ARSENIC 313
6. POTENTIAL FOR HUMAN EXPOSURE
6.1 OVERVIEW
Arsenic has been identified in at least 1,149 of the 1,684 hazardous waste sites that have been proposed
for inclusion on the EPA National Priorities List (NPL) (HazDat 2006). However, the number of sites
evaluated for arsenic is not known. The frequency of these sites can be seen in Figure 6-1. Of these sites,
1,134 are located within the United States and 11, 2, and 2 are located in the Commonwealth of Puerto
Rico, the Virgin Islands, and Guam (not shown).
Arsenic is widely distributed in the Earth's crust, which contains about 3.4 ppm arsenic (Wedepohl 1991).
It is mostly found in nature in minerals, such as realgar (As
4
S
4
), orpiment (As
2
S
3
), and arsenolite (As
2
O
3
),
and only found in its elemental form to a small extent. There are over 150 arsenic-bearing minerals
(Budavari et al. 2001; Carapella 1992). While arsenic is released to the environment from natural sources
such as wind-blown soil and volcanoes, releases from anthropogenic sources far exceed those from
natural sources. Anthropogenic sources of arsenic include nonferrous metal mining and smelting,
pesticide application, coal combustion, wood combustion, and waste incineration. Most anthropogenic
releases of arsenic are to land or soil, primarily in the form of pesticides or solid wastes. However,
substantial amounts are also released to air and water.
Arsenic found in soil either naturally occurring or from anthropogenic releases forms insoluble complexes
with iron, aluminum, and magnesium oxides found in soil surfaces, and in this form, arsenic is relatively
immobile. However, under reducing conditions, arsenic can be released from the solid phase, resulting in
soluble mobile forms of arsenic, which may potentially leach into groundwater or result in runoff of
arsenic into surface waters. In aquatic systems, inorganic arsenic occurs primarily in two oxidation states,
As(V) and As(III). Both forms generally co-exist, although As(V) predominates under oxidizing
conditions and As(III) predominates under reducing conditions. Arsenic may undergo a variety of
reactions in the environment, including oxidation-reduction reactions, ligand exchange, precipitation, and
biotransformation (EPA 1979, 1984a; Pongratz 1998; Welch et al. 1988). These reactions are influenced
by Eh (the oxidation-reduction potential), pH, metal sulfide and sulfide ion concentrations, iron
concentration, temperature, salinity, and distribution and composition of the biota (EPA 1979; Wakao et
al. 1988). Much of the arsenic will adsorb to particulate matter and sediment. Arsenic released to air
exists mainly in the form of particulate matter. Arsenic released from combustion processes will
ARSENIC 315
6. POTENTIAL FOR HUMAN EXPOSURE
generally occur as highly soluble oxides. These particles are dispersed by the wind and returned to the
earth in wet or dry deposition. Arsines that are released to the atmosphere as a result of microbial action
are oxidized to nonvolatile species that settle back to the ground.
Because arsenic is a natural component of the Earth's crust, low levels of the element are found in all
environmental media. Atmospheric levels of arsenic in remote locations (away from human releases)
range from 1 to 3 ng/m
3
, while concentrations in urban areas may range from 20 to 100 ng/m
3
.
Concentrations in water are usually <10 μg/L, although higher levels may occur near natural mineral
deposits or anthropogenic sources. Natural levels of arsenic in soil usually range from 1 to 40 mg/kg,
with a mean of 5 mg/kg, although much higher levels may occur in mining areas, at waste sites, near high
geological deposits of arsenic-rich minerals, or from pesticide application. Arsenic is also found in many
foods, at concentrations that usually range from 20 to 140 μg/kg. Total arsenic concentrations may be
substantially higher in certain seafoods. However, the general consensus in the literature is that about 85–
>90% of the arsenic in the edible parts of marine fish and shellfish is organic arsenic (e.g., arsenobetaine,
arsenochloline, dimethylarsinic acid) and that approximately 10% is inorganic arsenic (EPA 2003b).
Drinking water in the United States generally contains an average of 2 μg/L of arsenic (EPA 1982c),
although 12% of water supplies from surface water sources in the north Central region of the United
States and 12% of supplies from groundwater sources in the western region have levels exceeding
20 μg/L (Karagas et al. 1998). In January 2001, EPA adopted a new standard that arsenic levels in
drinking water were not to exceed 10 μg/L, replacing the previous standard of 50 μg/L. The date for
compliance with the new MCL was January 23, 2006 (EPA 2001).
For most people, diet is the largest source of exposure to arsenic. Mean dietary intakes of total arsenic of
50.6 μg/day (range of 1.01–1,081 μg/day) and 58.5 μg/day (range of 0.21–1,276 μg/day) has been
reported for females and males (MacIntosh et al. 1997). U.S. dietary intake of inorganic arsenic has been
estimated to range from 1 to 20 μg/day, with grains and produce expected to be significant contributors to
dietary inorganic arsenic intake (Schoof et al. 1999a, 1999b). The predominant dietary source of arsenic
is generally seafood. Inorganic arsenic in seafood sampled in a market basket survey of inorganic arsenic
in food ranged from <0.001 to 0.002 μg/g (Schoof et al. 1999a, 1999b). Intake of arsenic from air and
soil are usually much smaller than that from food and water (Meacher et al. 2002).
People who produce or use arsenic compounds in occupations such as nonferrous metal smelting,
pesticide manufacturing or application, wood preservation, semiconductor manufacturing, or glass
production may be exposed to substantially higher levels of arsenic, mainly from dusts or aerosols in air.
ARSENIC 316
6. POTENTIAL FOR HUMAN EXPOSURE
Exposure at waste sites may occur by a variety of pathways, including inhalation of dusts in air, ingestion
of contaminated soil or water, or through the food chain. The magnitude of the exposures can only be
evaluated on a site-by-site basis; however, exposures generally do not exceed background intakes from
food and drinking water.
Tables 4-1, 4-2, 4-3, and 4-4 summarize all of the names, abbreviations, and structures of the various
arsenic compounds that are discussed in Chapter 6.
6.2 RELEASES TO THE ENVIRONMENT
The Toxics Release Inventory (TRI) data should be used with caution because only certain types of
facilities are required to report (EPA 2005k). This is not an exhaustive list. Manufacturing and
processing facilities are required to report information to the TRI only if they employ 10 or more full-time
employees; if their facility is included in Standard Industrial Classification (SIC) Codes 10 (except 1011,
1081, and 1094), 12 (except 1241), 20–39, 4911 (limited to facilities that combust coal and/or oil for the
purpose of generating electricity for distribution in commerce), 4931 (limited to facilities that combust
coal and/or oil for the purpose of generating electricity for distribution in commerce), 4939 (limited to
facilities that combust coal and/or oil for the purpose of generating electricity for distribution in
commerce), 4953 (limited to facilities regulated under RCRA Subtitle C, 42 U.S.C. section 6921 et seq.),
5169, 5171, and 7389 (limited S.C. section 6921 et seq.), 5169, 5171, and 7389 (limited to facilities
primarily engaged in solvents recovery services on a contract or fee basis); and if their facility produces,
imports, or processes ≥25,000 pounds of any TRI chemical or otherwise uses >10,000 pounds of a TRI
chemical in a calendar year (EPA 2005k).
6.2.1 Air
Estimated releases of 4,800 pounds (~2.2 metric tons) of arsenic to the atmosphere from 58 domestic
manufacturing and processing facilities in 2004, accounted for about 0.52% of the estimated total
environmental releases from facilities required to report to the TRI (TRI04 2006). Estimated releases of
0.13 million pounds (~59 metric tons) of arsenic compounds to the atmosphere from 361 domestic
manufacturing and processing facilities in 2004, accounted for about 0.11% of the estimated total
environmental releases from facilities required to report to the TRI (TRI04 2006). These releases for
arsenic and arsenic compounds are summarized in Table 6-1 and 6-2, respectively.
ARSENIC 317
6. POTENTIAL FOR HUMAN EXPOSURE
Table 6-1. Releases to the Environment from Facilities that Produce, Process, or
Use Arsenic
a
Reported amounts released in pounds per year
b
Total release
State
c
RF
d
Air
e
Water
f
UI
g
Land
h
Other
i
On-site
j
Off-site
k
On- and
off-site
AL 1 51 162 0 110,264 0 110,425 52 110,477
AR 2 0 0 No data 0 0 No data 0 0
AZ 2 10 0 0 20,717 0 20,727 0 20,727
CA 3 13 14 0 5,482 0 13 5,497 5,510
FL 2 4 0 0 0 4,950 4 4,950 4,954
GA 4 8 10 0 1,603 5 13 1,613 1,626
IA 1 0 1 0 0 0 0 1 1
ID 1 39 0 0 361,252 0 361,291 0 361,291
IL 2 250 129 0 14,087 0 379 14,087 14,466
IN 1 5 5 0 13,250 250 5 13,505 13,510
KS 1 0 0 No data 0 0 No data 0 0
KY 1 0 1 0 0 6 1 6 7
MI 2 0 5 0 0 750 5 750 755
MN 1 15 47 0 14,504 0 15 14,551 14,566
MO 1 5 0 0 0 4,040 5 4,040 4,045
MS 2 0 0 0 0 0 0 0 0
NC 4 35 8 0 1 1 43 2 45
NV 1 0 0 0 0 0 0 0 0
NY 4 0 1 0 26,525 1 26,401 126 26,527
OH 2 13 0 0 0 0 13 0 13
OR 1 0 0 0 92,606 0 92,606 0 92,606
PA 5 166 8 0 14,362 26,140 199 40,477 40,676
SC 3 10 10 0 0 1,002 15 1,007 1,022
TN 3 3,988 0 0 0 0 3,988 0 3,988
TX 5 139 376 168,563 12,600 0 181,636 42 181,678
ARSENIC 318
6. POTENTIAL FOR HUMAN EXPOSURE
Table 6-1. Releases to the Environment from Facilities that Produce, Process, or
Use Arsenic
a
Reported amounts released in pounds per year
b
Total release
State
c
RF
d
Air
e
Water
f
UI
g
Land
h
Other
i
On-site
j
Off-site
k
On- and
off-site
WI 2 15 0 0 760 0 15 760 776
WV 1 0 0 0 10,135 0 10,135 0 10,135
Total 58 4,766 778 168,563 698,149 37,145 807,935 101,466 909,401
a
The TRI data should be used with caution since only certain types of facilities are required to report. This is not an
exhaustive list. Data are rounded to nearest whole number.
b
Data in TRI are maximum amounts released by each facility.
c
Post office state abbreviations are used.
d
Number of reporting facilities.
e
The sum of fugitive and point source releases are included in releases to air by a given facility.
f
Surface water discharges, waste water treatment-(metals only), and publicly owned treatment works (POTWs)
(metal and metal compounds).
g
Class I wells, Class II-V wells, and underground injection.
h
Resource Conservation and Recovery Act (RCRA) subtitle C landfills; other on-site landfills, land treatment, surface
impoundments, other land disposal, other landfills.
i
Storage only, solidification/stabilization (metals only), other off-site management, transfers to waste broker for
disposal, unknown
j
The sum of all releases of the chemical to air, land, water, and underground injection wells.
k
Total amount of chemical transferred off-site, including to POTWs.
RF = reporting facilities; UI = underground injection
Source: TRI04 2006 (Data are from 2004)
ARSENIC 319
6. POTENTIAL FOR HUMAN EXPOSURE
Table 6-2. Releases to the Environment from Facilities that Produce, Process, or
Use Arsenic Compounds
a
Reported amounts released in pounds per year
b
Total release
State
c
RF
d
Air
e
Water
f
UI
g
Land
h
Other
i
On-site
j
Off-site
k
On- and off-
site
AK 1 511 0 1,400,000 1,200,000 0 2,600,511 0 2,600,511
AL 19 4,299 18,127 0 853,469 7,555 875,876 7,574 883,450
AR 12 0 0 0 133 26,435 0 26,568 26,568
AZ 5 5,421 0 0 402,335 422 394,749 13,429 408,178
CA 5 65 14 0 355,660 86,396 160,673 281,461 442,134
CO 1 11 0 0 4,094 0 4,105 0 4,105
CT 1 0 0 0 0 0 No data 0 0
FL 15 3,208 503 0 343,508 4,057 346,310 4,966 351,276
GA 23 8,643 7,823 0 422,124 5,127 437,496 6,221 443,717
HI 1 0 0 0 0 0 No data 0 0
IA 4 1,291 482 0 0 35,324 1,773 35,324 37,097
ID 3 332 20 0 1,056,904 0 1,057,256 0 1,057,256
IL 11 3,960 3,110 0 96,093 21,038 71,819 52,382 124,202
IN 21 13,786 8,282 0 768,297 42,808 632,704 200,470 833,174
KS 4 924 0 0 12,082 1 13,006 1 13,007
KY 18 14,406 8,427 0 616,074 95,285 578,080 156,112 734,192
LA 7 265 23 0 25,426 0 25,563 151 25,714
MA 1 0 0 0 0 500 0 500 500
MD 8 1,870 291 0 34,130 114,115 2,661 147,745 150,406
MI 10 1,123 2,310 68,924 101,857 1,059 77,505 97,769 175,274
MN 2 10 130 0 19,270 0 19,410 0 19,410
MO 6 462 116 0 27,855 936 10,026 19,343 29,369
MS 6 61 121 0 11,676 46 11,228 676 11,904
MT 3 630 0 0 2,138,190 37 2,138,820 37 2,138,857
NC 15 5,626 4,732 0 168,030 2,429 178,388 2,429 180,818
ND 6 6,326 5 0 318,175 0 137,961 186,545 324,506
NE 2 180 0 0 11,000 0 11,180 0 11,180
NJ 2 0 1 0 0 8 0 9 9
NM 2 130 0 0 18,326 0 18,456 0 18,456
NV 10 3,041 30,017 0 98,894,564 0 98,927,328 294 98,927,622
NY 3 67 36 0 27,059 802 27,141 823 27,964
OH 17 8,595 8,352 81,024 741,730 274 668,157 171,818 839,975
OK 4 115 13 0 25,000 4,202 115 29,215 29,330
OR 4 0 5 0 0 4,012 5 4,012 4,017
PA 23 18,963 2,166 0 666,753 69,053 403,582 353,353 756,935
PR 3 0 0 0 0 0 No data 0 0
RI 1 0 8 0 0 1,006 8 1,006 1,014
ARSENIC 320
6. POTENTIAL FOR HUMAN EXPOSURE
Table 6-2. Releases to the Environment from Facilities that Produce, Process, or
Use Arsenic Compounds
a
Reported amounts released in pounds per year
b
Total release
State
c
RF
d
Air
e
Water
f
UI
g
Land
h
Other
i
On-site
j
Off-site
k
On- and off-
site
SC 13 2,178 1,443 0 25,817 22,705 29,438 22,705 52,143
SD 1 0 0 0 0 0 No data 0 0
TN 13 3,379 25,878 0 292,914 17,219 258,643 80,746 339,389
TX 17 4,616 199 33,148 196,385 31,557 226,751 39,155 265,906
UT 5 6,715 4,500 0 6,368,500 3,500 6,379,715 3,500 6,383,215
VA 11 1,911 2,773 0 160,154 8,463 164,789 8,512 173,301
WA 4 0 0 0 0 0 No data 0 0
WI 4 94 21 0 1,313 9,216 223 10,421 10,644
WV 12 2,693 2,417 0 536,628 10,000 441,237 110,501 551,738
WY 2 3,300 0 0 10,800 0 14,100 0 14,100
Total 361 129,205 132,347 1,583,096 116,952,326 625,588 117,346,787 2,075,775 119,422,562
a
The TRI data should be used with caution since only certain types of facilities are required to report. This is not an
exhaustive list. Data are rounded to nearest whole number.
b
Data in TRI are maximum amounts released by each facility.
c
Post office state abbreviations are used.
d
Number of reporting facilities.
e
The sum of fugitive and point source releases are included in releases to air by a given facility.
f
Surface water discharges, waste water treatment-(metals only), and publicly owned treatment works (POTWs) (metal
and metal compounds).
g
Class I wells, Class II-V wells, and underground injection.
h
Resource Conservation and Recovery Act (RCRA) subtitle C landfills; other on-site landfills, land treatment, surface
impoundments, other land disposal, other landfills.
i
Storage only, solidification/stabilization (metals only), other off-site management, transfers to waste broker for
disposal, unknown
j
The sum of all releases of the chemical to air, land, water, and underground injection wells.
k
Total amount of chemical transferred off-site, including to POTWs.
RF = reporting facilities; UI = underground injection
Source: TRI04 2006 (Data are from 2004)
ARSENIC 321
6. POTENTIAL FOR HUMAN EXPOSURE
Arsenic naturally occurs in soil and will be present in the atmosphere as airborne dust. It is also emitted
from volcanoes and in areas of dormant volcanism (e.g., fumaroles). Gaseous alkyl arsenic compounds
may be released from soil that has been treated with inorganic arsenic compounds as a result of biogenic
processes (Schroeder et al. 1987; Tamaki and Frankenberger 1992). Arsenic naturally occurs in sea water
and vegetation and is released into the atmosphere in sea salt spray and forest fires. Anthropogenic
sources of arsenic include nonferrous metal smelting, coal, oil and wood combustion, and municipal
waste incineration. Arsenic naturally occurs in coal and oil and therefore, coal- and oil-fired power plants
release arsenic to the atmosphere in their emissions (Pacyna 1987). Arsenic’s use in agriculture and
industrial processes also contributes to its emissions. One important source of arsenic emissions is cotton
ginning in which the cotton seeds are removed from the raw cotton.
The National Air Toxics Assessment reported that total anthropogenic emissions for arsenic compounds
in the United States in 1996 were 355 tons/year (EPA 2005b). EPA conducted a modeling study with the
Assessment System for Population Exposure Nationwide (ASPEN) in which estimates of emissions of
hazardous air pollutants were used to estimate air quality (Rosenbaum et al. 1999). Using 1990 data, the
total emissions of arsenic in the conterminous 48 states, excluding road dust or windblown dust from
construction or agricultural tilling was estimated to be 3.0 tons/day with 90% of emissions coming from
point sources and 5% each from area and mobile sources. U.S. emissions of arsenic to the atmosphere
were estimated as 3,300 metric tons per year between 1979 and 1986 (Pacyna et al. 1995). There is
evidence that anthropogenic emissions, at least from smelters, are lower than they had been in the early
1980s. It is likely that air releases of arsenic decreased during the 1980s due to regulations on industrial
emissions (EPA 1986f), improved control technology for coal-burning facilities, and decreased use of
arsenical pesticides.
Nriagu and Pacyna (1988) and Pacyna et al. (1995) estimated worldwide emissions of arsenic to the
atmosphere for 1983. Estimates of yearly emissions from anthropogenic sources ranged from 12,000 to
25,600 metric tons with a median value of 18,800 metric tons. Natural sources contributed 1,100–
23,500 metric tons annually. Chilvers and Peterson (1987) estimated global natural and anthropogenic
arsenic emissions to the atmosphere as 73,500 and 28,100 metric tons per year, respectively. Copper
smelting and coal combustion accounted for 65% of anthropogenic emissions. A U.S. Bureau of Mines
study on the flow of mineral commodities estimated that global emissions of arsenic from metal smelting,
coal burning, and other industrial uses ranged from 24,000 to 124,000 metric tons per year compared to
natural releases, mostly from volcanoes, ranging from 2,800 to 8,000 metric tons per year (Loebenstein
1994).
ARSENIC 322
6. POTENTIAL FOR HUMAN EXPOSURE
Pirrone and Keeler (1996) compared trends of trace element emissions from major anthropogenic sources
in the Great Lakes region with ambient concentrations observed in urban areas of the region. They found
that arsenic emissions increased about 2.8% per year from 1982 to 1988 and then decreased steadily by
about 1.4% per year to 1993. Coal combustion in electric utilities and in residential, commercial, and
industrial facilities was an important source of arsenic in the region, accounting for about 69% of the total
emissions. Iron-steel manufacturing accounted for about 13% of the region wide arsenic emissions and
nonferrous metals production for 17%.
Arsenic in the particulate phase is the predominant (89–98.6%) form of arsenic in the troposphere
(Matschullat 2000). Inorganic species, most commonly trivalent arsenic, is the dominant form of arsenic
in the air over emission areas; methylated forms of arsenic are probably of minor significance. Arsenic-
containing air samples of smelter or coal-fired power plant origin consist largely of trivalent arsenic in
both vapor and particulate form (Pacyna 1987). Oxides are the primary species evolved from fossil fuel
and industrial processes. Additionally, arsenic trisulfide has also been reported from coal combustion,
organic arsines from oil combustion, and arsenic trichloride from refuse incineration.
Arsenic has been identified in 59 air samples collected from 1,684 current or former NPL hazardous
waste sites where it was detected in some environmental media (HazDat 2006).
6.2.2 Water
Estimated releases of 780 pounds (~0.35 metric tons) of arsenic to surface water from 58 domestic
manufacturing and processing facilities in 2004, accounted for about 0.09% of the estimated total
environmental releases from facilities required to report to the TRI (TRI04 2006). Estimated releases of
1.3x105 pounds (~59 metric tons) of arsenic compounds to surface water from 361 domestic
manufacturing and processing facilities in 2004, accounted for about 0.11% of the estimated total
environmental releases from facilities required to report to the TRI (TRI04 2006). These releases for
arsenic and arsenic compounds are summarized in Tables 6-1 and 6-2, respectively.
Arsenic may be released to water from the natural weathering of soil and rocks, and in areas of vulcanism.
Arsenic may also leach from soil and minerals into groundwater. Anthropogenic sources of arsenic
releases to water include mining, nonferrous metals, especially copper, smelting, waste water, dumping of
sewage sludge, coal burning power plants, manufacturing processes, urban runoff, atmospheric deposition
and poultry farms (Garbarino et al. 2003; Nriagu and Pacyna 1988; Pacyna et al. 1995). A contributory
ARSENIC 323
6. POTENTIAL FOR HUMAN EXPOSURE
part of mining and coal burning power plants is leaching from abandoned mine tailing and fly ash waste
piles. Significant amounts of arsenic are released in liquid effluents from gold-milling operations using
cyanide (Environment Canada 1993). Nriagu and Pacyna (1988) and Pacyna et al. (1995) estimated
global anthropogenic inputs of arsenic into rivers, lakes, and oceans for 1983; annual estimated inputs
ranged from 11,600 to 70,300 metric tons with a median value of 41,800 metric tons. Arsenic was
detected in 58% of samples of urban storm water runoff from 8 of 15 cities surveyed in the National
Urban Runoff Program at concentrations ranging from 1 to 50.5 μg/L (Cole et al. 1984).
Leaching of arsenic from soil, landfills, or slag deposits is a source of arsenic in groundwater (Francis and
White 1987; Wadge and Hutton 1987). The arsenic in soil may be naturally-occurring or a result of the
application of arsenic-containing pesticides or sludge. Wood treated with CCA is used in piers, piling
and bulkheads and arsenic can leach from the treated wood (Breslin and Adler-Ivanbrook 1998; Brooks
1996; Cooper 1991; Sanders et al. 1994; Weis et al. 1998). Ammoniacal copper zinc arsenate (ACZA) is
another arsenic-containing waterborne preservative; however, it is not as widely used as CCA (Lebow et
al. 2000).
Arsenic has been identified in 846 groundwater and 414 surface water samples collected from 1,684 NPL
hazardous waste sites, where it was detected in some environmental media (HazDat 2006).
6.2.3 Soil
Estimated releases of 0.70 million pounds (~320 metric tons) of arsenic to soils from 58 domestic
manufacturing and processing facilities in 2004, accounted for about 77% of the estimated total
environmental releases from facilities required to report to the TRI (TRI04 2006). An additional
0.17 million pounds (~77 metric tons), constituting about 19% of the total environmental emissions, were
released via underground injection (TRI04 2006). Estimated releases of 117 million pounds
(~5.3x104 metric tons) of arsenic compounds to soils from 361 domestic manufacturing and processing
facilities in 2004, accounted for about 98% of the estimated total environmental releases from facilities
required to report to the TRI (TRI04 2006). An additional 1.6 million pounds (~720 metric tons),
constituting about 1.3% of the total environmental emissions, were released via underground injection
(TRI04 2006). These releases for arsenic and arsenic compounds are summarized in Tables 6-1 and 6-2,
respectively.
The soil receives arsenic from a variety of anthropogenic sources, including ash residue from power
plants, smelting operations, mining wastes, and municipal, commercial, and industrial waste. Ash from
ARSENIC 324
6. POTENTIAL FOR HUMAN EXPOSURE
power plants is often incorporated into cement and other materials that are used for roads and
construction. Arsenic may be released from such material into soil. Nriagu and Pacyna (1988) and
Pacyna et al. (1995) estimated global anthropogenic inputs of arsenic into soil for 1983. Excluding mine
tailings and smelter slag, annual estimated inputs ranged from 52,000 to 112,000 metric tons with a
median value of 82,000 metric tons. Mine tailings and smelter slag were estimated to add an additional
7,200–11,000 and 4,500–9,000 metric tons, respectively. Old abandoned mine tailings undoubtedly
contribute still more. Wood treated with CCA used in foundations or posts could potentially release
arsenic into the surrounding soil. CCA preservatives have been shown to leach to varying degrees from
wood, as well as through soils in both field and laboratory studies (Chirenje et al. 2003a; Hingston et al.
2001; Lebow et al. 2000; Rahman et al. 2004; Stilwell and Graetz 2001; USDA/USDT 2000). Arsenic
may also be released on land through the application of pesticides and fertilizer. Senesi et al. (1999)
reported the range of arsenic in 32 fertilizers as 2.2–322 ng/g. Roxarsone (3-nitro-4-hydroxyphenyl-
arsonic acid), which was used to treat poultry feed in approximately 70% of the broiler poultry operations
in 1999–2000, is excreted unchanged in the manure. Poultry litter (manure and bedding) is routinely used
as fertilizer to cropland and pasture. In 2000, assuming 70% of the 8.3 billion broiler poultry produced in
the United States were fed roxarsone-treated feed, the resulting manure would contain approximately
2.5x10
5
kg of arsenic (Garbarino et al. 2003). Land application of sewage sludge is another source of
arsenic in soil. Arsenic was detected in sewage sludge samples from 23 cities at concentrations of 0.3–
53 μg/g (Mumma et al. 1984).
Arsenic has been identified in 758 soil and 515 sediment samples collected from 1,684 NPL hazardous
waste sites, where it was detected in some environmental media (HazDat 2006).
6.3 ENVIRONMENTAL FATE
6.3.1 Transport and Partitioning
Arsenic in soil may be transported by wind or in runoff or may leach into the subsurface soil. However,
because many arsenic compounds tend to partition to soil or sediment under oxidizing conditions,
leaching usually does not transport arsenic to any great depth (EPA 1982c; Moore et al. 1988; Pantsar-
Kallio and Manninen 1997; Welch et al. 1988). Arsenic is largely immobile in agricultural soils;
therefore, it tends to concentrate and remain in upper soil layers indefinitely. Downward migration has
been shown to be greater in a sandy soil than in a clay loam (Sanok et al. 1995). Arsenic from lead
arsenate that was used for pest control did not migrate downward below 20 cm in one fruit orchard; in
another orchard, 15 years after sludge amendments and deep plowing, essentially all arsenic residues
ARSENIC 325
6. POTENTIAL FOR HUMAN EXPOSURE
remained in the upper 40 cm of soil (Merwin et al. 1994). Leaching of arsenic in polluted wetland soil
was low; leaching was correlated with the amount of dissolved organic matter in the soil (Kalbitz and
Wennrich 1998). The effect of soil characteristics, namely pH, organic matter content, clay content, iron
oxide content, aluminum oxide content, and cation exchange capacity (CEC), on the adsorption of various
metals, including the metalloid arsenic, to 20 Dutch surface soils was assessed by regression analysis
(Janssen et al. 1997). The most influential parameter affecting arsenic adsorption was the iron content of
the soil.
Arsenic that is adsorbed to iron and manganese oxides may be released under reducing conditions, which
may occur in sediment or flooding conditions (LaForce et al. 1998; McGeehan 1996; Mok and Wai
1994). In addition to reductive dissolution, when nutrient levels are adequate, microbial action can also
result in dissolution (LaForce et al. 1998). Interestingly, drying of the previously flooded soil increases
arsenic adsorption, possibly due to alterations in iron mineralogy (McGeehan et al. 1998).
Darland and Inskeep (1997) conducted a study to determine the effects of pH and phosphate competition
on the transport of arsenate (H
x
AsO
4
x-3
) through saturated columns filled with sand containing free iron
oxides. At pH 4.5 and 6.5, arsenate transport was strongly retarded, while at pH 8.5, it was rapid. The
enhanced transport of arsenate at pH 8 is consistent with the pH dependence of surface complexation
reactions describing arsenate sorption by metal oxide minerals that can be categorized as a ligand
exchange mechanism. Phosphate was shown to compete effectively with arsenate for adsorption sites on
the sand, but the competition was not sufficient to desorb all of the arsenate in batch column experiments,
even when the applied phosphate exceeded the column adsorption capacity by a factor of two. The
researchers concluded that arsenate desorption kinetics may play an important role in the transport of
arsenate through porous media. In a study looking at the effect of competing anions on the adsorption of
arsenite and arsenate on ferrihydrite, the effect of phosphate on arsenate adsorption was greater at higher
pH than at low pH and the opposite trend was observed for arsenite. While sulfate did not change the
affinity of arsenate for ferrihydrite, sulfate reduced the adsorption of arsenite at pHs below 7.0 (Jain and
Loeppert 2000).
Smith et al. (1999) investigated the sorption properties of both As(V) and As(III) in 10 Australian soils of
widely different chemistry and mineralogy at commonly found arsenic levels. Adsorption of both
arsenate and arsenite was rapid (1 hour). The amount of As(V) sorbed varied widely (1.7–62.0 L/kg);
soils with lower amounts of oxidic material adsorbed much less arsenic than those with higher amounts of
these minerals. Arsenate sorption was highly correlated with the iron oxide content of the soil and this
ARSENIC 326
6. POTENTIAL FOR HUMAN EXPOSURE
factor probably accounts for much of the variation in soil adsorptivity. Considerable leaching of arsenic
occurred at a separate site where cattle were treated with a dip containing arsenic (cattle dip site) and that
contained similar soil properties to that studied by Smith et al. (1999). Arsenite adsorption, which was
investigated in four of the Australian soils, was sorbed to a lesser extent than was arsenate. This was
attributed to soil minerology and the species of As(V) (arsenate) and As(III) (arsenite) present in solution;
-
at pH 5–7, the dominant As(V) species are H
2
AsO
4
and HAsO
4
2-
and neutral H
3
AsO
3
is the dominant
As(III) species. For soils containing low amounts of oxidic minerals, pH had little effect on As(V)
sorption, while for oxidic soils, a decrease in sorption was evident as the pH increased. In contrast,
As(III) sorption increased with increasing pH (Smith et al. 1999). Jain et al. (1999) reported similar
results where arsenite were both found to bind strongly to iron oxides; however, the adsorption of
arsenate decreases with increasing pH, while the adsorption of arsenite increases with increasing pH (Jain
et al. 1999). As(III), which exists in a neutral form as arsenous acid, H
3
AsO
3
(pK
a
=9.23, 12.13, 13.4), is
less strongly adsorbed on mineral surfaces than the oxyanions of arsenic acid, H
3
AsO
4
, (pK
a
=2.22, 6.98,
11.53) (NRC 1999). Based on its pK
a
values, arsenic acid would exist as a mixture of arsenate anions,
H
2
AsO
4
-
and HAsO
4
2-
, under most environmental conditions (pH 5–9).
The practice of liming to remediate contaminated soils and mine tailings has the potential to mobilize
arsenic. Experiments performed by Jones et al. (1997) indicate that the increased mobility appears to be
consistent with the pH dependence of sorption reactions of arsenic on iron oxide minerals rather than
dissolution-precipitation reactions involving arsenic. They recommend that remediation of acidic mine
tailings or other arsenic-contaminated soils be carefully evaluated with respect to potential arsenic
mobilization, especially at contaminated sites hydraulically connected to surface or groundwaters.
Transport and partitioning of arsenic in water depends upon the chemical form (oxidation state and
counter ion) of the arsenic and on interactions with other materials present. Soluble forms move with the
water, and may be carried long distances through rivers (EPA 1979). However, arsenic may be adsorbed
from water onto sediments or soils, especially clays, iron oxides, aluminum hydroxides, manganese
compounds, and organic material (EPA 1979, 1982c; Welch et al. 1988). Under oxidizing and mildly
reducing conditions, groundwater arsenic concentrations are usually controlled by adsorption rather than
by mineral precipitation. The extent of arsenic adsorption under equilibrium conditions is characterized
by the distribution coefficient, K
d
, which measures the equilibrium partitioning ratio of adsorbed to
dissolved contaminant. The value of K
d
depends strongly upon the pH of the water, the arsenic oxidation
state, and the temperature. In acidic and neutral waters, As(V) is extensively adsorbed, while As(III) is
relatively weakly adsorbed. Trivalent inorganic arsenic exists predominantly as arsenous acid (H
3
AsO
3
)
ARSENIC 327
6. POTENTIAL FOR HUMAN EXPOSURE
at environmental pH and is not strongly adsorbed to suspended solids and sediments in the water column.
-
Pentavalent inorganic arsenic exists predominantly as H
2
AsO
4
and HAsO
4
2-
in most environmental
waters, which has considerably greater adsorption characteristics than arsenous acid. While in acidic and
neutral waters, As(V) is more strongly adsorbed relative to As(III), in high-pH waters (pH >9) aquifer K
d
values are considerably lower for both oxidation states (Mariner et al. 1996). Sediment-bound arsenic
may be released back into the water by chemical or biological interconversions of arsenic species (see
Section 6.3.2).
Arsenic enters rivers from where mining operations occurred and is transported downstream, moving
from water and sediment into biofilm (attached algae, bacterial, and associated fine detrital material), and
then into invertebrates and fish. The source of arsenic in the water column may be resuspended sediment.
While arsenic bioaccumulates in animals, it does not appear to biomagnify between tropic levels (Eisler
1994; Farag et al. 1998; Williams et al. 2006).
Most anthropogenic arsenic emitted to the atmosphere arises from high temperature processes (e.g., coal
and oil combustion, smelting operations, and refuse incineration) and occurs as fine particles with a mass
median diameter of about 1 μm (Coles et al. 1979; Pacyna 1987). These particles are transported by wind
and air currents until they are returned to earth by wet or dry deposition. Their residence time in the
atmosphere is about 7–9 days, in which time the particles may be transported thousands of kilometers
(EPA 1982b; Pacyna 1987). Long-range transport was evident in analyzing deposition of arsenic in
countries like Norway; there was no indication that the marine environment contributed significantly to
the deposition (Steinnes et al. 1992). Atmospheric fallout can be a significant source of arsenic in coastal
and inland waters near industrial areas. Scudlark et al. (1994) determined the average wet depositional
flux of arsenic as 49 μg As/m
2
/year for two sites in Chesapeake Bay, Maryland from June 1990 to
July 1991. They found a high degree of spatial and temporal variability. The elemental fluxes derived
predominantly from anthropogenic sources. Golomb et al. (1997) report average total (wet + dry)
deposition rates to Massachusetts Bay of 132 μg/m
2
/year, of which 21 μg/m
2
/year was wet deposition
during the period September 15, 1992–September 16, 1993. Hoff et al. (1996) estimated the following
arsenic loadings into the Great Lakes for 1994 (lake, wet deposition, dry deposition): Superior,
11,000 kg/year, 3,600 kg/year; Michigan, 5,000 kg/year, 1,800 kg/year; Erie, 5,500 kg/year,
1,800 kg/year; and Ontario, 3,000 kg/year, 580 kg/year. The measured dry deposition fluxes of arsenic at
four sampling sites around Lake Michigan ranged approximately from 0.01 to 1.5 μg As/m
2
/day;
estimated inputs of arsenic into Lake Michigan were reported to be 1.4x10
3
kg/year (Shahin et al. 2000).
ARSENIC 328
6. POTENTIAL FOR HUMAN EXPOSURE
Terrestrial plants may accumulate arsenic by root uptake from the soil or by absorption of airborne
arsenic deposited on the leaves, and certain species may accumulate substantial levels (EPA 1982b). Yet,
even when grown on highly polluted soil or soil naturally high in arsenic, the arsenic level taken up by the
plants is comparatively low (Gebel et al. 1998b; Pitten et al. 1999). Kale, lettuce, carrots, and potatoes
were grown in experimental plots surrounding a wood preservation factory in Denmark where waste
wood was incinerated to investigate the amount and pathways for arsenic uptake by plants (Larsen et al.
1992). On incineration, the arsenate in the wood preservative was partially converted to arsenite; the
arsenic emitted from the stack was primarily particle bound. Elevated levels of inorganic arsenic were
found in the test plants and in the soil around the factory. Statistical analyses revealed that the
dominating pathway for transport of arsenic from the factory to the leafy vegetables (kale) was by direct
atmospheric deposition, while arsenic in the root crops (potatoes and carrots) was a result of both soil
uptake and atmospheric deposition. Arsenic accumulation by plants is affected by arsenic speciation.
Uptake of four arsenic species (arsenite, arsenate, methylarsonic acid, and dimethylarsinic acid) by
turnips grown under soilless culture conditions showed that while uptake increased with increasing
arsenic concentration in the nutrient, the organic arsenicals showed higher upward translocation than the
inorganic arsenical (Carbonell-Barrachina et al. 1999). The total amount of arsenic taken up by the turnip
plants (roots and shoots) followed the trend methylarsenate (MMA)<dimethylarsinic acid (DMA)
<arsenite<arsenate. In a similar experiment, conducted with tomato plants, the total amount of arsenic
taken up by the tomato plants followed the trend DMA<MMA<arsenate≈arsenite, with arsenic
concentrations in the plants increasing with increasing arsenic concentration in the nutrient solution.
Arsenic was mainly accumulated in the root system (85%) with smaller amounts translocating to the fruit
(1%). However, plants treated with MMA and DMA had higher arsenic concentrations in the shoots and
fruit than those treated with arsenite or arsenate (Burlo et al. 1999). Terrestrial plants growing on land
bordering arsenic-contaminated waters show relatively little arsenic content, even though the sediments
have arsenic concentrations as high as 200 μg/g (Tamaki and Frankenberger 1992). Arsenic
concentrations in vegetables grown in uncontaminated soils and contaminated soils containing arsenic, as
well as other metals and organic contaminants, were generally <12 μg/kg wet weight. A maximum
arsenic concentration of 18 μg/kg wet weight was found in unpeeled carrots grown in soil, which
contained a mean arsenic concentration of 27 mg/kg dry weight (Samsøe-Petersen et al. 2002).
In a study by Rahman et al. (2004), CCA-treated lumber was used to construct raised garden beds to
determine how far the components of CCA migrated in the soil and the uptake of these components by
crops grown in the soil. Arsenic was found to diffuse laterally into the soil from the CCA-treated wood,
with the highest concentrations found at 0–2 cm from the treated wood and a steady decline in
ARSENIC 329
6. POTENTIAL FOR HUMAN EXPOSURE
concentration with increased distance. The highest average arsenic concentrations found in soil closest
(0–2 cm) to the CCA-treated wood were 56 and 46 μg/g in loamy sand and sandy loam soils, respectively.
At a distance of 30–35 cm from the CCA-treated wood, arsenic concentrations were approximately 7 μg/g
in both soils. All samples were of the top 0–15 cm of soil. Crops grown in both soil types within 0–2 cm
of the CCA-treated wood contained higher concentrations of arsenic, 0.186 and 10.894 μg/g for carrots
without peal and bean leaves and stems, respectively, than those grown at 1.5 m from the CCA-treated
wood, 0.006 and 0.682 μg/g for bean pods and bean leaves and stems, respectively. However, based on
FDA guidelines on tolerance limits, these crops would be considered approved for human consumption.
Studies by Chirenje et al. (2003a) also showed that elevated arsenic concentrations were found in surface
(0–5 cm) soils immediately surrounding, within the first 0.3 m, of utility poles, fences, and decks made
with CCA-treated wood. Factors such as the preservative formula, fixation temperature, post treatment
handling, and timber dimensions of CCA-treated wood, as well as the pH, salinity, and temperature of the
leaching media can affect the leach rates from CCA-treated wood (Hingston et al. 2001). Studies of
leaching of the components of CCA- and ACZA-treated wood used to construct a boardwalk in wetland
environments reported elevated arsenic levels in soil and sediment below and adjacent to these structures.
Generally, these levels decreased with increasing distance from the structure (Lebow et al. 2000).
Increased concentrations of arsenic were also observed under CCA-treated bridges. Arsenic levels
declined with distance from the bridge and were near background levels at 1.8–3 m from the bridge’s
perimeter (USDA/USDT 2000).
In a study by Lebow et al. (2003), the use of a water repellent finish on CCA-treated wood significantly
reduces the amount of arsenic, as well as copper and chromium, in the run-off water. It was also observed
the exposure to UV radiation caused a significant increase in leaching from both finished and unfinished
samples of CCA-treated wood. Small amounts of arsenic can be transferred from CCA-treated wood to
skin from touching CCA-treated wood surfaces (Hemond and Solo-Gabriele 2004; Kwon et al. 2004;
Shalat et al. 2006; Ursitti et al. 2004; Wang et al. 2005).
Breslin and Adler-Ivanbrook (1998) examined the leaching of the copper, chromium, and arsenic from
CCA-treated wood in laboratory studies using samples of treated southern yellow pine in solutions
simulating estuarine waters. The tank leaching solutions were frequently sampled and replaced to
approximate field conditions. Initial 12-hour fluxes ranging from 0.2x10
-10
to 5.2x10
-10
mol/mm
2
d was
reported for arsenic. After 90 days, arsenic fluxes decreased to 0.5x10
-11
–3.1x10
-11
mol/mm
2
d. A study
by Cooper (1991) demonstrated that the buffer system used in leaching studies of components from CCA-
treated wood can significantly change the amount arsenic released from treated wood. Samples of four
ARSENIC 330
6. POTENTIAL FOR HUMAN EXPOSURE
species of CCA-treated wood were exposed to four acidic leaching solutions. In the samples exposed to
water adjusted to pHs of 3.5, 4.5, and 5.5, losses of arsenic after 13 days were generally <7%. However,
when a leaching solution of sodium hydroxide and citric acid buffer (pH 5.5) was used, the percent of
arsenic leached ranged from 27.4 to 46.7% (Cooper 1991).
Arsenic bioaccumulation depends on various factors, such as environmental setting (marine, estuarine,
freshwater), organism type (fish, invertebrate), trophic status within the aquatic food chain, exposure
concentrations, and route of uptake (Williams et al. 2006). Bioaccumulation refers to the net
accumulation of a chemical by aquatic organisms as a result of uptake from all environmental sources,
such as water, food, and sediment, whereas bioconcentration refers to the uptake of a chemical by an
aquatic organism through water (EPA 2003b). Biomagnification in aquatic food chains does not appear
to be significant (EPA 1979, 1982b, 1983e, 2003b; Mason et al. 2000; Williams et al. 2006).
Bioconcentration of arsenic occurs in aquatic organisms, primarily in algae and lower invertebrates. Both
bottom-feeding and predatory fish can accumulate contaminants found in water. Bottom-feeders are
readily exposed to the greater quantities of metals, including the metalloid arsenic, which accumulate in
sediments. Predators may bioaccumulate metals from the surrounding water or from feeding on other
fish, including bottom-feeders, which can result in the biomagnification of the metals in their tissues. An
extensive study of the factors affecting bioaccumulation of arsenic in two streams in western Maryland in
1997–1998 found no evidence of biomagnification since arsenic concentrations in organisms tend to
decrease with increasing tropic level (Mason et al. 2000). Arsenic is mainly accumulated in the
exoskeleton of invertebrates and in the livers of fish. No differences were found in the arsenic levels in
different species of fish, which included herbivorous, insectivorous, and carnivorous species. The major
bioaccumulation transfer is between water and algae, at the base of the food chain and this has a strong
impact on the concentration in fish. National Contaminant Biomonitoring data produced by the Fish and
Wildlife Service were used to test whether differences exist between bottom-feeders and predators in
tissue levels of metals and other contaminants. No differences were found for arsenic (Kidwell et al.
1995). The bioconcentration factors (BCFs) of bryophytes, invertebrates, and fish (livers) in Swedish
lakes and brooks impacted by smelter emissions were 8,700, 1,900–2,200, and 200–800, respectively
(Lithner et al. 1995). EPA (2003b) assessed a large dataset of bioaccumulation data for various fish and
invertebrate species. BCF values in this dataset ranged from 0.048 to 1,390.
Williams et al. (2006) reviewed 12 studies of arsenic bioaccumulation in freshwater fish, and proposed
that BCF and bioaccumulation factor (BAF) values are not constant across arsenic concentrations in
ARSENIC 331
6. POTENTIAL FOR HUMAN EXPOSURE
water. BCF or BAF values from these 12 studies ranged from 0.1 to 3,091. Williams et al. (2006) found
that BCF and BAF values appear to be the highest within the range of ambient arsenic concentrations, and
decline steeply to relatively low levels as the arsenic concentrations in water increase. Based on this
analysis, arsenic concentrations in tissue and BAF values may be a power function of arsenic
concentrations in water. EPA (2007b) also reported that for many nonessential metals, including arsenic,
accumulation is nonlinear with respect to exposure concentration.
6.3.2 Transformation and Degradation
6.3.2.1 Air
Arsenic is released into the atmosphere primarily as arsenic trioxide or, less frequently, in one of several
volatile organic compounds, mainly arsines (EPA 1982b). Trivalent arsenic and methyl arsines in the
atmosphere undergo oxidation to the pentavalent state (EPA 1984a), and arsenic in the atmosphere is
usually a mixture of the trivalent and pentavalent forms (EPA 1984a; Scudlark and Church 1988).
Photolysis is not considered an important fate process for arsenic compounds (EPA 1979).
6.3.2.2 Water
Arsenic in water can undergo a complex series of transformations, including oxidation-reduction
reactions, ligand exchange, precipitation, and biotransformation (EPA 1979, 1984a; Sanders et al. 1994;
Welch et al. 1988). Rate constants for these various reactions are not readily available, but the factors
most strongly influencing fate processes in water include Eh, pH, metal sulfide and sulfide ion
concentrations, iron concentrations, temperature, salinity, distribution and composition of the biota,
season, and the nature and concentration of natural organic matter (EPA 1979; Farago 1997; Redman et
al. 2002; Wakao et al. 1988). Organic arsenical pesticides, such as MSMA, DSMA, and DMA do not
degrade by hydrolysis or by aquatic photolysis (EPA 2006). No formation of arsine gas from marine
environments has been reported (Tamaki and Frankenberger 1992).
Inorganic species of arsenic are predominant in the aquatic environment. In the pH range of natural
-
waters, the predominant aqueous inorganic As(V) species are the arsenate ions, H
2
AsO
4
and HAsO
4
2-
; the
predominant inorganic As(III) species is As(OH)
3
(Aurillo et al. 1994; EPA 1982c). As(V) generally
dominates in oxidizing environments such as surface water and As(III) dominates under reducing
conditions such as may occur in groundwater containing high levels of arsenic. However, the reduction
of arsenate to arsenite is slow, so arsenate can be found in reducing environments. Conversely, the
oxidation of arsenite in oxidizing environments is moderately slow (half-life, 0.4–7 days in coastal
ARSENIC 332
6. POTENTIAL FOR HUMAN EXPOSURE
systems) and therefore, arsenite can be found in oxidizing environments (Mariner et al. 1996; Sanders et
al. 1994). The main organic species in fresh water are MMA and DMA; however, these species are
usually present at lower concentrations than inorganic arsenic species (Eisler 1994). (The toxicities of
MMA and DMA are discussed in Chapter 3.) Aquatic microorganisms may reduce the arsenate to
arsenite, as well as methylate arsenate to its mono- or dimethylated forms (Aurillo et al. 1994; Benson
1989; Braman and Foreback 1973; Edmonds and Francesconi 1987; Sanders et al. 1994). Methylated
species are also produced by the biogenic reduction of more complex organoarsenic compounds like
arsenocholine or arsenobetaine. Water samples from a number of lakes and estuaries, mostly in
California, show measurable concentrations of methylated arsenic (equivalent to 1–59% of total arsenic)
(Anderson and Bruland 1991). Within the oxic photic zone, arsenate and DMA were the dominant
species. A seasonal study of one lake demonstrated that DMA was the dominant form of arsenic in
surface waters during late summer and fall. Methylated species declined and arsenate species increased
when the lake turned over in late fall. Mono Lake, a highly alkaline body of water, and four rivers did not
have measurable concentrations of methylated arsenic. It was hypothesized that the reason why
methylated forms were not detected in Mono Lake was that the extremely high inorganic arsenic
concentrations in the lake, 230 μM (17 mg/L), could overwhelm the analysis of small amounts of organic
forms. Other possibilities are that the high alkalinity or very high phosphate levels in the water, 260 μM
(25 mg/L), are not conducive to biogenic methylation (Anderson and Bruland 1991). Both reduction and
methylation of As(V) may lead to increased mobilization of arsenic, since As(III), dimethylarsinates, and
monomethylarsonates are much less particle-reactive than As(V) (Aurillo et al. 1994). In the estuarial
Patuxet River, Maryland, arsenate concentrations peaked during the summer, at 1.0 μg/L in 1988–
1989 (Sanders et al. 1994). In contrast, winter to spring levels were around 0.1 μg/L. Arsenite
concentrations were irregularly present at low levels during the year. Peaks of DMA occurred at various
times, particularly in the winter and late spring and appeared to be linked with algal blooms. The DMA
peak declined over several months that was followed by a rise in MMA. The MMA was thought to be
occurring as a degradation product of DMA. A similar seasonal pattern of arsenic speciation was
observed in Chesapeake Bay. Arsenite methylation took place during the warmer months leading to
changes down the main stem of the bay; arsensite production dominated the upper reaches of the bay and
methylated species dominated the more saline lower reaches. In coastal waters, reduced and methylated
species are present in lower concentrations, around 10–20% of total arsenic (Sanders et al. 1994). In
groundwater, arsenic generally exists as the oxyanion of arsenate (H
x
AsO
4
3-x
) or arsenite (H
x
AsO
3
3-x
), or
both; however, the distribution between arsenite and arsenate is not always predictable based on
oxidation-reduction potential (Robertson 1989; Welch et al. 1988).
ARSENIC 333
6. POTENTIAL FOR HUMAN EXPOSURE
6.3.2.3 Sediment and Soil
In soil, arsenic is found as a complex mixture of mineral phases, such as co-precipitated and sorbed
species, as well as dissolved species (Roberts et al. 2007). The degree of arsenic solubility in soil will
depend on the amount of arsenic distributed between these different mineral phases. The dissolution of
arsenic is also affected by particle size. The distribution between these phases may reflect the arsenic
source (e.g., pesticide application, wood treatment, tanning, or mining operations), and may change with
weathering and associations with iron and manganese oxides and phosphate minerals in the soil (Roberts
et al. 2007; Ruby et al. 1999). Davis et al. (1996) reported that in soil in Anaconda, Montana, a smelting
site from 1860 to 1980, contained arsenic that is only in a sparingly soluble form, consisting of primarily
arsenic oxides and phosphates.
The arsenic cycle in soils is complex, with many biotic and abiotic processes controlling its overall fate
and environmental impact. Arsenic in soil exists in various oxidation states and chemical species,
depending upon soil pH and oxidation-reduction potential. Under most environmental conditions,
-
inorganic As(V) will exist as a mixture of arsenate anions, H
2
AsO
4
and HAsO
4
2-
, and inorganic As(III)
will exist as H
3
AsO
3
. The arsenate and arsenite oxyanions have various degrees of protonation depending
upon pH (EPA 1982b; McGeehan 1996). As(V) predominates in aerobic soils, and As(III) predominates
in slightly reduced soils (e.g., temporarily flooded) or sediments (EPA 1982b; Sanders et al. 1994).
As(III) commonly partitions to the aqueous phase in anoxic environments, and would be more mobile.
As(V) usually remains bound to minerals, such as ferrihydrite and alumina, limiting its mobility and
bioavailability (Rhine et al. 2006).
Arsenite is moderately unstable in the presence of oxygen; however, it can be found under aerobic
conditions as well (Sanders et al. 1994). While arsenate is strongly sorbed by soils under aerobic
conditions, it is rapidly desorbed as the system becomes anaerobic. Once it is desorbed, arsenate can be
reduced to arsenite, which exhibits greater mobility in soils (McGeehan 1996). Transformations between
the various oxidation states and species of arsenic occur as a result of biotic or abiotic processes
(Bhumbla and Keefer 1994). While degradation of an organic compound is typically considered
complete mineralization, in the case of organic arsenic compounds, the element arsenic itself cannot be
degraded. However, the organic portion of the molecule can be metabolized (Woolson 1976).
Arsenicals applied to soils may be methylated by microorganisms to arsines, which are lost through
volatilization, and organic forms may be mineralized to inorganic forms. Gao and Burau (1997) reported
ARSENIC 334
6. POTENTIAL FOR HUMAN EXPOSURE
that the overall percentage of DMA and MMA minerialized after 70 days ranged from 3 to 87% in air-dry
soil and a soil near saturation, respectively. The rate of demethylation of DMA increased with soil
moisture. Over the same 70-day period, arsenic losses as volatile arsines were much lower than
minerialization, ranging from 0.001 to 0.4%. Arsine evolution rates followed the order:
DMA>MMA>arsenite=arsenate (Gao and Burau 1997). Woolson and Kearney (1973) reported that
14
C-labeled DMA degraded differently in soils under aerobic and anaerobic conditions. Under anaerobic
conditions, 61% of the applied DMA was converted to a volatile alkyl arsine after 24 weeks, and lost
from the soil system. Under aerobic conditions, 35% was converted to a volatile organo-arsenic
compound, possibly dimethyl arsine, and 41% was converted to
14
CO
2
and arsenate after 24 weeks.
Similar to microorganisms in soils, Reimer (1989) reported that microorganisms found in natural marine
sediments and sediments contaminated with mine-tailings are also capable of methylating arsenic under
aerobic and anaerobic conditions. Von Endt et al. (1968) reported that the degradation of
14
C-labelled
monosodium methanearsonate (MSMA) was found to range from 1.7 to 10% in Dundee silty clay loam
soil and Sharkey clay soil after 60 days, respectively. MSMA decomposition to CO
2
was a slow process
without a lag period. Sterilized soils were found to produce essentially no
14
CO
2
(0.7%) after 60 days,
indicating that soil bacteria contributed to the decomposition of MSMA (Von Endt et al. 1968). Akkari et
al. (1986) studied the degradation of MSMA in various soils. At 20% water content, half-lives of 144, 88,
and 178 days were reported in Sharkey clay, Taloka silt loam, and Steele-Crevasse sand loam,
respectively. The Sharkey soil with the highest clay content was expected to have the greatest adsorptive
capacity for both water and MSMA, reducing the amount of MSMA available in the soil solution to
microorganisms that degrade the MSMA. The half-lives were 25, 41, and 178 days under anaerobic
(flooded) conditions in Sharkey clay, Taloka silt loam, and Steele-Crevasse sand loam, respectively.
Under flooded conditions, MSMA degradation occurs by reductive methylation to form arsinite and
alkylarsine gas. The authors attributed the longer half-lives for MSMA degradation in the Steele-
Crevasse sand loam soil to its low organic matter content, which may have supported fewer microbial
populations needed for oxidation demethylation under aerobic conditions. Under flooded conditions,
anaerobiosis is expected to be slowest in low organic matter sandy loam soils (Akkari et al. 1986).
Organic arsenical pesticides, such as MSMA, DSMA, and DMA, do not degrade by hydrolysis or by soil
photolysis (EPA 2006).
Roxarsone (3-nitro-4-hydroxyphenylarsonic acid) used in poultry feed is found excreted unchanged in
poultry litter (bedding and manure). Roxarsone found in poultry litter, which is used to amend
agricultural soil, was found to degrade to arsenate in approximately 3–4 weeks upon composting
ARSENIC 335
6. POTENTIAL FOR HUMAN EXPOSURE
(Garbarino et al. 2003). In addition, the arsenic in poultry litter was found to be easily mobilized by
water; however, its leach rate from amended soils was slow enough that it accumulated in soils
(Rutherford et al. 2003).
A sequential fractionation scheme was used to assess the chemical nature, and thus the potential
bioavailability, of arsenic at cattle dip sites in Australia where sodium arsenite was used extensively in
cattle dips from the turn of the century until the early 1950s (McLaren et al. 1998). Most sites contained
substantial amounts, 13% on the average, of arsenic in the two most labile fractions indicating a high
potential for bioaccessibility and leaching. The bulk of the arsenic appeared to be associated with
amorphous iron and aluminum minerals in soil. Similarly, arsenic in soil and mine waste in the Tamar
Valley in England was found to be concentrated in a fraction associated with iron and organic-iron
(Kavanagh et al. 1997). Laboratory studies were performed to assess the phase partitioning of trace
metals, including the metalloid arsenic, to sediment from the Coeur d’Alene River, a mining area of
Idaho, and the release of metals under simulated minor and major flooding events (LaForce et al. 1998).
Arsenic was primarily associated with the iron and manganese oxides as seen by its large release when
these oxides were reduced. Arsenic levels were comparatively low in the organic fraction and remaining
residual fraction and negligible in the extractible fractions.
6.3.2.4 Other Media
Carbonell-Barrachina et al. (2000) found the speciation and solubility of arsenic in sewage sludge
suspensions to be affected by pH and Eh. Under oxidizing conditions, the solubility of arsenic was low,
with a major portion of the soluble arsenic present as organic arsenic compounds, mainly dimethylarsinic
acid (approximately 74% of the total arsenic in solution). Under moderately reducing conditions (0–
100 mV), inorganic arsenic accounted for the majority (90%) of the total arsenic in solution, and the
solubility of arsenic was increased due to dissolution of iron oxyhydroxides. Under strongly reducing
conditions (-250 mV), arsenic solubility was decreased by the formation of insoluble sulfides. The pH of
the solution was also found to influence the speciation and solubility of arsenic. At neutral pH, the
solubility of arsenic was at its maximum, and decreased under acidic or alkaline conditions. Inorganic
arsenic species were the dominant species at pH 5.0; at pH 6.5, the major soluble forms were organic
arsenic species. The biomethylation of arsenic was limited at acidic pH, and was at its maximum at near
neutral pH (Carbonell-Barrachina et al. 2000).
ARSENIC 336
6. POTENTIAL FOR HUMAN EXPOSURE
6.4 LEVELS MONITORED OR ESTIMATED IN THE ENVIRONMENT
Reliable evaluation of the potential for human exposure to arsenic depends in part on the reliability of
supporting analytical data from environmental samples and biological specimens. Concentrations of
arsenic in unpolluted atmospheres and in pristine surface waters are often so low as to be near the limits
of current analytical methods. In reviewing data on arsenic levels monitored or estimated in the
environment, it should also be noted that the amount of chemical identified analytically is not necessarily
equivalent to the amount that is bioavailable. The analytical methods available for monitoring arsenic in
a variety of environmental media are detailed in Chapter 7.
6.4.1 Air
Arsenic in ambient air is usually a mixture of particulate arsenite and arsenate; organic species are of
negligible importance except in areas of substantial methylated arsenic pesticide application or biotic
activity (EPA 1984a). Mean levels in ambient air in the United States have been reported to range from
<1 to 3 ng/m
3
in remote areas and from 20 to 30 ng/m
3
in urban areas (Davidson et al. 1985; EPA 1982c;
IARC 1980; NAS 1977a). EPA conducted a modeling study with the Assessment System for Population
Exposure Nationwide (ASPEN) in which estimates of emissions of hazardous air pollutants were used to
estimate ambient concentrations (Rosenbaum et al. 1999). Using 1990 data to estimate total emissions of
arsenic in the conterminous 48 states, excluding road dust or windblown dust from construction or
agricultural tilling, the 25th percentile, median, and 75th percentile arsenic concentration were estimated
to be 9, 20, and 30 ng/m
3
, respectively. Maps illustrating the amount of toxic air pollutant emissions,
including arsenic compounds, by county in 1996 for the 48 coterminous states of the United States as well
as Puerto Rico and the Virgin Islands are available on the internet at http://www.epa.gov/ttn/atw/-
nata/mapemis.html, as of March 2005. Schroeder et al. (1987) listed ranges of arsenic concentrations in
air of 0.007–1.9, 1.0–28, and 2–2,320 ng/m
3
in remote, rural, and urban areas, respectively. The average
annual arsenic concentration in air at Nahant, Massachusetts, just north of Boston, between September
1992 and September 1993, was 1.2 ng/m
3
; 75% of the arsenic was associated with fine (<2.5 μm)
particles. The long-term means of the ambient concentrations of arsenic measured in urban areas of the
Great Lakes region from 1982 to 1993 ranged from 4.2 to 9.6 ng/m
3
(Pirrone and Keeler 1996). Large
cities generally have higher arsenic air concentrations than smaller ones due to emissions from coal-fired
power plants (IARC 1980), but maximum 24-hour concentrations generally are <100 ng/m
3
(EPA 1984a).
In the spring of 1990, aerosols and cloud water that were sampled by aircraft at an altitude of 1.2–3 km
above the Midwestern United States had a mean mixed layer arsenic concentration of
1.6±0.9 ng/m
3
(Burkhard et al. 1994). A mean arsenic concentration of 1.0±0.5 ng/m
3
was reported at
ARSENIC 337
6. POTENTIAL FOR HUMAN EXPOSURE
Mayville, New York, a site 400 km to the northwest of the sampling area and directly downwind on most
days.
Arsenic was monitored at an application site in the San Joaquin Valley, California and at four sites in
nearby communities in 1987 where sodium arsenite was used as a fungicide on tokay grapes (Baker et al.
1996). The maximum arsenic concentration measured 15–20 meters from the edge of the field was
260 ng/m
3
. The maximum arsenic concentration at four community sites in the area was 76 ng/m
3
. The
concentration at an urban background site was 3 ng/m
3
(Baker et al. 1996). Sodium arsenite is no longer
registered in California (Baker et al. 1996). The highest historic arsenic levels detected in the atmosphere
were near nonferrous metal smelters, with reported concentrations up to 2,500 ng/m
3
(IARC 1980; NAS
1977a; Schroeder et al. 1987).
Arsenic air concentrations measured in several indoor public places (e.g., cafeteria, coffee house, music
club, Amtrak train, and several restaurants) with environmental tobacco smoke (ETS) ranged from <0.1 to
1 ng/m
3
, with a mean of 0.4±0.3 ng/m
3
. Sites that were ETS-free (university office and library) had
arsenic concentrations <0.13 ng/m
3
(Landsberger and Wu 1995). The Toxic Exposure Assessment at
Columbia/Harvard (TEACH) study measured levels of various toxics in New York City air in 1999.
Exposures were assessed in a group of 46 high school students in West Central Harlem. Mean arsenic
concentrations in summer home outdoor, home indoor, and personal air of the participants were 0.37,
0.40, and 0.45 ng/m
3
, respectively (Kinney et al. 2002). Detected arsenic concentrations in indoor and
outdoor air collected as part of the National Human Exposure Assessment Survey (NHEXAS) in Arizona
ranged from 3.4 to 22.3 and from 3.5 to 25.7 ng/m
3
, respectively, with 71 and 68% below the detection
limit (1.8–14.3 ng/m
3
) (O'Rourke et al. 1999).
6.4.2 Water
Arsenic is widely distributed in surface water, groundwater, and finished drinking water in the United
States. A survey of 293 stations in two nationwide sampling networks on major U.S. rivers found median
arsenic levels to be 1 μg/L; the 75th percentile level was 3 μg/L (Smith et al. 1987). Arsenic was detected
in 1,298 of 3,452 surface water samples recorded in the STORET database for 2004 at concentrations
ranging from 0.138 to 1,700 μg/L in samples where arsenic was detected (EPA 2005c). Two streams in
western Maryland that were the focus of a major bioaccumulation study in 1997–1998 had arsenic
concentrations of 0.370±0.200 and 0.670±0.460 μg/L (Mason et al. 2000). Surface water will be
impacted by runoff from polluted sites. An average arsenic concentration of 5.12 μg/L was reported in
ARSENIC 338
6. POTENTIAL FOR HUMAN EXPOSURE
water from Moon Lake, a Mississippi River alluvial floodplain in northwest Mississippi. Intensive
cultivation has occurred in this area, including cotton, soybeans, and rice (Cooper and Gillespie 2001).
Hard-rock mining activities occurred in the southern part of Colorado and New Mexico north of Taos
since the latter part of the 19th century until recently, which have impacted the Rio Grande and its
tributaries. A mean arsenic concentration of approximately 0.8 μg/L was reported for the main stem of
the Rio Grande sampled in June and September 1994. Arsenic concentrations in the Alamosa River,
Colorado were 0.11 and 0.14 μg/L in June and September 1994, respectively, and 1.4 μg/L in Big Arsenic
Spring, New Mexico in September 1994 (Taylor et al. 2001). Arsenic concentrations in water from
watersheds in Black Hills, South Dakota, an area impacted by gold mining activities ranged from 2.5 to
55 μg/L and from 1.7 to 51 μg/L in unfiltered and filtered samples, respectively; concentrations from
reference areas ranged from 1.1 to 3.4 μg/L and from 0.9 to1.9 μg/L in unfiltered and filtered samples,
respectively (May et al. 2001). Arsenic concentrations ranged from 0.29 to 34.0 μg/L in water samples
from Wakulla River and St. Joseph Bay North, along the Florida Panhandle; arsenic contamination in this
area is likely to result from nonpoint source pollution (Philp et al. 2003).
Data on total arsenic in surface water from a number of seas and oceans show levels of <1 μg/L, except in
the Antarctic Ocean and Southwest Pacific Oceans where the levels are 1.1 and 1.2 μg/L, respectively.
Levels in coastal waters and estuaries are generally somewhat higher, in the range of 1–3 μg/L. However,
estuarine water in Salinas, California had arsenic levels of 7.42 μg/L (Francesconi et al. 1994). The
dissolved arsenic concentration in water at 40 sites in the Indian River Lagoon System in Florida ranged
from 0.35 to 1.6 μg/L with a mean of 0.89±0.34 μg/L (Trocine and Trefry 1996). Thermal waters
generally have arsenic levels of 20–3,800 μg/L, although levels as high as 276,000 μg/L have been
recorded (Eisler 1994).
Arsenic levels in groundwater average about 1–2 μg/L, except in some western states with volcanic rock
and sulfidic mineral deposits high in arsenic, where arsenic levels up to 3,400 μg/L have been observed
(IARC 1980; Page 1981; Robertson 1989; Welch et al. 1988). In western mining areas, groundwater
arsenic concentrations up to 48,000 μg/L have been reported (Welch et al. 1988). Arsenic concentrations
in groundwater samples collected form 73 wells in 10 counties in southeast Michigan in 1997 ranged
from 0.5 to 278 μg/L, with an average of 29 μg/L. Most (53–98%) of the arsenic was detected as arsenite
(Kim et al. 2002). The U.S. Geological Survey mapped concentrations of arsenic in approximately
31,350 groundwater samples collected between 1973 and 2001; the counties in which at least 25% of
wells exceed various levels are shown in Figure 6-2 (USGS 2007a). Most arsenic in natural waters is a
ARSENIC 339
6. POTENTIAL FOR HUMAN EXPOSURE
Figure 6-2. Counties in Which at Least 25% of Wells Exceed Different Arsenic
Levels
Source: USGS 2007a
ARSENIC 340
6. POTENTIAL FOR HUMAN EXPOSURE
mixture of arsenate and arsenite, with arsenate usually predominating (Braman and Foreback 1973; EPA
1982c, 1984a). Methylated forms have also been detected in both surface water and groundwater, at
levels ranging from 0.01 to 7.4 μg/L (Braman and Foreback 1973; Hood 1985), with most values below
0.3 μg/L (Hood 1985). In a survey of shallow groundwater quality in the alluvial aquifer beneath a major
urban center, Denver, Colorado, arsenic levels in the 30 randomly-chosen wells sampled had median
levels of <1 μg/L; the maximum level was 33 μg/L (Bruce and McMahon 1996). Arsenic levels in
groundwater sometimes exceeded the EPA maximum contaminant level (MCL), which was 50 μg/L at
the time, in the Willamette Valley, Oregon and a nine-county region of southeastern Michigan (USGS
1999b, 1999c).
Arsenic has also been detected in rainwater at average concentrations of 0.2–0.5 μg/L (Welch et al. 1988).
This range is consistent with that found in a 1997–1998 study in western Maryland, which was the focus
of a major bioaccumulation study (Mason et al. 2000). Arsenic levels in wet deposition in the watershed
as well as throughfall into the two streams were 0.345±0.392, 0.400±0.400, and 0.330±0.250 μg/L,
respectively. Median arsenic concentrations in 30-day rainwater composite samples collected
May-September 1994 from eight arctic catchments in northern Europe at varying distances and wind
directions from the emissions of a Russian nickel ore mining, roasting, and smelting industry on the Kola
Peninsula ranged from 0.07 to 12.3 μg/L (Reimann et al. 1997). Rain and snow samples were collected
during the fall of 1996 and winter of 1997 at eight locations in a semi-circular pattern radiating out (2–
15 km) in the direction of the prevailing wind from the Claremont incinerator located in New Hampshire.
This incinerator processes 200 tons of solid waste per day. Arsenic concentrations in rainwater and snow
ranged from 0.020 to 0.079 μg/L and from 0.80 to 1.28 μg/L, respectively (Feng et al. 2000).
Drinking water is one of the most important sources of arsenic exposure. Surveys of drinking water in the
United States have found that >99% of public water supplies have arsenic concentrations below the EPA
MCL, which was 50 μg/L at the time (EPA 1984a). In an EPA study of tap water from 3,834 U.S.
residences, the average value was 2.4 μg/L (EPA 1982c).
Before the MCL for arsenic in drinking water was lowered from 50 to 10 μg/L, studies were undertaken
to ascertain how different standards would affect compliance. One such survey sponsored by the Water
Industry Technical Action Fund was the National Arsenic Occurrence Survey (NAOS). NAOS was based
on a representational survey of public water systems defined by source type, system size, and
geographical location. Additionally, it included a natural occurrence factor, a stratifying variable that
could qualitatively describe the likelihood of arsenic occurrence in the supply. To predict finished water
ARSENIC 341
6. POTENTIAL FOR HUMAN EXPOSURE
arsenic concentrations, data on the water treatment options, efficiency, and frequency of use were
factored in. The results of the NAOS are presented in Table 6-3. The NAOS results are in general
agreement with two older and more limited national surveys, EPA’s National Inorganics and
Radionuclides Survey (NIRS) and the Metropolitan Water District of Southern California Survey
(MWDSC). The percentages of water systems that would be out of compliance are estimated to be 1.7,
3.6, 9.3, and 20.7% for arsenic MCLs of 20, 10, 5, and 2 μg/L, respectively. Arsenic concentrations were
determined in drinking in EPA Region V (Indiana, Illinois, Michigan, Minnesota, Ohio, and Wisconsin)
as part of the NHEXAS; mean arsenic concentration in flushed and standing tap water were both 1.1 μg/L
(Thomas et al. 1999). A review by Frost et al. (2003) of existing data from the EPA Arsenic Occurrence
and Exposure Database, as well as additional data from state health and environmental departments and
water utilities found that 33 counties in 11 states had estimated mean drinking water arsenic
concentrations of 10 μg/L or greater. Eleven counties had mean arsenic concentrations of ≥20 μg/L, and
two counties had mean arsenic concentrations of ≥50 μg/L (Frost et al. 2003).
The north central region and the western region of the United States have the highest arsenic levels in
surface water and groundwater sources, respectively. In a study of drinking water from New Hampshire,
arsenic concentrations ranged from <0.01 to 180 μg/L in the 793 households tested. More than 10% of
the private wells had arsenic concentrations >10 μg/L, and 2.5% had levels >50 μg/L (Karagas et al.
1998). In New Hampshire, 992 randomly selected household water samples were analyzed for arsenic
levels and the results for domestic well users were compared with those for users of municipal water
supplies (Peters et al. 1999). The concentrations ranged from <0.0003 to 180 μg/L, with water from
domestic wells containing significantly more arsenic than water from municipal supplies; the median
concentration of the former was about 0.5 μg/L and the latter was 0.2 μg/L. None of the municipal
supplies exceeded an arsenic concentration of 50 μg/L, and 2% of the domestic wells were found to have
arsenic concentrations that exceeded 50 μg/L. Approximately 2% of the municipal water users have
water with arsenic levels exceeding 10 μg/L compared with 13% of domestic wells. Twenty-five percent
of domestic wells and 5% of municipal supplies were found to have arsenic concentrations exceeding
2 μg/L. The highest arsenic levels in New Hampshire are associated with bedrock wells in the south
eastern and south central part of the state (Peters et al. 1999). In a study of arsenic in well water supplies
in Saskatchewan, Canada, 13% of samples were >20 μg/L and one sample exceeded 100 μg/L (Thompson
et al. 1999). It was noted that the samples with high arsenic levels were derived from sites that were in
near proximity to each other, indicating the presence of ‘hot spots’ with similar geological characteristics.
As part of an epidemiological study, Engel and Smith (1994) investigated the levels of arsenic in drinking
ARSENIC 342
6. POTENTIAL FOR HUMAN EXPOSURE
Table 6-3. Regional Occurrence of Arsenic in U.S. Water Sources and Finished
Drinking Water
Arsenic concentration in μg/L
Geographical region <1 1–5 5–20 >20
Occurrence in U.S. surface water sources
Region 1. New England 50 50 0 0
Region 2. Mid-Atlantic 84 12 4 0
Region 3. South East 93 7 0 0
Region 4. Midwest 24 76 0 0
Region 5. South Central 32 55 13 0
Region 6. North Central 33 22 33 0
Region 7. Western 42 58 0 0
Occurrence in U.S. groundwater sources
Region 1. New England 71 21 7 0
Region 2. Mid-Atlantic 81 4 11 4
Region 3. South East 82 14 2 0
Region 4. Midwest 40 40 15 5
Region 5. South Central 68 27 15 0
Region 6. North Central 30 40 30 0
Region 7. Western 24 34 28 14
Occurrence in U.S. finished surface water supplies
Region 1. New England 88 12 0 0
Region 2. Mid-Atlantic 92 8 0 0
Region 3. South East 100 0 0 0
Region 4. Midwest 73 27 0 0
Region 5. South Central 74 19 7 0
Region 6. North Central 44 44 0 12
Region 7. Western 42 58 0 0
Occurrence in U.S. finished groundwater supplies
Region 1. New England 79 21 0 0
Region 2. Mid-Atlantic 81 4 11 4
Region 3. South East 94 4 2 0
Region 4. Midwest 58 27 12 3
Region 5. South Central 61 27 12 0
Region 6. North Central 40 50 10 0
Region 7. Western 20 40 22 12
Source: National Arsenic Occurrence Survey (Frey and Edwards 1997)
ARSENIC 343
6. POTENTIAL FOR HUMAN EXPOSURE
water throughout the United States between 1968 and 1984. They found that 30 counties in 11 states had
mean arsenic levels of >5 μg/L, with a range of 5.4–91.5 μg/L; 15 counties had mean levels from 5 to
10 μg/L; 10 counties had mean levels from 10 to 20 μg/L; and 5 counties had levels >20 μg/L. The
highest levels were found in Churchill County, Nevada, where 89% of the population was exposed to a
mean arsenic concentration of 100 μg/L and 11% to a mean of 27 μg/L. A study by Frost et al. (2003)
identified 33 counties from 11 states in which the average arsenic concentration of at least 75% of public
wells was >10 μg/L. Arsenic concentrations in drinking water from these counties ranged from 10.3 to
90.0 μg/L in Pinal, Arizona and Churchill, Nevada, respectively (Frost et al. 2003).
Many communities have high levels of arsenic in their drinking water because of contamination or as a
result of the geology of the area. In Millard County, Utah, seven towns had median and maximum
arsenic levels of 18.1–190.7 and 125–620 μg/L, respectively, in their drinking water (Lewis et al. 1999).
The mean arsenic concentration in tap water from homes in Ajo, Arizona, about 2 miles from an open pit
copper mine and smelter was 90 μg/L (Morse et al. 1979). The town’s water was supplied from five deep
wells.
Countries such as Mexico, Bangladesh, India, Chile, Argentina, and Vietnam have highly elevated levels
of arsenic in drinking water in some regions (Bagla and Kaiser 1996; Berg et al. 2001; Tondel et al. 1999;
WHO 2001; Wyatt et al. 1998a, 1998b). In Bangladesh and West Bengal, the soil naturally contains high
levels of arsenic, which leaches into the shallow groundwater that is tapped for drinking water. In West
Bengal, India, it is estimated that more than one million Indians are drinking arsenic-laced water and tens
of millions more could be at risk in areas that have not been tested for contamination. Analysis of
20,000 tube-well waters revealed that 62% have arsenic at levels above the World Health Organization
(WHO) permissible exposure limit (PEL) in drinking water of 10 μg/L, with some as high as 3,700 μg/L
(Bagla and Kaiser 1996). Analysis of 10,991 and 58,166 groundwater samples from 42 and 9 arsenic-
affected districts in Bangladesh and West Bengal were found to have arsenic levels that were 59 and 34%,
respectively, above 50 μg/L (Chowdhury et al. 2000). Berg et al. (2001), studied the arsenic
contamination of the Red River alluvial tract in Hanoi, Vietnam and the surrounding rural areas. Arsenic
concentrations in groundwater from private small-scale tube-wells averaged 159 μg/L, ranging from 1 to
3,050 μg/L. Arsenic concentrations ranged from 37 to 320 μg/L in raw groundwater pumped from the
lower aquifer for the Hanoi water supply (Berg et al. 2001). Several investigators have noticed a
correlation between high levels of arsenic and fluoride in drinking water (Wyatt et al. 1998a, 1998b).
Arsenic concentrations in drinking water from four villages in Bangladesh ranged from 10 to 2,040 μg/L
(Tondel et al. 1999).
ARSENIC 344
6. POTENTIAL FOR HUMAN EXPOSURE
6.4.3 Sediment and Soil
Arsenic is widely distributed in the Earth's crust, which contains about 3.4 ppm arsenic (Wedepohl 1991).
It is mostly found in nature minerals, such as realgar (As
4
S
4
), orpiment (As
2
S
3
), and arsenolite (As
2
O
3
),
and only found in its elemental form to a small extent. There are over 150 arsenic-bearing minerals
(Budavari et al. 2001; Carapella 1992). Arsenic concentrations in soils from various countries can range
from 0.1 to 50 μg/g and can vary widely among geographic regions. Typical arsenic concentrations for
uncontaminated soils range from 1 to 40 μg/g, with the lowest concentrations in sandy soils and soils
derived from granites. Higher arsenic concentrations are found in alluvial soils and soils with high
organic content (Mandal and Suzuki 2002). Arsenic in soil may originate from the parent materials that
form the soil, industrial wastes, or use of arsenical pesticides. Geological processes that may lead to high
arsenic concentrations in rock and subsequently the surrounding soil include hydrothermic activity and
pegmatite formation (Peters et al. 1999). In the first case, thermal activity results in the dissolution and
transport of metals, including the metalloid arsenic, which are precipitated in fractures in rocks. In the
second process, cooling magmas may concentrate metals that are injected into rocks, crystallizing as
pegmatites. Areas of volcanic activity include large areas of California, Hawaii, Alaska, Iceland, and
New Zealand.
The U.S. Geological Survey reports the mean and range of arsenic in soil and other surficial materials as
7.2 and <0.1–97 μg/g, respectively (USGS 1984). The concentrations of arsenic in 445 Florida surface
soils ranged from 0.01 to 50.6 μg/g (Chen et al. 1999). The median, arithmetic mean, and geometric
mean were 0.35, 1.34±3.77, and 0.42±4.10 μg/g, respectively. Chirenje et al. (2003b) reported a
geometric mean arsenic concentrations of 0.40 (0.21–660) and 2.81 (0.32–110) μg/g in surface soil
samples (0–20 cm) collected in May–June 2000 from Gainesville and Miami, Florida, respectively. The
geometric mean arsenic concentration in 50 California soils was 2.8 μg/g (Chen et al. 1999). In the
Florida surface soils, arsenic was highly correlated (α=0.0001) with the soil content of clay, organic
carbon, CEC, total iron, and total aluminum. Arsenic tends to be associated with clay fractions and iron
and manganese oxyhydroxides. Soils of granitic origin are generally low in arsenic, about 4 μg/g,
whereas arsenic in soils derived from sedimentary rocks may be as high as 20–30 μg/g (Yan-Chu 1994).
Soils overlying arsenic-rich geologic deposits, such as sulfide ores, may have soil concentrations two
orders of magnitude higher (NAS 1977a).
ARSENIC 345
6. POTENTIAL FOR HUMAN EXPOSURE
Soils in mining areas or near smelters may contain high levels of arsenic. Arsenic concentrations up to
27,000 μg/g were reported in soils contaminated with mine or smelter wastes (EPA 1982b). Soils at an
abandoned mining site in the Tamar Valley in southwest England have arsenic concentrations that may
exceed 50,000 μg/g (Erry et al. 1999). The average arsenic levels in the top 2 cm of different soil types in
the vicinity of a former copper smelter in Anaconda, Montana, ranged from 121 to 236 μg/g; levels were
significantly related to proximity and wind direction to the smelter site (Hwang et al. 1997a). Smelter
fallout can contaminate land miles from the source. Soils in mainland southern King County were studied
for the presence of arsenic and lead (WSDOE 2005). Soil samples were collected in the fall of 1999 and
the spring of 2001 from locations around the ASARCO smelter, which operated in Ruston from the 1890s
to 1986. The study area ran roughly from the I-90 corridor south to the King-Pierce county line, from the
Puget Sound shore to the Cascade foothills. Almost all of the contamination was found was in the 0–
6-inch depths of the cores samples; 62 of the 75 samples were found to have arsenic levels above 20 ppm
(WSDOE 2005).
Soil on agricultural lands treated with arsenical pesticides may retain substantial amounts of arsenic. One
study reported an arsenic concentration of 22 μg/g in treated soil compared to 2 μg/g for nearby untreated
soil (EPA 1982b). Arsenic was measured in soil samples taken from 10 potato fields in Suffolk County
on Long Island, New York, where sodium arsenite had been used for vine control and fall weed control
for many years. Lead arsenate also may have been used as an insecticide in certain areas. The mean
arsenic levels taken at a depth of 0–18 cm from each of the 10 fields ranged from 27.8±5.44 μg/g dry
weight (n=10) to 51.0±7.40 μg/g dry weight (n=10). These levels were markedly higher than the level of
2.26±0.33 μg/g (n=10) for untreated control soils (Sanok et al. 1995). A survey was conducted in 1993 to
determine the concentrations of arsenic and lead in soil samples from 13 old orchards in New York State.
Lead arsenate was used for pest control in fruit orchards for many years, mainly from the 1930s to 1960s,
and residues remain in the soil. Concentrations of arsenic ranged from 1.60 to 141 μg/g dry weight
(Merwin et al. 1994). Arsenic and lead concentrations were also measured in former orchard soils
contaminated by lead arsenate from the Hanford site in Washington State. The mean arsenic
concentration in surface (5–10 cm) and subsurface (10–50 cm) soils were 30 (2.9–270) and 74 (32–
180) μg/g dry weight, respectively (Yokel and Delistraty 2003). Average arsenic concentration of 5.728,
5.614, and 6.746 μg/g were reported in soils, lake sediments, and wetland sediments, respectively, from
Moon Lake, a Mississippi River alluvial floodplain in northwest Mississippi. Intensive cultivation has
occurred in this area, including cotton, soybeans, and rice (Cooper and Gillespie 2001). A geometric
mean arsenic concentration of 20.6 mg/kg (range 4.6–340 mg/kg) was reported soil collected during the
ARSENIC 346
6. POTENTIAL FOR HUMAN EXPOSURE
summer and fall of 2003 from 85 homes in Middleport, New York, where historical pesticide
manufacturing was associated with arsenic in the soil (Tsuji et al. 2005).
The Washington State Area-wide Soil Contamination Project provides various data on arsenic
contamination in soils across Washington State (Washington State 2006). Arsenic concentrations within
areas affected by area-wide soil contamination are highly variable, ranging from natural background
levels to >3,000 ppm in smelter areas. Generally, average arsenic concentrations in soil at developed
properties are <100 ppm. Areas affected by smelter emissions in King, Pierce, Snohomish, and Stevens
Counties have a higher likelihood of arsenic soil contamination than other areas of the State due to
historical emissions from metal smelters located in Tacoma, Harbor Island, Everett, Northport, and Trail,
British Columbia. Areas where apples and pears were historically grown, such as Chelan, Spokane,
Yakima, and Okanogan Counties, also have a higher likelihood of arsenic soil contamination than other
areas due to the past use of lead arsenate pesticides. Generally, arsenic contamination in soils from
historical smelter emissions and historical use of lead-arsenate pesticides is found in the upper 6–
18 inches of soil (Washington State 2006).
The New Jersey Department of Environmental Protection (Historic Pesticide Contamination Task Force
1999) reported on the analysis of soil samples collected from 18 sites for various pesticide residues,
including arsenic, from current and former agricultural sites in New Jersey in order to assess
contamination from historic pesticide use. Arsenic was detected in all 463 samples, with concentrations
ranging from 1.4 to 310 ppm.
Natural concentrations of arsenic in sediments are usually <10 μg/g dry weight, but can vary widely
around the world (Mandal and Suzuki 2002). Sediment arsenic concentrations reported for U.S. rivers,
lakes, and streams range from about 0.1 to 4,000 μg/g (Eisler 1994; Heit et al. 1984; NAS 1977a; Welch
et al. 1988). During August through November 1992 and August 1993, bed sediment in the South Platte
River Basin (Colorado, Nebraska, and Wyoming) was sampled and analyzed for 45 elements, including
arsenic. The range of arsenic found was 2.8–31 μg/g dry weight and the geometric mean (n=23) was
5.7 μg/g (Heiny and Tate 1997). The arsenic concentration in surface sediment (0–2 cm) at 43 sites in the
Indian River Lagoon System in Florida ranged from 0.6 to 15 μg/g dry weight with a mean of
5.0±3.9 μg/g (Trocine and Trefry 1996). Arsenic levels were well correlated with those of aluminum.
Correlation with aluminum levels is used to normalize sediment level concentrations to natural levels in
Florida estuaries. Surficial sediments collected from 18 locations in 3 major tributaries to Newark Bay,
New Jersey, were analyzed for 7 toxic metals, including arsenic (Bonnevie et al. 1994). The highest
ARSENIC 347
6. POTENTIAL FOR HUMAN EXPOSURE
concentrations of arsenic were found in the Rahway River adjacent to a chemical plant, 58 μg/g dry
weight, and in the Hackensack River adjacent to a coal-fired power plant, 49 μg/g. The average arsenic
concentration for all sediments was 17±16 μg/g. Sediments collected from seven sites in Baltimore
Harbor, Maryland, at five seasonal periods between June 1987 and June 1988 had a geometric mean
maximum of 7.29 μg/g dry weight and a geometric mean minimum of 1.25 μg/g (Miles and Tome 1997).
This harbor is one of two sub-tributaries of the Chesapeake Bay where contaminants have been
discharged on a large scale.
The upper Clark Fork River basin in western Montana is widely contaminated by metals from past
mining, milling, and smelting activities. In a 1991 study, arsenic levels were determined in sediment
along the river and in a reservoir 205 km downstream. Total arsenic in sediments from Clark Fork River
decreased from 404 μg/g dry weight at the farthest upstream sampling station to 11 μg/g, 201 km
downstream. Sediment samples from the Milltown Reservoir had arsenic concentrations ranging from
6 to 56 μg/g (Brumbaugh et al. 1994). Total recoverable arsenic in nonfiltered pore water from the Clark
Fork River decreased from 1,740 μg/L at the farthest upstream sampling station to 31 μg/L at the 201 km
station (Brumbaugh et al. 1994). The Coeur d’Alene river basin in northern Idaho has been contaminated
with heavy metals from mining and smelting operations since 1885 (Farag et al. 1998). A 1994 study
determined the metal content of sediment, biofilm, and invertebrates at 13 sites in the basin, 10 with
historic mining activity and 3 reference sites. The mean arsenic levels in sediment at the mining sites
ranged from 8.3 to 179.0 μg/g dry weight, compared to 2.4–13.1 μg/g dry weight at the reference sites.
The mean arsenic levels in biofilm adhering to rock in the water at the mining sites ranged from 7.5 to
155.8 μg/g dry weight, compared to 7.2–27.3 μg/g dry weight at the reference sites. In Whitewood Creek,
South Dakota, where as much as 100 million tons of mining and milling waste derived from gold mining
activities were discharged between 1876 and 1977, mean and maximum sediment arsenic concentrations
were 1,920 and 11,000 μg/g, respectively (USGS 1987). Uncontaminated sediment had mean arsenic
levels of 9.2 μg/g. Arsenic concentrations in surface (0–5 cm) sediments from watersheds in Black Hills,
South Dakota, an area impacted by gold mining activities, ranged from 23 to 1,951 μg/g dry weight;
concentrations from reference areas ranged from 10 to 58 μg/g dry weight (May et al. 2001). Swan Lake,
a sub-bay of Galveston Bay in Texas is a highly industrial area that received runoff from a tin smelter in
the 1940s and 1950s. Surface sediments at 17 sites where oysters and mussels were collected ranged
from 4.53 to 103 μg/g (Park and Presley 1997). A site in the channel leading from the old smelter had
arsenic levels of 568 μg/g. Surface sediment was less contaminated than deeper sediment, indicating less
arsenic input recently than in the past as a result of the smelter closing.
ARSENIC 348
6. POTENTIAL FOR HUMAN EXPOSURE
It has been suggested that the wood preservative most commonly used in dock pilings and bulkheads,
CCA, can be toxic to estuarine organisms. Wendt et al. (1996) measured arsenic in surface sediments and
oysters from creeks with high densities of docks and from nearby reference creeks with no docks. The
average concentrations in the sediments ranged from 14 to 17 μg/g throughout the study area, which is
within the range of natural background levels. Weis et al. (1998) sampled sediments along a 10-m
transect from CCA-treated wood bulkheads from four Atlantic coast estuaries. Arsenic concentrations
were highest in the fine-grained portion of the sediments near the CCA-treated bulkhead (0–1 m); arsenic
concentrations were generally at reference levels at distances >1 m from the bulkheads (Weis et al. 1998).
Soils below and around play structures constructed from CCA-treated wood in the City of Toronto,
Canada were sampled and analyzed for inorganic arsenic (Ursitti et al. 2004). A mean arsenic
concentration of 2.1 μg/g (range 0.5–10 μg/g) was reported in soil samples taken within 1 m of the CCA-
treated wood for all play structures. Soil samples that were collected 10 m from the play structures served
as a background had arsenic concentration of 2.4 μg/g (range 0.5–13 μg/g). A mean arsenic concentration
of 6.2 μg/g (range 0.5–47.5 μg/g) was reported in soil samples taken below CCA-treated wood for all play
structures. Of the 217 play structures in the study, 32 had arsenic concentrations under the play structures
that exceeded the Canadian federal soil guidelines with arsenic concentrations ranging from 12.4 to
47.5 μg/g. From this study, the authors concluded that arsenic does not migrate laterally, but does
accumulate in soil under elevated platforms constructed from CCA-treated wood (Ursitti et al. 2004).
6.4.4 Other Environmental Media
Low levels of arsenic are commonly found in food; the highest levels are found in seafood, meats, and
grains. Typical U.S. dietary levels of arsenic in these foods range from 0.02 mg/kg in grains and cereals
to 0.14 mg/kg in meat, fish, and poultry (Gartrell et al. 1986). Shellfish and other marine foods contain
the highest arsenic concentrations and are the largest dietary source of arsenic (Gunderson 1995a; Jelinek
and Corneliussen 1977; Tao and Bolger 1999). Arsenic levels in various fish and shellfish are presented
in Table 6-4. In the U.S. Food and Drug Administration (FDA) Total Diet Study, 1991–1997, seafood
contained the highest levels of arsenic, followed by rice/rice cereal, mushrooms, and poultry.
Concentrations in canned tuna (in oil), fish sticks, haddock (pan-cooked), and boiled shrimp were 0.609–
1.470, 0.380–2.792, 0.510–10.430, and 0.290–2.681 mg/kg, respectively (Tao and Bolger 1999).
Typically, arsenic levels in foods in the Total Diet Study, 1991–1996 were low, <0.03 mg/kg; only 63 of
the 264 foods contained arsenic above this level. Similar results were reported in the Total Diet Study,
ARSENIC 349
6. POTENTIAL FOR HUMAN EXPOSURE
Table 6-4. Levels of Arsenic in Fish and Shellfish—Recent Studies
Arsenic concentration
a
Sample type (μg/g) Comments Reference
Yellowtail flounder
Muscle (n=8)
Liver (n=6)
Gonad (n=6)
Marine organisms
Ray (n=8)
Cod (n=8)
Plaice (n=8)
Sole (n=8)
Sea-bream (n=8)
Mussell (n=8)
Bluefin tuna (
Thunnus thynus)
(n=14)
Fish
Bottom feeding (n=2,020)
Predatory (n=12)
Oysters
<1 m from docks (n=10)
>10 m from docks (n=10)
Reference (no docks)
(n=10
Clams (n=22)
Marine organisms
Snails
Blue crab
Fish
Shrimp
Whole crab
Oysters (n=10, pooled)
Mussels (n=7, pooled)
Marine organisms
Blue crab
Fish
Oysters, two areas
n=78, pooled
n=874, pooled
Samples collected from Hellou et al. 1998
8–37
Northwest Atlantic 1993
7–60
1.2–9.4
Belgian fish markets in
Buchet and Lison
16.4
1991; inorganic arsenic
1998
4.7
ranged from 0.003 to
0.2 μg/g
19.8
5.1
2.4
3.5
3.2 Virgin Rocks, Grand Banks
Hellou et al. 1992
of Newfoundland, Canada,
1990
National Contaminant
Kidwell et al. 1995
0.16±0.23 wet weight
Biomonitoring Program,
0.16±0.140 wet weight
1984–1985, 112 stations
South Carolina, private
Wendt et al. 1996
8.3±1.1
residential docks on tidal
7.6±0.9
creeks, 1994
8.4±1.3
12±1.1 Indian River Lagoon, Trocine and Trefry
Florida, 22 sites, 1990 1996
Swan Lake, Galveston Bay, Park and Presley
13.3±17.0
Texas, 1993 1997
6.61
0.82
1.37±0.64
5.35±2.51
7.28±1.32
7.75±2.15
GPNEP, 1992, Galveston Park and Presley
2.31±2.15
Bay, Texas 1997
2.46
NOAA NS&T Program,
Park and Presley
1986–1990
1997
4.50±1.08 Galveston Bay
9.67±7.00 Gulf of Mexico
ARSENIC 350
6. POTENTIAL FOR HUMAN EXPOSURE
Table 6-4. Levels of Arsenic in Fish and Shellfish—Recent Studies
Arsenic concentration
a
Sample type
(μg/g)
Comments Reference
Marine crustaceans
Parapenaeus longirostris
(pink shrimp) (n=826,
10 pools
Aristeus antennatus (red
shrimp) (n=387, 8 pool)
Plesionika martia (shrimp)
(n=456, 7 pools)
Nephrops norvegicus
(Norway lobster) (n=270,
5 pools)
Freshwater fish
Sabalo (
Brycon
melanopterus)
(n=3)
Huazaco (
Hoplias
malabaracus)
(n=4)
Bagre (
Pimelodus ornatus)
(n=8)
Boquichio (
Prochilodus
nigricans)
(n=1)
Doncello (
Pseudo-
platystoma sp.
) (n=1)
Freshwater fish
Bowfin (n=59)
Bass (n=47)
Channel catfish (n=50)
Chain pickerel (n=19)
Yellow perch (n=51)
Black crappie (n=52)
American eel (n=24)
Shellcracker n=52)
Bluegill (n=52)
Redbreast (n=43)
Spotted sucker (n=35)
34.84±19.21
(12.01–62.60)
17.09±3.49
(10.45–20.82)
40.82±2.50
(36.37–44.06)
43.48±14.21
(35.63–69.15)
0.015–0.101
nd–0.005
nd–0.201
0.063
0.055
0.32±0.04 wet weight
0.03±0 wet weight
0.09±00.02 wet weight
0.05±0.01 wet weight
0.05±0.01 wet weight
0.04±0.01 wet weight
0.04±0.01 wet weight
0.06±0 wet weight
0.05±0.02 wet weight
0.07±0.01 wet weight
0.03±0 wet weight
Commercial crustaceans
Storelli and
from the Mediterranean
Marcotrigiano 2001
Sea (Italy)
Fish samples (muscle)
Gutleb et al. 2002
were collected in August
1997 from the Candamo
River, Peru; a pristine
rainforest valley prior to the
start of oil-drilling activities
Savannah River, along and
Burger et al. 2002
below the Department of
Energy’s Savannah River
Site (SRS); samples
analyzed were edible fillets
ARSENIC 351
6. POTENTIAL FOR HUMAN EXPOSURE
Table 6-4. Levels of Arsenic in Fish and Shellfish—Recent Studies
Arsenic concentration
a
Sample type
(μg/g)
Comments Reference
Horseshoe crabs
Apodeme (n=74) 7.034±0.65 wet weight
Overall mean in tissues of Burger et al. 2003
Egg (n=63)
Leg (n=74)
5.924±0.345 wet weight
14.482±0.685 wet
crabs collected from New
Jersey in 2000
weight
Apodeme (n=40) 7.513±0.835 wet weight
Overall mean in tissues of
Egg (n=35)
Leg (n=40)
6.766±0.478 wet weight
18.102±1.489 wet
crabs collected from
Delaware in 2000
weight
a
Concentrations are means±standard deviation, unless otherwise stated. Concentrations are in a dry weight basis,
unless otherwise stated.
GM = geometric mean; GPNEP = Galveston Bay National Estuary Program; nd = not detected;
NOAA NS&T = National Oceanic and Atmospheric Administration National Status and Trends
ARSENIC 352
6. POTENTIAL FOR HUMAN EXPOSURE
1991–1997, where the mean arsenic concentration in all foods was 0.036 mg/kg dry weight and arsenic
was not detectable in about 88% of the foods and was detected at trace levels in another 7.8% of foods.
The foods with the highest mean arsenic levels were haddock, canned tuna, fish sticks, shrimp, and fish
sandwiches, with arsenic concentrations ranging from 5.33 to 0.568 mg/kg dry weight (Capar and
Cunningham 2000). Nriagu and Lin (1995) analyzed 26 brands of wild rice sold in the United States and
found arsenic levels ranging from 0.006 to 0.142 μg/g dry weight. Arsenic concentrations ranging from
0.05 to 0.4 μg/g are typically reported for rice from North America, Europe, and Taiwan (Meharg and
Rahman 2003).
During a comprehensive total diet study extending from 1985 to 1988, foods were collected in six
Canadian cities and processed into 112 composite food samples (Dabeka et al. 1993). The mean, median,
and range of total arsenic in all samples were 0.0732, 0.0051, and <0.0001–4.840 μg/g, respectively.
Food groups containing the highest mean arsenic levels were fish (1.662 μg/g), meat and poultry
(0.0243 μg/g), bakery goods and cereals (0.0245 μg/g), and fats and oils (0.0190 μg/g). Of the individual
samples, marine fish had the highest arsenic levels, with a mean of 3.048 μg/g for the cooked composites
and 2.466 μg/g for the raw samples. Canned fish (1.201 μg/g) and shellfish (2.041 μg/g) also contained
high means. Cooked poultry, raw mushrooms, and chocolate bars contained 0.100, 0.084, and
0.105 μg/g, respectively.
National monitoring data from the Food Safety and Inspection Service National Residue Program (NRP)
(1994–2000) found that the mean total arsenic concentration in livers of young chickens ranged from
0.33 to 0.43 μg/g, with an overall mean of 0.39 μg/g (Lasky et al. 2004). The mean arsenic
concentrations in liver for mature chickens, turkeys, hogs, and all other species over the same time period
ranged from 0.10 to 0.16 μg/g. Lasky et al. (2004) used the NRP arsenic data in livers of young chickens
to estimate the concentrations of arsenic in muscle tissue, the most commonly consumed part of the
chicken. Assuming that 65% of the arsenic in poultry and meat is inorganic, at a mean level of chicken
consumption of 60 g/person/day, people may ingest an estimated 1.38–5.24 μg/day of inorganic arsenic
from chicken.
A Danish study (Pedersen et al. 1994) reports the arsenic levels in beverages as the mean (range) in μg/L
as follows: red wine, 9 (<2–25); white wine, 11 (<2–33); fortified wine, 5 (<2–11); beer, 7 (4–11); soft
drinks, 3 (<2–8); miscellaneous juices, 8 (3–13); instant coffee, 4 (0.7–7); and instant cocoa, 5.6 (1.6–
12.8).
ARSENIC 353
6. POTENTIAL FOR HUMAN EXPOSURE
In a study of dietary arsenic exposure in the Indigenous Peoples of the western Northwest Territories,
Canada, fish contained the highest arsenic concentrations in foods consumed by the Dene and Métis
populations with the highest concentration, 1.960 μg/g, found in smoked/dried cisco (fish). Other foods
derived from land mammals, birds, and plants contained lower arsenic concentrations. A mean arsenic
intake of <1.0 μg/kg/day was reported for this population (Berti et al. 1998).
The general consensus in the literature is that about 85–>90% of the arsenic in the edible parts of marine
fish and shellfish is organic arsenic (e.g., arsenobetaine, arsenochloline, dimethylarsinic acid) and that
approximately 10% is inorganic arsenic (EPA 2003b). However, the inorganic arsenic content in seafood
may be highly variable. For example, a study in the Netherlands reported that inorganic arsenic
comprised 0.1–41% of the total arsenic in seafood (Vaessen and van Ooik 1989). Buchet et al. (1994)
found that, on the average, 3% of the total arsenic in mussels was inorganic in form. Some commercially
available seaweeds, especially brown algae varieties, may have high percentages of the total arsenic
present as inorganic arsenic (>50%) (Almela et al. 2002; Laparra et al. 2003). Arsenic concentrations
ranging from 17 to 88 mg/kg dry weight were found in commercially available seaweeds (van Netten et
al. 2000). Other arsenic compounds that may be found in seafood are arsenic-containing ribose
derivatives called arsenosugars. Arsenosugars are the common organoarsenicals found in marine algae;
they are also found in mussels, oysters, and clams (Le et al. 2004). Less information about the forms of
arsenic in freshwater fish is known at this time (EPA 2003b).
Schoof et al. (1999a) reported on the analysis of 40 commodities anticipated to account for 90% of dietary
inorganic arsenic intake. In this study, the amount of inorganic arsenic was measured in these foods.
Consistent with earlier studies, total arsenic concentrations were highest in the seafood sampled (ranging
from 160 ng/g in freshwater fish to 2,360 ng/g in marine fish). In contrast, average inorganic arsenic in
seafood ranged from <1 to 2 ng/g. The highest inorganic arsenic concentrations were found in raw rice
(74 ng/g), followed by flour (11 ng/g), grape juice (9 ng/g), and cooked spinach (6 ng/g).
Tobacco contains an average arsenic concentration of 1.5 ppm, or about 1.5 μg per cigarette (EPA 1998j).
Before arsenical pesticides were banned, tobacco contained up to 52 mg As/kg, whereas after the ban,
maximum arsenic levels were reduced to 3 μg/g (Kraus et al. 2000). An international literature survey
reports arsenic yields of 0–1.4 μg/cigarette for mainstream (inhaled) cigarette smoke (Smith et al. 1997).
The wide range of arsenic yields for flue-cured cigarettes suggests that the field history, soil, and fertilizer
conditions under which the tobacco is grown will affect the arsenic concentration (Smith et al. 1997).
ARSENIC 354
6. POTENTIAL FOR HUMAN EXPOSURE
Arsenic emission factors of 0.015–0.023 μg/cigarette (mean 0.018±0.003 μg/cigarette) have been
measured for sidestream smoke from a burning cigarette (Landsberger and Wu 1995).
A median arsenic concentration of 2.1 μg/g and a deposition rate of 0.008 μg/m
2
/day was reported in
house dust in homes evaluated as part of the German Environmental Survey in 1990–1992. A mean
arsenic concentration of 7.3 μg/g was reported in house dust from 48 residences in Ottawa, Canada (Butte
and Heinzow 2002). These arsenic concentrations are expected to be representative of background levels.
In general high arsenic concentrations were found in household dust collected from homes in areas with
known arsenic contamination. Mean arsenic concentrations of 12.6 (2.6–57) and 10.8 (1.0–49) μg/g were
reported in house dust collected from the entryway and child play areas, respectively, from homes in a
community in Washington State with a history of lead arsenate use (Wolz et al. 2003). Arsenic was
detected in all 135 indoor floor dust samples collected as part of the NHEXAS from Arizona mining
communities, ranging from 0.3–50.6 μg/g, (O'Rourke et al. 1999). A geometric mean arsenic
concentration of 10.8 μg/g (range 1.0–172 μg/g) was reported in house dust from 96 homes in Middleport,
New York, with historical pesticide manufacture, collected during the summer and fall of 2003 (Tsuji et
al. 2005).
Arsenic has also been detected in several homeopathic medicines at concentrations up to 650 μg/g (Kerr
and Saryan 1986). Some Asian proprietary medicines that are manufactured in China, Hong Kong, and
other Asian countries have been reported to contain levels of inorganic arsenic ranging from 25 to
107,000 μg/g (Chan 1994). Fifty medicinally important leafy samples that were analyzed for elemental
concentrations contained arsenic at levels ranging from 0.12 to 7.36 μg/g, with a mean of 2.38±1.2 μg/g
(Reddy and Reddy 1997). Arsenic concentrations ranged from 0.005 to 3.77 μg/g in 95 dietary
supplements purchased from retail stores in the Washington, DC area in 1999 (Dolan et al. 2003).
Commercially available samples of Valarian, St. John's Wort, Passion Flower, and Echinacea were
purchased in the United States and analyzed for various contaminants; arsenic concentrations were
0.0016–0.0085, 0.0065–0.017.8, 0.0024–0.0124, and 0.0021–0.0102 μg/g, respectively, in these samples
(Huggett et al. 2001). Concentrations of heavy metals including the metalloid arsenic were evaluated in
54 samples of Asian remedies that were purchased in stores in Vietnam and Hong Kong that would be
easily accessible to travelers, as well as in health food and Asian groceries in Florida, New York, and
New Jersey. Four remedies were found to contain daily doses exceeding 0.1 mg. Two of these contained
what would have been a potentially significant arsenic dose, with daily doses of 16 and 7.4 mg of arsenic
(Garvey et al. 2001).
ARSENIC 355
6. POTENTIAL FOR HUMAN EXPOSURE
The possible presence of toxic compounds in waste materials has raised concerns about the fate of these
compounds either during the composting process or when the composted product is applied to soils.
Three waste compost products generated at the Connecticut Agricultural Experiment Station had arsenic
levels of 12.8, 9.8, and 13 μg/g dry weight, respectively (Eitzer et al. 1997). The arsenic levels in
municipal solid waste composts from 10 facilities across the United States ranged from 0.9 to 15.6 μg/g
dry weight with a mean of 6.7 μg/g (He et al. 1995). These are lower than the EPA 503 regulatory limit
for arsenic of 41 μg/g for agricultural use of sewage sludge (EPA 1993b). Concentrations of arsenic in
U.S. sewage sludges, which are sometimes spread on soil, were <1 μg/g. Arsenic is a common impurity
in minerals used in fertilizers. A comprehensive Italian study found that the arsenic content in a number
of mineral and synthetic fertilizers ranged from 2.2 to 322 mg/kg with a sample of triple superphosphate
having the highest level (Senesi et al. 1999). Arsenic naturally occurs in coal and crude oil at levels of
0.34–130 and 0.0024–1.63 ppm, respectively, which would account for its presence in flue gas, fly ash,
and bottom ash from power plants (Pacyna 1987).
Background arsenic levels in living organisms are usually <1 μg/g wet weight (Eisler 1994). Levels are
higher in areas with mining and smelting activity or where arsenical pesticides were used. Eisler (1994)
has an extensive listing of arsenic levels in terrestrial and aquatic flora and fauna from literature sources
to about 1990. The U.S. Fish and Wildlife Service’s National Contaminant Biomonitoring Program have
analyzed contaminants in fish at 116 stations (rivers and the Great Lakes) across the United States. The
geometric mean concentration of arsenic for the five collection periods starting in 1976 were (period,
concentration wet weight basis): 1976–1977, 0.199 μg/g; 1978–1979, 0.129 μg/g; 1980–1981,
0.119 μg/g; 1984, 0.106 μg/g; and 1986, 0.083 μg/g (Schmitt et al. 1999). In 1986, the maximum and
85th percentile arsenic levels were 1.53 and 0.24 μg/g, respectively. The highest concentrations of
arsenic for all five collection periods were in bloaters from Lake Michigan at Sheboygan, Wisconsin.
Arsenic levels declined by 50% at this site between 1976–1997 and 1984. The major source of arsenic
into Lake Michigan was a facility at Marinette, Wisconsin, which manufactured arsenic herbicides.
Table 6-4 contains arsenic levels in aquatic organisms from more recent studies. The Coeur d’Alene river
basin in northern Idaho has been contaminated with heavy metals from mining and smelting operations
since 1885 (Farag et al. 1998). A 1994 study determined the metal content of sediment, biofilm, and
invertebrates at 13 sites in the basin, 10 with historic mining activity, and 3 reference sites. The mean
arsenic levels in benthic macroinvertebrates at the mining sites ranged from 2.2 to 97.0 μg/g dry weight,
compared to 2.1–2.4 μg/g dry weight at the reference sites. A study of aquatic organism in Swan Lake, a
highly polluted sub-bay of Galveston Bay, Texas showed that arsenic concentrations were in the order
snail>oyster>crab>shrimp>fish (Park and Presley 1997). In contrast to metals like silver, cadmium,
ARSENIC 356
6. POTENTIAL FOR HUMAN EXPOSURE
copper, and zinc, arsenic concentrations in oysters and mussels were less than in the sediment from which
they were collected. No significant correlation was found between levels of arsenic in clams in the Indian
River Lagoon in Florida with those found in sediment or water samples (Trocine and Trefry 1996). Small
animals living at mining sites ingest more arsenic in their diet and have higher arsenic levels in their
bodies than those living on uncontaminated sites (Erry et al. 1999). Seasonal variations in both arsenic
intake and dietary composition may affect the amount of arsenic taken up by the body and transferred to
predator animals. Tissue arsenic content of wood mice and bank voles living on both arsenic-
contaminated mining sites and uncontaminated sites were greater in autumn than spring. The lower tissue
arsenic levels in spring of rodents living on contaminated sites suggest that there is no progressive
accumulation of arsenic in overwintering animals.
6.5 GENERAL POPULATION AND OCCUPATIONAL EXPOSURE
Exposure to arsenic may include exposure to the more toxic inorganic forms of arsenic, organic forms of
arsenic, or both. While many studies do not indicate the forms of arsenic to which people are exposed,
this information may often be inferred from the source of exposure (e.g., fish generally contain arsenic as
arsenobetaine). Yost et al. (1998) reported that the estimated daily dietary intake of inorganic arsenic for
various age groups ranged from 8.3 to 14 μg/day and from 4.8 to 12.7 μg/day in the United States and
Canada, respectively, with 21–40% of the total dietary arsenic occurring in inorganic forms.
Drinking water may also be a significant source of arsenic exposure in areas where arsenic is naturally
present in groundwater. While estimates of arsenic intake for typical adults drinking 2 L of water per day
average about 5 μg/day (EPA 1982c), intake can be much higher (10–100 μg/day) in geographical areas
with high levels of arsenic in soil or groundwater (see Figure 6-2). It is assumed that nearly all arsenic in
drinking water is inorganic (EPA 2001).
In the United States, food intake of arsenic has been estimated to range from 2 μg/day in infants to
92 μg/day in 60–65-year-old men (see Table 6-5) (Tao and Bolger 1999). The average intake of
inorganic arsenic are estimated to range from 1.34 μg/day in infants to 12.54 μg/day in 60–65-year-old
men. Tao and Bolger (1999) assumed that 10% of the total arsenic in seafood was inorganic and that
100% of the arsenic in all other foods was inorganic. The greatest dietary contribution to total arsenic
was seafood (76–96%) for all age groups, except infants. For infants, seafood and rice products
contributed 42 and 31%, respectively. Adult dietary arsenic intakes reported for other countries range
ARSENIC 357
6. POTENTIAL FOR HUMAN EXPOSURE
Table 6-5. Mean Daily Dietary Intake of Arsenic for Selected U.S. Population
Groups
Date of study
Mean daily intake (μg/kg body weight/day)
1984–1986
a
1986–1991
b
1991–1997
c
Provisional tolerable daily intake (PTDI)
d
2.1 2.1 2.1
6–11 months 0.82 0.5 0.31
2 years 1.22 0.81 1.80
14–16 years, female 0.54 0.36 0.41
14–16 years, male 0.60 0.39 0.24
25–30 years, female 0.66 0.44 0.44
25–30 years, male 0.76 0.51 0.72
60–65 years, female 0.71 0.46 1.08
60–65 years, male 0.74 0.48 1.14
a
Gunderson 1995a
b
Gunderson 1995b
c
Tao and Bolger 1999
d
No agreement has been reached on a maximum acceptable intake for total arsenic; the FAO/WHO has assigned a
PTDI for inorganic arsenic of 2.1 μg/kg body weight for adults. Data from FDA studies. FDA does not recommend
daily intake levels for Arsenic.
ARSENIC 358
6. POTENTIAL FOR HUMAN EXPOSURE
from 11.7 to 280 μg/day (Tao and Bolger 1999). Schoof et al. (1999b) estimated that intake of inorganic
arsenic in the U.S. diet ranges from 1 to 20 μg/day, with a mean of 3.2 μg/day. In contrast, these
estimates of inorganic arsenic intakes are based on measured inorganic arsenic concentrations from a
market basket survey.
The FDA conducted earlier Total Diet Studies in 1984–1986 and 1986–1991. For the sampling period of
June 1984 to April 1986, the total daily intake of arsenic from foods was 58.1 μg for a 25–30-year-old
male with seafood contributing 87% of the total (Gunderson 1995a). For the sampling period from July
1986 to April 1991, the total daily intake of arsenic from foods was lower, 38.6 μg for a 25–30-year-old
male. Seafood again was the major source of arsenic, contributing 88% of the total (Gunderson 1995b).
Results of the two Total Diet Studies for selected population groups are shown in Table 6-5. The Total
Diet Study for the sampling period from September 1991 to December 1996, shows that arsenic, at
≥0.03 μg/g, was found in 55 (21%) of the 261–264 foods/mixed dishes analyzed. The highest
concentrations again were found in seafood, followed by rice/rice cereal, mushrooms, and poultry. The
estimated total daily intake of arsenic from foods was 56.6 μg for a 25–30-year-old male. Seafood was
the major contributor, accounting for 88–96% of the estimated total arsenic intake of adults.
Average daily dietary exposures to arsenic were estimated for approximately 120,000 U.S. adults by
combining data on annual diet, as measured by a food frequency questionnaire, with residue data for
table-ready foods that were collected for the annual FDA Total Diet Study. Dietary exposures to arsenic
were highly variable, with a mean of 50.6 μg/day (range, 1.01–1,081 μg/day) for females and 58.5 μg/day
(range, 0.21–1,276 μg/day) for males (MacIntosh et al. 1997). Inorganic arsenic intake in 969 men and
women was assessed by a semi-quantitative food frequency questionnaire in combination with a database
for total arsenic content in foods and by toenail concentrations of arsenic. The mean estimated average
daily consumption of inorganic arsenic was 10.22 μg/day with a range of 0.93–104.89 μg/day. An
assumption of 1.5% of the total arsenic in fish and 20% of the total arsenic in shellfish was inorganic
arsenic was used in this assessment (MacIntosh et al. 1997).
During a comprehensive total diet study extending from 1985 to 1988, the estimated daily dietary
ingestion of total arsenic by the average Canadian was 38.1 μg and varied from 14.9 μg for the 1–4 year-
old-age group to 59.2 μg for 20–39-year-old males (Dabeka et al. 1993). Daily intakes of arsenic from
food by women in the Shiga Prefecture, Japan, were investigated by the duplicate portion method and by
the market basket method. In 1991 and 1992, the daily intakes determined by the duplicate portion
ARSENIC 359
6. POTENTIAL FOR HUMAN EXPOSURE
method were 206 and 210 μg, respectively. Those determined by the market basket method were 160 and
280 μg, respectively (Tsuda et al. 1995b).
Arsenic concentrations in human breast milk have been reported to range from 4 to <10 μg/L in pooled
human milk samples from Scotland and Finland to 200 μg/L in samples from Antofagasta, Chile, where
there is a high natural environmental concentration of arsenic (Broomhall and Kovar 1986). The arsenic
concentration in the breast milk of 35 women in Ismir, Turkey, a volcanic area with high thermal activity
ranged from 3.24 to 5.41 μg/L, with a median of 4.22 μg/L (Ulman et al. 1998). Sternowsky et al. (2002)
analyzed breast milk from 36 women from three different regions in Germany. These regions included
the city of Hamburg, a rural area, Soltau, Lower Saxony, and Munster, the potentially contaminated area.
Arsenic was not detected (<0.3 μg/L) in 154 of 187 samples, with the highest concentration, 2.8 μg/L,
found in a sample from the rural area. The geometric means from the three areas were comparable.
The mean arsenic levels in three groups of cows in the region that grazed on land impacted by lava and
thermal activity were 4.71, 4.46, and 4.93 μg/L, compared to 5.25 μg/L for cows kept in sheds and fed
commercial pellet feed and municipal water (Ulman et al. 1998). Mean arsenic concentrations in cow's
milk ranging from 18.6 to 17.1 μg/L and from 16.7 to 18.0 μg/L were reported for cow's grazing in
nonindustrial and an industrial regions, respectively, in Turkey (Erdogan et al. 2004).
A Danish study found that carrots grown in soil containing 30 μg/g of arsenic, which is somewhat above
the 20 μg/g limit for total arsenic set by Denmark for growing produce, contained 0.014 μg/g fresh weight
of arsenic, all in the form of inorganic As(III) and As(V) (Helgesen and Larsen 1998). An adult
consuming 376 grams of vegetables a day (90
th
percentile) represented solely by carrots would consume
5.3 μg of arsenic a day. The study concluded that the estimated intake of arsenic from produce grown in
soil meeting regulatory limits was low compared with other food sources and water.
If vegetables are grown in planters made of wood treated with CCA, arsenic may leach out of the wood
and be taken up by the vegetables. In a study by Rahman et al. (2004), arsenic was found to diffuse into
the soil from the CCA-treated wood, with the highest concentrations found at 0–2 cm from the CCA-
treated wood and a steady decline in concentration with increased distance from the wood. Crops grown
within 0–2 cm of the CCA-treated wood contained higher concentrations of arsenic than those grown at
1.5 m from the treated wood. However, the concentrations are below U.S. FDA tolerance limits that have
been set for arsenic in select food items. In addition, food grown in this manner is unlikely to constitute a
significant part of a person’s diet.
ARSENIC 360
6. POTENTIAL FOR HUMAN EXPOSURE
In 2003, U.S. manufacturers of arsenical wood preservatives began a voluntary transition from CCA to
other wood preservatives in wood products for certain residential uses, such as play structures, picnic
tables, decks, fencing, and boardwalks. This phase out was completed on December 31, 2003; wood
treated prior to this date could still be used and structures made with CCA-treated wood would not be
affected. CCA-treated wood products continue to be used in industrial applications (EPA 2003a). EPA’s
Consumer Awareness Program (CAP) for CCA is a voluntary program established by the manufacturers
of CCA products to inform consumers about the proper handling, use, and disposal of CCA-treated wood.
Additional information about this program can be found from EPA (2007a).
The arsenic content in the human body is 3–4 mg and tends to increase with age. Arsenic concentrations
in most tissues of the human body are <0.3 to 147 μg/g dry weight, excluding hair, nails, and teeth.
Mammals tend to accumulate arsenic in keratin-rich tissues such as hair and nails. The normal
concentrations of arsenic range from about 0.08 to 0.25 μg/g in hair, and 0.34 μg/g in nails. The normal
concentration of arsenic in urine can range from 5 to 40 μg per day (total) (Mandal and Suzuki 2002).
Table 6-6 contains arsenic levels in various human tissues.
A German study investigated the transfer of arsenic from the environment to humans in the northern
Palatine region, a former mining area characterized by high soil levels of arsenic (<2–605 μg/g) in
residential areas compared to a region in southern lower Saxony with nonelevated levels of arsenic in soil
(Gebel et al. 1998a). None of the residents were occupationally exposed to arsenic and the arsenic levels
in drinking water were generally below 0.015 mg/L. The mean levels of arsenic in urine and hair were
lower in the reference area than in the former mining area (see Table 6-6), although within the mining
area, there was a slight increase in arsenic levels in hair and arsenic excreted in urine with increasing
arsenic content in soil. Children in the Palatine region did not have higher contents of arsenic in their hair
or urine. The most significant factor contributing to elevated levels of arsenic in hair and urine was
seafood consumption. In the combined population of people living in mining areas containing high levels
of arsenic in soil and other areas, the level of arsenic in urine was positively associated with the extent of
seafood consumption. However, the study also showed that seafood consumption does not lead to an
extreme increase in excretion of arsenic in the urine. There are apparently other, unidentified factors
affecting the urine levels. Only arsenic in urine, not in hair, was significantly correlated with age. The
level of arsenic in urine was very slightly, but significantly correlated with the consumption of home-
grown produce. Tobacco smoking had no correlation with the arsenic content of either hair or urine
(Gebel et al. 1998a).
ARSENIC 361
6. POTENTIAL FOR HUMAN EXPOSURE
Table 6-6. Levels of Arsenic in Human Tissue and Urine—Recent Studies
Concentration
Site population Sample Mean
a
Range Units Reference
Fort Valley, Georgia, Pesticide manufacturing facility (Superfund site)
40 workers (samples collected
Urine, random 11.6 <1–57 μg/L Hewitt et al.
at end of work week)
Urine, 24-hour 11.0 <1–54 μg/L
1995
Hair 0.78 <0.01–6.3 μg/g
Fingernails 0.79 <0.01–6.1 μg/g
Hermosa, Sonora, Mexico
Children, ages 7–11, exposed
Urine, 24-hour Wyatt et al.
to arsenic in water (mean
1998a,
concentration [mean dose]):
1998b
9 μg/L [0.481 μg/kg/day] 10.26 4.05–19.68 μg/day
15 μg/L [0.867 μg/kg/day] 10.54 2.82–20.44 μg/day
30 μg/L [1.92 μg/kg/day] 25.18 5.44–93.28 μg/day
Glasgow, Scotland
Adults, normal (n=1,250) Hair 0.650 0.20–8.17 μg/g
Raie 1996
Adults, postmortem (n=9) Liver 0.048 [0.024] 0.011–0.152 μg/g
Infants, postmortem (n=9) Liver 0.0099 [0.007] 0.0034–0.019 μg/g
Adults, postmortem (n=8) Lung 0.044 [0.022] 0.0121–0.125 μg/g
Infants, postmortem (n=9) Lung 0.007 [0.0055] 0.0011–0.015 μg/g
Adults, postmortem (n=9) Spleen 0.015 [0.008] 0.001–0.063 μg/g
Infants, postmortem (n=8) Spleen 0.0049 0.0011– μg/g
[0.0045] 0.0088
Palatinate Region, Germany (high As)
b
Residents (n=199) Urine, 24-hour 3.96 [3.21] <0.1–18.32 μg/g Gebel et al.
Residents (n=211) Hair 0.028 [0.016] <0.005–0.154 μg/g
1998a
Saxony, Germany (low As—reference)
b
Residents (n=75) Urine, 24-hour 7.58 [6.20] 0.29–23.78 μg/g Gebel et al.
Residents (n=74) Hair 0.069 [0.053] 0.013–0.682 μg/g
1998a
Ismir, Turkey, (volcanic area with high thermal activity)
Nonoccupationally exposed Breast milk 4.23 [4.26] 3.24–5.41 μg/L Ulman et al.
women (n=35) 1998
Erlangen-Nurenberg Germany 1/92–12/93
Nonoccupationally exposed Lung 5.5 <1–13.0 ng/g Kraus et al.
people (n=50) ww 2000
28.4 <1–73.6 ng/g
dw
ARSENIC 362
6. POTENTIAL FOR HUMAN EXPOSURE
Table 6-6. Levels of Arsenic in Human Tissue and Urine—Recent Studies
Concentration
Site population Sample Mean
a
Range Units Reference
Tarragona (Catalonia, Spain) 1997–1999
Nonoccupationally exposed Lung <0.05 μg/g
Garcia et al.
people (n=78) ww
2001
Bone <0.05
Kidney <0.05
Liver <0.05
Lung <0.05
West Bengal, India
Residents consuming arsenic- Fingernail 7.32 2.14–40.25
μg/g Mandal et
contaminated water (n=47)
al. 2003
Hair 4.46 0.70–16.17
Residents consuming
Fingernail 0.19 0.11–0.30
nonarsenic-contaminated water
(n=15)
Hair 0.07 0.03–0.12
Middleport, NY, USA
Children <7 years (n=77) Urine 15.1
c
2.1–59.6
μg/L Tsuji et al.
Children <13 years (n=142) Urine 15.7
c
2.1–59.9
2005
Children
≥7 years and adults
Urine 15.8
c
3.9–773
(n=362)
All participants Urine 15.7
c
2.1–773
a
Medians, if reported, are in brackets.
b
The reference group (Saxony) had significantly higher levels of arsenic in urine and hair. However, data from both
groups correspond to normal range reference data.
c
Geometric mean, total arsenic
dw = dry weight; ww = wet weight
ARSENIC 363
6. POTENTIAL FOR HUMAN EXPOSURE
A study was performed to look at the arsenic levels, as well as the arsenic species present, in hair and nail
samples from individuals in an arsenic-affected area in West Bengal, India. Mean arsenic concentrations
in hair and fingernails of the chronically arsenic exposed population were 4.46 and 7.32 μg/g, respectively
and were 0.07 and 0.19 μg/g in a control population. Fingernail samples were found to contain mostly
inorganic arsenic (>80%) as a mixture of As(III) and As(V), as well as DMA(III) and DMA(V). Hair
samples also mostly contained inorganic arsenic (>90%), as well as MMA(V) and DMA(V) (Mandal et
al. 2003).
Arsenic in soil in communities surrounding former smelters is a public health concern, especially for
infants and children who may consume significant quantities of soil. Since lead arsenate was used in
apple and other fruit orchards, often at very high application rates, and this compound would be expected
to accumulate and persist in surface soil, there are concerns to human health when these when old
orchards are converted into subdivisions or when they are used to grow food crops or forage. However,
arsenic in soil may be imbedded in minerals or occur as insoluble compounds such as sulfides and
therefore, not be taken up by the body from the gastrointestinal tract. In addition, oxidation of mineral
surfaces may result in armoring the primary mineral grain by a secondary reaction product. Arsenic-
bearing solids are often encapsulated in insoluble matrices such as silica, further diminishing arsenic
availability (Davis et al. 1992).
Sarkar and Datta (2004) examined the bioavailability of arsenic from two soils with different arsenic
retention capacities. In this study, Immokalee (Florida) and Orelia (Texas) soils were incubated after
spiking with sodium arsenate for 4 months. The Immokalee soil is a sandy spodosol with low Fe/Al,
Ca/Mg, and P contents and is likely to have minimal arsenic retention capacity. The Orelia soil is a sandy
clay that is expected to have strong arsenic retention capacity. Arsenic speciation and bioavailability
were studied immediately after spiking and after 4 months of incubation. Approximately 85% of the total
arsenic (soluble and exchangeable fractions) was considered bioavailable and phytoavailable immediately
after pesticide application for the Immokalee soil; after 4 months of incubation, this decreased to
approximately 46%. Immediately after pesticide application, the amounts of arsenic extracted in the
soluble/exchangeable and Fe/Al-bound fractions were similar that of the Immokalee soil. After 4 months,
the soluble arsenic decreased to approximately 45% and the Fe/Al-bound arsenic increased to about 40%.
Experiments looking at the bioavailability of arsenic from these two soils indicated that the potentially
irreversible adsorption of arsenic by the Orelia soil rendered a significant portion of the total arsenic
unavailable for absorption by the human gastrointestinal system. Initially after pesticide application,
ARSENIC 364
6. POTENTIAL FOR HUMAN EXPOSURE
100% of the arsenic was bioavailable; after 4 months, the bioavailable fraction was found to decrease to
88 and 69% in the Immokalee and Orelia soils, respectively (Sarkar and Datta 2004).
Hamel et al. (1998) used synthetic gastric juice to estimate the bioaccessible fraction of metals in the
stomach with varying liquid to solid ratios. They found that the bioaccessibility may vary in different
soils and with varying liquid to solid ratios. Bioaccessibility was defined as the amount of metal that is
soluble in synthetic gastric juice and therefore, potentially available for uptake across the intestinal lumen,
while bioavailability was defined as the amount that was actually taken across the cell membranes.
Arsenic bioaccessibility for National Institute of Standards and Technology (NIST) Montana Soil SRM
2710, with a certified arsenic concentration of 626 μg/g, was fairly consistent across the liquid-to-solid
ratios and ranged from 41.8±18 to 56±21%. The extractability of a hazardous waste contaminated soil
from Jersey City, New Jersey, was different than that observed for the Montana NIST soil. For the Jersey
City soil, which had an arsenic concentration of 1,120 μg/g, there was an increase in the bioaccessible
arsenic as the liquid-to-solid ratio increased. Bioaccessible arsenic ranged from 4.5±0.8 (at a liquid-to-
solid ratio of 100:1) to 25±9% (at a ratio of 5,000:1). Similarly, smelter impacted soils from Anaconda,
Montana contain metal-arsenic oxides and phosphates whose bioaccessibility is limited by solubility
restraints for residence times typical of the gastrointestinal tract (Davis et al. 1992, 1996).
Inhalation of arsenic from ambient air is usually a minor exposure route for the general population. For
example, the dose to a person who breathes 20 m
3
/day of air containing 20–30 ng/m
3
(see Section 6.4.1)
would be about 0.4–0.6 μg/day. However, smokers may be exposed to arsenic by inhalation of
mainstream smoke. Assuming that 20% of the arsenic in cigarettes is present in smoke, an individual
smoking two packs of cigarettes per day would inhale about 12 μg of arsenic (EPA 1984a). However, a
German study of the arsenic levels in lung tissue of 50 unexposed deceased people (see Table 6-6) found
no significant difference in lung arsenic concentrations of smokers versus nonsmokers, nor were there any
significant age- or sex-related differences (Kraus et al. 2000). Before arsenical pesticides were banned,
tobacco contained up to 52 μg As/g, whereas after the ban, maximum arsenic levels were reduced to
3 μg/g.
Occupational exposure to arsenic may be significant in several industries, mainly nonferrous smelting,
arsenic production, wood preservation, glass manufacturing, and arsenical pesticide production and
application. Since arsenic compounds are used as a desiccant for cotton, workers involved in harvesting
and ginning cotton may be exposed to arsenic. Occupational exposure would be via inhalation and
dermal contact. Should any arsenic be retained in the cotton, workers handling the fabric and the general
ARSENIC 365
6. POTENTIAL FOR HUMAN EXPOSURE
public would be exposed. The electronics industry is expanding the use of gallium arsenide in the
production of electro-optical devices and integrated circuits, and workers in the industry where gallium
arsenide is used may be exposed to hazardous substances such as arsenic, arsine, and various acids
(Sheehy and Jones 1993). Occupational exposure to arsenic is generally assessed by measuring urinary
excretion of arsenic. Past exposure is commonly assessed by arsenic levels in hair. Different types of
occupational exposures may result in different uptakes of arsenic because of the bioavailability of the
form of arsenic to which workers are exposed. For example, maintenance workers at a Slovak coal-fired
power plant exposed to 8-hour TWA arsenic air concentrations of 48.3 μg/m
3
(range, 0.17–375.2) had
urinary total arsenic levels of 16.9 μg As/g creatinine (range, 2.6–50.8), suggesting that bioavailability of
arsenic from airborne coal fly ash is about one-third that from in copper smelters and similar settings
(Yager et al. 1997). Approximately 90% of the arsenic-containing particulates were ≥3.5 μm. Apostoli et
al. (1999) monitored 51 glass workers exposed to arsenic trioxide by measuring dust in the breathing
zone. The mean concentration of arsenic in air was 82.9 μg/m
3
(1.5–312 μg/m
3
); exposure was higher for
workers involved in handling the particulate matter. The occupation exposures to principal contaminants,
including arsenic, at five coal-fired power plants were evaluated during June–August 2002. Eight air
samples were collected per similar exposure group at four of the five facilities; inorganic arsenic
concentrations in all samples were below the limit of detection (0.37–0.72 μg/m
3
), as well as being below
the OSHA permissible exposure limit (PEL) of 10 μg/m
3
(Bird et al. 2004).
NIOSH researchers conducted a study of arsenic exposures and control systems for gallium arsenide
operations at three microelectronics facilities during 1986–1987 (Sheehy and Jones 1993). Results at one
plant showed that in all processes evaluated but one, the average arsenic exposures were at or above the
OSHA action level of 5 μg/m
3
, with a maximum exposure of 8.2 μg/m
3
. While cleaning the Liquid
Encapsulated Czochralski (LEC) pullers, the average potential arsenic exposure of the cleaning operators
was 100 times the OSHA PEL of 10 μg/m
3
. Area arsenic samples collected at the plant in break-rooms
and offices, 20–60 feet from the process rooms, had average arsenic concentrations of 1.4 μg/m
3
. At the
other two plants, personal exposures to arsenic were well controlled for all processes evaluated.
A study has been conducted to examine the relationship between total arsenic levels in hair of employees
in a semiconductor fabrication facility and job responsibility, a surrogate variable for arsenic exposure
(de Peyster and Silvers 1995). Airborne arsenic was found in areas where equipment was cleaned but not
in administrative areas. The highest airborne arsenic level found in the study, 15 μg/m
3
, was collected
from the breathing zone of a maintenance employee who was cleaning a source housing over a period of
2 hours in an area with local exhaust ventilation. A concentration of 2 μg/m
3
was found during the
ARSENIC 366
6. POTENTIAL FOR HUMAN EXPOSURE
remainder of the cleaning period (~53 minutes). Workers in maintenance who were regularly assigned to
cleaning equipment, and therefore presumed to have the highest exposure potential, had a mean hair
arsenic level of 0.042 μg/g. This was higher than the mean of 0.033 μg/g observed in administrative
controls, but the difference was not significant. Maintenance workers who only occasionally cleaned and
maintained arsenic-contaminated equipment had a mean hair arsenic level of 0.034 μg/g, which was
comparable to the controls. The highest group mean hair arsenic level of 0.044 μg/g, surprisingly, was
found in supervisors and engineers who were presumed to have the lowest exposure potential of all
workers in the process areas. However, the highest concentrations of hair arsenic in engineers, 0.076 and
0.106 μg/g, were observed in two heavy smokers who smoked 1–2 packs of cigarettes per day. A 2-way
analysis of variance indicated that smoking appeared to be a significant contributing factor whereas
occupational exposure was not.
Hwang and Chen (2000) evaluated arsenic exposure in 21 maintenance engineers (exposed group) and
10 computer programmers (control group) at 3 semiconductor manufacturing facilities. Samples of air,
wipe, and urine, as well as used cleaning cloths and gloves were collected to determine arsenic exposure.
Arsenic was undetectable in 46 of the 93 air samples, and most samples were generally below the
recommended occupational exposure limit (10 μg/m
3
) in work areas during ion implanter maintenance.
Arsine was detectable in 22 of the 45 area air samples and in 15 of the 35 personal air samples; however,
all concentrations were well below the occupational exposure limit of 50 ppb (160 μg/m
3
). Mean arsine
concentrations ranged from not detected to 4.0 ppb (15 μg/m
3
) in area air samples, and the mean arsine
concentration of personal air for maintenance engineers was 4.3 ppb (14 μg/m
3
). Arsenic concentrations
in wipe samples, used cleaning cloths, and gloves, varied from not detected to 146 μg/cm
2
. During ion
implanter maintenance, urinary arsenic levels were found to increase (1.0–7.8 μg/g creatinine) in the
maintenance engineers, from a mean baseline concentration of 3.6 μg/g creatinine. The average urinary
arsenic level for the computer programmers was 3.8 μg/g creatinine (Hwang and Chen 2000). Mean
arsenic concentrations in blood of 103 workers in the optoelectronic industry and 67 controls were
8.58 and 7.85 μg/L, respectively (Liao et al. 2004).
Concentrations of various metals, including arsenic, were measured in autopsy tissues (liver, lung, kidney,
brain, and bone) collected from 78 nonoccupationally exposed subjects from Tarragona County, Spain
between 1997 and 1999. In general, arsenic concentrations were under the analytical detection limit
(0.05 μg/g wet weight) in all tissues (Garcia et al. 2001).
ARSENIC 367
6. POTENTIAL FOR HUMAN EXPOSURE
CCA preservatives are commonly used for treating timber used in constructions in marine and other
humid environments or in contact with the ground. Exposure to CCA compounds may occur through
dermal contact and inhalation of dust while working with the treated timber. Nygren et al. (1992)
investigated the occupational exposure to airborne dust, chromium, copper, and arsenic in six joinery
shops in Sweden where impregnated wood was used for most of their production. The mean airborne
concentration of arsenic around various types of joinery machines ranged from 0.54 to 3.1 μg/m
3
. No
increased concentrations of arsenic were found in the workers’ urine. A study was carried out in
Denmark to evaluate arsenic exposure in taxidermists, workers impregnating wood with CCA solutions,
fence builders, construction workers, and workers impregnating electric pylons with arsenic solution
(Jensen and Olsen 1995). Airborne arsenic exposure was documented in 19 of 27 individuals working
with products containing arsenic. The maximum exposure concentration was 17.3 μg/m
3
, found for a
single worker who was filling an impregnation container with CCA paste. Median exposures for indoor
workers producing garden fences and weekend cottages were 3.7 and 0.9 μg/m
3
, respectively. The
maximum urine concentration reported in the study was 294.5 nanomoles arsenic per millimole creatinine
(195 μg As/g creatinine) and was from the injector impregnating electric pylons. The median
concentration in workers on electric pylons was 80 nanomoles arsenic per millimole creatinine (53 μg
As/g creatinine), which was 6 times the concentration in reference individuals. Urine arsenic levels in
workers producing garden fences and in taxidermists were 2.9 and 1.8 times the reference level,
respectively.
The NIOSH National Occupational Exposure Survey (NOES) conducted in 1981–1983 estimated that
about 55,000 workers were potentially exposed to arsenic (NOES 1990). The NOES was based on field
surveys of 4,490 facilities that included virtually all workplace environments, except mining and
agriculture, where eight or more persons are employed. The principal exposure pathway is probably
inhalation of arsenic adsorbed to particulates, but ingestion and possibly dermal exposure may also be
common. Since arsenic is no longer produced in the United States (see Section 5.1) and many arsenical
pesticide uses have been banned (see Chapter 8), it is likely that the number of workers occupationally
exposed to arsenic has decreased markedly in more recent years.
6.6 EXPOSURES OF CHILDREN
This section focuses on exposures from conception to maturity at 18 years in humans. Differences from
adults in susceptibility to hazardous substances are discussed in Section 3.7, Children’s Susceptibility.
ARSENIC 368
6. POTENTIAL FOR HUMAN EXPOSURE
Children are not small adults. A child’s exposure may differ from an adult’s exposure in many ways.
Children drink more fluids, eat more food, breathe more air per kilogram of body weight, and have a
larger skin surface in proportion to their body volume. A child’s diet often differs from that of adults.
The developing human’s source of nutrition changes with age: from placental nourishment to breast milk
or formula to the diet of older children who eat more of certain types of foods than adults. A child’s
behavior and lifestyle also influence exposure. Children crawl on the floor, put things in their mouths,
sometimes eat inappropriate things (such as dirt or paint chips), and spend more time outdoors. Children
also are closer to the ground, and they do not use the judgment of adults to avoid hazards (NRC 1993).
As with adults, most children are exposed to arsenic largely through their diet. Since the greatest dietary
intake of arsenic is from fish and seafood, infants and young children for whom a substantial part of their
food is milk, would not be exposed to arsenic from dietary sources as much as older children. Even when
mothers consumer large amounts of seafood, there does not appear to be any major transfer of
arsenobetaine, the major form of arsenic in seafood, from seafood to milk (Grandjean et al. 1995).
Arsenic concentrations were very low in human milk sampled from 88 mothers in the Faroe Islands,
where the seafood diet includes pilot whale meat and blubber. The total arsenic concentrations ranged
from 0.1 to 4.4 μg/kg, with a median of 1.6 μg/kg (Grandjean et al. 1995). The arsenic concentration in
the breast milk of 35 women in Ismir, Turkey, a volcanic area with high thermal activity ranged from
3.24 to 5.41 μg/L, with a median of 4.22 μg/L (Ulman et al. 1998). The mean arsenic levels in three
groups of cows in the region that grazed on land impacted by lava and thermal activity were 4.71, 4.46,
and 4.93 μg/L, compared to 5.25 μg/L for cows kept in sheds and fed commercial pellet feed and
municipal water. The arsenic levels in the urine of pregnant women and the cord blood of their infants
were 0.625±0.027 and 0.825±0.079 μg/L, respectively. The authors concluded that there was no harmful
exposure to arsenic in volcanic areas with high arsenic levels from suckling infants or feeding them local
cow’s milk, nor was there harm to the newborns from their mother’s diet. Sternowsky et al. (2002)
analyzed breast milk from 36 women from three different regions in Germany. These regions included
the city of Hamburg, a rural area, Soltau, Lower Saxony, and Munster, the potentially contaminated area.
Arsenic was not detected (<0.3 μg/L) in 154 of 187 samples, with the highest concentration, 2.8 μg/L,
found in a sample from the rural area. The geometric mean arsenic concentrations from the three areas
were comparable. Calculated oral intakes of arsenic were between 0.12 and 0.37 μg/day for an infant at
3 months of age and weighing 6 kg.
According to the FDA study of 1986–1991, the mean daily intakes of arsenic are 0.5 and 0.81 μg/kg body
weight per day for a 6–11-month-old infant and 2-year-old child, respectively (Gunderson 1995b). This
ARSENIC 369
6. POTENTIAL FOR HUMAN EXPOSURE
can be compared to a mean daily intake of 0.51 μg/kg-body weight per day for a 25–30-year-old male
(see Table 6-5). A Total Diet Study, from September 1991 to December 1996, estimated that the average
inorganic arsenic intake for children of various age/sex groups were (age-sex group, total arsenic intake in
μg/day, inorganic arsenic intake in μg/day): 6–11 months, 2.15, 1.35; 2 years, 23.4, 4.41; 6 years, 30.3,
4.64; 10 years, 13.3, 4.21; and 14–16 years (females), 21.8, 5.15; 14–16 years (males), 15.4, 4.51 (Tao
and Bolger 1999). The greatest dietary contribution (76–96%) of total arsenic intake for all age groups
other than infants was seafood. For infants, 41 and 34% of the estimated total arsenic intakes are from
seafood and rice/rice cereals, respectively (Tao and Bolger 1999). Only for toddlers does the intake
approach the World Health Organization’s (WHO) provisional tolerable daily intake (PTDI) for inorganic
arsenic (see Table 6-5). A 1985–1988 Canadian total diet study estimated that 1–4-year-olds ingested
14.9 μg of total arsenic per day compared with 38.1 μg by the average Canadian and 59.2 μg for 20–
39-year-old males (Dabeka et al. 1993). Yost et al. (2004) estimated the mean dietary intake for inorganic
arsenic for children (1–6 years of age) to be 3.2 μg/day, with a range of 1.6–6.2 μg/day for the 10
th
and
95
th
percentiles, respectively. Inorganic arsenic intake was predominantly contributed by grain and grain
products, fruits and fruit juices, rice and rice products, and milk (Yost et al. 2004). Total arsenic and
arsenobentaine concentrations were measured in 16 baby food samples obtained from manufactures in
Spain; total arsenic concentrations ranged from 2.042 to 0.270 μg/g in plaice with vegetables and sole
with white sauce, respectively. Arsenobetaine, which is the arsenical commonly found in fish, accounted
for essentially 100% of the arsenic present in the samples (Vinas et al. 2003).
Arsenic exposure from drinking water may be elevated especially in groundwater from areas where
arsenic occurs naturally in soil such as the western and north central sections of the United States (see
Table 6-3 and Figure 6-2).
Arsenic exposure in communities near mining and smelting facilities or where arsenic had formerly been
applied to agricultural land are a public health concern, especially for infants and children. Since arsenic
remains in the surface soil indefinitely and long past land uses may be forgotten, people may not realize
that they are living in areas where high levels of arsenic may occur in soil. Contaminated soils pose a
particular hazard to children because of both hand-to-mouth behavior and intentional ingestion of soil
(pica) that contains metals and other contaminants (Hamel et al. 1998). In these communities, arsenic
may contaminate carpeting or may have been tracked in from outside. Children may be exposed to this
arsenic while crawling around or playing on contaminated carpeting. Exposure may also result from
dermal contact with the soil, or by inhaling the dust and then swallowing it after mucociliary transport up
ARSENIC 370
6. POTENTIAL FOR HUMAN EXPOSURE
out of the lungs. Because much of the arsenic in soil is embedded in or adsorbed to soil particles or
insoluble, it may not be in a form accessible for uptake by the body.
Hwang et al. (1997b) studied the arsenic exposure of children in Anaconda, Montana, in the vicinity of a
former copper smelter from the summer of 1992 through the summer of 1993. Environmental samples
and first morning voided urine samples from 414 children <72 months old were collected. Attention was
focused on that fraction of the environmental source that was thought to be of the greatest risk to the child
(i.e., arsenic in small particles [<250 μm]) that could most readily adhere to hands and toys and could be
inadvertently ingested. Average arsenic levels in different types of soil ranged from 121 to 236 μg/g.
Several studies have reported mean soil ingestion values for children ranging from 9 to 1,834 μg /day.
Assuming that high arsenic exposure areas have average arsenic levels in soil from 60 to 150 μg/g, the
resulting daily arsenic intake from soil could range from 1 to 275 μg/day per child. The geometric mean
of speciated urinary arsenic (combined As(III), As(V), MMA and DMA) was 8.6±1.7 μg/L (n=289) in the
Hwang study. A nationwide survey on arsenic exposure in the vicinity of smelter sites revealed that
children without excess arsenic exposure had average total urinary arsenic levels ranging from 5 to
10 μg/L (Hwang et al. 1997a). Compared to these values, the mean total urinary arsenic values found in
the Hwang study were markedly higher, but they were still well below the WHO-recommended
maximum excretion level for total arsenic of 100 μg/L as an action level for intervention. The
investigators hypothesized that the relatively low urinary arsenic levels found in the study were probably
a reflection of the low bioavailability of some forms of arsenic in contaminated soil. Hwang et al.
(1997a) stated that arsenic intake through skin contact is insignificant and may be neglected in the
assessment of childhood arsenic exposure. They recommend that parents or guardians pay more attention
to their children’s activity, especially hand-to-mouth behavior, even though the environmental
contaminants might be elevated only slightly. Children in the northern Palatine region of German study,
a former mining area characterized by high levels of arsenic (<2–605 μg/g) in residential areas did not
show higher arsenic levels in their hair or urine than children from a reference area of Germany (Gebel et
al. 1998a).
While CCA registrants voluntarily canceled the production of CCA-treated wood for residential use in
2003, there is a potential for exposure to arsenic from existing structures (Zartarian et al. 2006). Based on
a review of existing studies, Hemond and Solo-Gabriele (2004) estimated that children with contact with
CCA-treated wood may be subjected to doses in the range of tens of micrograms of arsenic per day. The
most important route of exposure appeared to be by hand-to-mouth activities after contact with the CCA-
treated wood. Kwon et al. (2004) compared the amounts of water-soluble arsenic on hands of children in
ARSENIC 371
6. POTENTIAL FOR HUMAN EXPOSURE
contact with CCA-treated wood structures or sand in playgrounds. The mean amount of water-soluble
arsenic on children’s hands from playgrounds without CCA-treated wood was 0.095 μg (range 0.011–
0.41 μg). A mean amount of water-soluble arsenic on children’s hands from playgrounds with CCA-
treated wood was 0.5 μg (range 0.0078–3.5 μg) (Kwon et al. 2004). Additional data from the study by
Kwon et al. (2004) showed that total arsenic collected in hand-washing water (insoluble arsenic on the
filter combined with the water-soluble arsenic in the filtrate) was 0.934 and 0.265 μg for the CCA
playgrounds and the non-CCA playgrounds, respectively (Wang et al. 2005). Two wood surface swab
samples collected from 217 play structures constructed from CCA-treated wood in the City of Toronto,
Canada were sampled and analyzed for inorganic arsenic (Ursitti et al. 2004). Dislodgeable arsenic
concentrations were found to vary widely from nondetectable (0.08–0.25 μg/100 cm
2
) to
521 μg/100 cm
2
(mean = 41.6 μg/100 cm
2
), and were found to not be a useful predictor of soil arsenic
levels (Ursitti et al. 2004).
Shalat et al. (2006) evaluated postexposure hand rinses and urine for total arsenic for 11 children (13–
71 months) in homes in Miami-Dade County, Florida, with and without CCA-treated playgrounds. Seven
playgrounds were included in this study, and five of these contained either CCA-treated or partially CCA-
treated wood. In addition, samples of wood, soil (5–8 cm from the base of the playground structure),
mulch (when present), and synthetic wipes were analyzed for total arsenic. Wood and soil arsenic
concentration were <2.0 and <3 mg/kg for the non-CCA-treated playgrounds, respectively. Mean arsenic
concentrations of 2,380 mg/kg (range 1,440–3,270 mg/kg) and 19 mg/kg (4.0–42 mg/kg) were reported
for wood and soil, respectively, in the playgrounds with CCA-treated wood. An arsenic concentration in
mulch at one playground without CCA-treated wood was 0.4 mg/kg, and arsenic concentrations were
0.6 and 69 mg/kg in mulch at two of the playgrounds with CCA-treated wood. The amount of arsenic
removed by synthetic wipes from the non-CCA-treated wood was <0.5 μg, while the mean amount of
arsenic removed from the CCA-treated wood was 117 μg (range 1.0–313 μg). The amount of arsenic in
hand rinses from children who played at the playgrounds with non-CCA-treated wood and at the
playgrounds with CCA-treated wood were <0.2 and 0.6 μg (range <0.2–1.9 μg), respectively. The mean
urinary total arsenic concentration was 0.0136 μg/L (range 0.0072–0.0231 μg/L) for all children. No
association between assess to CCA-treated playgrounds and urinary arsenic levels was found (Shalat et al.
2006).
The potential exposure children may receive from playing in play structures constructed from CCA-
treated wood is generally smaller than that they would receive from food and water. For comparison,
Yost et al. (2004) estimated the mean dietary intake for inorganic arsenic for children (1–6 years of age)
ARSENIC 372
6. POTENTIAL FOR HUMAN EXPOSURE
to be 3.2 μg/day, with a range of 1.6–6.2 μg/day for the 10
th
and 95
th
percentiles, respectively. In a Total
Diet Study, from September 1991 to December 1996, estimated average total intakes for children aged 6–
11 months, 2 years, 6 years, and 10 years were 2.15, 23.4, 30.3, and 13.3 μg/day, respectively. Average
inorganic arsenic intakes for the same age groups were estimated as 1.35, 4.41, 4.64, and 4.21 μg/day,
respectively, based on data for total arsenic in foods and the assumption that 10% of the total arsenic in
seafood was inorganic and that 100% of the arsenic in all other foods was inorganic (Tao and Bolger
1999). Hand washing after play would reduce the potential exposure to children to arsenic after playing
on play structures constructed with CCA-treated wood, since most of the arsenic on the children’s hands
was removed with water (Kwon et al. 2004).
Concentrations of several toxic metals, including the metalloid arsenic, were measured in the placentas of
200 women in two urban cities in Ukraine, Kyiv and Dniprodzerzhinsk. Arsenic was detected in only 5%
of the samples with concentrations ranging from <0.156 to 0.378 μg/g. In a study in Bulgaria, placental
arsenic concentrations of 7 and 23 μg/g were reported in a control and smelter area, respectively. A
placental arsenic concentration of 34 μg/g was reported in a region of Argentina with high concentrations
of arsenic in drinking water (Zadorozhnaja et al. 2000).
Parents can inadvertently carry hazardous materials home from work on their clothes, skin, hair, and
tools, and in their vehicles (DHHS 1995). Falk et al. (1981b) reported a case of hepatic angiosarcoma in a
child that could be associated with arsenic contamination of a parent’s clothing, the water supply, and the
environment. The father worked in a copper mine and smelter area where his clothing was contaminated
with dust containing arsenic. His daughter, who exhibited a high degree of pica, ate soil from the yard,
and licked soil off her father’s shoes. In a study of arsenic levels in homes in Hawaii, Klemmer et al.
(1975) found higher levels in homes of employees of firms that used arsenic for pesticides or wood
preservation, compared to homes where residents’ work did not involve arsenic. The concentration of
arsenic in dust from the homes of workers exposed to arsenic ranged from 5.2 to 1,080 μg/g, compared to
concentrations of 1.1–31 μg/g in dust from control homes.
While the harmful effects of many components of tobacco smoke are well known, those due to heavy
metals in the smoke have not been sufficiently emphasized. The adverse health effects of these toxic
metals on the fetus through maternal smoking are of special concern (Chiba and Masironi 1992). The
concentration of arsenic in tobacco is relatively low, usually below detectable limits (<1 μg/g). Although
the concentrations of inorganic and organic arsenic in the urine of adults do not appear to be influenced
by smoking, a positive association was found between urinary arsenic levels in children and parental
ARSENIC 373
6. POTENTIAL FOR HUMAN EXPOSURE
smoking habits. As detailed in a WHO report, the mean arsenic level in the urine of children of
nonsmoking parents was 4.2 μg/g creatinine, in children with one smoking parent, it was 5.5 μg/g, and in
children with both parents smoking, it was 13 μg/g (Chiba and Masironi 1992). Tsuji et al. (2005)
reported geometric mean concentrations of total arsenic of 15.1 and 15.7 μg/L in children <7 and
<13 years old, respectively, from households in Middleport, New York, where historical pesticide
manufacture was associated with arsenic in soil. Geometric mean concentrations of inorganic arsenic,
MMA, and DMA were 0.81, 0.54, and 2.5 μg/L, respectively, in children <7 years old and 0.83, 0.55, and
3.0 μg/L, respectively, in children <13 years old (Tsuji et al. 2005).
The use of Chinese herbal medicines (CHM) appears to be common among Chinese women. Both CHM
and Chinese proprietary medicines (CPM) are used for treatment of minor ailments in babies and
children. Herbal medicines are available in capsule or tablet form in drug stores, supermarkets, and by
mail. The CPM “Sin Lak Pill,” “Lu Shen Wan,” and other anti-asthma preparations have been found to
contain inorganic arsenic levels ranging from 25 to 107,000 μg/g, and cases of acute arsenic poisoning
have been found in children and adults using these CPM (Chan 1994). Babies and children are
particularly at risk because they may be given higher doses of these preparations per kg of body weight
than adults would normally consume. They may also lack the hepatic enzymes responsible for drug
biotransformation and detoxification (Chan 1994). Concentrations of heavy metals, including arsenic,
were evaluated in 54 samples of Asian remedies that were purchased in stores in Vietnam and Hong Kong
that would be easily accessible to travelers, as well as in health food and Asian groceries in Florida, New
York, and New Jersey. One remedy that was recommended to treat children’s fever would expose a
15 kg child to approximately 5.0 mg of arsenic per day (Garvey et al. 2001). A folk remedy, purchased in
California, for the treatment of chicken pox, flu-like symptoms, and nasal congestion, which had been
given to two children in Wisconsin, was found to contain 36% arsenic acid. One-half teaspoon of this
powder (about 500 mg of arsenic) was dissolved in hot water and taken 2–3 times per day (Werner et al.
2001).
Various metallic pigments and colors in the form of salts or lakes are used in toy production. Therefore,
children may be exposed to toxic metals while playing with toys, especially when they lick, suck, or
swallow a toy or a piece of a toy. Toys produced in European Union Markets must conform to
restrictions concerning the bioavailability of toxic metals, including arsenic. The maximum limit for
bioavailability of arsenic from the accessible parts of a toy is set to 0.1 μg/day. This corresponds to an
arsenic migration limit of 25 μg/g for all toy material, including modeling clay and paints (Rastogi and
Pritzl 1996). A study was carried out to determine whether crayons, water colors, and water-based paints
ARSENIC 374
6. POTENTIAL FOR HUMAN EXPOSURE
conform with the migration limits for toxic metals (Rastogi and Pritzl 1996). For the analysis, 94 samples
representing 48 products were obtained from China, Taiwan, Japan, the United States, and European
countries. Fifty-two samples showed migration of arsenic, ranging from 0.01 to 3.75 μg/g.
6.7 POPULATIONS WITH POTENTIALLY HIGH EXPOSURES
In addition to individuals who are occupationally exposed to arsenic (see Section 6.5), there are several
groups within the general population that have potentially high exposures (higher than background levels)
to arsenic. These populations include individuals living in proximity to sites where arsenic was produced,
used (e.g., as a pesticide), or disposed, and individuals living near one of the 1,684 NPL hazardous waste
sites where arsenic has been found at elevated levels in some environmental media (HazDat 2006). It also
includes point sources such as smelters, coal-fired power plants, and municipal incinerators. People
living in areas of volcanic activity may be exposed to higher levels of arsenic since high levels are more
likely to be present in the environment. Other populations at risk of potentially high levels of exposure
include those whose water supply contains high levels of arsenic and those consuming large amounts of
seafood or seaweed. However, as pointed out previously (see Section 6.4.4), arsenic in fish and shellfish,
is largely in the form of the less harmful organic arsenical, arsenobetaine; however, some commercially
available seaweeds, especially brown algae varieties, may have high percentages of the total arsenic
present as inorganic arsenic (>50%) (Almela et al. 2002; Laparra et al. 2003). While elevated urinary
arsenic excretion levels have been associated with the consumption of fish and seafood, in a study of
32 sport fish consumers from Lakes Erie, Huron, and Michigan, only 6 (19%) had detectable urine arsenic
concentrations, >4 μg/L, and 5 of these consumed fish from Lake Huron (Anderson et al. 1998).
Exposure of high levels of arsenic in drinking water is more apt to be absorbed by the body and be
harmful than exposure to arsenic in seafood. For example, a group of 36 people in Zimapán, Mexico who
consumed water from an aquifer with 1.0 mg As/L had hair arsenic levels of 2.6–14.1 μg/g (10 μg/g
average), compared with 2.4–13.9 μg/g (6.19 μg/g average) for a reference population that consumed
bottled water with <0.014 mg/L arsenic (Armienta et al. 1997).
A study was conducted to determine if significant arsenic exposure was occurring at a Superfund site in
Fort Valley, Georgia (Hewitt et al. 1995). Random urine, 24-hour urine, hair, and fingernail samples
were collected at the end of the workweek from 40 employees at an active pesticide manufacturing
facility where arsenical pesticides had been produced for over 50 years prior to the mid-1970s.
Measurement of arsenic in the urine is considered to be the best method for monitoring recent exposure in
industrial populations. Hair and fingernail analyses may provide an indication of exposures that occurred
ARSENIC 375
6. POTENTIAL FOR HUMAN EXPOSURE
up to several months prior to testing, but both can adsorb and strongly retain arsenic from external
sources. Since arsenic is rapidly cleared from the blood (half-life of 3–4 hours), blood arsenic levels are
not considered suitable for monitoring populations for chronic low-level arsenic exposure. Results of the
Hewitt study are summarized in Table 6-6. Urinary arsenic levels for all workers were well within the
commonly accepted normal range of <100 μg/L.
As noted above, workers in a number of industries may have high exposures to arsenic, especially if
proper safety procedures are not followed. For members of the general population, above-average
exposure to arsenic from drinking water is possible in areas of high natural arsenic levels in groundwater
or elevated arsenic levels in drinking water due to industrial discharges, pesticide applications, or leaching
from hazardous waste facilities. Individuals living in the vicinity of large smelters and other industrial
emitters of arsenic may be exposed to above-average arsenic levels both in the air, and as a result of
atmospheric deposition, in water and soil and subsequent uptake into crops.
People sawing or drilling arsenic-treated wood without protective masks or burning this wood may be
exposed to elevated levels of arsenic in air.
Recreational and subsistence fishers who consume appreciably higher amounts of locally caught fish from
contaminated bodies of water may be exposed to higher levels of arsenic associated with dietary intake.
Arsenic contamination has triggered the issuance of several human health advisories (EPA 1998g). As of
December 1997, arsenic was identified as the causative pollutant in a restricted consumption advisory for
the general population for all fish in a 7-mile area including Devil’s Swamp Lake and Bayou Baton
Rouge in Louisiana. A public health advisory has been issued for consumption of fish and shellfish from
the Duwamish River, Seattle, Washington due to arsenic and other chemicals (WSDOE 2005).
6.8 ADEQUACY OF THE DATABASE
Section 104(i)(5) of CERCLA, as amended, directs the Administrator of ATSDR (in consultation with the
Administrator of EPA and agencies and programs of the Public Health Service) to assess whether
adequate information on the health effects of arsenic is available. Where adequate information is not
available, ATSDR, in conjunction with NTP, is required to assure the initiation of a program of research
designed to determine the health effects (and techniques for developing methods to determine such health
effects) of arsenic.
ARSENIC 376
6. POTENTIAL FOR HUMAN EXPOSURE
The following categories of possible data needs have been identified by a joint team of scientists from
ATSDR, NTP, and EPA. They are defined as substance-specific informational needs that if met would
reduce the uncertainties of human health assessment. This definition should not be interpreted to mean
that all data needs discussed in this section must be filled. In the future, the identified data needs will be
evaluated and prioritized, and a substance-specific research agenda will be proposed.
6.8.1 Identification of Data Needs
Physical and Chemical Properties. The chemical and physical properties of the arsenic species of
chief toxicological and environmental concern are sufficiently well characterized to allow estimation of
the environmental fates of these compounds. However, more information regarding the K
ow
and K
oc
values of the organic arsenicals would help predict the fate of these compounds in the environment.
Production, Import/Export, Use, Release, and Disposal. According to the Emergency Planning
and Community Right-to-Know Act of 1986, 42 U.S.C. Section 11023, industries are required to submit
substance release and off-site transfer information to the EPA. The TRI, which contains this information
for 2004, became available in May of 2006. This database is updated yearly and should provide a list of
industrial production facilities and emissions.
While arsenic has not been produced in the United States since 1985, the United States is the largest
consumer of arsenic and substantial quantities of arsenic are imported, primarily as arsenic trioxide
(USGS 2006a). The agricultural use of inorganic arsenic pesticides have been discontinued in the United
States. However, some organic arsenicals still may be used in agriculture. Current production and use
data for individual arsenical pesticides and other arsenic compounds would help to estimate human
exposure to the various arsenic species. Because arsenical pesticides are so persistent, a more complete
picture of past use of these products would enable us to predict what areas may contain high levels of
arsenic in soil.
Comprehensive estimates on emissions of arsenic date to the early 1980s (Nriagu and Pacyna 1988). The
industrial picture has changed considerably since then and emission controls are being mandated more
and more. For example, emission factors for Canadian smelters calculated in 1993 were grossly lower
than those estimated in 1983 (Skeaff and Dubreuil 1997). There is a need for accurate and up-to-date
measurements of atmospheric arsenic releases from both natural and anthropogenic sources to better
assess human exposure to arsenic and guide environmental protection measures.
ARSENIC 377
6. POTENTIAL FOR HUMAN EXPOSURE
Environmental Fate. The interconversion of the various arsenic species and transport among the
environmental media is complex and not all aspects are well-studied. Additional quantitative data on the
rates of oxidation, reduction, and biotransformation reactions of arsenic compounds, and how these
depend on environmental conditions would be useful in evaluating and predicting the fate and transport of
arsenic at hazardous waste sites and other areas.
Bioavailability from Environmental Media. Toxicokinetic and toxicity studies establish that
bioaccessible (e.g., soluble, not strongly adsorbed to soil or embedded in minerals) arsenic is highly
absorbed following inhalation and oral exposure (see Sections 3.4.1.2 and 3.4.1.1). Some work has been
done on the effect of environmental matrix (soil, food) on accessibility and absorption of arsenic (Davis et
al. 1992, 1996; Hamel et al. 1998; Roberts et al. 2002, 2007), but additional data would be valuable.
Limited data suggests that dermal absorption of arsenic is very low (see Section 3.4.1.3) (Lowney et al.
2005), further data would be useful to establish whether arsenic uptake occurs from contact with
contaminated soil or water, since humans may be exposed by these routes near hazardous waste sites.
Food Chain Bioaccumulation. Bioconcentration factors have been measured for several freshwater
and marine species. While some species (mainly marine algae and shellfish) tend to bioconcentrate
arsenic (EPA 1980a; Roper et al. 1996), it is not biomagnified through the food chain (Eisler 1994; EPA
1979, 1982b, 1983e, 2003b; Williams et al. 2006).
Carrots growing on land containing somewhat more than the permissible of arsenic in crop land did not
contain levels of arsenic that were harmful (Helgesen and Larsen 1998). However, further research on the
uptake of arsenic by a variety of plants in a wide range of arsenic polluted sites (e.g., mining area,
orchards previously treated with lead arsenate) would be valuable in assessing human exposure near such
sites through the consumption of vegetables from home gardens.
Exposure Levels in Environmental Media. Additional reliable monitoring data for the levels of
arsenic in contaminated media at hazardous waste sites are needed or need to be made available, so that
the information obtained on levels of arsenic in the environment can be used in combination with the
known body burden of arsenic to assess the potential risk of adverse health effects in populations living in
the vicinity of hazardous waste sites.
ARSENIC 378
6. POTENTIAL FOR HUMAN EXPOSURE
Extensive monitoring data are available for total arsenic in all environmental media. Additional
monitoring studies, specifically those that include identification of arsenic species, would allow more
precise estimation of current exposure levels and possible human health risks.
Exposure Levels in Humans. Arsenic has been detected in human tissues, including blood, urine,
hair, nails, and internal organs. Data are available for populations exposed in the workplace and for the
general population (de Peyster and Silvers 1995; Jensen and Olsen 1995; Nygren et al. 1992), and some
studies have been published on exposures near waste sites (Hwang et al. 1997a; Tsuji et al. 2005).
Additional biomonitoring studies of residents near waste sites that contain arsenic would be helpful in
evaluating the likely human health risks from these sites.
While some data are available on the speciation of arsenic in food, additional data on the particular
species of arsenic, rather than just the total arsenic concentration, present in foods, especially seafood, are
needed to better estimate the potential hazards to human health by the consumption of these foods (Ryan
et al. 2001).
This information is necessary for assessing the need to conduct health studies on these populations.
Exposures of Children. Contaminated soils pose a particular hazard to children because of pica and
hand-to-mouth activities. Some studies have been performed on exposure and body burden (Hwang et al.
1997a), but additional studies, including investigations of unique pathways for exposures of children and
the amount of soil a child ingests, would provide valuable data. Small amounts of arsenic were found to
be transferred to hands of children playing on play structures constructed from CCA-treated wood
(Hemond and Solo-Gabriele 2004; Kwon et al. 2004; Shalat et al. 2006; Ursitti et al. 2004; Wang et al.
2005). Based on a review of existing studies, Hemond and Solo-Gabriele (2004) estimated that children
with contact with CCA-treated wood may be subjected to doses in the range of tens of micrograms of
arsenic per day and suggested that exposure by this route warrants further study. The PTDI assigned by
the Food and Agriculture Organization of the United Nations and the World Health Organization
(FAO/WHO) applies to adults. Studies are needed to assess whether children are different in their weight
adjusted intake of arsenic. No childhood-specific means for reducing exposure were identified.
Child health data needs relating to susceptibility are discussed in Section 3.12.2, Identification of Data
Needs: Children’s Susceptibility.
ARSENIC 379
6. POTENTIAL FOR HUMAN EXPOSURE
Exposure Registries. No exposure registries for arsenic were located. This substance is not
currently one of the compounds for which a sub-registry has been established in the National Exposure
Registry. The substance will be considered in the future when chemical selection is made for sub-
registries to be established. The information that is amassed in the National Exposure Registry facilitates
the epidemiological research needed to assess adverse health outcomes that may be related to exposure to
this substance.
6.8.2 Ongoing Studies
The Federal Research in Progress (FEDRIP 2006) database provides additional information obtainable
from a few ongoing studies that may fill in some of the data needs identified in Section 6.8.1. These
studies are summarized in Table 6-7.
The U.S. Geological Survey, along with other federal and state agencies, industry, and academia, is
conducting the National Geochemical Survey (NGS) in order to produce a body of geochemical data for
the United States based primarily on stream sediments that have been analyzed using a consistent set of
analytical methods. The goal of the NGS is to analyze at least one stream sediment sample in every
289 km
2
area by a single analytical method across the entire United States (USGS 2007b).
EPA is conducting a 4-year (2000–2003) national screening-level study of contaminants in freshwater
fish, referred to as the National Fish Tissue Study (EPA 2004c). This study will allow the EPA to
develop national estimates of the mean concentrations of 268 chemicals in tissues of fish from lakes and
reservoirs of the coterminous United States. EPA analysis of the data from this study was scheduled to
begin in January 2005, with the final report scheduled to be released in 2006. Interim raw data have been
released each year, and are available from EPA. Fish samples have been analyzed for total inorganic
arsenic (As(III) and As(V) combined), arsenic(III), arsenic(V), MMA(V), and DMA(V). Analysis for
total arsenic was not performed as part of this study.
The American Water Works Association Research Foundation (AWWARF) supports research on arsenic
in drinking water.
ARSENIC 380
6. POTENTIAL FOR HUMAN EXPOSURE
Table 6-7. Ongoing Studies on the Environmental Fate and Exposure of Humans
to Arsenic
Investigator Affiliation Research description Sponsor
Basta, NT Ohio State University,
School of Natural
Resources, Columbus, Ohio
Heavy metal and trace element
biogeochemistry in soils;
chemical speciation,
bioavailability, and toxicity
USDA
Blum, CB Columbia University Health
Sciences, New York, New
Bioavailability lead and arsenic
in soil to humans
NIEHS
York
Hamilton, JW Dartmouth College,
Hanover, New Hampshire
Toxic metals—biological and
environmental implications
NIEHS
Hoppin, J Not specified Monitoring of arsenic and other
compounds in the blood and
urine of a cohort of pregnant
women in Norway
NIEHS
Kpomblekou, AK;
Ankumah, RO
Tuskegee University,
Agriculture and Home
Economics, Tuskegee,
Alabama
Biochemical processes in soils
treated with trace-element-
enriched broiler litter; to
determine total arsenic and
USDA
other metal concentrations and
the distribution of their chemical
forms in soils under long-term
broiler litter treatments
Loeppert, RH Texas A&M University, Soil
and Crop Sciences, College
Station, Texas
Inorganic chemical processes
influencing soil and water
quality
USDA
Peryea, FJ Wenatchee Tree Fruit
Research & Extension
Center Washington State
University, Pullman,
Washington
Quantification of
biogeochemical processes in
lead arsenate-contaminated
orchard soils and development
of soil and plant management
practices to minimize the
toxicity risks that these soils
impose on agricultural crops
and to human and
USDA
environmental health
Miller, DM; DeLaune,
P; Miller, WP
University of Arkansas,
Crop, Soil and
Environmental Sciences,
Arsenic levels in soils of
northwest Arkansas
USDA
Fayetteville, Arkansas
van Geen, A Columbia University,
Lamont-Doherty Earth
Observatory, Palisades,
New York
Studies on arsenic in
groundwater in Bangladesh
NSF
ARSENIC 381
6. POTENTIAL FOR HUMAN EXPOSURE
Table 6-7. Ongoing Studies on the Environmental Fate and Exposure of Humans
to Arsenic
Investigator Affiliation Research description Sponsor
Walker, MJ et al. University of Nevada,
Natural Resources and
Arsenic in Churchill County,
Nevada domestic water
USDA
Environmental Sciences,
Reno, Nevada
supplies
Zheng, Y Columbia University Health
Sciences, New York, New
York
Arsenic mobilization in
Bangladesh groundwater
NIEHS
NIEHS = National Institute of Environmental Health Sciences; NSF = National Science Foundation; USDA = U.S.
Department of Agriculture
Source: FEDRIP 2006
