Health risk assessment for arsenic contaminated soil Brian L. Murphy and Amy P. Toole Gradient Corp. 44 Brattle Street. Cambridge, MA 02138, USA
Paul D. Bergstrom ARCO Coal Company, Denver, Co 82002, USA
Abstract
This paper describes risk assessment methods for two chronic exposure pathways involving arsenic contaminated soil, namely inhalation of fugitive dust emissions over a lifetime, and inadvertent soil/house dust ingestion. The endpoint in the first case is assumed to be lung cancer and in the second case skin cancer. In order to estimate exposures, inhalation rates and soil/dust ingestion rates are estimated for different age groups; indoor/outdoor time budgets for different age groups are developed; and indoor surface dust and air arsenic concentrations are estimated based on outdoor concentration measurements. Differences observed in indoor/outdoor ratios and arsenic containing dust particle size among different types of communities are noted, as well as possible relationship of particle size to bioavailability. Calculations of risk are presented using cancer potency factors developed by the U.S. Environmental Protection Agency, and uncertainties in these toxicity estimates are described based on: (1) evidence that arsenic may be neither a cancer initiator nor promotor, but may act instead as a late stage carcinogen and (2) evidence that the arsenic dose-response relationship for ingestion may be nonlinear at low doses due to increasing methylation of inorganic arsenic. The first of these considerations influences the relative importance ascribed to arsenic doses in different age groups. The latter consideration indicates that the risk estimates described here are probably very conservative.
Introduction The risks associated with chronic exposure to low doses of arsenic is a relevant problem today in some mining communities, where ore extraction, beneficiation, and smelting have given rise to elevated soil arsenic concentrations. There is evidence that at chronic low doses, arsenic exposure may lead to lung and skin cancer (Doull et al. 1980) The purpose of this paper is to determine the potential health risks associated with chronic exposure to arsenic contaminated soils, and to evaluate the relative importance of different exposure pathways. Risk is calculated for two communities, one a mining community, the other a smelting community, in the northwest region of the United States. A mining community as discussed here, is one in which the extraction of ore, or beneficiation of ore, has resulted in contamination of soil by mill tailings and waste rock. A smelting community is one significantly affected by wind carried smelter emissions. The risk evaluation is based on the assumption that risk is a product of exposure and toxicity. The emphasis in this assessment is to make all exposure pathway and toxicity estimates equivalent in terms of degree of conservatism. Special attention is paid in the assessment to
differences in transport properties of arsenic in mining versus smelting communities, and to differences between indoor and outdoor exposure. Calculations are made specifically for lifetime risk of lung cancer from inhaled arsenic, and skin cancer from ingested arsenic.
Exposure via Soil and Dust Ingestion Consider first, exposure to arsenic v i a ingestion of contaminated soils. For residential areas of the study region, relatively complete soil arsenic concentration data are available. A fairly typical surface soil concentration, 500 mg arsenic/kg soil, is taken here for illustrating soil exposure and risk calculations. Extensive indoor housedust concentration measurements for arsenic are not available; however, indoor dust arsenic concentrations may be estimated from outdoor soil arsenic levels based on an analysis of indoor-outdoor soil metal concentration ratios found in other communities. In evaluating indoor-outdoor ratios, a distinction must be made, between the mining and the smelting community. Review of the data indicates that while in both types of communities there is a relationship between indoor and outdoor dust contaminant concentrations, this relationship differs between the two types of communities. These
164
Health risk assessment f o r arsenic containated soil
differences may be due, at least in part, to different physical and chemical properties of the dust particles produced by the different operations. It has been suggested that the surface properties and moisture content of the smelter particles allow them more readily to adhere to shoes, clothing, and pets than other particles, and thus to be tracked indoors from outside more readily than other particles. The principal study on which the indoor-outdoor arsenic concentration ratio for mining communities is based, was completed at Park City, Utah where indoor house dust and outdoor garden soil arsenic concentrations were measured and compared for a number of residences (Franzen et al., 1988). Park City is an area in which there are many persons living in and around a region with widespread arsenic tailing contamination. Analysis of the Park City dam reveals the following linear relationship between indoor and outdoor dust arsenic concentrations ~ranzen et al., 1988): Si -- 0.13So + 4.5 mg kg"1
(1)
where Si and So are respectively the concentrations of arsenic in indoor house dust and in outdoor soil in mg/kg. The constant 0.13, indicating the indoor dust arsenic concentration is almost 15% that of outdoor soil, may be thought of as a "transfer coefficient." The constant 4.5 mg/kg in equation (1), suggests there is a small contribution to the indoor dust arsenic concentrations, that is not accounted for by the transfer coefficient. This could be due to the indoor generation of arsenic containing dust. Tobacco smoke and emissions from gas stoves, for example, may contain low levels of arsenic. Interestingly, a similar relationship is obtained by looking at indoor and outdoor, dust/soil lead concentrations in mining communities. Barltrop et al. (1975, 1988), measured lead concentrations in garden soil and house dust for residences in two mining communities in England; Derbyshire, and Shipham/North Petherton. After aggregating the soil values in each community into "low", "medium" and "high categories representative of, below 1,000, between 1,000 and 10,000 and greater than 10,000 mg kgl lead in soil, respectively, Steel et al. (1989) plotted the data and by linear regression obtained the equation: Si = 0.15So + 500 mg kg1
(2)
Similar to the results for arsenic, the indoor-outdoor dust transfer coefficient determined for lead is 0.15. Not surprisingly, this equation also indicates, that the indoor generation of lead containing dust is more significant than the indoor generation of arsenic; approximately 500 mg lead/kg housedust is attributable to indoor sources versus 4.5 mg arsenic/kg housedust. Dust from lead containing paint chips may be responsible for this large background concentration. In contrast to mining communities, the relationship between outdoor soil and indoor dust concentrations of arsenic in smelting communities appears to be approximately one to one; that is Si = So. While relatively few researchers have actually examined indoor-outdoor
dust arsenic concentrations, several have provided evidence that the concentrations of lead in housedust reflect those in soil in smelting communities. This relationship between lead concentrations in dusts and soils has been fairly extensively examined by the U.S. Environmental Protection Agency (U.S. EPA, 1986) in both smelting and urban communities. EPA's summary of the observed concentrations of lead in soil and housedust suggest that the indoor-outdoor concentration ratio in smelter communities is approximately 1:1. In other studies near lead smelters, both Roberts (1974) and Diemel et al. (1981) have reported housedust lead concentrations to be similar to soil lea concentrations. Additional evidence is derived from a comparison of mean lead levels measured in air, soil and dust near a smelter in Shoshone County, Idaho (Yankel et al., 1977; Schilling et al., 1984) before and after the closing of the smelter. Analysis of the data shows that air concentrations measured in three different areas decreased by a factor of two after smelter closing, while soil concentrations remained relatively constant. The average lead concentrations in housedust remained similar to those of soil, though somewhat more scattered geographically. Data collected from a smelter community in Montana provides support for the assumption that arsenic follows the same indoor-outdoor relationship as that observed for lead. The community, which actually no longer exists, was at one time situated very close to the base of smelter stacks. Garden soil and housedust arsenic concentration measurements from this community (U.S. EPA, 1986a) while limited, suggests that the estimate Si = So is a relatively good one for these homes. Having obtained concentration estimates for outdoor soil and indoor dust, the next step in calculating exposure to arsenic via soil/dust ingestion, is to estimate intake rates. Ingestion of soil occurs mainly by the transfer of dust from hands to mouth and is generally agreed to be a major exposure route, especially for young children who have frequent hand to mouth contact. Age-specific rates of soil/dust ingestion are estimated as in Table 1. These rates are based on recent so-called "diaper studies" by Calabrese et al. (1988). The diaper studies measured trace elements in children's feces as a key to how much soil children eat. Although two other studies of this nature also exist in the the literature (Binder et al., 1986; Clausing et al., 1987), the Calabrese work is the most recent and corrects some of the errors in methodology made in the earlier studies. Calabrese and coworkers used hospitalized children as a control group, and adult volunteers who ingested soil in capsules, as a check on the assUmptions regarding amounts ingested and excreted. The only value in Table 1 which is directly from Calabrese et al. (1988) is that for 1.5-3 year olds. Other rates, extrapolated from the Calabrese data, are U.S. Environmental Protection Agency (EPA) recommendations as specified in the Superfund Exposure Assessment Manual (U.S. EPA, 1988a). Decreased rates for older children are consistent with observations of dramatically decreased mouthing behavior with age (Barltrop, 1966; Mahaffey, 1977). Note that all estimates are for inadvertent soil ingestion; these are not estimates for children with soil pica.
B. L. Murphy, A. P. Toole and P. D. Bergstrom Table 1 Soil ingestion rates based on U.S. EPA recommendations from the Superfund Exposure Assessment Manual (1988a), with the exception of age range 1.5-3.5 yr, based on Calabrese et al. (1988). As an adult estimate (18-70 yr) was not included in the EPA recommendations, average adult intake is assumed comparable to that of a 5-18 year old individual. Age
mg Total soil ingested~day
165
Table 2 Fraction of waking hours spent indoors and outdoors, by age category, for the 247 days per year during which the ground in the study area is assumed to be non-frozen and devoid of snowpack.
Age (yr)
Fraction of waking Fraction of waking hours outdoors hours indoors
0-1
0.00
1.00
0-9 m
0
1-6
0.25
0.75
9-18 m
50
6-11
0.21
0.79
1.5-3.5 yr
50
11-70
0.06
0.94
3.5-5 yr
50
5-18 yr
10
18-70 yr
10
Preliminary direct evidence that adult soil ingestion rates are significantly lower than those of children, comes from measurements of urinary arsenic levels for residents of a United States smelter community who were relocated from their homes by the EPA in the summer of 1986 (Binder, 1987). Thirty-two people provided urine samples before and after relocation. Results showed that the average pre-move urinary arsenic level for persons greater than 8 years was 17.2 lag L l , and the mean level post-move was 14.6 lag L 1. Six children younger than 8 years had an average pre-move urinary arsenic level of 76 lag L 1 and an average post-move level of 15.3 lag Lq. The excess level of arsenic in the urine, above the 15 lag background level measured after relocation, was 60.7 lag L-l, and 2.6 lag L-1, for children and adults respectively. Correcting for urine volume (i.e. children 1 L/day and adults 2 L/day), the ratio of these differences reveals that the young children had 15 times higher arsenic intake levels than the adults. These excess levels of urinary arsenic in children are attributed to intake of arsenic via soil ingestion. While a quantitative estimate of soil ingestion rates for children and adults can not be confirmed from these results, the significant differences in excess urinary arsenic levels between children and adults, suggests that the soil ingestion rate of older children (8 years) and adults, is much less than half that of younger children. The last piece of information necessary to complete the exposure estimate for soil ingestion, is an estimate of what fraction of total soil ingested each day is attributable to indoor ingestion of house dust and what fraction to outdoor ingestion. Calabrese et al. (1988) report it is experimentally very difficult to distinguish the fraction of outdoor versus indoor soil ingested. Thus, it is assumed that the fraction of soil ingested either indoors or outdoors is proportional to the fraction of waking hours spent in either location. Based on time activity-pattern data (Anderson et al., 1985) and some basic sleeping pattern assumptions
Estimates are based on activity pattern data contained in U.S. EPA (1986b), Anderson et al. (1985), and International Commission on Radiological Protection (1984).
(U.S Department of Health and Human Services, 1988; International Commission on Radiological Protection, 1984) these fractions are estimated as shown in Table 2. In temperate latitudes, these fractions actually only describe the fraction of soil ingested indoors versus outdoors for some number of days in which the ground is non-frozen and devoid of snow pack. In the study region, it is estimated that this condition holds only for 247 days per year CO.S. EPA, 1986a) and that for 118 days per year essentially 100% of an individual's time is spent indoors.
Exposure via Inhalation Similar concentration, intake rate, and activity pattern data to that described for soil ingestion calculations, are necessary to calculate exposure to arsenic via inhalation. The availability of arsenic concentration data for indoor-outdoor air in the study region is analogous to that for soil; outdoor data is relatively complete, but little or no data for indoor air arsenic concentrations exisL A typical, average annual outdoor air concenl~ation of 0.01 lag arsenic/m 3 air is taken here for demonstrating outdoor exposure and risk calculations, and, as done for soil, the indoor arsenic concentration is estimated from that outdoors. Indoor air arsenic concentrations may be thought of as originating primarily from two sources: penetration of outdoor airborne arsenic indoors, and resuspension of house dust containing arsenic. Contributions by these sources to indoor air arsenic concentrations, Ci (lag arsenic/m3), are represented respectively, by the fin'st and second terms in the equation: Ci = 0.3Co + 0.5 Cp (Six 106) Penetration Resuspension
(3)
where Co is the concentration of arsenic in outdoor air ~ g arsenic/m 3 of air), Cp is the indoor resuspended particulate
166
Health risk assessment for arsenic containated soil
Table 3 Indoor and outdoor activity-weighted average ventilation rates as a function of age category and activity level.
Table 4 Fraction of a 24 hour day spent indoors and outdoors, by age category, for the 247 days per year during which the ground in the study area is assumed to be non-frozen and devoid of snowpack.
Activity-weighted ventilation rate (L rain"! ) Age (yr)
Indoor
Outdoor
0-1
0.84
0.84
1-6
10.7
22.7
6-11
13.0
26.8
11 - 18
17.2
28.0
18 - 70
13.1
24.5
Age (yr)
Fraction of day spent outdoors
Fraction of day spent indoors
0-1
0.00
1.00
1-6
0.13
0.87
6-11
0.12
0.88
11 - 70
0.05
0.95
Estimates are based on activity pattern contained in U.S. EPA (1986b), Anderson et al. (1985), and International Commission on Radiological Protection (1984).
Values are based on data contained in Anderson et al. (1985).
(or resuspended housedust) concentration (l.tg suspended housedust/m3 air), Si is the arsenic concentration in housedust (mg arsenic/kg housedust), and 10 -6 is a conversion factor. The constant in the penetration term of equation (3), 0.35, is an average estimate from Freed et al. (1983) based on studies by Alzona et al. (1979) and Cohen and Cohen (1980), of the fraction of outdoor suspended particles that are transported, and remain suspended indoors. There are two major processes that reduce this fraction from 1.0. One is the physical inability of larger particles to penetrate the shell of the house, the other is the greater tendency of particulates to be deposited on surfaces indoors than outdoors. The constant, 0.5, in the resuspension term of equation (3), is based on several studies that suggest approximately 50% of the indoor suspended particulate concentration is due to the resuspension of house dust. Lefcoe and Inculet (1975), for example, measured indoor particulate levels over a two month period in a centrally air-f'dtered, air conditioned home. Outdoor air cleaning equipment was used on a 24-hour-on, 24-hour-off basis with five air changes per hour when the equipment was on. A decrease by about a factor of two, observed for indoor air particulate counts above 0.3 larn when the filtration system was on, is attributed to indoor sources, including resuspension, being comparable to outdoor sources. A similar conclusion is obtained from a study by Moschandrea et al. (1979). During a four day period at a Denver residence, outdoor airborne particulate concentrations of lead and bromine dropped to near zero compared with their levels before and after this period. From their paper, it is estimated that the indoor air levels of lead and bromine, however, stayed at respectively, 0.5 and 0.6 of the average values before and after. As the concentration of lead and bromine in housedust or resuspended housedust, would not be expected to change significantly over this
short time period, the 50 percent decrease in indoor air levels must be due almost entirely to the change in outdoor levels. Thus, it can be concluded that approximately 50 percent of the indoor suspended particulates are from the penetration of outdoor particulates, and approximately 50 percent are from the resuspension of house dust. A value of 50 ~tg m -3 is taken as a default value for the concentration of total indoor suspended particulates, Cp, based on a 39-66 I~g m -3 concentration range measured by Yocom et al. (1977) in residences, and a 50 I.tg m -3 estimate made by Hawley (1985) for Niagara Falls homes. The indoor housedust arsenic concentration, Si for mining and smelting communities was calculated in the previous section. Note that it is multiplied here by a factor of 10-6 to convert the units from mg kg 1 to ktg m -3. Clearly, since Si varies between smelting and mining communities, the indoor air concentration of arsenic may be expected to differ between the two types of communities. Age specific daily intake rates for arsenic in air may be estimated from ventilation rates. Although available data on ventilation rates, especially for children, are incomplete, a partial list o f mean ventilation rates for different ages, sexes and activity levels has been compiled by Anderson et al. (1985). Based on these data, mean ventilation rates for five age categories and four ventilation rates, were calculated. Then, using data contained in the activity data files developed by SRI for EPA's Office of Planning and Standards for the amount of time spent at low, medium and high activity levels indoor and outdoors (Anderson et al., 1985) activity-weighted average ventilation rates as shown in Table 3 were determined. Note that indoor activity-weighted ventilation rates are significantly lower than outdoor rates, as a large fraction of indoor time is spent resting or at low activity levels. Finally, to calculate total daily exposure to arsenic in air, it is necessary to calculate the fraction of time spent indoors versus outdoors. The same assumptions as used in the previous section to calculate indoor versus outdoor soil
B. L. Murphy, A. P. Toole and P. D. Bergstrom
exposure are used here. In contrast to soil exposure, however, individuals are exposed to arsenic in air both while awake and asleep, thus the fraction of time spent indoors and outdoors must be based on a 24 hour day rather than on the number of waking hours. These values are given in Table 4.
Arsenic Toxicity Having evaluated exposure, the second part of the arsenic risk assessment for skin cancer via soil ingestion, and for lung cancer via inhalation, is an evaluation of the toxicity of arsenic. The midpoint of standard EPA recommended arsenic potency factors for ingestion, 1.5 (mg/kg/day) 1 (range 1-2 (mg/kg/day)-1 (U.S. EPA, 1989a), and an EPA potency factor of 50 (mg/kg/day)"1 recommended for arsenic inhalation (U.S. EPA 1984), are used in order to compare calculated risks in this example to EPA allowable risk. There are three aspects of these arsenic potency factors, however, to be kept in mind when using the values. The first point regarding arsenic potency is that there appears to be a strong correlation between the effects of arsenic and the age at which individuals are exposed. Brown and Chu (1983) analysed smelter data for lung cancer in workers laboring in the smelter environment. All workers had been employed for I0 years, but started at different ages, i.e. 20 years of age, 30 years etc. (Brown and Chu (1983) observed a pronounced increase in cancer rates for persons starting work in an arsenic enriched environment at older ages than those starting at younger ages. Lee et al. (1988) have hypothesized that arsenic is neither an initiator nor promoter of cancer, but acts instead in the progression of precancerous cells to malignant ceils, as a gene amplifier. The evidence for the hypothesis of Lee and associates comes from results of their study showing that sodium arsenite and sodium arsenate induced a high frequency of methotrexate-resistant mouse cells which were shown to have amplified copies of the dihydrofolate reductase gene. Oncogene amplification in some cancers has been shown to correlate with the degree of progression of a tumor to malignancy (Lee et al., 1988). If in fact it is in this late stage that the effect of arsenic occurs, this would explain Brown and Chu's findings. Older people who have a higher probability of the initiation and promotion stages of cancer being complete, would have a higher probability of displaying an effect from arsenic exposure. A possible increased effect of arsenic with age is an extremely important point, as it makes the doses later in life more significant in determining risk than doses earlier in life. Although age dependent effective potency could be developed, this has not been done to date. Commonly, as will be done here, lifetime dose is calculated and divided by the adult body weight. An alternative, is to compute a dose for each age category as a function of body weight for that category. In the calculation of skin cancer risk, calculating dose as a function of body weight would produce a risk which is higher than that calculated assuming an adult body weight, by almost a factor of two. The second point regarding arsenic toxicity that bears underscoring is that while the EPA potency factors assume only 30% of inhaled arsenic is bioavailable, due to larger
167
particles being trapped in the nasal and bronchial passages, they assume virtually 100% of ingested arsenic is bioavailable. The assumption that 100% of ingested arsenic is bioavailable is based on studies with animals ingesting arsenic dissolved in water. In the present assessment the ingested matrix is soil for which there is little information on the bioavailability, and some doubt as to whether it is the same as that in water. One piece of evidence, that suggests the bioavailability of arsenic in soil may be significantly less than 100%, is an unpublished study conducted by Tetra Tech (1987). In this study, soil containing arsenic in water adjusted to pH 2, was slurried with hydrochloric acid to simulate leaching of arsenic from soil in the stomach. After six hours, only 60 + 24% of the arsenic had been leached from the soil. While the stomach model is undoubtedly simplified, it nevertheless may be indicative that the estimated 100% bioavailability of ingested arsenic in soil is high. Finally, there is some evidence that the arsenic dose response is not linear, yet the EPA toxicity factors are linearly extrapolated from relatively high doses. The ingestion potency factor is based on an epidemiological study of a Taiwanese population, the validity of which itself has been questioned. (One major concern regarding the study stems from the fact that the drinking water to which the Taiwanese were exposed and which contained high concentrations of arsenic, also contained other chemicals (i.e. ergot alkaloids) capable of producing the same effect (U.S. EPA, 1988b).) The inhalation potency factor is based on occupational exposure studies with smelter workers. The reason for questioning the linear extrapolation of data from high dose exposures, is that a greater fraction of arsenic in urine is methylated at low does than at high doses (Vahter, 1983). Since methylation of arsenic appears to be a detoxification pathway (Marafante and Vahter, 1984) this suggests that the dose/toxic response curve is somewhat steeper at lower doses of arsenic than at higher doses.
Risk Calculations As stated at the outset of this paper, risk is calculated as the product of exposure and toxicity. Using the estimates put forth in preceding sections, daily exposure can be calculated as the product of intake rate and concentration. Toxicity is estimated by potency factors recommended by the EPA. Calculations of lifetime risk may then be completed based on an assumed average lifetime of 70 years, an average adult body weight of 70 kg, and arsenic concentrations in outdoor soil and outdoor air of 500 mg kg 1 and 0.01 [tg m3, respectively. Although it is a very conservative assumption, the calculation assumes that people reside in one location for a lifetime. In fact, an individual in the United States moves an average of once every 9 years (U.S. EPA, 1989b). The risks for skin cancer are discounted by a factor of ten as allowed by the EPA for arsenic induced non-fatal skin cancers in these calculations. Table 5 shows the calculated lifetime risks for indoor and outdoor inhalation and ingestion of arsenic, both in smelter and in mining communities. Several interesting results are evident from a comparison of the numbers in
168
Health risk assessment for arsenic containated soil Table 5 Calculated lifetime risks for exposure to 500 mg kg "1 arsenic in soil and 0.01 ~tg m -3 arsenic in air. Risk Exposure pathway
Smelter community Indoor Outdoor
Mining community Indoor Outdoor
Ingestion
1.2 x 10-5
9.7 x 10-7
1.9 x 10-6
9.7 x 10-7
Inhalation
5.7 x 10-5
2.6 x 10 -6
2.6 x 10-5
2.6 x 10 -6
Total
6.9 x 10-5
3.6 x 10 -6
2.8 x 10-5
3.6 x 10 -6
this table. First, and perhaps most importantly, indoor risk is larger than outdoor risk in both communities, for both types of exposure. This is largely a result of the fact that people spend significantly more time indoors than outdoors. Indoor risks from ingestion are almost seven times lower in the mining community than in the smelter community. Risk is reduced in the mining community, because the equations relating indoor-outdoor concentrations estimate that only 15% of the outdoor arsenic concentration is tracked indoors in the mining community, while 100% is estimated tracked indoors in the smelting community. Indoor inhalation risk is also lower in the mining case, since resuspended housedust exposure constitutes a large portion of the indoor air arsenic concentrations. Almost 81% of the indoor air risk results from resuspension of housedust, in the smelting community; 19% results from material that has been transported airborne through the building shell. In contrast, in the mining community, the percentage of risk coming from resuspended house dust is 40%, while 60% results from transport of airborne material. Primarily as a function of the fact that waste rock and tailings are not tracked indoors as easily as the smaller smelter particles, when indoor and outdoor risks are combined to give an overall risk picture, the mining community risks appear to be somewhat smaller than those in the smelter community. Total lifetime risks due to arsenic exposure in the mining community is 3.2 x 10"5; total risk in the smelter community 7.3 x 10-5 Both risks, however, fall within EPA allowable risk levels of 10-4 to 10-7
Summary and Conclusions The risks associated with arsenic contaminated soils are of special concern to persons living in mining and smelting communities. Careful evaluation of these risks includes assessment of both indoor and outdoor concentrations and intake rates, as well as of the differences between mining and smelting communities. The calculations completed here, using typical outdoor soil and air arsenic concentrations for a mining region in the northwest United States, demonstrate that indoor exposures dominate outdoor. They also indicate that total ingestion risks associated with tailings (mining community) are almost five times less than those associated with smelter emissions
and that inhalation risks associated with tailings are about one-half those of the smelter. Although risk assessments such as this rarely distinguish between the types of sources for contaminant, these results show that such a distinction is extremely important in estimating risk. In this assessment, resuspended surface soils tracked indoors emerge as a primary risk pathway in the smelting community, more so than airborne emissions. The community used for demonstration purposes, however, no longer maintains an operable smelter. This relationship would not necessarily hold in an operating smelter community where concentrations are considerably higher. Although the exposure estimates were not intended to be overly conservative, use of EPA toxicity factors for use in comparing the risk numbers with EPA allowable cancer risks, ultimately incorporated some conservatism. In light of the issues of bioavailability and the body's ability to methylate low doses, risk as calculated here, may actually be somewhat overestimated. Nevertheless, the numbers calculated fall within the ranw of EPA's allowable cancer risk guidelines of 10 -4 to 10" lifetime risk. References Alzoua, J., Cohen, B.L., Rudolph, H., Jow, H.N. and Frohlinger, J.O. 1979. Indoor-outdoor relationships for particulate matter of outdoor origin. Atmospheric Environment, 13, 55-60. Anderson E., Browne, N., Duletsky, S., et al. 1985 Development of Statistical Distributions or Ranges of Standard Factors Used in Exposure Assessments. EPA-600/8-85/010. Report to U.S. EPA Office of Health and Environmental Assessment. Washington, D.C. Barltrop, D. 1966. The prevalence of pica. American Journal of Disease in Children, 112, 116-123. Bafltrop, D., Thorton, I., Strehlow, C.D., and Webb, J.S. 1975. Absorption of lead from dust and soil. Postgraduate Medical Journal, 5, 801-804. Barltrop, D. and Strehlow, C.D. 1988. The contribution from soil and house dust lead to lead burden in childhood. Presented at Conference on Lead in Soil: Issues and Guidelines, March 7-9, at Chapel Hill, N.C. Binder, S., Sokal, D., and Maughan, D. 1986. Estimafng soil ingestion: the use of tracer elements in estimating the amount of soil ingested by young children. Archives of Environmental Health, 41, 341-345. Binder, S. (U.S. Department of Health and Human Services) March 3, 1987. Memo regarding Mill Creek pre- and post-move urinary arsenic levels. Brown, C.C. and Chu, K.C. 1983. A new method for the analysis of cohort studies: implications of the multistage theory of
B. L. Murphy, A. P. Toole and P. D. B e r g s t r o m carcinogenesis applied to occupational arsenic exposure. Environmental Health Perspectives, 50, 293-308. Calabrese, E.J., Pastides, H., Barnes, R. et al. 1988. How Much Soil Do Young Children Ingest: an Epidemiologic study. Draft Report to Syntex Corp. Amherst: University of Massachusetts. Clausing P., Brunekreff, B. and van Wijnen, J.H. 1987. A method for estimating sol] ingestion by children, lnterational Archives for Occupational and Environmental Health, 59, 73-82. Cohen, A.F. and Cohen, B.L. 1980. Protection from being indoors against inhalation of suspended particulate matter of outdoor origin. Atmospheric Environment, 14, 183-184. Diemel, LA.L., Brunekreef, B., Boley, J.S.M., Biersteker, K. and Veenstra, S.J. 1981. The Amhem lead study, lI: indoor pollution and indoor/outdoor relationship. Environmental Research, 25, 449. Donll, J., Klaasen, C.D. and Amdur, M.0. 1980. Casarett and Doull's Toxicology. The Basic Science of Poisons. Macmillan Publishing Co., Inc., New York. Freed, J.R., Chambers, T., Christie, W.N. and Carpenter, C.E. 1983. Method for assessing exposure to chemical substance. U.S. EPA Office of Toxic Substances, EPA 560/5-83- 015. Franzen,D., Sackman, A., Oale, R. and Chapin, M. 1988. Analytical Results Report for Ambient Air and Residental Characterisation at Prospect Square, Park City, Utah. Report Prepared for EPA Hazardous Site Evaluation Division. Hawley, J.K. 1985. "Assessment of Health Risk from Exposure to Contaminated Soft." Risk Analysis, 5(4), 289. International Commission on Radiological Protection. 1984. Report of the Task Group on Reference Man. Oxford: Pergamon Press. Lee, T.C., Tanaka, N., Lamb, P.W., Gilmer, T.M. and Barrett, J.C. 1988. Induction of gene amplification by arsenic. Science, 241, 79. Lefcoe, N.M. and Inculet, I.I. 1975. Particulates in domestic premises. Archives of Environmental Health, 30, 565. Mahaffey, K.R. 1977. Quantities of lead producing health effects in humans: sources and bioavallability. Environmental Health Perspectives, 19, 285-295. Marafante, E. and Vahter, M. 1984. The effect of methyltransferase inhibition on the metabolism of [74As] arsenite in mice and rabbits. Chemical and Biological Interactions, 50, 49-57. Moschandreas, D.J., J.W. Winchester, J.W. Nelson and R.M. Burton. 1979. Fine particle residential indoor air pollution. Atmospheric Environment, 13, 1413. Roberts, T.M. 1974. Lead contamination around secondary smelters: estimation of dispersal and accumulation. Science, 186, 1120. Schilling R., Rass, D., Sokal, D., Inez R., Brokopp, C. and Maughan, A.D. 1984. Children's exposure to smelter-associated lead,
169
Montana and Idaho. Proceedings of the 5th National Conference on Management of Uncontrolled Hazardous Waste Sites, November 7-9, Washington, D.C.p. 239. Steele, M.J., Beck, B.D., Murphy, B.L. and Strauss, H.S. 1989. Assessing the contribution from lead in mining wastes to blood lead. Submitted to Regulatory Toxicology and Pharmacology. Tetra Tech. 1987. Unpublished data. U.S. Department of Health and Human Services (Agency of Toxic Substances and Disease Registry). 1988. The Nature andExtent of Lead Poisoning in Children in the United States: A Report to Congress. U.S. Environmental Protection Agency (Office of Health and Environmental Assessment). 1984. Health Assessment Document for Arsenic. EPA-6OO/8-83-021F. U.S. Environmental Protection Agency (Region VIII). 1986a. Endangerment Assessment: Mill Creek~Montana Smelter Site. Revised Final Report. U.S. Environmental Protection Agency (Office of Air Quality Planning and Standards). 1986b. Rev&w of National Ambient Air Quality Standards for Lead: Assessment of Scientb'ic and Technical Information. Draft Report. U.S. Environmental Protection Agency (Office of Remedial Response). 1988a. Superfund Exposure Assessment Manual. EPA 540/1-88/001. U.S. Environmental Proteaion Agency. 1988b. Special report on ingested inorganic arsenic: Skin cancer Nutritional Essentiality. EPA-625/3-87/013P. U.S. Environmental Protection Agency. 1989a. Integrated Risk Information System (IRIS) (U.S. EPA Online Chemical Potency Information). U.S. Environmental Protection Agency 1989b. Exposure Factors Handbook. EPA-600/8-89/043. Vahter,M. 1983. Metabolism of arsenic. In: Fowler, B.A. (ed.), Biological and Environmental Effects of Arsenic. Elsevier Science Publishers, B.V. Yankel, AJ., Von Lindem, I.H. and Walter, S.D. 1977. The Silver Valley lead study: the relationship between childhood blood levels and environmental exposure. Journal of the Air Pollution Control Association, 27, 766. Yocom, I.E., W.A Cote, and F.B. Benson. 1977. As cited in: Walden R.A. and Scheu P.A., (eds.) ".Indoor Air Pollution. John Wiley & Sons. New York. 1983. (Manuscript No. 193: received July 3, and accepted for publication September 25, 1989)
Paul D. Bergstrom ARCO Coal Company, Denver, Co 82002, USA
Abstract
This paper describes risk assessment methods for two chronic exposure pathways involving arsenic contaminated soil, namely inhalation of fugitive dust emissions over a lifetime, and inadvertent soil/house dust ingestion. The endpoint in the first case is assumed to be lung cancer and in the second case skin cancer. In order to estimate exposures, inhalation rates and soil/dust ingestion rates are estimated for different age groups; indoor/outdoor time budgets for different age groups are developed; and indoor surface dust and air arsenic concentrations are estimated based on outdoor concentration measurements. Differences observed in indoor/outdoor ratios and arsenic containing dust particle size among different types of communities are noted, as well as possible relationship of particle size to bioavailability. Calculations of risk are presented using cancer potency factors developed by the U.S. Environmental Protection Agency, and uncertainties in these toxicity estimates are described based on: (1) evidence that arsenic may be neither a cancer initiator nor promotor, but may act instead as a late stage carcinogen and (2) evidence that the arsenic dose-response relationship for ingestion may be nonlinear at low doses due to increasing methylation of inorganic arsenic. The first of these considerations influences the relative importance ascribed to arsenic doses in different age groups. The latter consideration indicates that the risk estimates described here are probably very conservative.
Introduction The risks associated with chronic exposure to low doses of arsenic is a relevant problem today in some mining communities, where ore extraction, beneficiation, and smelting have given rise to elevated soil arsenic concentrations. There is evidence that at chronic low doses, arsenic exposure may lead to lung and skin cancer (Doull et al. 1980) The purpose of this paper is to determine the potential health risks associated with chronic exposure to arsenic contaminated soils, and to evaluate the relative importance of different exposure pathways. Risk is calculated for two communities, one a mining community, the other a smelting community, in the northwest region of the United States. A mining community as discussed here, is one in which the extraction of ore, or beneficiation of ore, has resulted in contamination of soil by mill tailings and waste rock. A smelting community is one significantly affected by wind carried smelter emissions. The risk evaluation is based on the assumption that risk is a product of exposure and toxicity. The emphasis in this assessment is to make all exposure pathway and toxicity estimates equivalent in terms of degree of conservatism. Special attention is paid in the assessment to
differences in transport properties of arsenic in mining versus smelting communities, and to differences between indoor and outdoor exposure. Calculations are made specifically for lifetime risk of lung cancer from inhaled arsenic, and skin cancer from ingested arsenic.
Exposure via Soil and Dust Ingestion Consider first, exposure to arsenic v i a ingestion of contaminated soils. For residential areas of the study region, relatively complete soil arsenic concentration data are available. A fairly typical surface soil concentration, 500 mg arsenic/kg soil, is taken here for illustrating soil exposure and risk calculations. Extensive indoor housedust concentration measurements for arsenic are not available; however, indoor dust arsenic concentrations may be estimated from outdoor soil arsenic levels based on an analysis of indoor-outdoor soil metal concentration ratios found in other communities. In evaluating indoor-outdoor ratios, a distinction must be made, between the mining and the smelting community. Review of the data indicates that while in both types of communities there is a relationship between indoor and outdoor dust contaminant concentrations, this relationship differs between the two types of communities. These
164
Health risk assessment f o r arsenic containated soil
differences may be due, at least in part, to different physical and chemical properties of the dust particles produced by the different operations. It has been suggested that the surface properties and moisture content of the smelter particles allow them more readily to adhere to shoes, clothing, and pets than other particles, and thus to be tracked indoors from outside more readily than other particles. The principal study on which the indoor-outdoor arsenic concentration ratio for mining communities is based, was completed at Park City, Utah where indoor house dust and outdoor garden soil arsenic concentrations were measured and compared for a number of residences (Franzen et al., 1988). Park City is an area in which there are many persons living in and around a region with widespread arsenic tailing contamination. Analysis of the Park City dam reveals the following linear relationship between indoor and outdoor dust arsenic concentrations ~ranzen et al., 1988): Si -- 0.13So + 4.5 mg kg"1
(1)
where Si and So are respectively the concentrations of arsenic in indoor house dust and in outdoor soil in mg/kg. The constant 0.13, indicating the indoor dust arsenic concentration is almost 15% that of outdoor soil, may be thought of as a "transfer coefficient." The constant 4.5 mg/kg in equation (1), suggests there is a small contribution to the indoor dust arsenic concentrations, that is not accounted for by the transfer coefficient. This could be due to the indoor generation of arsenic containing dust. Tobacco smoke and emissions from gas stoves, for example, may contain low levels of arsenic. Interestingly, a similar relationship is obtained by looking at indoor and outdoor, dust/soil lead concentrations in mining communities. Barltrop et al. (1975, 1988), measured lead concentrations in garden soil and house dust for residences in two mining communities in England; Derbyshire, and Shipham/North Petherton. After aggregating the soil values in each community into "low", "medium" and "high categories representative of, below 1,000, between 1,000 and 10,000 and greater than 10,000 mg kgl lead in soil, respectively, Steel et al. (1989) plotted the data and by linear regression obtained the equation: Si = 0.15So + 500 mg kg1
(2)
Similar to the results for arsenic, the indoor-outdoor dust transfer coefficient determined for lead is 0.15. Not surprisingly, this equation also indicates, that the indoor generation of lead containing dust is more significant than the indoor generation of arsenic; approximately 500 mg lead/kg housedust is attributable to indoor sources versus 4.5 mg arsenic/kg housedust. Dust from lead containing paint chips may be responsible for this large background concentration. In contrast to mining communities, the relationship between outdoor soil and indoor dust concentrations of arsenic in smelting communities appears to be approximately one to one; that is Si = So. While relatively few researchers have actually examined indoor-outdoor
dust arsenic concentrations, several have provided evidence that the concentrations of lead in housedust reflect those in soil in smelting communities. This relationship between lead concentrations in dusts and soils has been fairly extensively examined by the U.S. Environmental Protection Agency (U.S. EPA, 1986) in both smelting and urban communities. EPA's summary of the observed concentrations of lead in soil and housedust suggest that the indoor-outdoor concentration ratio in smelter communities is approximately 1:1. In other studies near lead smelters, both Roberts (1974) and Diemel et al. (1981) have reported housedust lead concentrations to be similar to soil lea concentrations. Additional evidence is derived from a comparison of mean lead levels measured in air, soil and dust near a smelter in Shoshone County, Idaho (Yankel et al., 1977; Schilling et al., 1984) before and after the closing of the smelter. Analysis of the data shows that air concentrations measured in three different areas decreased by a factor of two after smelter closing, while soil concentrations remained relatively constant. The average lead concentrations in housedust remained similar to those of soil, though somewhat more scattered geographically. Data collected from a smelter community in Montana provides support for the assumption that arsenic follows the same indoor-outdoor relationship as that observed for lead. The community, which actually no longer exists, was at one time situated very close to the base of smelter stacks. Garden soil and housedust arsenic concentration measurements from this community (U.S. EPA, 1986a) while limited, suggests that the estimate Si = So is a relatively good one for these homes. Having obtained concentration estimates for outdoor soil and indoor dust, the next step in calculating exposure to arsenic via soil/dust ingestion, is to estimate intake rates. Ingestion of soil occurs mainly by the transfer of dust from hands to mouth and is generally agreed to be a major exposure route, especially for young children who have frequent hand to mouth contact. Age-specific rates of soil/dust ingestion are estimated as in Table 1. These rates are based on recent so-called "diaper studies" by Calabrese et al. (1988). The diaper studies measured trace elements in children's feces as a key to how much soil children eat. Although two other studies of this nature also exist in the the literature (Binder et al., 1986; Clausing et al., 1987), the Calabrese work is the most recent and corrects some of the errors in methodology made in the earlier studies. Calabrese and coworkers used hospitalized children as a control group, and adult volunteers who ingested soil in capsules, as a check on the assUmptions regarding amounts ingested and excreted. The only value in Table 1 which is directly from Calabrese et al. (1988) is that for 1.5-3 year olds. Other rates, extrapolated from the Calabrese data, are U.S. Environmental Protection Agency (EPA) recommendations as specified in the Superfund Exposure Assessment Manual (U.S. EPA, 1988a). Decreased rates for older children are consistent with observations of dramatically decreased mouthing behavior with age (Barltrop, 1966; Mahaffey, 1977). Note that all estimates are for inadvertent soil ingestion; these are not estimates for children with soil pica.
B. L. Murphy, A. P. Toole and P. D. Bergstrom Table 1 Soil ingestion rates based on U.S. EPA recommendations from the Superfund Exposure Assessment Manual (1988a), with the exception of age range 1.5-3.5 yr, based on Calabrese et al. (1988). As an adult estimate (18-70 yr) was not included in the EPA recommendations, average adult intake is assumed comparable to that of a 5-18 year old individual. Age
mg Total soil ingested~day
165
Table 2 Fraction of waking hours spent indoors and outdoors, by age category, for the 247 days per year during which the ground in the study area is assumed to be non-frozen and devoid of snowpack.
Age (yr)
Fraction of waking Fraction of waking hours outdoors hours indoors
0-1
0.00
1.00
0-9 m
0
1-6
0.25
0.75
9-18 m
50
6-11
0.21
0.79
1.5-3.5 yr
50
11-70
0.06
0.94
3.5-5 yr
50
5-18 yr
10
18-70 yr
10
Preliminary direct evidence that adult soil ingestion rates are significantly lower than those of children, comes from measurements of urinary arsenic levels for residents of a United States smelter community who were relocated from their homes by the EPA in the summer of 1986 (Binder, 1987). Thirty-two people provided urine samples before and after relocation. Results showed that the average pre-move urinary arsenic level for persons greater than 8 years was 17.2 lag L l , and the mean level post-move was 14.6 lag L 1. Six children younger than 8 years had an average pre-move urinary arsenic level of 76 lag L 1 and an average post-move level of 15.3 lag Lq. The excess level of arsenic in the urine, above the 15 lag background level measured after relocation, was 60.7 lag L-l, and 2.6 lag L-1, for children and adults respectively. Correcting for urine volume (i.e. children 1 L/day and adults 2 L/day), the ratio of these differences reveals that the young children had 15 times higher arsenic intake levels than the adults. These excess levels of urinary arsenic in children are attributed to intake of arsenic via soil ingestion. While a quantitative estimate of soil ingestion rates for children and adults can not be confirmed from these results, the significant differences in excess urinary arsenic levels between children and adults, suggests that the soil ingestion rate of older children (8 years) and adults, is much less than half that of younger children. The last piece of information necessary to complete the exposure estimate for soil ingestion, is an estimate of what fraction of total soil ingested each day is attributable to indoor ingestion of house dust and what fraction to outdoor ingestion. Calabrese et al. (1988) report it is experimentally very difficult to distinguish the fraction of outdoor versus indoor soil ingested. Thus, it is assumed that the fraction of soil ingested either indoors or outdoors is proportional to the fraction of waking hours spent in either location. Based on time activity-pattern data (Anderson et al., 1985) and some basic sleeping pattern assumptions
Estimates are based on activity pattern data contained in U.S. EPA (1986b), Anderson et al. (1985), and International Commission on Radiological Protection (1984).
(U.S Department of Health and Human Services, 1988; International Commission on Radiological Protection, 1984) these fractions are estimated as shown in Table 2. In temperate latitudes, these fractions actually only describe the fraction of soil ingested indoors versus outdoors for some number of days in which the ground is non-frozen and devoid of snow pack. In the study region, it is estimated that this condition holds only for 247 days per year CO.S. EPA, 1986a) and that for 118 days per year essentially 100% of an individual's time is spent indoors.
Exposure via Inhalation Similar concentration, intake rate, and activity pattern data to that described for soil ingestion calculations, are necessary to calculate exposure to arsenic via inhalation. The availability of arsenic concentration data for indoor-outdoor air in the study region is analogous to that for soil; outdoor data is relatively complete, but little or no data for indoor air arsenic concentrations exisL A typical, average annual outdoor air concenl~ation of 0.01 lag arsenic/m 3 air is taken here for demonstrating outdoor exposure and risk calculations, and, as done for soil, the indoor arsenic concentration is estimated from that outdoors. Indoor air arsenic concentrations may be thought of as originating primarily from two sources: penetration of outdoor airborne arsenic indoors, and resuspension of house dust containing arsenic. Contributions by these sources to indoor air arsenic concentrations, Ci (lag arsenic/m3), are represented respectively, by the fin'st and second terms in the equation: Ci = 0.3Co + 0.5 Cp (Six 106) Penetration Resuspension
(3)
where Co is the concentration of arsenic in outdoor air ~ g arsenic/m 3 of air), Cp is the indoor resuspended particulate
166
Health risk assessment for arsenic containated soil
Table 3 Indoor and outdoor activity-weighted average ventilation rates as a function of age category and activity level.
Table 4 Fraction of a 24 hour day spent indoors and outdoors, by age category, for the 247 days per year during which the ground in the study area is assumed to be non-frozen and devoid of snowpack.
Activity-weighted ventilation rate (L rain"! ) Age (yr)
Indoor
Outdoor
0-1
0.84
0.84
1-6
10.7
22.7
6-11
13.0
26.8
11 - 18
17.2
28.0
18 - 70
13.1
24.5
Age (yr)
Fraction of day spent outdoors
Fraction of day spent indoors
0-1
0.00
1.00
1-6
0.13
0.87
6-11
0.12
0.88
11 - 70
0.05
0.95
Estimates are based on activity pattern contained in U.S. EPA (1986b), Anderson et al. (1985), and International Commission on Radiological Protection (1984).
Values are based on data contained in Anderson et al. (1985).
(or resuspended housedust) concentration (l.tg suspended housedust/m3 air), Si is the arsenic concentration in housedust (mg arsenic/kg housedust), and 10 -6 is a conversion factor. The constant in the penetration term of equation (3), 0.35, is an average estimate from Freed et al. (1983) based on studies by Alzona et al. (1979) and Cohen and Cohen (1980), of the fraction of outdoor suspended particles that are transported, and remain suspended indoors. There are two major processes that reduce this fraction from 1.0. One is the physical inability of larger particles to penetrate the shell of the house, the other is the greater tendency of particulates to be deposited on surfaces indoors than outdoors. The constant, 0.5, in the resuspension term of equation (3), is based on several studies that suggest approximately 50% of the indoor suspended particulate concentration is due to the resuspension of house dust. Lefcoe and Inculet (1975), for example, measured indoor particulate levels over a two month period in a centrally air-f'dtered, air conditioned home. Outdoor air cleaning equipment was used on a 24-hour-on, 24-hour-off basis with five air changes per hour when the equipment was on. A decrease by about a factor of two, observed for indoor air particulate counts above 0.3 larn when the filtration system was on, is attributed to indoor sources, including resuspension, being comparable to outdoor sources. A similar conclusion is obtained from a study by Moschandrea et al. (1979). During a four day period at a Denver residence, outdoor airborne particulate concentrations of lead and bromine dropped to near zero compared with their levels before and after this period. From their paper, it is estimated that the indoor air levels of lead and bromine, however, stayed at respectively, 0.5 and 0.6 of the average values before and after. As the concentration of lead and bromine in housedust or resuspended housedust, would not be expected to change significantly over this
short time period, the 50 percent decrease in indoor air levels must be due almost entirely to the change in outdoor levels. Thus, it can be concluded that approximately 50 percent of the indoor suspended particulates are from the penetration of outdoor particulates, and approximately 50 percent are from the resuspension of house dust. A value of 50 ~tg m -3 is taken as a default value for the concentration of total indoor suspended particulates, Cp, based on a 39-66 I~g m -3 concentration range measured by Yocom et al. (1977) in residences, and a 50 I.tg m -3 estimate made by Hawley (1985) for Niagara Falls homes. The indoor housedust arsenic concentration, Si for mining and smelting communities was calculated in the previous section. Note that it is multiplied here by a factor of 10-6 to convert the units from mg kg 1 to ktg m -3. Clearly, since Si varies between smelting and mining communities, the indoor air concentration of arsenic may be expected to differ between the two types of communities. Age specific daily intake rates for arsenic in air may be estimated from ventilation rates. Although available data on ventilation rates, especially for children, are incomplete, a partial list o f mean ventilation rates for different ages, sexes and activity levels has been compiled by Anderson et al. (1985). Based on these data, mean ventilation rates for five age categories and four ventilation rates, were calculated. Then, using data contained in the activity data files developed by SRI for EPA's Office of Planning and Standards for the amount of time spent at low, medium and high activity levels indoor and outdoors (Anderson et al., 1985) activity-weighted average ventilation rates as shown in Table 3 were determined. Note that indoor activity-weighted ventilation rates are significantly lower than outdoor rates, as a large fraction of indoor time is spent resting or at low activity levels. Finally, to calculate total daily exposure to arsenic in air, it is necessary to calculate the fraction of time spent indoors versus outdoors. The same assumptions as used in the previous section to calculate indoor versus outdoor soil
B. L. Murphy, A. P. Toole and P. D. Bergstrom
exposure are used here. In contrast to soil exposure, however, individuals are exposed to arsenic in air both while awake and asleep, thus the fraction of time spent indoors and outdoors must be based on a 24 hour day rather than on the number of waking hours. These values are given in Table 4.
Arsenic Toxicity Having evaluated exposure, the second part of the arsenic risk assessment for skin cancer via soil ingestion, and for lung cancer via inhalation, is an evaluation of the toxicity of arsenic. The midpoint of standard EPA recommended arsenic potency factors for ingestion, 1.5 (mg/kg/day) 1 (range 1-2 (mg/kg/day)-1 (U.S. EPA, 1989a), and an EPA potency factor of 50 (mg/kg/day)"1 recommended for arsenic inhalation (U.S. EPA 1984), are used in order to compare calculated risks in this example to EPA allowable risk. There are three aspects of these arsenic potency factors, however, to be kept in mind when using the values. The first point regarding arsenic potency is that there appears to be a strong correlation between the effects of arsenic and the age at which individuals are exposed. Brown and Chu (1983) analysed smelter data for lung cancer in workers laboring in the smelter environment. All workers had been employed for I0 years, but started at different ages, i.e. 20 years of age, 30 years etc. (Brown and Chu (1983) observed a pronounced increase in cancer rates for persons starting work in an arsenic enriched environment at older ages than those starting at younger ages. Lee et al. (1988) have hypothesized that arsenic is neither an initiator nor promoter of cancer, but acts instead in the progression of precancerous cells to malignant ceils, as a gene amplifier. The evidence for the hypothesis of Lee and associates comes from results of their study showing that sodium arsenite and sodium arsenate induced a high frequency of methotrexate-resistant mouse cells which were shown to have amplified copies of the dihydrofolate reductase gene. Oncogene amplification in some cancers has been shown to correlate with the degree of progression of a tumor to malignancy (Lee et al., 1988). If in fact it is in this late stage that the effect of arsenic occurs, this would explain Brown and Chu's findings. Older people who have a higher probability of the initiation and promotion stages of cancer being complete, would have a higher probability of displaying an effect from arsenic exposure. A possible increased effect of arsenic with age is an extremely important point, as it makes the doses later in life more significant in determining risk than doses earlier in life. Although age dependent effective potency could be developed, this has not been done to date. Commonly, as will be done here, lifetime dose is calculated and divided by the adult body weight. An alternative, is to compute a dose for each age category as a function of body weight for that category. In the calculation of skin cancer risk, calculating dose as a function of body weight would produce a risk which is higher than that calculated assuming an adult body weight, by almost a factor of two. The second point regarding arsenic toxicity that bears underscoring is that while the EPA potency factors assume only 30% of inhaled arsenic is bioavailable, due to larger
167
particles being trapped in the nasal and bronchial passages, they assume virtually 100% of ingested arsenic is bioavailable. The assumption that 100% of ingested arsenic is bioavailable is based on studies with animals ingesting arsenic dissolved in water. In the present assessment the ingested matrix is soil for which there is little information on the bioavailability, and some doubt as to whether it is the same as that in water. One piece of evidence, that suggests the bioavailability of arsenic in soil may be significantly less than 100%, is an unpublished study conducted by Tetra Tech (1987). In this study, soil containing arsenic in water adjusted to pH 2, was slurried with hydrochloric acid to simulate leaching of arsenic from soil in the stomach. After six hours, only 60 + 24% of the arsenic had been leached from the soil. While the stomach model is undoubtedly simplified, it nevertheless may be indicative that the estimated 100% bioavailability of ingested arsenic in soil is high. Finally, there is some evidence that the arsenic dose response is not linear, yet the EPA toxicity factors are linearly extrapolated from relatively high doses. The ingestion potency factor is based on an epidemiological study of a Taiwanese population, the validity of which itself has been questioned. (One major concern regarding the study stems from the fact that the drinking water to which the Taiwanese were exposed and which contained high concentrations of arsenic, also contained other chemicals (i.e. ergot alkaloids) capable of producing the same effect (U.S. EPA, 1988b).) The inhalation potency factor is based on occupational exposure studies with smelter workers. The reason for questioning the linear extrapolation of data from high dose exposures, is that a greater fraction of arsenic in urine is methylated at low does than at high doses (Vahter, 1983). Since methylation of arsenic appears to be a detoxification pathway (Marafante and Vahter, 1984) this suggests that the dose/toxic response curve is somewhat steeper at lower doses of arsenic than at higher doses.
Risk Calculations As stated at the outset of this paper, risk is calculated as the product of exposure and toxicity. Using the estimates put forth in preceding sections, daily exposure can be calculated as the product of intake rate and concentration. Toxicity is estimated by potency factors recommended by the EPA. Calculations of lifetime risk may then be completed based on an assumed average lifetime of 70 years, an average adult body weight of 70 kg, and arsenic concentrations in outdoor soil and outdoor air of 500 mg kg 1 and 0.01 [tg m3, respectively. Although it is a very conservative assumption, the calculation assumes that people reside in one location for a lifetime. In fact, an individual in the United States moves an average of once every 9 years (U.S. EPA, 1989b). The risks for skin cancer are discounted by a factor of ten as allowed by the EPA for arsenic induced non-fatal skin cancers in these calculations. Table 5 shows the calculated lifetime risks for indoor and outdoor inhalation and ingestion of arsenic, both in smelter and in mining communities. Several interesting results are evident from a comparison of the numbers in
168
Health risk assessment for arsenic containated soil Table 5 Calculated lifetime risks for exposure to 500 mg kg "1 arsenic in soil and 0.01 ~tg m -3 arsenic in air. Risk Exposure pathway
Smelter community Indoor Outdoor
Mining community Indoor Outdoor
Ingestion
1.2 x 10-5
9.7 x 10-7
1.9 x 10-6
9.7 x 10-7
Inhalation
5.7 x 10-5
2.6 x 10 -6
2.6 x 10-5
2.6 x 10 -6
Total
6.9 x 10-5
3.6 x 10 -6
2.8 x 10-5
3.6 x 10 -6
this table. First, and perhaps most importantly, indoor risk is larger than outdoor risk in both communities, for both types of exposure. This is largely a result of the fact that people spend significantly more time indoors than outdoors. Indoor risks from ingestion are almost seven times lower in the mining community than in the smelter community. Risk is reduced in the mining community, because the equations relating indoor-outdoor concentrations estimate that only 15% of the outdoor arsenic concentration is tracked indoors in the mining community, while 100% is estimated tracked indoors in the smelting community. Indoor inhalation risk is also lower in the mining case, since resuspended housedust exposure constitutes a large portion of the indoor air arsenic concentrations. Almost 81% of the indoor air risk results from resuspension of housedust, in the smelting community; 19% results from material that has been transported airborne through the building shell. In contrast, in the mining community, the percentage of risk coming from resuspended house dust is 40%, while 60% results from transport of airborne material. Primarily as a function of the fact that waste rock and tailings are not tracked indoors as easily as the smaller smelter particles, when indoor and outdoor risks are combined to give an overall risk picture, the mining community risks appear to be somewhat smaller than those in the smelter community. Total lifetime risks due to arsenic exposure in the mining community is 3.2 x 10"5; total risk in the smelter community 7.3 x 10-5 Both risks, however, fall within EPA allowable risk levels of 10-4 to 10-7
Summary and Conclusions The risks associated with arsenic contaminated soils are of special concern to persons living in mining and smelting communities. Careful evaluation of these risks includes assessment of both indoor and outdoor concentrations and intake rates, as well as of the differences between mining and smelting communities. The calculations completed here, using typical outdoor soil and air arsenic concentrations for a mining region in the northwest United States, demonstrate that indoor exposures dominate outdoor. They also indicate that total ingestion risks associated with tailings (mining community) are almost five times less than those associated with smelter emissions
and that inhalation risks associated with tailings are about one-half those of the smelter. Although risk assessments such as this rarely distinguish between the types of sources for contaminant, these results show that such a distinction is extremely important in estimating risk. In this assessment, resuspended surface soils tracked indoors emerge as a primary risk pathway in the smelting community, more so than airborne emissions. The community used for demonstration purposes, however, no longer maintains an operable smelter. This relationship would not necessarily hold in an operating smelter community where concentrations are considerably higher. Although the exposure estimates were not intended to be overly conservative, use of EPA toxicity factors for use in comparing the risk numbers with EPA allowable cancer risks, ultimately incorporated some conservatism. In light of the issues of bioavailability and the body's ability to methylate low doses, risk as calculated here, may actually be somewhat overestimated. Nevertheless, the numbers calculated fall within the ranw of EPA's allowable cancer risk guidelines of 10 -4 to 10" lifetime risk. References Alzoua, J., Cohen, B.L., Rudolph, H., Jow, H.N. and Frohlinger, J.O. 1979. Indoor-outdoor relationships for particulate matter of outdoor origin. Atmospheric Environment, 13, 55-60. Anderson E., Browne, N., Duletsky, S., et al. 1985 Development of Statistical Distributions or Ranges of Standard Factors Used in Exposure Assessments. EPA-600/8-85/010. Report to U.S. EPA Office of Health and Environmental Assessment. Washington, D.C. Barltrop, D. 1966. The prevalence of pica. American Journal of Disease in Children, 112, 116-123. Bafltrop, D., Thorton, I., Strehlow, C.D., and Webb, J.S. 1975. Absorption of lead from dust and soil. Postgraduate Medical Journal, 5, 801-804. Barltrop, D. and Strehlow, C.D. 1988. The contribution from soil and house dust lead to lead burden in childhood. Presented at Conference on Lead in Soil: Issues and Guidelines, March 7-9, at Chapel Hill, N.C. Binder, S., Sokal, D., and Maughan, D. 1986. Estimafng soil ingestion: the use of tracer elements in estimating the amount of soil ingested by young children. Archives of Environmental Health, 41, 341-345. Binder, S. (U.S. Department of Health and Human Services) March 3, 1987. Memo regarding Mill Creek pre- and post-move urinary arsenic levels. Brown, C.C. and Chu, K.C. 1983. A new method for the analysis of cohort studies: implications of the multistage theory of
B. L. Murphy, A. P. Toole and P. D. B e r g s t r o m carcinogenesis applied to occupational arsenic exposure. Environmental Health Perspectives, 50, 293-308. Calabrese, E.J., Pastides, H., Barnes, R. et al. 1988. How Much Soil Do Young Children Ingest: an Epidemiologic study. Draft Report to Syntex Corp. Amherst: University of Massachusetts. Clausing P., Brunekreff, B. and van Wijnen, J.H. 1987. A method for estimating sol] ingestion by children, lnterational Archives for Occupational and Environmental Health, 59, 73-82. Cohen, A.F. and Cohen, B.L. 1980. Protection from being indoors against inhalation of suspended particulate matter of outdoor origin. Atmospheric Environment, 14, 183-184. Diemel, LA.L., Brunekreef, B., Boley, J.S.M., Biersteker, K. and Veenstra, S.J. 1981. The Amhem lead study, lI: indoor pollution and indoor/outdoor relationship. Environmental Research, 25, 449. Donll, J., Klaasen, C.D. and Amdur, M.0. 1980. Casarett and Doull's Toxicology. The Basic Science of Poisons. Macmillan Publishing Co., Inc., New York. Freed, J.R., Chambers, T., Christie, W.N. and Carpenter, C.E. 1983. Method for assessing exposure to chemical substance. U.S. EPA Office of Toxic Substances, EPA 560/5-83- 015. Franzen,D., Sackman, A., Oale, R. and Chapin, M. 1988. Analytical Results Report for Ambient Air and Residental Characterisation at Prospect Square, Park City, Utah. Report Prepared for EPA Hazardous Site Evaluation Division. Hawley, J.K. 1985. "Assessment of Health Risk from Exposure to Contaminated Soft." Risk Analysis, 5(4), 289. International Commission on Radiological Protection. 1984. Report of the Task Group on Reference Man. Oxford: Pergamon Press. Lee, T.C., Tanaka, N., Lamb, P.W., Gilmer, T.M. and Barrett, J.C. 1988. Induction of gene amplification by arsenic. Science, 241, 79. Lefcoe, N.M. and Inculet, I.I. 1975. Particulates in domestic premises. Archives of Environmental Health, 30, 565. Mahaffey, K.R. 1977. Quantities of lead producing health effects in humans: sources and bioavallability. Environmental Health Perspectives, 19, 285-295. Marafante, E. and Vahter, M. 1984. The effect of methyltransferase inhibition on the metabolism of [74As] arsenite in mice and rabbits. Chemical and Biological Interactions, 50, 49-57. Moschandreas, D.J., J.W. Winchester, J.W. Nelson and R.M. Burton. 1979. Fine particle residential indoor air pollution. Atmospheric Environment, 13, 1413. Roberts, T.M. 1974. Lead contamination around secondary smelters: estimation of dispersal and accumulation. Science, 186, 1120. Schilling R., Rass, D., Sokal, D., Inez R., Brokopp, C. and Maughan, A.D. 1984. Children's exposure to smelter-associated lead,
169
Montana and Idaho. Proceedings of the 5th National Conference on Management of Uncontrolled Hazardous Waste Sites, November 7-9, Washington, D.C.p. 239. Steele, M.J., Beck, B.D., Murphy, B.L. and Strauss, H.S. 1989. Assessing the contribution from lead in mining wastes to blood lead. Submitted to Regulatory Toxicology and Pharmacology. Tetra Tech. 1987. Unpublished data. U.S. Department of Health and Human Services (Agency of Toxic Substances and Disease Registry). 1988. The Nature andExtent of Lead Poisoning in Children in the United States: A Report to Congress. U.S. Environmental Protection Agency (Office of Health and Environmental Assessment). 1984. Health Assessment Document for Arsenic. EPA-6OO/8-83-021F. U.S. Environmental Protection Agency (Region VIII). 1986a. Endangerment Assessment: Mill Creek~Montana Smelter Site. Revised Final Report. U.S. Environmental Protection Agency (Office of Air Quality Planning and Standards). 1986b. Rev&w of National Ambient Air Quality Standards for Lead: Assessment of Scientb'ic and Technical Information. Draft Report. U.S. Environmental Protection Agency (Office of Remedial Response). 1988a. Superfund Exposure Assessment Manual. EPA 540/1-88/001. U.S. Environmental Proteaion Agency. 1988b. Special report on ingested inorganic arsenic: Skin cancer Nutritional Essentiality. EPA-625/3-87/013P. U.S. Environmental Protection Agency. 1989a. Integrated Risk Information System (IRIS) (U.S. EPA Online Chemical Potency Information). U.S. Environmental Protection Agency 1989b. Exposure Factors Handbook. EPA-600/8-89/043. Vahter,M. 1983. Metabolism of arsenic. In: Fowler, B.A. (ed.), Biological and Environmental Effects of Arsenic. Elsevier Science Publishers, B.V. Yankel, AJ., Von Lindem, I.H. and Walter, S.D. 1977. The Silver Valley lead study: the relationship between childhood blood levels and environmental exposure. Journal of the Air Pollution Control Association, 27, 766. Yocom, I.E., W.A Cote, and F.B. Benson. 1977. As cited in: Walden R.A. and Scheu P.A., (eds.) ".Indoor Air Pollution. John Wiley & Sons. New York. 1983. (Manuscript No. 193: received July 3, and accepted for publication September 25, 1989)
