OCR error (C:\Conversion\JobRoot\00000CEL\tiff\20013NEV.tif): Unspecified error

                         ARSENIC AND FISH CONSUMPTION

                                   Table of Contents


       Abstract                                                                 1

 1.0    Introduction                                                              1

2.0    Arsenic in Fish                                                            2

3.0    Arsenic in Shellfish                                                       5

'410 -'   Speicatioh of Arsenic in Fish and Shellfish                                   6

5.0    Estimated Arsenic in Fish and Shellfish from Water Containing 50 ppb Arsenic   8

       5.1     EPA Bioconcentrations Factors for Arsenic                            8
       5.2     Estimated total Arsenic                                              9
       5.2     Estimated Inorganic Arsenic                                         9

6.0    Estimated Inorganic Arsenic Exposure from Fish/Shellfish Consumption         10

7.0    Inorganic Arsenic Exposure from Fish, Shellfish and Potable Water

8.0    Uncertainty                                                               15

9.0    References                                                               16

                        ARSENIC AND FISH CONSUMPTION

                          Health and Ecological Effects Division
                            Office of Science and Technology
                                     Office of Water

This report summarizes available data on human intake of inorganic arsenic by consumption of
fish and shellfish and arsenic in drinking water. It estimates total exposure to inorganic arsenic
from these vehicles under a variety of exposure scenarios.

Much of the arsenic in fish and shellfish is present in the form of organic compounds rather than
as inorganic arsenic. For this report, EPA utilized published data on the concentrations of total
arsenic and inorganic arsenic in a variety offish and shellfish species.  These data along with
data from the U.S. Department of Agriculture Food Consumption Survey on fish/shellfish intake
of consumers and non consumers were used to generate estimates for the inorganic arsenic intake
for several exposure scenarios. The scenario for the group with the highest potential exposure
(90 yug/day) was individuals consuming a diet high in fish and shellfish and having a preference
for shellfish. An scenario for the average fish consumer estimated an intake of 4 /zg/day and the
scenario for the general consumer with only occasional fish/shellfish intake estimated an intake
of 0.6 Aig/inorganic arsenic per day. Scenarios for consumer groups with other fish/shellfish
consumption patterns were evaluated as well.

The evaluation of inorganic arsenic exposure from fish and shellfish provides support for
utilizing the existing MCL of 50 ppb  for arsenic as an ambient uater criterion in some areas until
EPA updates its risk assessment for arsenic and revises the MCL. The exposure evaluation also
illustrates a need for site specific criterion when high consumption offish and shellfish is
coupled with arsenic contamination of drinking water. Toxicity concern related to the arsenic in
marine fish and shellfish is mitigated by the fact that it is largely present as arsenobetaine, a
metabolically stable compound that is rapidly excreted. Additional studies of the forms of
organic arsenic in fresh water species are needed.


The U.S. Environmental Protection Agency's (EPA) guidance for arsenic under the Safe
Drinking Water Act and the Clean Water Act are different. EPA's drinking water standard, or
maximum contaminant level (MCL) for arsenic is 50 ppb and was developed by the Public
Health Service in the 1940's. The Ambient Water Quality Criterion under the Clean  Water Act
is 0.018 ppb based on an estimated one in a million  cancer risk (EPA,  1980). EPA has
                                   Decembers, 1997

 recognized that there is considerable uncertainty in the cancer risk value and is presently in the
 process of developing a new risk analysis in order to propose a new MCL. Under legislati\ c
 requirements, the EPA will issue a proposal for an arsenic MCL in the year 2000. Until that time
 the 50 ppb will remain in effect for public potable water sources.

 The question has been raised as to whether the 50 ppb MCL for arsenic in drinking water can
 also serve as an Ambient Water Quality Criterion for arsenic until the risk assessment revision is
 complete and whether fish/shellfish consumption from the same waters adds significantly to the
 inorganic arsenic exposure. According to the Clean Water Act Criteria Document for arsenic
 (EPA, 1980), bivalent and pentavalent, inorganic, arsenic compounds are the most toxic species.
This document accepts that premise and updates the Criteria Document in terms of the forms of
arsenic in fish and shellfish.

The following report examines the available quantitative data on arsenic in fish and shellfish as
well as its speciation (inorganic vs. organic).  Estimates are made for human exposures to
inorganic arsenic from fish/shellfish and drinking water under several exposure scenarios that
apply to the average and high end of the distribution curve for fish/shellfish consumption.  An
exposure assessment for the average consumer within the general population (which includes
nonconsumers) is  also presented. The document is a technical summary of the available data on
the arsenic in fish  and shellfish as well as an exposure evaluation for inorganic arsenic.


The quantitative data on arsenic concentrations and speciation in fish are limited but are
generally consistent with the hypothesis that most and sometimes all of the arsenic in fish is
organic rather than inorganic. The available analytical data on arsenic in fish are presented
below. The data are grouped by the source of the fish and the type of arsenic. All data are
 reported in terms of \\et mass. The following presents data on total arsenic in marine species and
 then freshwater species followed by inorganic arsenic for marine species and then freshwater

 Total Arsenic in Marine and Estuarine Species. Ballin et al. (1994) analyzed samples of 13
 species of marine  fish for total arsenic. Average concentrations ranged from 0.6 to 37 ppm.
 Only  two of 20 samples had values greater than 10 ppm, one of three plaice samples and a catfish
 sample. In cases where samples for a given species  originated from different source waters.
 there  was considerable variability in total arsenic concentration. Total arsenic concentrations
 from  three different herring samples ranged from 0.7 to 4 ppm; there was also  variability among
 the results for 5 samples from the same fish. The two species with the highest average levels of
 total arsenic were plaice from Fladenground (32 ppm; standard deviation 14) and catfish from
 Gr. Fisherbank (37 ppm; standard deviation 28).
                                    December 3, 1997

 In another study, Lawrence et al. (1986) obtained samples offish muscle from different areas in
 Canada. Both Atlantic and Pacific fish species were evaluated.  The average total arsenic
 concentrations for replicate samples ranged from 1.1 ppm (herring) to 13.2 ppm (sole) for
 Atlantic species and 0.31 (salmon) to 7.4 ppm (cod) for Pacific Species. A total of 6 Atlantic
 species and 5 Pacific species were evaluated. Slight differences were apparent in the Atlantic
 versus Pacific samples of the same species. For example the sole sample from the Atlantic had
 13.2 ppm while that from the Pacific had 5.2 ppm and the cod sample from the Atlantic had 5.2
ppm while that froni the Pacific had 7.4 ppm. Based on the sample variability observed by
Ballin et al. (1994), these differences are most likely a reflection of variability in samples rather
than differences that result because of the arsenic in the source water. Additional support for this
conclusion is provided by analysis of a sample of sole purchased locally in Ottawa which had
only 0.10 ppm total arsenic, a value far lower than that for either the Atlantic or Pacific samples.
The low value for this sample may represent loss during storage and shipping. Le et al. (1994)
found that up to 48% of the total arsenic could be released in defrost liquid.

In order to evaluate microwave assisted distillation with atomic absorption spectrometry as a
method for determining inorganic arsenic, Lopes et al. (1994) analyzed commercially purchased
canned or frozen samples of anchovies, tuna, sardines, hake and sole for total as well as inorganic
arsenic. Total arsenic concentrations ranged from 0.82 ppm (tuna) to  7.76 ppm (sole)

In a report developed for Region 10 of the U.S. EPA, Chew (1996) summarized data from the
published literature on the concentrations of total arsenic and inorganic arsenic in fish and
shellfish. The data apply primarily to samples from Japan and all but one sample came from the
Pacific Ocean. Total average arsenic concentrations ranged from values less than 1 to  10 ppm
for  fish with two outliers: skate (64 ppm) and stingray (17 ppm).

Total Arsenic in Fresh Water Species.  Ballin et al. (1994) examined the total arsenic in fresh
\\ater fish from rivers in Northern Germany, the River Elbe or from a fish hatcher.'.  Seven
species were examined; the average total arsenic concentrations were  lower than those for marine
fish and ranged from the detection limit to 1.5 ppm.  The highest concentration was present in
rainbow trout specimens from a fish hatchery and the lowest concentration was found in perch
from a river in Northern Germany. The trout sample was the only one with a concentration of
greater than 0.1  ppm.

The lower levels of total arsenic in fresh water fish are substantiated by  analysis of the muscle
tissue of several  species collected in Ontario and Alberta, Canada (Lawrence et al.. 1986).
Concentrations ranged from 0.007 ppm (yellow perch) to 0.24 ppm (striped perch). Nine
replicate samples were analyzed. Among the 23 samples of fresh water fish analyzed by Ballin
et al (1994) and Lawrence et al. (1986), the total arsenic was less than 0.3 ppm for 22.
                                    December3, 1997

 Inorganic Arsenic in Marine and Estuarine Species. In the data summarized by Chew (1996).
 average inorganic arsenic concentrations ranged from the detection limit to 0.2 ppm. The
 species with the highest average percentage of arsenic as inorganic arsenic were shark (9.5%),
 sturgeon (6.9%) and sucker (8.5%). In all other cases the percent of inorganic arsenic was less
 than 4 %. The two species with the highest average concentrations of total arsenic (skate,
 stingray) had none of their arsenic present as inorganic arsenic and the species with the highest
 concentrations of inorganic arsenic (shark, sturgeon, sucker) had low average total arsenic
 concentrations (2.1,0.6 and 0.2 ppm respectively).  In the study by Lopez et al. (1994), the
 percent inorganic arsenic in the 5 fish samples analyzed was less than 5% in all samples when
 analyzed by microwave-assisted distillation and atomic absorption spectrometry.

 Ballin et al. (1994) did not analyze the fish for either inorganic arsenic or total organic arsenic.
 They did analyze the tissues for arsenobetaine and phospholipid arsenic, the major organic forms
of arsenic in fish.  In marine fish, the arsenobetaine accounted for 96% to 100% of the total
 arsenic and the phospholipid arsenic for 0.17 to 4.12 % of the total. Assuming, no other organic
 arsenic forms were present in the fish examined, the maximum amount of inorganic arsenic
present was 2%.

Lawrence et al. (1986) analyzed 11 replicate samples offish muscle for arsenobetaine and
arsenocholine using purification by high performance liquid chromatography (HPLC) and
identification with atomic absorption spectrometry and fast atom bombardment mass
spectrometry. In addition to arsenobetaine and arsenocholine, the method was able to identify
two unknown organic compounds containing arsenic.  In all fish samples except salmon, the only
 compound identified was arsenobetaine which accounted for 78 to 88% of the total arsenic. In
 salmon an unidentified form of organic arsenic was present in addition to arsenobetaine. The
 arsenobetaine was 41% of the total arsenic and the unknown compound was 42 % of the total

 Caution must be used in evaluating the Lawrence et al. (1986) data because the percent recovery
 from samples spiked with arsenobetaine was 80 to 84% quite similar to the percent of arsenic
 reported to be present as arsenobetaine in most samples. Thus, this method could under report
 the amount of arsenic present as organic arsenic because of recovery problems.  The authors
 point out that when the results are corrected for recovery they indicate that arsenobetaine
 accounted for "essentially all  of the arsenic present in the marine samples analyzed". It must also
 be remembered that Lawrence et al. (1986) did not analyze the samples for phospholipid
 containing arsenic which Ballin et al. (1994) demonstrated to be present in marine fish tissues.

 Inorganic Arsenic in Fresh Water Species. In analysis of 9 replicate samples of fresh water
 fish, Lawrence et al. (1986) did not identify either arsenobetaine  or arsenocholine. An unknown
 organoarsenic compound  was present in all samples. This compound accounted for 71 to 85% of
 the total arsenic. Caution must be used in evaluating these data since it was not possible to
                                    Decembers, 1997

quantify the recovery for the unidentified compound. The same unknown compound was present
in all fresh fish samples whether they were from Manitoba or Alberta.  Elution of the
unidentified compound from the HPLC column indicated that it was more hydrophilic than


Total Arsenic.  Total arsenic concentrations in shellfish tend to be higher than those for finfish.
In the data summarized by Chew (1996), average total arsenic concentrations for shellfish
ranged from 0.2 to 126 ppm. The highest concentrations were seen in two mollusk samples.
However,  mollusk values were highly variable among the 20 samples tested (range: 1-126 ppm;
Chew, 1996). In the study by Ballin et al. (1994), the total arsenic in 4 species of shellfish
ranged from 2.6 to 21 ppm; the highest concentration was found in lobster. The average value
from pooled samples of blue mussels ( 40 samples) was 2.6 ppm (Ballin et al., 1994). Lawrence
et al. (1986) found the total arsenic in replicate lobster, scallop and shrimp samples to be 5.2
ppm, 0.68  ppm and 20.8 ppm respectively. The concentrations of total arsenic in a samples of
lobster purchased commercially in Ottawa was 4.7 ppm and in a shrimp sample was 7.2 ppm.
Lopez et al. (1994) found values of 4.01 ppm, 0.34 ppm and 2.95 ppm for commercially
purchased  samples of cockles, prawns, and mussels respectively.

Inorganic Arsenic. The average amount of arsenic present as inorganic arsenic  in shellfish was
less than 3% in all cases but one among the data summarized by Chew (1996). In the case of
Borneo dilatata, the one exception, 98% of the arsenic was inorganic but the total arsenic was
low (0.2 ppm) making the net exposure to inorganic arsenic low despite the high percentage
present. Lopez  et al. (1994) found the inorganic arsenic to account for 8% of the total in cockles
and 11% in muscles.  The amount of inorganic arsenic in  the prawns was below  the limit of
detection (0.023 ppm).

The data by Ballin et al. (1994) are of minimal value for deriving an estimate of the inorganic
arsenic in shellfish because samples were not analyzed for inorganic arsenic and the authors felt
that two organic forms of arsenic monitored (arsenobetaine and phospholipid arsenic) did not
account for all of the organic arsenic.

Ballin et al. (1994) evaluated shrimp, lobster, mussels and oysters  for their arsenobetaine and
arsenic containing phospholipids. Oysters and blue mussels had the lowest concentrations of
arsenobetaine plus phospholipid arsenic (30 and 40%). The authors hypothesized that water
soluble arsenocholine, the metabolic precursor to arsenobetaine, accounted for most of the
difference between the arsenobetaine and total arsenic concentrations rather than inorganic
arsenic. Oysters and muscles had a fair amount of their arsenic present in the fat soluble
phospholipid phase (10 to 20%) suggesting that some arsenocholine had been incorporated in the
choline-containing phospholipids or sphingolipids.
                                    Decembers, 1997

 In shrimp, Ballin et al. (1994) found that 87% of the arsenic was present as arsenobetaine and 4%
 as phospholipid arsenic while Lawrence et al.(1986) found that 76% was present as
 arsenobetaine and 15% as arsenocholine. In lobster, 59% of the arsenic was present as
 arsenobetaine and 2 % as phospholipid arsenic according to Ballin et al. (1994) and Lawrence et
 al. (1886) found 87% as arsenobetaine and none as arsenocholine. As discussed above, one
 cannot infer from these data that the remaining arsenic is present in inorganic compounds. Le et
 al. (1994) found arsenobetaine to be the primary organic arsenic compound in shrimp and

Table I summarizes the data on total, organic and inorganic concentrations of arsenic in fish and
shellfish. The organic arsenic data are extrapolated from either the total arsenic and inorganic
arsenic data (Chew, 1996) or the total arsenic plus  the arsenobetaine, arsenocholine and/or
phospholipid arsenic data (Ballin et al., 1994; Lawrence et al., 1986). In general, the data
support the conclusion, that, in fish, less than 10% of the total arsenic is inorganic (Chew, 1996 ).
Indeed, it was 4% or less for all fish species other than shark, sturgeon and sucker evaluated by
Chew (1996); a total of 40 species were evaluated.
Table I
Arsenic in Fish and Shellfish


Inorganic Arsenic
ppm or %.
0.1-64 ppm
78-88 %
Ballin etal., 1994
Chew. 1996
Lawrence et al., 1986
Lopez etal.. 1994
Ballin etal., 1994
Chew, 1996
Lawrence et al., 1986
Lopez etal., 1994
 Weights expressed as ppm wet weight
 ND = Not Determined
 NA = Not Applicable
 DL = Detection Limit

                                   Decembers, 1997

 The predominant organic arsenic compounds in marine fish  and shellfish are arsenobetaine and
 arsenocholine (Ballin et al., 1994). Some of the arsenocholine is found in tissue phospholipids.
 The amount of arsenobetaine exceeds the arsenocholine (Chew, 1996). In both compounds,
 arsenic has replaced the nitrogen of the natural metabolite (choline; betaine).  Arsenic is
 incorporated into the betaine molecule by microorganisms, phytoplankton, zooplankton and
 algae (Ballin et al., 1994). The fish obtain arsenobetaine from their food supply.  The principle
 organic form of arsenic in freshwater fish is neither arsenobetaine or arsenocholine according to
 data collected by Lawrence wt al. (1986). A single compound was isolated from all samples and
 accounted at least 70 to 85% of all the arsenic present if recovery was complete. The
 composition of this compound was not identified but it appeared to be more hydrophilic than

 Betaine is formed metabolically from choline through oxidation and becomes an excretory
 nitrogen metabolite.' Betaine excretions vary between species. In addition, betaine can serve as a
 methyl donor in biological systems, becoming N,N-dimethyl glycine (Montgomery, 1990).
 Betaine acts as an osmolyte in marine species (Neufield and Wright, 1996) and as a
chemosensory agent (Knutsen, 1992). A study of feeding behavior in North Sea rurbot and
Dover sole suggest that betaine is one of a number of water-soluble, nitrogen-containing
compounds that stimulate feeding behavior in fish larvae and may help to attract the larvae to the
plankton layer (Knutsen, 1992).  Betaine is probably released to water in plankton-rich areas
 producing an betaine-enriched microenvironment.

 Arsenobetaine is metabolically inert in mammalian systems. Almost all of the radiolabeled
arsenic in arsenobetaine administered orally or intravenously to rats, mice or guinea pig was
 excreted in three days (Vahter et al., 1983; Yamauchi et al. 1986). In rats and mice, more than
 99% of the excreted label was found in the urine as arsenobetaine. In comparable  studies using
arsenocholine, there was greater label retention with 70-80% excreted in three days (Marafante et
al., 1984). Extracts from mouse urine showed that more than 90% of the water soluble arsenic
excreted was present as arsenobetaine.  It can be assumed that some of the retained arsenocholine
 is incorporated in membranes as phosphatidyl choline compounds or in lipoprotein complexes
and, thus, will have little tendency to bioaccumulate as inorganic arsenic.

Small amounts of methylarsonic acid and dimethylarsinic acid have been  identified in fish and
shellfish (Buchet et al., 1994; Chew 1996).  Chew (1996) reported data from a study offish at the
ASARCO Tacoma Smelter Site in Washington state.  The amount of methylarsonic acid in
 striped sea perch was 0.02 ppb and that in rock sole was 0.002 ppb. The dimethylarsinic acid in
these two species was 0.02 ppb and < 6.6 ppb, respectively. In mussels there was  0.02 ppb for
 both the methylarsonic and dimethylarsinic acid. The total arsenic concentration was only
 available for the rock sole and the total arsenic data were not internally consistent  with the
 speciation data. Therefore, it is difficult to evaluate the significance of these results.  Buchet et
 al. (1994) found that the recovery of methylarsonic acid and dimethylarsinic acid varied with the
                                    December 3, 1997

extraction technique and between samples for the same fish.
The data available on the speciation of arsenic in fish and shellfish mitigate some of the
concerns generally associated with arsenic exposure. In most cases, more than 95% of the
arsenic is present as organic rather than inorganic compounds (Chew, 1996).  The most prevalent
of the organic species, especially in marine fish, is arsenobetaine (Ballin et al., 1994; Lawrence et
al. 1986), a compound with minimal tissue retention in the animals species studied and a
compound that is excreted without metabolic alteration.  The compound present in the next
highest concentration is arsenocholine. The arsenocholine is estimated to be less than 1% of the
total arsenic (Edmonds and Francesconi, 1993). Most of the arsenocholine is converted to
arsenobetaine and excreted. A small portion may become incorporated in phospholipids  and
retained; another small amount may be converted to trimethylarsine oxide (Chew, 1996).

Arsenobetaine and arsenocholine and the unidentified organic arsenic compound in freshwater
fish are hydrophilic and have little tendency to bioaccumulate in edible fish tissues.  They are
unlikely to be present in adipose deposits due to their hydrophilic nature although some
arsenocholine may be present in membrane phospholipids. Organic arsenicals, especially
arsenobetaine, appear to be significantly less toxic than inorganic arsenic species (Edmonds and
Francesconi, 1993). Each of these factors diminishes human health concerns related to exposure
to organic arsenic compounds in fish and shellfish. Additional support for a conclusion that the
organic arsenic compounds from fish and shellfish do not bioconcentrate is provided by data
showing that samples of human milk from 88 mothers from the Faroe Islands did not show
elevated arsenic in their transition milk despite consumption of diets rich in seafoods (Grandjean
etal., 1995).


5.1     EPA Bioconcentration Factors for Arsenic

The EPA bioconcentration factor for total arsenic in a fish and shellfish is 44 (EPA, 1980, 1984).
It applies to bioconcentration from a mixture offish and shellfish (roughly 10-15% shellfish).
The arsenic concentration by shellfish exceeds that for fish by nearly two orders of magnitude
(EPA, 1980). The shellfish value (350) came from a 112 day test of a saltwater oyster species
exposed to trivalent arsenic while the bioconcentration factor for bluegill was 4 after a 28 day
exposure (EPA,  1980).

The EPA bioconcentration factors were derived from laboratory studies where the water  was
spiked with trivalent inorganic arsenic. Thus, they may not be representative of what happens in
a natural ecosystem where inorganic arsenic is processed through a number of trophic levels
                                    Decembers, 1997

 before it reaches the fish or shellfish.

 The data presented in Sections 2.0 and 3.0 of this report do not include any information on the
 amount of arsenic in the source waters from which samples were obtained. Thus, it is not
 possible to evaluate the bioaccumulation that lead to the tissue levels of arsenic measured in the
 fish or shellfish. In general, the average total arsenic in clean costal and ocean waters is low,
 about 1 to 3 ug/L. Levels are much higher in estuary systems receiving arsenic discharges (Neff,
 1997). The arsenic concentration for most lakes and rivers is less than 5 ug/L (Crecelius, 1997).

A comparison of the amounts of total arsenic in some of the fish and shellfish samples collected
from the marine environment with normal background levels of inorganic arsenic in the water,
suggests that, for at least the marine environment, arsenic accumulated to a greater extent than
suggested by laboratory bioconcentration factor of 4 measured for a freshwater species in a
laboratory study. The data on the accumulation of arsenic in fresh water species is conceptually
closer to laboratory bioconcentration factors.

5.2    Estimated Total Arsenic

For this report, concentrations of total and inorganic arsenic in edible tissues from fish and
shellfish are estimated using the EPA bioconcentration factor for arsenic and the arsenic
concentration in ambient water. Deficiencies in the bioconcentration factor are part of the
uncertainty for the analysis.

In cases where the fish and shellfish come from water containing 50 ppb arsenic, the estimated
total arsenic concentration in edible tissues is 2.2 mg/kg or 2.2 ppm when calculated using the
EPA bioconcentration factor.

       50 ug/L x 44 L/kg(BCF) x 1 mg/1000 /zg = 2.2 mg/kg total arsenic

This estimate is within the range observed for fish and shellfish in the most recent USFDA Total
Diet Study (0.75 ppm for fish sticks to 2.8 for cod/haddock; Macintosh, 1997; personal
communications). In the data for fish collected by Chew (1996), 70% of the samples had total
arsenic concentrations below 2.2 ppm.  In the studies by Ballin et al. (1994 and Lawrence et al.
(1986) over 60% of the combined marine and fresh water fish data  set fell below this value.
Because nothing is known concerning the arsenic concentrations in the source water for the field
sample data presented by Ballin et al. (1994), Chew (1996) Lawrence et al. (1986) or Lopez et al.
(1994), the comparison of the calculated concentration for fish and shellfish with the field data
merely supports the calculated  value as plausible.

5.3    Estimated Inorganic Arsenic
                                    Decembers, 1997

 The maximum inorganic arsenic in fish and shellfish used for this estimate is 4% as discussed in
 Sections 2.0 and 3.0 above.  The median inorganic arsenic value for the fish and shellfish data
 reported by Chew (1996) is 0.4 %. No inorganic arsenic was detected in 23 of 42 fish samples

 and  18 of 50 shellfish samples.  Therefore, the median value reflects the higher inorganic arsenic
 concentrations found in shellfish and is a conservative value.

 Using a 4% maximum inorganic arsenic value for a mixed fish and shellfish diet, 2.2 ppm total
 arsenic in fish/shellfish  is equivalent to 0.09 ppm inorganic arsenic.
       2.2 mg/kg As ^  x 4 g As^^/lOO g As ^ = 0.09 mg/kg As
Using 0.4% as the median inorganic arsenic concentration for a mixed fish and shellfish diet, 2.2
ppm total arsenic in fish/shellfish is equivalent to 0.01 ppm inorganic arsenic.

       2.2 mg/kg As toul  x 0.4 g Asinorganic/100 g As tolal = 0.01 mg/kg As inorganic


Inorganic arsenic exposure estimates for high and average fish and shellfish consumers can be
derived using the inorganic arsenic concentrations estimates above and information on
population fish/shellfish consumption. The following exposure scenarios are presented for this
report based on available data.

       High fish/high arsenic  - The 99.9th percentile fish/shellfish consumer and the estimated
       maximum inorganic arsenic concentration for a mixed fish/shellfish diet (4%).  This
       group eats the maximum amount offish and shellfish on a daily basis and consistently
       chooses species that have the higher percentages of inorganic arsenic.

       High fish/average arsenic -The 99.9th percentile fish/shellfish consumer and the median
       inorganic arsenic concentration (0.4%). This group eats the maximum amount offish and
       shellfish on a daily basis and chooses species with a variety of inorganic arsenic
       concentrations over a range from low to high.

       Average fish/high arsenic - The 50th percentile fish/shellfish consumer and the
       estimated maximum inorganic arsenic concentration for a mixed fish /shellfish diet.  This
       group has a diet that includes fish and/or shellfish frequently but is not totally dependant
       on fish/shellfish as a dietary protein source. The  group preferences tend towards the
       fish/shellfish species that have the higher concentrations of inorganic arsenic.
                                    December3, 1997

       Average fish/ average arsenic - The 50th percentile fish/shellfish consumer and the
       median inorganic arsenic concentration. This group has a diet that includes fish and/or
       shellfish frequently but is not totally dependant on fish/shellfish as a dietary protein
       The group preferences include a variety offish and shellfish species with inorganic
       arsenic concentrations that range from low to high.

The fish consumption values used for these calculation were derived from the 1989-1991 dietary
records from the USDA Continuing Survey of Food Intake (USEPA,  1995).  The USD A food
consumption survey collects data on three consecutive days of food intake.  Data for one day is
provided through a 24-hour recall interview and data for two days through food intake records
kept by the respondent. Fish-consumers were segregated from all respondents based on their
consumption offish at least once during the three day survey period.  The population data used
for the exposure estimates described above apply only to the fish-consuming population rather
than the entire population Therefore the data may represent a rather skewed distribution. For the
purpose of defining a exposure on the high end of the  distribution curve the data are appropriate
and useful.

The 99.9th percentile value for females (461  g/day) is used to simulate the eating habits of
subsistence fishers such as the Eskimos and other native Indian tribes that consume a diet that is
very high in fish and shellfish. This is the highest intake value reported.  The 99.9th percentile
value or maximum  reported value is more than 4 times the 95th percentile value (USEPA, 1995).
The 50th percentile value from the USDA data for males was used for the exposure estimate
rather than that for females because it is a higher value.

Eskimo's and other native Indian tribes in Alaska have the highest consumption of fish and
shellfish within the United States (Wolfe, 1996).  In one study of 351  Eskimos, Indians and
Aleuts, average fish and shellfish consumption was 109 g/day (Nobmann et aL 1992). This
intake lies between the 75th and 95th percentiles offish consumers  in the country as a whole
(EPA, 1995). Wolfe (1996) found that the average intake of wild foods by the subsistence
populations in Alaska was slightly greater than one pound of wild food per day with 61% of this
total contributed by fish and shellfish. In some areas the average consumption of wild foods was
two pounds per day.  In studies of the Tulalip and Squaxin Island tribes of Puget Sound the-mean
fish/shellfish consumption for a 70 Kg adult was 71 g/day and the 95% percentile value was 226
g/day  (Toy et al., 1995). In a study of the Native Tribes of the Columbia River Basin, the mean
value  for the adult fish consuming population was 63  g/day and the 99th percentile value was
389 g/day (CRITFC,  1994). These data support the fish/shellfish intake values used for the
exposure estimates.

Two addition exposure estimates are also included in this report.  These groups are identified as
                                    December 3, 1997

       High fish (shellfish prefercncc)/high arsenic -The 99.9% fish/shellfish consumer who
       is also at the 99% for shellfish consumption (125 g/day). These individuals consume a

       high percent of their fish/shellfish in the form of shellfish and select species with the high
       concentrations of inorganic arsenic.

       General population/high arsenic - The average person who consumes fish or shellfish
       only occasionally but selects species with the high concentrations of inorganic arsenic.

The first of the added groups cover subsistence users offish and shellfish who have a greater
than average intake of shellfish. The arsenic exposure for this group includes separate
calculations for the arsenic in fish and shellfish using a fish bioconcentration factor of 4 for fish
and a factor of 350 for shellfish (USEPA, 1980, 1985).  The 99th percentile shellfish
consumption from the USD A data (125 g/day) was used for the shellfish portion of the diet and
the difference between the 99.9th percentile fish/shellfish value (461 g/day) and the shellfish
value was used for fish consumption (336 g/day).

The first five exposure scenarios defined above apply to those individuals who routinely
consume fish and/or shellfish as a dietary protein source. However, most of the general
population consumes fish and shellfish only  occasionally, and some individuals never eat fish or
shellfish. Thus, the general population has a lower exposure averaged over time. The EPA uses
a daily fish intake of 6.5 g/day to represent these individuals (EPA,  1989).  This is a normalized
concentration which recognizes that, on the days that fish and/or shellfish are consumed, the
intake will be higher than 6.5 grams but there will also be many days in the course of a year that
there is no consumption of either fish or shellfish. This group is identified as "General
Population " in  subsequent tables. As a worst case, the higher inorganic arsenic concentration
was used for the general population arsenic exposure calculation. Individuals who consume fish
or shellfish only occasionally tend to have a fe\v species they favor (e.g. tuna, shrimp) and the
species of preference may be among the higher arsenic species.

Inorganic arsenic exposures from fish and shellfish under the different exposure scenarios listed
above are summarized in Table 2. The fish/shellfish consumption values apply to total
fish/shellfish consumption and include marine, estuarine and freshwater species.  They are
calculated using the following equation:

       2.2 mg As y,^/ kg fish/shellfish  x fish/shellfish intake (kg/day) x
           As ,norganic/100 mg Astoul - mg As ,norgamc/day
 Units have been adjusted so that the inorganic arsenic concentrations are expressed in /^g/ day in
 Table 2. The USEPA limitation on arsenic in marine waters is 36 ppb (USEPA, 1992).
 Therefore, the estimates based on all fish and shellfish being from waters containing 50 ppb is an
                                     Decembers, 1997

overestimate in situations where a mixture of marine, estuarine and freshwater species are
Table 2
Inorganic Arsenic Exposure from Fish and Shellfish Consumption
Consumer Category
High Fish - High Arsenic
High Fish - Average Arsenic
Average Fish - High Arsenic
Average Fish -Average Arsenic
High Fish (shellfish preference)-
High Arsenic
General Population
Inorganic Arsenic
125 (shellfish)
336 (fish)
87.5 shellfish
2.7 fish
90 total
       1989-1991 data from the USDA Continuing Survey of Food Intake (USEPA, 1995)
**     1973-1974 data from the National Purchase Dairy Survey (USEPA, 1989)

Other dietary components can have an impact on the net inorganic arsenic exposure for the for all
consumer groups. Macintosh et al. (1996) found that other foods in the Total Diet Study such as
chicken and rice contributed to the totai arsenic exposure.  The form of arsenic in food may also
contribute to the effect of a eiven food material on the total bod\ arsenic load.

The data on inorganic arsenic from fish and shellfish from Table 2 can be combined with data on
arsenic concentrations in potable water to obtain a profile for net inorganic arsenic exposures in
the population groups characterized above. The high estimate for the arsenic concentration in
potable water is 20 ppb and the average estimate is 5 ppb.  These values were obtained from a
study of arsenic in potable water sources conducted by the University of Colorado at Boulder and
Malcolm Pimie, Inc. (1997).  Water consumption is estimated as 2 liters/day.

In the University of Colorado at Boulder/ Malcolm Pirnie Study (1997), 88 % or more of the
ground water samples analyzed in the  Western United States in three separate surveys had
                                   Decembers, 1997

concentrations of 20 ppb or less.  Concentrations were 5 ppb or less in more than 50% of the
ground water systems evaluated. The highest arsenic concentration from ground water sources
were in the Western region.  None of the surface water systems surveyed in the Western United
States contained greater than 5 ppb arsenic. In one survey, there were a few surface water
systems in the North Central region of the country that exceeded 20 ppb arsenic (12%).   Most
arsenic  in potable water is inorganic and ground water sources contain higher arsenic
concentrations than surface water sources.  The highest arsenic concentrations in the country are
concentrated on the west coast.

Table 3 presents the estimates for net exposure to inorganic arsenic when the estimates from fish
and shellfish consumption are combined with the data on the average and high concentrations of
arsenic in Public Water supplies across the country.  The value used as the high arsenic
concentrations is 20 ppb based on survey data rather than the Arsenic MCL and is exceeded by
only about 10% of Public Water Systems. Each of the exposure estimates for inorganic arsenic
intakes from fish, shellfish and water is less than the exposure that results from ingesting 2 L of
water containing the 50 ppb arsenic except for the High fish/shellfish preference scenarios.
Table 3
Total Inorganic Arsenic Exposures Fish/Shellfish and Water
for High and Average Fish Consuming Populations
Consumer Category
High Fish - High Arsenic

High Fish - Average Arsenic

Average Fish - High Arsenic

Average Fish - Average Arsenic

High Fish (Shellfish Preference)- High

General Population

                                    Decembers, 1997

 *       Based on a maximum 4% of the total arsenic being inorganic arsenic in an mixed
        fish/shellfish diet
 **     Water consumption is estimated as 2 liters/ day.
There are some regions in the western part of the country where arsenic levels in potable water
from public systems are equal to the MCL and other private systems where the arsenic may
exceed the MCL.  For Public Water Systems where the potable water concentration is equal to
the MCL, fish and shellfish consumption by the general populations does not increase the risk
from arsenic exposure since it represents a less than a 1 n% (1%) increase in the net arsenic
exposure. However, for regions where high levels of arsenic in the potable water are
accompanied by high levels offish and shellfish consumption the net increase in inorganic
arsenic exposure would be greater and site-specific criteria can be developed for surface waters
and for fish consumption.

In developing site-specific criteria the state should characterize the size and location of the
population of concern and determine their fish/shellfish and water intake rate. The fish and
shellfish consumption should consider the species and dietary intake per species.  Actual total
arsenic and inorganic arsenic values for the species consumed and actual concentrations in
drinking water should be used in the exposure calculations wherever possible. Other sources of
arsenic exposure should also be considered and quantified.


There are a number of uncertainties in the preceding exposure assessment for inorganic arsenic
from fish and shellfish originating from water containing 50 ppb arsenic.  The exposure estimates
assume a mixed  fish and shellfish diet in which average inorganic arsenic concentration is no
greater than  4% of the total  arsenic. This would not apply to any diet with high consumption of
shark, sturgeon and sucker.  However, these species are not used by the Eskimo and other
northern Indian tribes that serve as an example of a 99.9 percentile fish/shellfish-consuming
population.  Species that constitute the fish component of the diet for Eskimo's and other
northern Indians are salmon, halibut, herring, whitefish, sheefish, blackfish and cod (Wolfe,
The  exposure assessments for all but the High Fish (Shellfish Preference) group are also based
on a bioconcentration factor that applies to a mixed fish/shellfish diet.  It does not apply to a diet
that  is heavily weighted towards shellfish, particularly mollusks. In the Eskimo and other
northern Indian tribes, shellfish is a maximum of 9% of the diet (Wolfe, 1996) a value that is
representative of the fish/shellfish biconcentration factor used for the inorganic arsenic exposure
calculations. However, shellfish consumption for the Tulalip tribe of Washington State is about
60% of the fish/shellfish intake (Toy et  al., 1996). Thus, the inorganic arsenic exposure
estimates presented above would not apply to this group.
                                    Decembers, 1997

There is also some uncertainty in the bioconcentration factor since it results from laboratory
studies in which the water was spiked with inorganic.arsenic.  The values obtained may not be
representative of natural ecosystems where arsenic can passes through various trophic levels
before entry into fish tissues. The data suggest that bioaccumulation through the food chain is
more complex in marine species than in fresh water species.

There is some uncertainty in the toxicological assessment for organic arsenic compounds. To the
extent that most of the organic arsenic species in the fish are trimethylated species such as
arsenobetaine, arsenocholine and trimethylarsine oxide, toxicokinetic data support the conclusion
that there is little, if any interaction of the arsenic metabolite with other biomolecules. Thus, the
toxicity of these compounds is low. However, in cases where dimethylarsinic acid is found in
fish/shellfish species, low toxicity cannot be assumed because there are some data that suggest
that dimethylarsinic acid is a tumor promoter (Chew, 1996). If dimethylarsinic acid is a
promotor, it could become a risk factor for carcinogenicity. A weight-of-evidence determination
for the promoting properties of dimethylarsinic  acid has not been established. Lack of data on
the nature of the organoarsenic compound or compounds present in freshwater fish contributes
additional uncertainty in cases where most fish consumed are freshwater species.


Ballin, U., Kruse, R., and H.S. Russel. 1994. Determination of total arsenic and speciation of
arseno-betaine in marine fish by means of reaction-headspace gas chromatography utilizing
flame-ionization detection and element specific spectrometric detection. Fresenius J Anal Chem

Buchet, J.P., Pauwels, J., and R.  Lauwerys. 1994. Assessment of exposure to inorganic arsenic
following ingestion of marine organisms by volunteers. Environ. Res. 66:44-51.

Chew, C.  1996. Toxicity and exposure concerns related to arsenic in seafood: An arsenic
literature review for risk assessments. Prepared for USEPA Region 10 ESAT. Seattle
Washington as Technical Instruction Document 10-9601-815 under Work Unit document 4038.

Columbia River Inter-Tribal Fish commission (CRITFC). 1994. A fish consumption survey of
the Umatilla, Nez Perce, Yakama and Warm Springs Tribes of the Columbia River Basin.
CRITFC Technical Report No. 94-3.  Portland, OR.

Crecelius, E.A.. 1997. Review of EPA-822-R-97-002 (May 15, 1997) draft report entitled
"Arsenic and Fish Consumption'" Letter to Dennis McCauley, Great Lakes Environmental
Center, Traverse Center, MI. August 18.

Edmonds, J.S.D. and KA Francesconi. 1993. Arsenic in seafood:  human health aspects and
                                    Decembers, 1997

 regulations. Marine Pollution Bulletin 26:665-674.

 Grandjean, P., Weihe, P., Needham, L.L., Burse, V.W., Patterson, D.G., Sampson, E.J., Jorgesen,
 P.J., and M. Vahter. 1995. Relation of a seafood diet to mercury, selenium, arsenic, and
 polychlorinated biphenyl and other organochlorine concentrations in human milk. Environ. Res.

 Knutsen, J.A. 1992. Feeding behavior of North Sea turbot (Scophhalmus maximus) and Sover
 sole (Solea soled) larvae elicited by chemicals stimuli. Marine Biology 113:543-548.

 Lawrence, J.F., Michalik, P., Tarn, G., And H.B.S. Conacher. 1986. Identification of
 arsenobetaine and arsenocholine in candian fish and shellfish by high-performance liquid
 chromatography with atomic absorption detection and confirmation by fast atom bombardment
mass spectrometry. J. Agric. Food Chem. 34:315-319.

Le, Sean X.C., Cullen, W.R., and K.J. Reimer. 1994. Speciation of arsenic compounds in some
marine organisms. Environ. Sci. Technol. 28:1598-1604.

Lopez. J.C., Montoro, R., Cervera, M.L. and M. de la Guardia. 1994. Determination of inorganic
arsenic in sea food products by microwave-assisted distillation and atomic aabsorption
spectrometry. J. Analyt. Atom. Spec. 9:651-656.

Macintosh, D. L., Spengler, J.D., Ozkaynak, H., Tsai, L, and P.B. Ryan.  1996. Dietary
exposures to selected metals and pesticides. Environ Health Perspect 104:202-209.

Macintosh, D.L. 1997. Personal Communication. Rollins School of Public Health, Emory
University, Atlanta, GA 30322.

Marafante, E., Vahter, M., and L. Dencker. 1984. Metabolism of arsenocholine in mice, rats, and
rabbits. The Science of the Total Environment 34:223-240.

Montgomery, R., Conway, T.W. and A.A. Spector. 1990. Biochemistry: A Case Oriented
Approach. The C.V. Mosby Company. St. Louis, MO.

Neff, J.M.. 1997. Ecotoxicology of arsenic in the marine environment. Environ. Toxicol. Chem.

Neufield, D.S., and S.H. Wright. 1996.  Salinity change and cell volume: the response of tissues
 from the estuarian mussel Geukensia  demissa. J. Exp. Biol. 199:1619-1630.

Nobmann, E.D., Beyers, T., Lanier, A.P., Hankin, J.H, and M.Y. Jackson. 1992. The diet of
                                   Decembers, 1997

 Alaska Native adults. 1977-1988. Am. J. Clin. Nut. 55:1024-1032.

 Toy, K.A. Polissar, N.L., Liao, S. and G. D. Mittelstaedt. 1996. A fish consumption survey of the
 Tulalip and Squaxin Island tribes of the Puget Sound region. Tulalip Tribes, Department of
 Environment, 7615 Totem Beach Road, Marysville, Washington 98271.

 University of Colorado at Boulder and Malcolm Pirnie, Inc. 1997. National Compliance
 Assessment and Costs for Regulation of Arsenic in Drinking Water. Water Industry Technical
 Action Fund.

 USEPA. 1980.. Ambient Water Quality Criteria for Arsenic. United States Environmental
 Protection Agency. Office of Water Regulations and Standards (OWRS). EPA 440/5-80-021.

USEPA. 1984. Ambient Water Quality Criteria for Arsenic. United States Environmental
Protection Agency. Office of Water Regulations and Standards (OWRS). EPA 440/5-84-033.

USEPA. 1989. Exposure Factors Handbook. United States Environmental Protection Agency.
Office of Health and Environmental Assessment. EPA 600 8-89-043.

USEPA. 1992. National Toxics Rule: Criterion continuous concentration. United States
Environmental Protection Agency. Federal Register 57:60848.

USEPA. 1995 Draft. Mercury Study Report to Congress, Volume III: An Assessment of
 Exposure from Anthropogenic mercury Emissions in the United States. United States
 Environmental Protection Agency, Office of Air Quality Planning and Standards; Office of
 Research and Development.

 Vahter, M.. Marafante, E., and L. Dencker. 1983. Metabolism  of arsenobetaine in mice, rats and
 rabbits. Sci. Tot. Environ, 30:197-211.

 Wolfe, R.J.. 1996. Subsistence-food harvests in rural Alaska and food safety issues. Paper
 presented to the Institute of Medicine, National Academy of Sciences Committee on
 Environmental Justice. Spokane, Washington. August 13.
                                   December3, 1997