactive hydroxxl radieal.42 mereary appears 10
aa eatiier a rne Raton reaction chain, catalyzing ihe
produerioa oi H.O-.
Seoaad, mercury has a *ary high affinity to sulrhydryt
jfOttft,1* which In plasma proteins have been estimated
lo accewjt for as mueb as 10% to 50% of UM antjoxida-
tive capacity of pJasma.« By binding to sulfliydryl
group*, marcury inaetiviites anrJoxidative thiolic com-
pounds such a the gtanchicne.*1 Chitubioae has a
central tole in tba regeneration of the tocapheraxyl
radical m tocopheroL Mercury poiwwfng. which is asso-
ciated with inoeased lipid peroiddation in the over and
in tha kidneys, al
-------
JUN-04-1996 13:12
US EPfl/EflD Washington DC
202 260'7185 P.10,/12
the risk of AMI through the promotion of lipid psroxi-
dation receives empirical support from our finding that
both high hair mercury conceal and high urinary mer-
cury excretion weje associated with eJcvaied Tilers of
immune complexes coataiuiag oxidized LDL in a sub-
sample of our study subjects. Also, as we have reported
previously,14 uuum selenium concentration associated
inversely with autoanubodica against oxidized LDL
An additional possible arherogenic effsct uf mercury
would be the stimulatory effect on proliferation of
arterial smooth Bjusde ceils in » ceQ culture study,39
Mtrouy compounds hss*e also potentiated ADP-ift-
duced platelet aggregation." Seems* oxidized Hpids
promote arterial smooth muscle cell proliferation^ and
platelet activity*3 both of these achcrogeoic effects of
mercury could be consequences of enhanced free radical
stress and lipid peroaddaiion.
Mercury is a poisonous metallic clement that is re-
leased by manufacturing and burning fueto and minerals
as well as by industrial and Household wastes.* Eventu-
ally, it settles in waterways, where it joins naturally
occurring mercury. There, bacteria and algae convert it
to the tone mecbyi mercury.** According to die Harriis-
Hohenemjwr cornparaiiv* hazard index, mercury is by
far the most dangerous environmental poison of all
heavy metals.5*
The mercury content in Finnish lakes k high.5* Ia the
late 1970s, it was observed to be wpsciaSy high ia
ground WMHS ia Eastern Finland. There are several
both geological and man-made reasons for this*: The
Finnish lakes are shallow «nd bave brg« catchment
ateas. Both the soil and iha take *nters are relatively
acidic, which enhances the availability of soil mercury
and increases the level of irnthyl mercury in the edible
tissues of fish.*1 The ncklity of take waters is most
pronouncad in Eastern Finland.*7 Water levels m most
Finnish water systems vary widely between seasons, and
buouc material and mercury Cram the soil surface are ,
washed to the lake waters. Drying of swamps by ditching;
has lead to drainage of the high-mercury swamp water
into lakes. Finally, some Finnish watercourses may still
bear the consequences of discharges from chloralksli
plants that used mercury electrodes and the use of
Salonen tt al Mercurr* fish, «nd Coronary Dlsase 653
- Al least four prospective population studio have
reported an inverse association between fish, intake and
coronary mortality.1-* In the study by Kromhout and
covorkers.1 mortality from CHD was >509V less uaoag
men who consumed at least 30 g/d of fish than among
those who did aoc &at fish, &s asscnsd in 1960. These
findings provided persuasive evidence in tavor of a CHD
mortality-reducing effect of ev«i moderate consump-
tion of fatry sea water fica.
We retesiBd the Duteb finding m our study by using
the same cut point for daily fish intake. In our study
cohort, however, die dairy intake of fish of %50 g w«s
associated with a 24-fold risk of AMI aad a 2.4.feld risk
of coronary monaJiry compared wtu> men who con-
sumed <30 fif'd or fish. The mean daily 5sh intake m our
subjects was 47 g. more than double that in Zutphen
nwo-20 g ia I960. Thu majority o^ fish consumed in
Extern finland hi the 1950s was lean local freshwater
fish. The intake of these fish species was associated with
elevated hair mercury levels, whereas none of the fatty
fishes consumed (rainbow trout, salmon, herring, tuna)
ware associated with increased hair mercury. Thus, in
addition U increased n-3 polyunqacurated fairy acids.
these fatty ftshea have a lower mercury content.
The role of mercury as a risk {actor for CUD can
explain a number of previous findings for which a
jriausibto biological explanation has been lacking. First,
the association of nigh fish intake with decreased CHD
risk was not observed in two prospective studies in
populations consuming high amount* of flesh, u> Hawaii1
and Norway * end in a large prospective US study, ch«
Physicians' Health Study.7 Apart frntn measurement
imprecision and threshold effect a* possible explana-
tions, the lack of association could be due to differences
ia she mean mercury intake. Alternatively, mercury from
if*.
fish may affect Itpul pcrcoodation and CHD risk only *t
intakes (hat exceed an unknown threshold.
Second, ifl a cane-control study** and in a prospective
population study.*" an association was observed between
dental disease and the risk of CHD. In the Finnish
case-control study, die number of dental fillmip
associated with the risk of AMI.** The authors
lated that the increanrd risk would be due to bacteria
thai cause periodontius.** No mercury measurements
- nf the two stttdJe*, and the
-------
JUN-ea-1996 13U3 US EPfVEPD Uashington DC
654 Circulation Vol 91. No 3 Fdmary 1, 199S
202 260 7185 P.11/12
reason. the mean hair mercury content in our subjects
(1.92 nf/%) cannot be directly compared wjia value*
reported earlier. la previous studios in selected groups
of subjects, the mean hair nwrcury contents have varied
between 0-50 and 3.4 pgfg in Europe and between 3.5
and 4.3 u#z in Alaska.'*
The likelihood of bias or confounding as an explana-
tion /or our findings is vary small First, the dietary
iatato of mercury and toe hair mercury content did not
covary with any other of the mote than 60 nutrients
measured, except selenium. Other than the interactions
^^jileahun, thare a aoloiawii metabolic regulaiieo of
either the absorption, accumulation, or enaction of
mercury. The hair content of mercury is a direct measure
of the accumulation of methyl mercury in The l»dy over
a period of several month*.*1 Urinary mercury oxeretioa
at considered an indicator of the amount of elemental
aereury in the body, dewmd from both dental amalgams
and the djet.*1 Second, in extensive exploratory analyses,
the only poMniul confounding factor* that we found
*«« a low socioeconomk status and living » rani
areas. The socioeconcmie status was assessed compre-
hensively covering six different aspects.* Both of these
factors were .associated with bigh hair mercury content
and an increased risk of AMI and death. However, the
association between dietary intake, hair content, and
urinary excretion of mercury with the risk of AMI and
death persisted after a statistical control for these factors
and in urban and rural subjects separately. Abo, a low
socioeconoraic status and rural living should fee consid-
ered in this context wore as determinants of high fish
intake than of confounding factors. The weak positive
association between the amount of cigarette smoking
and the hair mercury content was most likely due 10 the
intake of mercury from cigarette smoke, and the juris-
tical control for smoking fa oar analysis may represent
Qvcradjuiuncnt.
Theoretically, oar findings could be specific only lor
men in Eastern Finland who traditionally have a high
intake of meat, fish, and saturated animal fat and a low
intake of selenium and vitamin O'°-*'-c* and, most nicely,
other vegetable-derived anticnrMants. Nevertheless, our
fadings in men in Eastern Finland provide new infer-
matton about the etiology of CHO. Pathological mech-
anisms involved most appjy to humans in general, even
though, the consequences of the high mercury intake for
the cardiovascular system may vary among populations
due to various effect modifying factors.
The results of the present study are significant m
several ways. First, although contumpdon of Jub may be
healthy in general, some fish may contain agents that are
not healthy for the human cardiovascular system. More
important our findings suggest thai mercury, e*ea in
subtojoe amounts, is a risk jacnr for coronary and fatal
CVD. Our observation in a subset of sabjects provides
evidence suggesting that high mercury intake would
increase the risk of CHD by promoting tipid pereodda-
finn. Because in many countries fish and seafood are
contaminated with mercury and becanee fish is an im-
portant fbodstaff aD over the world, oar findings could
be of enormous public health importance. For that
reason, they need to be retested urgently in oiber
population studio* tn which hair, nail, blood cell, or
urine samples have been collected and stored for min-
eral measurements. Q ouz observations are confirmed in
subsequent studies, the measurement and labeling of
mercury content of various foods, the prevention of
additional pollution of the environment by mercury, and,
eventually, the elimination of environmental mercury
amid be can*id«red new measures to advance the
precaution of cardiovascular diseases.
Acknowledgments
This work «M fupperiMi fey grant* from IJM Academy George A. Kaplan, PhD. and Oai V. LmJqvis^
PhD, fat wiewing lie rrnnuscrioi and to Raincr Rauramaa.
MO. PhD, tor the egrticip«ien of the Kuc>pio Ressarcfc
InstiiuM of 5xtt«iM MtdiciM in data snUeenon. w« jiw {h«nk
M«rja IhiBiieen. MSc. for read mcortfacc Pr Jaakka Efiaea
for eoeleg Ow cRretca ECCsr Dr Eska Taskincn. &r Juha
Venilenwa, and Dr Hanrw Limaaan (or theif purtxapatien m
the (upcrriciea of the maximal cieicae restr, and Kimmo
Hoekaiaeii. MSc, tor tarrying out the data arouses.
Rttfcrances
1. KrMriWHt D, BoMCtMicr EH. Oft LMMUK Co«ha4er C'TKc
ftMnc rotMioit bMwecii nib cnwmip»n« »md 20-y««r motfaUt
bva Wtoroiy ban ditto*, H Ki^il Nt*. JWSd 12:1305- 1209.
i Sbtfeiik Rft, MteU LU, Paul 0. Srayoek AM. Snnlur J. Fka
eeeiMB|mM wt mertnlity frwn careaaiy tican «»ow. * CaglJ
$«, Alioma A, feyehiinj M,
SU«TWOII and mernfirjr Aem coronary heat
Nt_
Mtt f.
«. DoWw* TA. CnaAi G. DUtarji pniyunmvnteQ falty art* Jtvj
ngraK* in ike Xoftipk Bik Aoer Inni««mi«n Trig! (MRHT],
lot SiBOfoalm AT. Kifcr KM. Mwtia RE. Bario* SIX «i». tfatlOt
of On&S A^MMwn
Nun Did. BMI, MBcrtMA S few*
5. CUf* JD. Rtot OM. TtA ooBRUKpxion »od monaiiw (row
(. Venm S2. H«dl fc BJelhc
* from *artfMMk* di«M.
7. MMrM MC HMPM iS. fa*a 9, Bvriog JB. Vfllea we. H«n-
Bdwni CH. A pfmnqii* a*4r of hk cnrKHmnOoM ON ardfa.
««mbr rtiin^. O.y«ii>un. 19KEM(5Wppl H J-*^3. Abunec.
IfciMieea K. SUonm R. S«fm»mm B. Stlonen Jf. NM»HHH« d»u
»>o bdnone Hem Dl*ea*e Risk
ef akkfle-aea) E»iatn Kaonh
R WHO MMBM Preim. WHO Moniek mije« ««™rtntCHD mar-
uiqr «id mermfr. h»/fA*'«W, l»*:i«s.VM«J.
ia Sdanen JT. AiftDai C. Hn»»e« JK. KManliwn J, Pucks p.
>, Uxcc,
lerem wJenww • e •ached pair
WW79.
U. SMtfoMi r. Hcta HO. OyMettMrg r hnim nimfcm eoua*.
Mb'm and rUh of ae1iwi«e Imn *'»•»»» in i r>rospeeij«e
KodPJ. Hgfnuii A. WincniM JC D> Brazil AM. KMy«wn OH, de
" ' i M. Vilhcnhurg HA. Dctn««n) Mte*«ni teveta in »eute
f. 19893W:) 161.1 16X
13. Sdene* JT. Salonen R, Seppine* K, KMII>)» K SunilaincA &
JCacpdt-K.Imef*cl)Ons erfierem mpp«r. ttteoium, mrf low de*a>
choteturol to ydKrosaedi. *• MM /.
SiMea It. YIi-H«niuaia S, Yamanwco R. « it
and pn>gr«»i»»
IS. SriteenJT, Nmtdnu K. Karpela H. Twwttchw I. ScveaMnR.
SotooM R, Ktghnered tan hveltaraanneiMed «Wi eteea wk at
injmeiiltd iatuaiOH im Entera FbioMi IOM. O»«Jii/wi, J99ZS61
J8M1J.
It
304 ce, (Mbrtf, UK Chnmotot Pmc 19W.
17. Syaauiiiie F» fc Meak m Up* parejd^Uafc^oi f**>*nnt
Tbtfcoi. tfM0i(idae« TrJufrtfS.
IS. Onlbe KB. Go*i« C Sunde ML. XngeeVy MI. Wiper K
jelenhuK lehnon m deertmd laabty of meOiyfatareuiy *dd«l to
dies oauinins runt. Jowtsf. (•U'l^lUZ-lIS^
19. SdftBen JT. b tJicrj * emwiuiaK need for tongnrtwiit
oiogk ntatreh? THe Xuepio bchuBit. Heart Daeati tUk Fiewr
Rose GA, BIwib«m H, Gafmn HP. Wocaj W.
A
-------
•n»
ah>
W.
Fa*
•*/
ran*
.rf/.
in4
•m
aMl
n
I.J«
AKU
«&
xdf
:r -
idw
JUN-94-1996 13:14
Zl. Wmrani PC. Tauber F, £ste»*»jcr H. Km ML Stake DK, Morm
CJ. pRjencr >i (Mm edb eonni/ucj e4dba4 low enmity
Spepratm i» te «jiww< nembimc torn patiMU «iik rim
ouuid intofBL-4*» (Uttum Oil. IWS'-SJ'.tn.tta.
a. fafcaen XT. Safcftea JR. SeppSnai X. Rxuimii R. Itounfehia J.
HDL, HULj, ana HDL, MjMr«a»on«, |nd Urn riik of Mm mjra.
ifttocttov I tracpeeture population JBidjr « E*w«n»
US EPQ/EPD Ulash i ngton DC
232 268 7185
litfa TA. &J*WB JT. Mwfcr»l« to high inlnuity eooUktMMIf
Kriiwt nm* ohyjfcal aetWty *U hSjH etrdSompireUiy Kwta v»
MMfctrt •* Kducc* pfaM ikttaacM to E«Kr» FlmtftJi ma
; Cfe iM/uiifei, iTO«fciii*ii27.
Satan JT. UMM It JCoepeia & SUOIMM S, TuoraiUiB J.
md thi rafc o( »ntt* vymmlU mhtrtiaai a pp.
t«Mfl* 4M*)« (•/ KM UMMMIR
21 Jtapbs QA. S«I«M* JT. SoenMaaomiC eondWem ia
: J 2».
31.
31
SUK» tt te mJt SO. ThB PIMM** «fc« of vinnt"
Mri «rid«n» *in» in -i^ «iul to ««• Ot
Solemn a si Mtmwy, Fish. aM Coronary oRiiR E. orate, a*eorbatr ud
SB)MM« JT. T1.« nal«
..
. Our
44. Mpeona A. XdJPMi* Y. Inmn N. Bchxrror of (naUrylmerairy is
«i. B«H* LC BcariMv 1C. Memmdt OH. TWai uuitfain cffret an
llptrf 5«ionteW(i en eeRuqr^aitau4 rats. CAem S
47. Sitoncn JT.Srtwwi USeppimft K, « IL JUUonmlrtp of iwnm
uhMtma ari nuinrifenti n ptarai lipoprotemtj pbttlii atang*-
Wwmto hewt dieiM IB Canim Knokk tnvn.
aM *• iminM
JTS-UL
*». IbUM SV. Boon PR.
*. fto qfnt JApS*
offend
SO. Ui KT. ffl« SH. *M« 01 Tin MbmbionP rite nn««y nwai
«T AJpp.|»4ge«ri platdn
Ji
. P^nhanmbjr 1 C»n« TE. ICbM 1C. Witznia JU
dvriccrenil: madHkatieMi of to»wtejniQr (ipopncem UM<
Silwm JT. SilaM R. Scpyja«a K tt »l
M ptMilx At«ai«a:
. duoMc^Knt) wial In men *Kh tow oninridunt
54, Hsrim RC H»lmi»»«T C M«wwr-"«i*a'f»j ami
- UM rilK. JPwinwrNMl. I9W9-.25-I6.
& Wilttmwi SC Conkfae KH. Rnkiiun. ?lt Merany iioaimlatwa
19W-.U-JMI464-
HKdnU. Ftah
J*«
UI /lift.
ItM Wattr aM bvwwMNM Aaeatdi (wtUwit. Naikwiai Buna u/
WUin wK UM SirrininnMic FMuwt. NO. A.
57. KMKM1. faaiM M. (bwMbiMii P. Maw** J. v«m M.
Uas •ot>; MiaMiK Matun ol wJrflfantai. /i«b««e«i
4aiMt fcc.li* •«< KU«I mynamal iotoai
T»»-T«U
W. DeStcfJOO F. Adda RT.
HS.-WiWa«WM Of, gvn*H CM.
«r
MM A
f I. Sakmn JT. Scbniwa M Hdaemie BeJifl HaeMO. (nr / £p
------- |