Exposures: Sources and Dynamics of Hormonally Active Agents in the Environment
This chapter discusses how exposure to hormonally active agents (HAAs) can occur and how exposure to HAAs in the environment is measured or estimated. For the purposes of this chapter, exposure is defined as the ''condition of a chemical contacting the outer boundary of [an organism]" (EPA 1992). Exposure of an organism to an HAA requires that the HAA be released into the environment and that it persist for a sufficient amount of time to bring about contact with an organism. Many of the HAAs are relatively persistent chemicals, and therefore, exposure to one or several of them is virtually guaranteed for most organisms.
Exposure assessment of some chemicals might be a difficult task, particularly when they remain in an organism for a long time. For females, that might be especially important because of exposure of offspring to chemicals absorbed or mobilized during pregnancy and particularly during lactation.
Understanding the relationship between exposure, absorption, disposition, metabolism, clearance, and repair and response is the key to predicting when exposure to an agent might result in a harmful dose.
General usage often confuses dose and exposure. To be precise about the terms, exposure generally is expressed in terms that describe concentration in a given medium (e.g., air or water) and, in some cases, duration (milligrams per liter, milligrams per cubic meter, parts per million, milligrams per day). Dose is based on exposure intake and body weight or surface area of the target organism (milligrams per kilogram of body weight per day or milligrams per square meter per hour).
This chapter has two major sections. The first section discusses the many sources from which HAAs can enter the environment, their persistence, and what concentrations have been found in environmental and biologic media. In thecontinue
second section, examples are given of some of the exposure concentrations for natural and anthropogenic HAAs for the general population and highly exposed subpopulations. The dosimetry and metabolism of HAAs are covered in Chapter 4.
Sources and Releases
HAAs encountered in the environment can be produced synthetically or naturally, and exposure can occur from a variety of sources, both involuntary and voluntary. For example, virtually all humans and many animals are exposed to some phytoestrogens in their diet by eating plant products that contain these natural substances. Humans and animals can be exposed involuntarily to synthetic HAAs as a result of drinking contaminated water, breathing contaminated air, ingesting food, or contacting contaminated soil. Humans can also be exposed voluntarily to many synthetic HAAs by using HAA-containing commercial products, such as cleaners, pesticides, and food additives (domesticated animals can also be exposed to these products, but such exposure is assumed to be involuntary). Finally, many humans voluntarily ingest or apply chemicals to their skin for a specific beneficial or therapeutic purpose, and they might expose animals to these chemicals as well. These chemicals, which might act as HAAs, include such pharmaceuticals as birth control pills, herbal supplements, cosmetics, and pesticide products. Natural HAAs are generally considered to consist of plant-produced estrogens, the phytoestrogens. Synthetically produced HAAs have been used and are used in pesticides (e.g., dichlorodiphenyltrichloroethane (DDT) and endosulfan), plastics (e.g., bisphenol A), and other industrial applications (e.g., polychlorinated biphenyls (PCBs)). In either case, an exogenous HAA can supplement, inhibit, or be without measurable effect compared with typical concentrations of endogenous hormones. As discussed in Chapter 2, numerous chemicals have been identified that are known or suspected to mimic or inhibit hormones in humans and wildlife. Although this chapter discusses the potential for exposure to many of these chemicals, it is by no means exhaustive, and the extensive literature on the environmental concentrations of these chemicals should be consulted for a more comprehensive review.
Many plants contain HAAs, particularly the legumes. Some are produced by the plant itself (phytoestrogens), and others come from fungi that infect the plants (mycotoxins). Phytoestrogen compounds, such as lignins (e.g., matairesinol and secoisolariciresinol (SECO)) and isoflavonoids (e.g., isoflavones and coumestans), are common in human and animal food. Isoflavones are found in most plant tissues and include the estrogenic compounds genistein, diadzein, biochanin A, and formononetin, all of which have been detected in human urine (Adlercreutz and Mazur 1998). In addition to the legumes that contain relatively high concentra-soft
tions of phytoestrogens (up to 84,000 µg/100 g of dry weight in soybeans), oilseeds and nuts also contain highly variable concentrations of these substances (4-370,000 µg/100 g of dry weight in flaxseed). Significant concentrations of lignins, but not isoflavonoids, are also found in cereals, grains, berries, vegetables, and teas. Most vegetables do not contain isoflavones but do have high concentrations of lignins. Cruciferous vegetables, such as broccoli, also contain low but measurable quantities of isoflavones, in spite of having high concentrations of indole-3-carbinol, an anticarcinogen. Such fruits as apples, plums, and bananas have very low concentrations of isoflavonoids and lignins, the possible exception being exotic fruits, which might have high concentrations of isoflavones. Beer has detectable concentrations of isoflavones (Adlercreutz and Mazur 1998).
Concentrations of phytoestrogens can vary dramatically in plants. For example, peas and green beans contain coumestrol-0.40 µg/g and 1 µg/g of dry weight, respectively (Price and Fenwick 1985). Concentrations of diadzein and genistein in soybeans ranged from 22 to 1,915 µg/g and 69 to 1,897 µg/g of wet weight, respectively. Concentrations of both isoflavones generally exceeded 200 µg/g in most samples (Reinli and Block 1996).
Numerous synthetic chemicals have been implicated as HAAs. Many of these chemicals are no longer widely used in commerce; however, that is not true for all of them. A one-time common pesticide, DDT, was banned from use in the United States in 1992, and it is a banned or restricted pesticide in most developed countries. However, it continues to be produced and used in several developing countries, such as India, because of its effectiveness and relatively inexpensive production.
Historical use and disposal of PCBs have resulted in the presence of these chemicals worldwide. Current releases of PCBs should be confined to industrial accidents, improper disposal, and dissipation from environmental sites (e.g., sediments) containing high concentrations of PCBs. Minor releases of natural PCBs have been associated with volcanic activity.
It is estimated that over 4 billion pounds of PCBs have been produced worldwide (Hooper et al. 1990) since the early 1930s. Gunderson (1995) notes a significant decline in chlorinated pesticides and PCBs over the past two decades in most environmental media. Production of PCBs was banned in the United States in 1977, although products containing PCBs are still in use not only in the United States but also in other countries around the world.
In Table 3-1, the use and production of several representative HAAs are described. Many of the organochlorine pesticides that have been cited as HAAs are no longer used in the United States (e.g., DDT, endosulfan, dieldrin, and toxaphene), although they continue to be found in various environmental media.continue
However, the lack of U.S. production of these chemicals does not mean that they are not produced by other countries that have a need for inexpensive and effective pesticides, regardless of their long-term environmental effects. Other HAAs, such as various phenolics and phthalates, continue to be used in a variety of industrial applications in the United States and worldwide.
Manufacturing processes, such as textile wet processing, make wide use of alkylphenols (APEs) (Naylor 1995). Nonylphenol ethoxylates (NPEs) account for 80% of APE volume (Naylor 1995). U.S. production and use of APEs in 1990 exceeded 450 million pounds (Naylor et al. 1992).
Dioxins are unique among the HAAs considered in this report. They are not naturally produced by plants for any beneficial purpose, nor are they purposefully made by any industrial process. Rather they are formed naturally by the combustion of plant matter and are by-products of many industries, such as pulp and paper production, and other processes that use chlorine. In the United States, forest fires account for approximately 20 kg of dioxins per year, residential wood burning results in 20 kg/yr, agricultural burning releases 10 kg/yr, and incinerators contribute approximately 350 kg/yr for a total annual dioxin release of 400 kg (Spiro and Thomas 1994).
Global deposition of various persistent organic chemicals has been observed for years, even in areas thought to be pristine. Concentrations of persistent organic chemicals, such as DDT, in rural areas are comparable to or higher than those in more-populated and industrialized regions of North America and Europe (Fellin et al. 1996). That is, to a large degree, due to atmospheric transport and condensation of these compounds at low temperatures. As a result, seasonal variations occur in deposition of these compounds from the atmosphere (D.J. Thomas et al. 1992), and deposition peaks in wintertime. Benzo[a]pyrene was detected at 20 pg/m3 in the winter but only 1 pg/m3 in the summer (Fellin et al. 1996). DDT concentrations measured in the water of the St. Lawrence River in 1991 were highest (mean concentration 3 ng/L) in April but had fallen by September (Pham et al. 1996). The principal DDT source was thought to be runoff from spring melt from the watersheds after atmospheric deposition during the winter. Concentrations were lower at other times of the year.
PCBs are very stable in the environment. They volatilize from water and thus can be transported in the atmosphere; however, most PCBs are associated with organic environmental components, such as soil and sediments. Environmental degradation of PCBs is dependent on the chlorine content of the various isomers. The greater the chlorine content of the PCB, the less degradation will occur in soil but the greater the potential for photodegradation from water surfaces and in the atmosphere (Delzell et al. 1994).
The half-life of relatively small, less chlorinated PCBs in the atmosphere hascontinue
been estimated to be 10-25 hr of direct noon-time sun (i.e., several actual days) (Delzell et al. 1994). For other more highly chlorinated PCBs, the atmospheric half-life (over the Great Lakes) is approximately 6 yr, in water the half-life is 3 to 9 yr, and in soil 3 to 17 yr (Hillery et al. 1997).
Kepone (chlordecone), a chlorinated pesticide, was released to the James River estuary in Virginia for 9 yr, beginning in 1966. It degrades very slowly and is resistant to destruction. Kepone concentrations in the estuary ranged from a high of 4.8 µg/g in zooplankton to 0.11 µg/g in bed sediments in 1977. The concentration in suspended material in the water column averaged 0.09 µg/g. Kepone accumulation was greatest in bed sediments of the middle estuary, located approximately 40 km downstream from the source, where concentrations ranged from 60 to 200 ppb. Approximately 42-90% of Kepone input was retained in the river system by entrapment in estuarine circulation and seasonal refluxing (Nichols 1990). Other persistent chemicals, such as PCBs, are also subject to recycling in the environment from sediments into the overlying water column.
HAAs have been detected in all environmental media as well as biologic tissue, but the concentrations can vary dramatically. Detecting the concentration and, in particular, determining changes in the concentration require monitoring of the various environmental media. Monitoring can be a one-time event (e.g., point sampling), periodic (e.g., annually), or continuous. It is used to show trends in deposition, production, or use. Monitoring may be determined for air, water (both ground and surface waters), soils, sediments, and biologic tissues, such as fish and human blood.
In the United States today, there are lower concentrations of PCBs and some chlorinated pesticides, such as DDT, than there were in the 1970s, before regulatory actions were taken to curb production and use of DDT and PCBs. However, during the 1990s, the rate of decline of these chemicals in some environments, such as the Great Lakes, has slowed because of recycling or continued input. These chemicals are found in rain water (Rapaport et al. 1985), providing a mechanism for their deposition into the Great Lakes and other regions of the United States because of their continued use elsewhere in the world (Iwata et al. 1993). This finding demonstrates the global nature of exposure and has raised concerns that it will be difficult to achieve further rapid decreases in these globally dispersed chemicals (Giesy et al. 1994b; Loganathan and Kannan 1994; Stow 1995; Williams et al. 1995).
Organochlorine compounds are virtually ubiquitous in the environment. Analysis of tree bark from over 90 sites worldwide indicated that DDT, endosulfan, chlordane, dieldrin, and hexachlorocyclohexanes (HCHs) were present at all sites at measurable concentrations, albeit at very low concentrations (0-10 ng/g of lipid) at some sites, regardless of how remote the site. Although DDT has beencontinue
banned from use in the United States since 1973, concentrations of DDE (a degradate of DDT) ranging from 1,000 to 10,000 ng/g of lipid continue to be found in tree bark taken from the midwestern United States. Simonich and Hites (1995) concluded that the more volatile the organochlorine (e.g., HCHs) the more readily it would move through the atmosphere from warmer climates and be distilled out onto vegetation, soil, and water in colder climates; however, this distillation process does not appear to operate as effectively for less volatile organochlorines, such as DDT and endosulfan, whose concentrations reflect current or past local usage.
Organochlorines have been measured and have been shown to be persistent in the environment. Iwata et al. (1993) measured concentrations of HCHs, DDT, chlordane, and PCBs in 1989-1990. HCHs showed higher ocean-air and surface-seawater concentrations in the northern hemisphere than in the southern hemisphere and higher concentrations closer to the poles than at more-central latitudes. Total DDT concentrations were highest near tropical Asia. The concentrations of chlordanes and PCBs were more uniform. HCH concentrations in water were in the range of tens of picograms to nanograms per liter; the other chemicals were detected in the range of picograms to tens of picograms per liter. It is generally accepted that these chemicals are transported in water vapor for long distances and then deposited as the water condenses in cold regionsexplaining the higher concentrations found at the poles (Iwata et al. 1993; Lode et al. 1995: Muir et al. 1995).
Toxaphene, endosulfan, p,p'-DDT, and p,p'-DDE were measured in the air above seawater and in seawater of Resolute Bay, Canada (Bidleman et al. 1995). Chlorinated bornanes, including the insecticide toxaphene, were detected in the air at 6.9 pg/m3 endosulfan was measured at 4.0 pg/m3, and p,p'-DDT was found at less than 0.3 pg/m3.
Studies in the southern hemisphere (India and Australia) show interesting patterns in DDT contamination of air, water, agricultural soil, sediment, and fish (Kannan et al. 1995). In India, where DDT continues to be used, air and water concentrations were 3.5 ng/m3 and 17.5 ng/L, respectively. In Australia, the concentrations were 0.017 ng/m3 and 0.17 ng/L, respectively, demonstrating the tremendous difference that can be associated with continued DDT use.
PCBs are present in the atmosphere primarily in the vapor phase; however, a small portion also exists in the particulate phase. The proportion of PCBs in the particulate phase is dependent on the ambient temperature and the vapor pressure of the specific PCB. A greater concentration of PCBs in the particulate phase develops with lower temperatures and vapor pressures (Delzell et al. 1994). The total amount of atmospheric PCBs has been estimated to be between 10,000 and 100,000 kg. Atmospheric deposition might form the greatest source of PCBcontinue
contamination for surface waters, as these compounds can be transported unchanged for thousands of kilometers. Delzell et al. (1994) found that concentrations of PCBs in ambient air did not vary substantially between urban and rural areas (0.13 to approximately 10 ng/m3), although urban areas with point sources for PCBs might have higher concentrations (up to 1.26 mg/m3 at a contaminated site). However, Hillery et al. (1997) found that atmospheric gas-phase concentrations of PCBs around the Great Lakes ranged from 89 to 370 pg/m3 in the early 1990s, with concentrations at a station near Buffalo, New York, an industrialized area, generally 2-3 times higher than those in more remote, less populated areas. PCB concentrations were also temperature dependent, with lower concentrations occurring during the colder months.
Indoor air generally has higher concentrations of PCBs than does outdoor air. Concentrations in office buildings, including research facilities and a shopping complex, ranged from 44 to 240 ng/m3; buildings with electrical equipment containing PCBs had up to a threefold greater concentration. Houses were found to have PCB concentrations ranging from 39 to 400 ng/m3. The authors concluded that PCB concentrations in indoor air in public buildings and homes were approximately 10 times greater than concentrations in outdoor air (McLeod 1981, as cited in Delzell et al. 1994).
Studies of the Great Lakes region in 1980 reported environmental concentrations of PCBs that were approximately twice as high as those found in 1991 (Pearson et al. 1996). Total PCB concentrations in the open waters of Lake Michigan dropped markedly (p < 0.05), from 1.2 ng/L in 1980 to 0.47 ng/L in 1991. The amounts of PCBs measured over the past few decades are based on different methods of analysis, with isomer-specific analytic methods being used over the last decade. Much work since that time has been on the highly toxic coplanar PCBs (Patterson et al. 1994). The PCB concentration in the water of remote Siskiwit Lake in 1984 was 2.3 ng/L (Swackhamer et al. 1988). Surface-water concentrations are generally less than 10 ng/L, although some concentrations in urban areas might exceed that (Delzell et al. 1994). Huestis et al. (1996) and others have discussed the difficulty in comparing concentrations of PCBs reported in the 1970s or early 1980s (based on total PCB concentrations) with values reported over the last decade (based on measurement of selected congeners). PCB concentrations in the surface waters of Siskwit Lake were 2.3 ng/L (Delzell et al. 1994).
In seawater, p,p'-DDE was found at 1.0 pg/L and chlorinated bornanes, including toxaphene, were detected at 48 pg/L (Bidleman et al. 1995). Toxaphene was found at 23 pg/L in Lake Laberge, Canada (Kidd et al. 1995).
The alkylphenol ethoxylates (APEs) constitute another class of synthetic compounds that are being examined as possible HAAs. Nonylphenol (NP) iscontinue
estrogenic (Soto et al. 1991). The APEs are a group of widely used surfactants and detergents. A study of APE concentrations in 30 U.S. rivers was sponsored by the U.S. Environmental Protection Agency and the Chemical Manufacturers Association (Naylor et al. 1992). The studies measured NP and various NP ethoxylates (NPEs) in rivers receiving municipal or industrial wastewater discharges. The NPEs consist of an NP molecule with between 1 and 100 ethoxylate chains (NPE1 to NPE100) almost exclusively in the para position.
Most of the water samples (60-75%) were below the limit of detection (0. 1 ppb for NP, NPE1, and NPE2; 1.6 ppb for NPE3-17). The highest concentrations for NP, NPE1, and NPE2, were about 1 ppb; the highest concentration for NPE3-17 was 15 ppb. Sediment samples had higher concentrations (18 ppb for NPE and 162 ppb for NP) (Naylor et al. 1992).
Bennie et al. (1997) analyzed NP in the Laurentian Great Lakes and reported that 58% of the surface-water samples contained 4-NP, with values ranging from less than 0.02 to 7.8 µg/L.
Dioxin concentrations in the sediments of remote Siskiwit Lake on an island in Lake Superior peaked between 1940 and 1970, the period of greatest expansion in the industrial use of chlorine. Sedimentation rates for dioxins decreased by 30% between 1970 and 1983, as did industrial use of chlorine (Spiro and Thomas 1994). PCB concentrations in the lake were 48 ppb in 1983 (Swackhamer et al. 1988). PCB concentrations in this lake, which has no point sources, were 48 ng/g of sediment.
In general, PCB concentrations in sediment have reflected the increased use of chlorine chemicals between 1940 and 1970. Sediments from Dark Lake, Wisconsin contained PCBs at 2.2 ppb between 1935 and 1948, but concentrations increased to 19-20 ppb between 1962 and 1981. Sediments from larger bodies of water (e.g., Lake Michigan) showed similar trends; concentrations of PCBs between the 1960s and the 1980s were greater than 90 ppb (Swackhamer and Armstrong 1986). Highly contaminated sediments, such as those from Waukegan Harbor, Illinois, near Chicago, might contain PCBs in excess of 500,000 ppm (NRC 1997).
As part of the U.S. Geological Survey's nationwide assessment of contaminants in carp, Goodbred et al. (1997) analyzed stream-bed sediments for phthalates (including diethylphthalate and diethylhexylphthalate) and phenols (including alkylphenol). Phenol concentrations in 22 samples ranged from nondetectable in 3 samples to greater than 1,000 µg/kg of dry weight in 1 sample; 13 samples contained less than 100 µg/kg. Phthalate concentrations ranged from nondetectable in 1 sample to greater than 2,300 µg/kg of dry weight in another sample (South Platte River at Denver); 8 samples contained less than 100 µg/kg and 12 samples contained between 100 and 500 µg/kg.break
Bennie et al. (1997) examined sediments from the Great Lakes and upper St. Lawrence River and found that 66% of the samples contained detectable concentrations of NP and NPEs at concentrations of up to 38 µg/g and 6.0 µg/g of dry weight, respectively.
Coumestrol, an isoflavonoid, is found in many plants. Coumestrol concentrations range from 0.1 µg/g of dry weight for spinach to 71.1 µg/g for fresh soybean sprouts; most vegetables contain less than 1 µg/g (Verdeal and Ryan 1979). Analysis of human urine showed excretion rates of 4,700-34.000 nmol/d for lignins and 7,400 nmol/d for isoflavones. Zearalenone dietary intake was estimated to be 0.05-0.1 µg/kg/d (Bennett and Shotwell 1979; Kuiper-Goodman et al. 1987; Warner and Pestka 1987).
The Food and Drug Administration (FDA) monitors for pesticides and other potential food contaminants that could affect the human endocrine system. The FDA Total Diet Study (conducted on an annual basis) is conducted to determine typical intake of pesticides and other chemicals in the United States (Gunderson 1995). From 1986 to 1991, nearly 5,000 finished food samples were analyzed for hundreds of pesticides and other agents. Intake was estimated by age and body weight. Concentrations were typically below the level of detection. Based on the annual Total Diet Study conducted between 1965 and 1984 by FDA, average daily intake of DDT by teenage males decreased from a high of 31 µg in 1965 to 2.5 µg in 1984, and PCB intake decreased from 1.4 µg in 1971 to 0.03 gg in 1984 (Tao and Bolger 1998). The Total Diet Study conducted between 1985 and 1991 indicated that of 4,914 food items analyzed, DDE was found in 16%, dieldrin in 8%. and lindane (hexachlorocyclobenzene) in 4% of the samples (Tao and Bolger 1998).
The continued use of DDT in developing countries does not necessarily result in greater contamination of biota. For example, a study of DDT contamination of fish in India and Australia showed similar values for both countries: The concentrations in fish were 15 ng/g in India and 22 ng/g in Australia (Kannan et al. 1995).
The U.S. Geological Survey examined 578 male and female carp at 25 river sites in the United States (Goodbred et al. 1997). Organochlorine pesticide concentrations in fish tissue ranged from nondetectable (5-10 µg/kg of wet weight) at two sites to 1,310 µg/kg of wet weight for fish taken from the South Platte River in Colorado. Total PCB concentrations ranged from nondetectable at 10 sites to 72,000 µg/kg from fish taken from the Housatonic River in Massachusetts. PCB concentrations in fish were highest in the northeastern United States, where tissue concentrations generally exceeded 1,000 µg/kg. The onlycontinue
other site with fish that exceeded that concentration was a section of the South Platte River near Denver that has been affected by urban development.
Examples of the content of persistent organochlorinesDDT or PCBsin biota from various regions, including those far from known sources, are shown in Table 3-2. The data in the table show ranges in the concentrations of total DDT usually measured as DDE (a persistent DDT metabolite) and either total PCBs or the sum of multiple PCB congeners. The concentrations of individual estrogenic. proestrogenic, or antiestrogenic PCB congeners are not known or recorded, or they are below the limits of detection. In a Swedish study, Andersson et al. (1988) found that the highest concentrations of DDT and PCBs were in fish predators, such as raptorial birds and seals. The DDT concentrations were 0.1457 µg/g of extractable lipid in fish, 5.5-400 µg/g in muscles from fish-eating birds (e.g., guillemot), 20-835 µg/g in eggs from high-trophic-level birds (e.g., peregrine falcon and sea eagle), and 1.7-66 µg/g in seals. PCB concentrations were 0.7-24 µg/g of extractable lipid in fish, 12.0-31 0 µg/g in muscles from fish-eating birds, 34-987 µg/g in eggs from high-trophic-level birds, and 1.9-75 µg/g in seals.
Geographic-distribution patterns, similar to those described above, are inferred from many studies by many investigators. Kannan et al. (1995) examined the concentrations of organochlorine residues in fish from many locations in tropical and subtropical Asia. As in other studies, the researchers reported that the concentrations of PCBs were higher near major urban areas (21-32 ng/g of wet weight for urban Australia compared with 2.4-7.6 ng/g for rural areas). Nevertheless, some specieseven in remote areasexhibit significant concentrations of persistent chemicals. The position of those animals in the food chain and the long biologic half-lives of the chemicals (see Table 4-1) explain the concentrations. The species include top carnivores (e.g., whales and polar bears) and organisms with abundant lipid content, such as some birds. The elevated concentrations of DDT found in more-remote regions, as well as urban areas, reflect its continued use in malaria-eradication programs for control of anopheles mosquitoes. The same geographic-use patterns probably contribute to higher concentrations of DDT in humans in such regions. DDT was the predominant organochlorine found in fish tissue from tropical countries (0.43-28 ng/g of wet tissue), whereas concentrations of the other organochlorines (PCBs and hexachlorocyclohexanes) were relatively low. Kannan et al. (1995) concluded that PCBs and DDT, being less volatile, persist longer closer to emission sources rather than being transported to more temperate climates.
As of 1994, data on organochlorine (and metal) contaminants in tissues of baleen whales (Mvsticetes) had been published for approximately 1,000 whales in 10 species from the world's oceans. Toothed whales (Odontocetes) have also been examined worldwide. As summarized by O'Shea and Brownell (1994). the concentrations of contaminants in tissues of the filter-feeding baleen whales gen-soft
erally are low (less than 0. 10-587 ppm for total DDT; 1.9-27.8 ppm for PCBs) in comparison with those found in the carnivorous toothed whales (2.4-2695 ppm for total DDT; 1.4-800 ppm for PCBs). Concentrations of total DDTs and total PCBs in baleen whales are greater in the northern hemisphere than they are in the southern oceans, apparently because of greater contamination of northern ecosystems (O'Shea and Brownell 1994).
Given the concentrations of many contaminants that accumulate in cetaceans. these animals could be considered marker species for determining the geographic extent of HAA effects. Comparisons of different populations of beluga whales are particularly relevant. Male animals from a population in the St. Lawrence estuary had concentrations of PCBs and DDT that were 25- and 32-fold higher, respectively, than those in populations from other arctic locations (Muir et al. 1990a,b). The data in Table 3-2 show the range of concentrations of DDTs and PCBs in these animals. Several species of marine mammals from the east coast of Canada generally had substantially higher concentrations than did Arctic animals (Muir et al. 1992). Muir et al. (1996) found a 1.5- and a 1.9-fold decrease in DDT and PCB concentrations, respectively, in male beluga whales in the St. Lawrence estuary between 1982-1985 and 1993-1994.
However, studies of PCB and DDT residues and other organochlorines in marine mammals from around the globe also suggest that concentrations in animals are declining in some regions. In ringed seals from the Canadian arctic, for example, PCB concentrations decreased by more than 60% from 1972 to 1981 (Addison et al. 1986); in the same animals, total DDT decreased by 30%. Concentrations of PCBs in polar bears in some regions actually increased between 1969 and 1984 (Muir et al. 1988a). Interpretations of the significance of changes in contaminant concentrations are open to question if the number of animals is small. The residues of PCBs and of DDT and its metabolites detected in the blubber of harbor porpoise and bottlenose dolphin from the north and east coasts of Scotland between 1988 and 1991 ranged from quite low (0.28 µg/g for total PCBs and 0.14 µg/g for total DDT) to relatively high (23 µg/g for total PCBs and 10.2 µg/g for total DDT) (Wells et al. 1994). The concentrations of these compounds were found to be highly dependent on the age and sex of the animals; higher concentrations were found in males and older animals of both sexes. Females are thought to transfer up to two-thirds of the organochlorine concentrations normally found in their blubber to their offspring during gestation and lactation. The range of concentrations seen in this study emphasizes that data from single or small numbers of animals might be of limited value when comparing information on organochlorine residues in marine mammals within or between regions.
The concentrations of some regulated halogenated organic compounds have decreased since the 1970s. For many other chemicals, there are inadequate data upon which to evaluate trends. The most studied chemicals are PCBs and DDT, and the production of these compounds has been banned in the United States forcontinue
the past 20 yr, resulting in declines in environmental concentrations. Examples of declines in other areas include a progressive and substantial decline in PCBs and DDT found in eggs taken from bird colonies in the Canadian Atlantic region between 1972 and 1978 and a decrease in PCBs and DDT in Bering Sea fish from 1982 to 1992 (Kannan et al. 1995). Concentrations of these compounds in salmonries have ceased declining during the late 1980s and early 1990s, most likely due to atmospheric transport of these chemicals from their use in other countries (Miller et al. 1993). However, Stow (1995) reported that herring gull eggs collected from the Great Lakes are now showing steady-state concentrations of PCBs, associated with continued evidence of reproductive impairments in fish-eating birds (see also Giesy et al. 1995).
Sources of HAA exposure were discussed above. The resulting exposures to these environmental contaminants are reviewed below. The examples illustrate the significant potential for diverse routes and modes of exposure. They also demonstrate the range of possible concentrations. Some exposures may lead to large doses of HAAs. Each exposure to a given HAA or to a mixture of substances (e.g, all PCB congeners) must be considered separately to determine whether it poses a significant health risk, but the committee recognizes that wildlife and humans are exposed to complex mixtures of chemicals and that interactions between chemicals in mixtures cannot always be predicted by examining each chemical individually.
Daily human exposures have been estimated for some chemicals that have been reported to have HAA activity (Table 3-3). People who either work with HAAs or whose diets are very high in HAAs can receive substantially high doses. People who live in highly contaminated areas also can experience above average exposures.
The exposure estimates in Table 3-3 are based on several assumptions, many of which tend to overestimate exposure, such as water intake or dietary intake. Other assumptions can overestimate or underestimate potential biologic activity of the agents. Factors that influence response, such as receptor affinity, bioavailability, protein binding, and potency are described in detail in Chapter 4. One example of a likely overestimate is that DDT concentrations are typically reported for total-DDT-related isomers. Although many of those isomers have much lower estrogenic activity than does o,p'-DDT, the relative estrogenicity of DDT is fixed at the estrogenicity of o,p'-DDT. On balance, DDT is more persistent in body fat than are most natural estrogens. Tables 3-3 and 3-4 illustrate the orders-of-magnitude difference between the exposures to the estrogenic drugs and the exposures to the environmental chemicals reported to have estrogenic activity. Table 3-4 lists ranges for daily production of endogenous estrogen by humans.break
Many exposure values are the result of simplistic assumptions, but they demonstrate several important points about exposure estimates for HAAs. First, one route of exposure often provides the dominant contribution to total dose. Oral exposure through food is generally greater than other exposures, in part because of a large daily intake of synthetic HAAs in fish and because of increased consumption of plant matter containing phytoestrogens by some populations, such as vegetarians. Second, inhalation is rarely a significant contributor to the dose of these chemicals from environmental exposures.
Human Background Concentrations
Steroid hormones, such as corticosteroids, androgens, and estrogens, are under feedback control from the pituitary gland. Therefore, as endogenous steroid hormone concentrations increase, the pituitary feedback control signals the endocrine organ (the adrenal gland or gonads), through pituitary hormones, to cease production or release of the endogenous steroid or to stimulate the release of opposing hormones. This homeostatic control in response to endogenous hormones is critical for maintaining proper hormone concentrations.
There is significant endogenous hormone production in males and females. The concentrations of endogenous estrogen hormones in humans change rapidly over short periods. Endogenous hormones can be active at concentrations as low as picograms per milliliter (pg/mL) in the blood, and the concentrations change with the reproductive cycle in females and episodically in males. As shown in Table 3-4, for example, adult men and prepubertal boys and girls have between 0.08 and 40 pg/mL of estradiol (Greenspan and Strewler 1997; Wilson et al. 1998). Nonpregnant women produce between 60 and 700 pg/mL of estradiol (DeGroot et al. 1995); the wide variation is due to the normal reproductive cycle. Estrogen concentrations increase dramatically during pregnancy. Pregnant women have between 500 and 15,000 pg/mL (DeGroot et al. 1995).
Circulating concentrations of estradiol range from 10 to 175 pg/mL during the female menstrual cycle, with the highest concentrations at the late follicular phase (Thomeycroft et al. 1971). Progesterone ranges from less than 1 to 10 ng/mLcontinue
(Thorneycroft et al. 1971). Studies of the concentrations of estrone and estradiol show that both hormones are present in the follicular phase of premenopausal women at about 50 pg/mL: whereas, postmenopausal women have concentrations of 10-30 pg/mL (Yen 1977; Slemenda et al. 1996). Blood concentrations of estradiol and estrone in normal adult men are similar to those in postmenopausal women (Zumoff et al. 1982).
Human exposure to PCBs from various environmental media in an urban setting has been determined (Table 3-5) using the assumption that an individual spends 56% of his or her time outdoors and 44% indoors (a substantial overestimation of outdoor time but illustrative nonetheless). This exposure assessment does not include intake from diet, which may be substantial as PCBs tend to bioaccumulate in higher trophic levels, such as in fish, dairy products, and meat. Exposure to PCBs from diet decreased from a high of 6.9 µg/d in 1971 to 0.05 µg/d in 1989. Combining dietary exposure with other routes of exposure yields a total daily exposure, on a body-weight basis, of 0.11 µg/kg of body weight per day for 1971 and 0.008 µg/kg of body weight per day in 1989, a decrease of 138-fold (Delzell et al. 1994).
An analysis of human adipose tissue taken from surgical patients and cadavers between 1972 and 1983 found that 95.3% of the U.S. population had detectable concentrations of PCBs, 28.9% had PCBs at greater than 1 ppm, and 5.1% had concentrations greater than 3 ppm. Concentrations increased with age, with children younger than 14 having less than half the concentrations of adults, possibly reflecting the banning of PCBs in 1976. More than 95% of all the U.S. population had detectable concentrations of PCBs, regardless of age, sex, race, or geographic location (Robinson et al. 1990).break
Highly Exposed Populations
Although many of the synthetic HAAs discussed in this report are no longer in general commerce, some of them are still produced in developing countries. As a result, there is continuing exposure for production workers and waste handlers, as well as the general population. Even for HAAs that are not longer manufactured, exposure might occur as a result of earlier production and use. For example, although PCBs are no longer produced in the United States, they are still found in older electrical equipment, and maintenance workers can be routinely exposed to them.
The issue of detection limits of the assays used is critical with regard to presenting information about the incidence of exposure to specific chemicals. The issue is covered in more detail in Chapter 11. Hill et al. (1989) analyzed 12 chlorinated phenols in 197 children living in Arkansas using a detection limit of 1 ppb. They reported a much higher percentage of exposure to the organochlorine chemicals measured than some other studies that used higher detection limits. For example, 54% of the children were found to have the organochlorine 2,4,5-trichlorophenol in their urine with a detection limit of 1 ppb, but with a detection limit of 5 ppb, only 9% of the children tested positive, a result similar to prior studies conducted with the higher detection limit.
Maternal milk is discussed in Chapter 4 as a route of excretion for nursing women. Breast milk is an important mode of exposure for nursing infants. Organochlorines tend to concentrate in fat (Rogan et al. 1991; Jensen and Slorach 1996), and human milk is about 3.3% fat. Organochlorine concentrations in milk might vary substantially, depending on the exposure of the mother. However, the benefits of breast milk are typically seen as outweighing the potential for chemical exposure, even for those at the high end of the exposure range (Rogan et al. 1991; Lederman 1996).
Although the dose to the fetus is of critical importance in determining HAA effects, the dose is dependent on the exposure of the mother. An analysis of PCBs in umbilical cord blood of women who had and had not consumed Great Lakes fish, known to have high PCB concentrations, indicated that the average mean total PCB concentration in the cord blood was approximately 1.0 ppb for all women. However, those women who ate Great Lakes fish had higher absolute concentrations of highly chlorinated PCBs in the cord blood of their neonates, and the concentrations were correlated with consumption relative to the time of pregnancy (Stewart et al. in press). (For more information on fetal doses, see Chapter 4.)
Based on an inhalation exposure factor of 22.8 m3/d (ICRP 1981) and DDT concentrations in air of 3.5 ng/m3 and 0.017 ng/m in India and Australia, respectively, the DDT exposure by inhalationmaking no assumptions about deposition in the lung, absorption, or dispositionwould be 79.8 ng/d in India andcontinue
0.388 ng/d in Australia. Assuming water intake of 2 L/d, the DDT exposure from consumption of water containing 17.5 ng/L in India and 0.17 ng/L in Australia would be 35 ng/d and 0.34 ng/d, respectively. Assuming a fish intake of 100 g/d and assuming that the analyzed part of the fish is also the part consumed, DDT intake would be 1,500 ng/d in India and 2,200 ng/d in Australia. The sum of exposures would be 1,615 ng/d for India and 2,201 ng/d for Australia.
Investigators have examined the dietary exposures of people living in polar regions, especially those who consume foods high in animal fat. It is reasonable to believe that such populations would receive high dietary exposures to organochlorines because of the higher concentrations of such chemicals in polar regions and because of accumulation of such chemicals through the food chain. One study compared Inuit women from Baffin Island in the eastern arctic and Sahtu Dene/Metis women from the western arctic. The Inuit women ate a diet high in ringed seal meat and blubber and high in walrus, mattak, and narwhal blubber, and the Sahtu Dene/Metis women ate a diet of caribou, whitefish, trout, and duck (Kuhnlein et al. 1995). Inuit women exceeded acceptable daily intakes of chlordane-related compounds and toxaphene more than 50% of the time and frequently exceeded acceptable intakes of dieldrin and PCBs. Inuit women had a mean intake of 24.2 µg of DDT per day, compared with 0.51 µg of DDT per day by Sahtu Dene/Metis women.
Other studies showed greater concentrations of organochlorines in Inuit women than in Caucasian women from urban areas south of Quebec (Dewailly et al. 1993a). DDE in human milk fat averaged 1,212 ng/g and 336 ng/g in Inuit and Caucasian women, respectively. Total PCB concentrations in maternal milk were 1,052 ng/g and 157 ng/g in Inuit and Caucasian women, respectively. Assays of foodstuffs showed that polar-bear fat had total PCB concentrations of 7.002 ng/g; beluga blubber, seal blubber, and arctic char muscle had total PCB concentrations of 1,002 ng/g, 527 ng/g, and 152 ng/g, respectively. These data show various sources for dietary exposure that could explain the higher doses and concentrations of organochlorines found in the Inuit population.
A comprehensive review of PCB exposure for humans living in the Great Lakes region was conducted by the Agency for Toxic Substances and Disease Registry (ATSDR 1999). It found that people who consumed fish taken from the Great Lakes had PCB body burdens that were 2-4 times greater than those of the general U.S. population. Furthermore, PCB concentrations in breast milk of women who ate Great Lakes fish were almost twice those of a control group (Fitzgerald et al. 1998, as cited in ATSDR 1999). PCB concentrations in blood were correlated with the number of years an individual had consumed Great Lakes fish. Individuals who consumed less than 6 pounds of fish per year had a geometric mean for PCB blood concentrations of 6.8 ppb, whereas those who consumed more than 24 pounds of Great Lake fish had mean blood PCB concentrations of 19 ppb (Hovinga et al. 1993).break
Human and animal exposures to the phytoestrogens. particularly isoflavones. can be very high, because these compounds are found in many foods. Genistein. daidzein, formononetin, and equol are all present in clover. Infertility in sheep. ''clover disease," has been traced to isoflavone concentrations as high as 5% of the dried weight of clover (Verdeal and Ryan 1979).
The recent practice of feeding infants soy-based formula has raised concerns with regard to the long-term health effects of exposure during development (Setchell et al. 1997; Irvine et al. 1998). For example, it has been recognized for some time that feeding infants soy-based formula was associated with goiter (thyroid enlargement associated with thyroid hormone deficiency) in animals and human infants (Shepard et al. 1960). One mechanism by which isoflavonoids, such as genistein, reduce thyroid hormone concentrations and result in goiter is by inhibiting thyroid peroxidase activity; this enzyme catalyzes thyroid hormone biosynthesis (Divi and Doerge 1996). The concentration of soy phytoestrogens that inhibited thyroid hormone biosynthesis is within the range of exposure of infants maintained on soy formula. Soy-based formulas contained isoflavones at 32-47 µg/mL, which corresponded to a daily exposure to total isolfavones of 4.58.0 mg/kg of body weight per day for a 4-mo-old infant. That concentration is 6-to 11-fold higher than concentrations known to cause hormonal effects in adults. (Divi et al. 1997; Setchell et al. 1997). In a study by Irvine et al. (1998). the phytoestrogen content of soy-based formulas and cereals were compared with dairy-based formulas and human breast milk. Again, infants received approximately 3 mg/kg of body weight per day from the soy-based formula, but a single daily serving of infant cereal could increase the isoflavone intake by more than 25%. Dairy-based formula and human breast milk contained isoflavones below the limit of detection. Human breast milk had undetectable concentrations of phytoestrogens regardless of the diet of the mother, including women who were vegetarians and consumed greater than 50 g of soy products in a 48-hr period before sampling.
Potential exposure to plant estrogens found in wood has been assessed by various in vitro and in vivo bioassays. Wood-derived estrogens, such as betasitosterol, could represent environmental hormone exposures, particularly from pulp and paper mill effluents, downstream of wood-processing facilities. Mellanen et al. (1996) used two breast-cancer cell lines in vitro (MCF7 and T47D) and expression of the vitellogenin gene in rainbow-trout livers to estimate estrogenic activity of wood-derived compounds. Some compounds, such as betasitosterol, were estrogenic in human and fish bioassays, but some phytoestrogens, such as betulin and pinosylvin, were estrogenic only in humans.break
Not all HAAs are pesticides, nor are all pesticides considered to be HAAs. However, many organochlorine compounds do have pesticidal uses and, therefore, have a significant exposure potential for nontarget populations.
Pesticides are listed among the common examples of HAAs. The organochlorine pesticides DDT, endosulfan, dieldrin, and vinclozolin have all been reported as having hormone-mimicking or hormone-inhibiting activity (Kupfer and Bulger 1980; Soto et al. 1994; Kelce et al. 1995). Pesticide exposures occur by various routes. Some routes of exposure are considered voluntary, such as dermal or inhalation exposure for someone who mixes, loads, or applies a pesticide. Other exposures are involuntary or unwitting, such as ingestion of food stuffs that have pesticide residues. Involuntary exposure can also occur as a result of aerial pesticide spray that drifts to nearby areas.
Sources of the greatest consumer exposure to dietary pesticide residues have been listed by the U.S. Department of Agriculture (USDA) (Kuchler et al. 1996). Exposures are summarized for on-site farm use, postharvest use of pesticides. pesticides used on imported foods, and canceled pesticides that are environmentally persistent (Kuchler et al. 1996). In that report, the USDA Agricultural Market Survey (AMS) used food-consumption data and residue data for 50 pesticides found on 12 fruits and vegetables to estimate dietary exposure. The data were analyzed as "commodity-pesticide pairs"; the residue value was determined for each of the 50 pesticides paired with each of the 12 commodities. Particular attention was paid to special categories, such as exposure for children of various ages. Samples were collected from markets. Before the residues were analyzed. the food was handled (washed, peeled, or cooked) as a typical consumer would prepare it. Several of the 50 pesticides have been implicated as HAAs (DDT, atrazine, dicofol, endosulfan, lindane, methoxychlor, and vinclozolin).
The AMS makes several important points about dietary exposure to pesticides (Kuchler et al. 1996):
· Residues on many samples were below the limits of detection.
· Postharvest application accounted for a large proportion of the residues detected.
· Canceled pesticides (such as DDT) were present on the foodstuffs.
· Detection does not necessarily equate to the potential for adverse effects. The concentration on the commodity, the amount ingested, and the inherent toxicity of the pesticide all must be considered.
· Consumption patterns are different for different groups. For example, children generally eat substantially more fruit (such as bananas or grapes) andcontinue
weigh considerably less than adults do, and children probably eat fewer brussels sprouts than adults do. Therefore, children would ingest more of the pesticides, on a body-weight basis, than adults for some crops and less for others.
The FDA Total Diet Study determined intake of selected pesticides, synthetic chemicals, and radionuclides from 1982 to 1991. The study focused on table-ready food rather than unwashed raw agricultural commodities or partially prepared fruits and vegetables (Gunderson 1995). It allowed FDA to determine exposures after various food preparations had been done. A market basket of 234 foods was collected periodically between 1986 and 1991 and tested for some 300 organic chemicals. The mean daily intake for these compounds was calculated based on milligrams of substance per kilogram of body weight per day. Eight age and gender groups were assessed for daily exposure. Compounds that have been identified as potential HAAs were included in the group of chemicals for which there were detectable residues (chlordane, chlorbenzilate, DDT, dicofol, dieldrin, endosulfan, endrin, heptachlor, methoxychlor, PCBs, pentachlorophenol. toxaphene, and vinclozolin). The largest exposure when corrected for body weight was almost always seen in the 2-yr-old group. Even in that age group, dietary exposure to DDT was less than 0.05 to 0.0001 µg/kg/d. PCBs were about 0.002 µg/kg/d. The Total Diet Study showed that even though use of DDT is prohibited and its concentrations were low, it was still detected in 16% of the items analyzed. Dieldrin was found in 8%, and endosulfan was found in 7%. The Total Diet Study also demonstrated that dietary exposures to banned pesticides (DDT and dieldrin) dropped by 50% or more from 1982-1984 to 1986-1991 in all age and gender groups; no such drop was seen for malathion or parathion, which remain in use (Table 3-6). Most (61-80%) of the dietary contribution of DDT and dieldrin came from meat and dairy products. Foods with higher lipid content tended to have higher concentrations of contaminants. The Total Diet Study included imported foods that might have been from countries using some U.S. banned pesticides; however, these imported foods were analyzed for the presence of banned chemicals. In 1997, DDT was found in 24% (244 samples) of 1,036 food samples analyzed for pesticides and other chemical residues. Endosulfan was found in 11 (13%) of the baby-food samples at concentrations of 0.00040.0145 ppm, according to the FDA 1997 Pesticides Monitoring Database.
Exposure to synthetic chemicals also can present risk of potential HAA activity. Several organochlorine chemicals, in addition to the pesticides mentioned above, have potential HAA activity. 2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD) (and related halogenated dibenzodioxins and dibenzofurans) and PCBs are persistent in the environment; some congeners are toxic at low concentrations, and some have hormonal (thyroid, antiestrogenic) activity (EPA 1994a,continue
1996). These compounds have been detected in various environmental and biologic media.
There is a potential for exposure to other synthetic chemicals that have measurable hormonal activity. One example is bisphenol A (BPA). BPA is a plasticizer used in epoxy resins that line various food and beverage cans and in polycarbonate resins (Table 3-1 ), which are used in microwave ovenware, returnable water and milk containers, refrigerator crisper drawers, and other food storage applications, including baby bottles. Bisphenol A is one of the top 50 chemicals produced in the United States, with over 1.6 billion pounds being produced in 1995 (Jennings 1994'; Kirschner 1996). BPA is not usually used for disposable containers because of its cost.
The potential for exposure to BPA from these applications has been studied by Brotons et al. (1995) and by the Society for the Plastics Industry (SPI) (Howe and Borodinsky 1998; Howe et al. 1998; Wingender et al. 1998). These groups have investigated the potential for migration of BPA from epoxy-lined cans and beverage containers into the liquid in the cans (Brotons et al. 1995). Neither group measured the migration of BPA into the solid food in the cans. Brotons et al. (1995) found that a 300-g can of peas contained 50 mL of fluid with a BPA concentration of 23 µg, and the fluid was found to have weak but measurable estrogenic activity as determined by the human MCF7-cell-line bioassay. They also ran tests to determine whether high temperatures (as one might use for sterilization) would increase migration of BPA. Estrogenic activity was detected in such cases. The authors reported no estrogenic activity in fresh vegetables used as controls; however, some activity might have been expected considering that many of the vegetables tested contain phytoestrogens (MAFF 1996a). It has been reported that the MCF7 cell line is sensitive to dietary or plant estrogens and will detect coumestrol with a 50% response at 0.48 ng/mL (Welshons et al. 1990).
SPI also measured BPA concentrations (Howe and Borodinsky 1998; Ho e et al. 1998; Wingender et al. 1998). SPI ran studies using FDA protocols to develop a worst-case potential exposure. The analytic data from the two groupscontinue
were quite similar; SPI originally found BPA concentrations at a range of undetectable (less than 5 ppb) to 121 ppb, with an average of 63 ppb for the food cans. All samples for the beverage cans were below the limit of detection (less than 5 ppb) (Howe and Borodinsky 1998; Howe et al. 1998; Wingender et al. 1998). SPI calculated a worst-case exposure based on the analytic-data assumptions about food-consumption patterns and food-type (as it relates to container type) distribution factors. Subsequent studies by SPI, however, demonstrated that its analytic method did not eliminate analytic interference, and its method over estimated the BPA concentrations in food cans in some analyses (Howe and Borodinsky 1998; Howe et al. 1998; Wingender et al. 1998). The SPI method was tested in nine laboratories. After elimination of interference, average BPA residues from food cans dropped to 36 ppb, indicating that the analytic method used to determine the presence and concentration of HAAs in various environmental samples can be an important factor when assessing the exposure concentrations.
SPI made a worst-case dietary exposure assessment, assuming that BPA residues were present at 5 ppb for beverages, even though none was detected, and present at 37 ppb for other foods. That assessment resulted in a total dietary concentration of 2.1 ppb. Using FDA assumptions for dietary intake (3,000 g/d) and body weight (60 kg), a worst-case oral exposure to BPA at 6.3 µg/d or 0.105 µg/kg of body weight per day was calculated.
Polycarbonate resins were not found to release detectable (less than 5 ppb) concentrations of BPA under high-temperature, worst-case conditions (Howe and Borodinsky 1998; Howe et al. 1998; Wingender et al. 1998). Using FDA assumptions and models for single-use and repeat-use items that come into contact with food, SPI concluded that the dietary concentration was less than 0.25 ppb or 0.75 µg per person per day.
Another product group that has been studied for its potential contribution to BPA exposure is resin-based dental sealants and composites. These products are used to seal and repair teeth. Olea et al. (1996) did a study in which approximately 50 mg of the sealant was applied to the teeth of 18 subjects. One hour later, BPA and other products were measured in the subjects' saliva. BPA was found at 90-931 µg. That 1-hr exposure is 10-100 times greater than the worst-case daily exposures for BPA in food cans (9.6 µg/d). One would expect the exposure in the first hour after application to be fairly extreme because it represents 0.2-2% of the total applied dose in 1 hr. Such release could not continue for long, because the resin would be lost in as little as 2 d.
Alkylphenols are widely used synthetic chemicals. With such large use, releases to the environment and, therefore, exposure to alkylphenol ethoxylates (APEs) could be considerable. Exposure to aquatic organisms in rivers that receive APE-containing effluent is the major area of concern, although exposure is mitigated by wastewater treatment, which removes 92.5-99.8% (Naylor 1995). The upper 95% confidence limit of nonylphenol (NP) in river water for a 30-rivercontinue
study was 0.35 ppb (Weeks et al. 1996). Assuming no reduction through treatment of drinking water and an average consumption of water at 2 L/d, human exposure to NP in drinking water would be 0.7 µg/d. Although NP is estrogenic (Soto et al. 1991 ), its activity is approximately 3 orders of magnitude below that of estradiol in vivo (Jobling et al. 1996). Thus, the estrogenicity produced from NP-contaminated drinking water would be equivalent to that produced by a daily estradiol exposure of 0.007 µg/d. Plastics, such as polystyrene and polyethylene, the other major plastic materials in addition to polycarbonate, also contain NP. which is used as an antioxidant to prevent discoloration.
Another class of synthetic chemicals that are putative HAAs are the phthalate esters. Ministry of Agriculture, Fisheries, and Food (MAFF) investigated dibutyl phthlate (DBP) and diethylhexyl phthalate (DEHP) for their presence in paper and board packaging, food, and infant formula (MAFF 1995, 1996a,b). DBP was found in 98% of the packaging materials at concentrations of 5-5,860 ppm. DEHP was present in 95% of the packaging at 5-3,030 ppm. Food samples stored in the packaging were analyzed; DBP was present in 27 of 31 samples at 0.04-62 ppm and DEHP was found in 30 of 31 samples at 0.1-25 ppm (MAFF 1995). The dose concentrations from dietary intake of total phthalates ranged from 0.1 to 0.8 mg per person per day (97.5% upper confidence limit is 0.4-1.6 mg per person per day) (MAFF 1996a). Exposure of infants through formula was 0.10-0.13 mg/kg of body weight per day.
Women of child-bearing age and postmenopausal women might have substantial voluntary exposure to estrogenic compounds. Current oral contraceptives contain a daily dose of 20-50 µg of a potent orally active estrogen, such as ethinyl estradiol or mestranol (Hardman et al. 1996). Estrogenic compounds are administered by prescription as daily contraceptive preparations, in hormone-replacement therapy, and as postcoital emergency contraceptives (the "morning-after" pill. Birth-control preparations use 20-50 µg of ethinyl estradiol or mestranol per day (Gerstman et al. 1991: Hardman et al. 1996). The postcoital contraceptives use one dose, which is several times greater than the typical daily dose for oral contraceptives, within 3 d of intercourse, followed 12 hr later by the same dose (Trussell et al. 1996).
Oral-contraceptive use might be a potential HAA exposure for the mother and the developing child. Studies have been done on the developmental effects of oral-contraceptive use that occurs before or in early pregnancy (Kallen et al. 1991). That case-control study investigated whether hypospadias in infants were associated with maternal use of oral contraceptives before or in early pregnancy. There was no association demonstrated in that case. Another potential HAA exposure is the use of oral contraceptives during lactation. Such exposure was studied using an exposed group of children whose mothers used ethinyl estradiolcontinue
and a control group whose mothers did not (Nilsson et al. 1986). Intellectual and psychologic behavior were studied, and weight and height increases were also investigated. Although the study did not demonstrate changes from maternal exposures to oral contraceptives, such study designs illustrate the importance of evaluating this potential HAA exposure.
Many women receive various forms of postmenopausal estrogen replacement. Estrone and estradiol are absorbed in equal amounts transdermally. Estrone is absorbed 2-4 times more than estradiol when taken orally (Scott et al. 1991). Estrogen exposures of 50-200 mg/d are effective in decreasing postmenopausal bone loss (Lindsay 1987). Shoff et al. (1998) suggest that consumption of foods high in phytoestrogens, particularly dark bread, might inversely affect testosterone concentrations in postmenopausal women.
Each pharmaceutical exposure to estrogen is designed to produce a desired effect. The hormones are administered under the care of a physician, who has control over the dose and can observe the outcome. The use of diethylstilbestrol during a critical period of gestation is now known to have produced adverse effects without evidence of a benefit (Dieckmann et al. 1953). In contrast to environmental exposures, the amount and circumstances of pharmacologic exposures are controlled.
The presence of vitellogenin, a female fish-specific protein, in male fish has been used as a biomarker for exposure of fish to estrogenic compounds. Jobling et al. (1998) found that exposure of roach (Rutilis rutilis) to sewage-treatment-plant effluents resulted in a higher than normal incidence of intersex fish (i.e., fish displaying both male and female gonadal characteristics). The number of intersex fish was positively correlated with vitellogenin concentrations in these fish as well as the concentration of sewage effluents in the river water.
In another study, vitellogenin concentrations were shown to be elevated in fish caged downstream from a sewage treatment plant, and the hypothesis was that the cause was an APE release into the environment (Harries et al. 1996). Further study on sewage-treatment-plant effluents indicated that 17ß-estradiol and estrone were present in these effluents at concentrations of up to tens of nanograms per liter. Vitellogenin synthesis was observed in male rainbow trout and roach exposed for 3 weeks to estradiol concentrations of 1-10 ng/L and 100 ng/L, respectively. Estrone was approximately 2-5 times less effective than estradiol at eliciting vitellogenin response in the rainbow trout. Exposure of both fish to 4-tert-octylphenol also significantly increased vitellogenin production at concentrations of less than 10 µg/L for rainbow trout and 100 µg/L for roach (Routledge et al. 1998).break
Summary and Conclusions
Exposure to some natural or synthetic HAAs is ubiquitous for humans and wildlife. Although most exposure to synthetic HAAs is involuntary, voluntary exposure can occur as a result of using pharmaceuticals and commercial products containing HAAs and manufacturing and using pesticides containing HAAs.
Many plants, particularly cultivated crops such as soy, contain significant quantities of phytoestrogens. Exposure is dependant on the diet of animals and humansit might be low for obligatory carnivores and high for herbivores or vegetarians. The amount of phytoestrogens present in plants can also vary from almost nonexistent in most fruits to more the 2,000 µg/g of wet weight in soybeans.
Among the synthetic HAAs to which humans and wildlife might be exposed, some, such as DDT and PCBs, have been banned from commercial production in the United States. However, other HAAs, such as alkylphenols and bisphenol A, continue to be manufactured and used in common commercial products. For many of these chemicals, their persistence in the environment, often for years in covered soils and sediments, means that exposure will continue into the future. PCBs have a half-life in soil of up to 17 yr. Although environmental concentrations of most HAAs have decreased as their use is curtailed, the recycling of contaminated sediments and soils can result in occasional "hot spots" of exposure.
Fish are frequently used as indicators of environmental contamination. Concentrations of PCBs in fish in some parts of the United States exceed 1,000 µg/ kg. Furthermore, HAAs are known to bioaccumulate up the aquatic food chain. As a result, concentrations of some HAAs, such as PCBs and DDT, are greatest in organisms at the top of the food chain, particularly in marine mammals, such as whales and seals, and fish-eating birds.
Human dietary intake of synthetic HAAs remains substantial, even intake of HAAs that have not been used commercially for many years. For example, a recent survey of the U.S. diet found detectable residues of DDT in 16% of the food samples. Human exposure is further demonstrated by concentrations of DDT in the adipose (fatty) tissue. Over 95% of adipose tissue samples taken from the U.S. population contained detectable concentrations of some HAA. Although the concentrations were found to be greatest in older individuals, even children were not immune from exposure.
Some populations are known to have extremely high exposure to HAAs. In particular, some aboriginal groups, such as the 32 Inuits of northern Canada and the United States, have diets high in synthetic HAAs as a result of consumption of contaminated marine mammals and fish.
Some infants and young children might also be exposed to high concentrations of HAAs, although in this case, natural phytoestrogens. The use of soy-based formula for infants has been shown to result in high concentrations of phytoestrogens in the blood of these children. Children, because they tend to eatcontinue
proportionately greater quantities of fruits than adults do, also might have greater exposure to pesticide residues, including banned pesticides.
The presence of natural estrogens in many plants makes it difficult to establish a baseline for exposure to HAAs, whether natural or synthetic. Given the variety of the diet of most Americans, establishing a background concentration for HAA ingestion is difficult, and little research, particularly research on phytoestrogen exposure, has been done in this area.
Long-term monitoring of known HAA-contaminated media (e.g., sediments) should be conducted to assess the persistence and recycling of HAAs in and between media.
Monitoring of environmental media should be expanded to include fish and other aquatic organisms taken from contaminated surface waters.
Intake of phytoestrogens and synthetic HAAs by humans and other biota should be studied to establish a baseline for typical HAA exposure. Predominant routes of exposure, particularly diet and drinking water, should be assessed to determine background intake concentrations for all HAAs.
Monitoring efforts should include subpopulations that are known or suspected of having high exposures, such as aboriginal populations consuming marine mammals, vegetarians, and infants consuming soy-based formulas.
Exposure assessments should include potential exposure to all possible HAAs, not just individual chemicals.break