5. BACKGROUND EXPOSURES TO CDD, CDF, AND PCB CONGENERS 5-1

5.1. INTRODUCTION 5-1

5.2. PREVIOUS ASSESSMENTS OF BACKGROUND EXPOSURES 5-1

5.3. UPDATED ASSESSMENT OF BACKGROUND EXPOSURES 5-8

5.3.1. North American Exposures 5-12

5.3.2. Comparison of Previous North American Studies to This Study 5-12

5.3.3. Comparison of Previous European Studies to this Study 5-15

5.4. ASSESSMENT OF BACKGROUND EXPOSURES ON THE BASIS OF BODY BURDEN DATA 5-18

5.4.1. Human Adipose Tissue and Blood Data 5-18

5.4.2. Dermal Exposure 5-27

5.5. HIGHLY EXPOSED POPULATIONS 5-28

5.5.1. Nursing Infants 5-29

5.5.2 Subsistence Fishers 5-33

5.5.3. Subsistence Farmers 5-34

The purpose of this chapter is to assess background exposures to the dioxin-like compounds. Recent assessments of background exposures cited in the scientific literature are summarized, and background exposures that have been estimated from the data presented in Section 4 of this report are presented. The term "background," as applied to exposure, can be used to represent different concepts. Two common definitions are (1) the level of exposure that would occur in an area without known point sources of the contaminant of concern or (2) the average level of exposure occurring in an area whether sources are present or not. For the purposes of this document, "background" is defined as suggested in the first definition above. To the extent possible, background exposures estimated in this chapter are based on monitoring data obtained from sites removed from known contaminant sources (or food data representative of the general food supply). These data are considered to be the most useful for describing background exposure levels.

Several researchers have published quantitative assessments of human exposures to CDDs and CDFs. Some of the more recent assessments are discussed below (Travis and Hattemer-Frey, 1991; Fürst et al.,1990; Fürst et al., 1991; Henry et al., 1992; Theelen, 1991; and Gilman and Newhook, 1991). It is generally concluded by these researchers that dietary intake is the primary pathway of human exposure to CDDs and CDFs. Over 90 percent of human exposure occur through the diet, with foods from animal origins being the predominant sources.

Travis and Hattemer-Frey (1991) estimated that the average daily intake of 2,3,7,8-TCDD by the general population of the United States is 34.8 pg/day. Ingestion exposures were estimated by multiplying the concentration of 2,3,7,8-TCDD in beef, milk, produce, fish, eggs, and water (estimated using the Fugacity Food Chain model) times the average U.S. adult consumption values for these products reported by Yang and Nelson (1986). The calculations assume that 100 percent of the 2,3,7,8-TCDD ingested are absorbed through the gut. Intake via inhalation was estimated by multiplying the concentration in air times the amount of air inhaled per day (20 m³) assuming that 100 percent of inhaled 2,3,7,8-TCDD are absorbed through the lung. The results of their assessment, summarized in Table 5-1, indicate that foods from animal origins comprise 95 percent of the estimated total daily exposure. These foods include milk and dairy products, beef, fish, and eggs. Exposure resulting from consumption of vegetables and other produce was estimated to account for 3.4 percent of the total intake. Exposure from ingestion of water, ingestion of soil, and inhalation of air together accounted for about 1 percent of the total daily intake.

Fürst et al. (1990) estimated human exposure to CDD/Fs based on the analysis of 107 food samples collected in the Federal Republic of Germany. The average daily TEQ intake was estimated to be 85 pg/person/day or 1.2 pg/kg body weight/day. Fürst et al. (1990) concluded that foods of animal origin contribute significantly to the human body burden of CDD/Fs. In a subsequent study, Fürst et al. (1991) assessed human exposure to CDDs and CDFs from foods using data from more than 300 randomly selected food samples and food consumption data reflective of consumption habits of the German population. These authors estimated that the German population's average daily intake of CDDs and CDFs from food is 158 pg TEQ per person of which 25 pg is 2,3,7,8-TCDD. Dairy products, meat and meat products (primarily beef), and fish and fish products each contribute about 32 to 36 percent of the daily intake of TEQ. Based on the levels of CDD/Fs observed in human samples, the average daily intake via food was estimated to be in the range of 1 to 3 pg TEQ/kg body weight.

Henry et al. (1992) of the U.S. Food and Drug Administration estimated the average exposure to the U.S. population from 2,3,7,8-TCDD through the food supply using the following assumptions: (1) all dairy products have background lipid 2,3,7,8-TCDD levels equivalent to those found in milk and half-and-half, i.e., about 55 ppq (whole dairy food levels were estimated using percent fat in each food); (2) levels averaging 35 ppq in beef tissue are present in all meat products; (3) ocean fish with tissue levels equal to half of the detection limit (about 0.5 ppt) are the sole fish source in the diet; (4) average food consumption figures (total-sample-basis) available from nationally representative data bases were used for frequency of eating (Market Research Corporation of America's (MRCA) Menu Census VI (1977-78)) and for serving sizes (U.S. Department of Agriculture's 1977-78 National Food Consumption Survey). FDA's estimates of 2,3,7,8-TCDD intake were derived by multiplying the food dioxin levels by the average amounts of food consumed per day. The results of its assessment, summarized in Table 5-2, indicate an average daily exposure of 15.9 pg/day of 2,3,7,8-TCDD of which 4 percent are due to dairy and milk products, 41 percent are due to meats, and 54 percent are due to ocean fish.

Theelen (1991), of the Netherlands National Institute of Public Health and Environmental Protection, estimated the average daily intake of 2,3,7,8-TCDD and total dioxin TEQ by residents of the Netherlands for various possible routes of exposure. The results, summarized in Table 5-3, indicate an average intake of 20 pg/day of 2,3,7,8-TCDD and 115 pg/day of total TEQ from food and 0.08 pg/day (2,3,7,8-TCDD) and 3.2 pg/day (TEQ) from combined direct air and soil exposure. Milk and dairy products make up about one-third of the total daily exposure. Animal fat in meat, poultry, and fish (i.e., fish oil) also contribute about one-third. Fish consumption represents 18.5 percent of total daily exposure. In a later study, Theelen et al. (1993) reported a median daily intake for adults of 1 pg TEQ/kg body weight, and a 95th percentile rate of 2 pg TEQ/kg body weight. These values were based on CDD/F residue levels in food products and food consumption survey data.

Gilman and Newhook (1991), of the Canadian Department of National Health and Welfare and the Ontario Ministry of the Environment, respectively, estimated an average lifetime daily intake of 140 to 290 pg of TEQ for the typical Canadian. Their results, summarized in Table 5-4, indicate that between 94 and 96 percent of the estimated intake are from food sources. No breakdown of intake by food type is provided in the report.

As reported in Section 4.6.1, CDD/Fs can migrate from bleached paper packaging and paper food-contact articles to foods. Some investigators have included this pathway in estimates of background exposure. U.S. EPA (1990a) estimated that TEQ intake due to leaching from paper products into food from paper packaging was in the range of 5.5 to 12.7 pg/d. Henry et al. (1992) estimated that daily intake of 2,3,7,8-TCDD due to migration from paper to food could amount to 12 pg/d, almost as much as the daily intake from unaffected food of 16 pg/d. (See Table 5-2.) As shown in Table 5-3, Theelen (1991) estimated that out of a total of about 120 pg of TEQ/d, 9 pg of TEQ/d could be due to migration from paper. These estimates are based on levels in paper before recent changes in industry practices that are expected to substantially reduce dioxin levels in paper. As discussed in Section 4.6.1, these reductions are expected to have significantly lowered the CDD/CDF levels currently found in food due to any leaching of dioxin-like compounds from paper.

Background exposures to CDD/CDFs in North America were estimated using (1) the TEQ data on arithmetic mean levels in environmental media and food from Table 4-11, (2) the standard contact rates for ingestion of soil, water, and food and inhalation of ambient air, and (3) the appropriate unit conversion factors. The estimated exposures and assumptions made concerning ingestion or contact rates are presented in Table 5-5.

The background exposures reported here were estimated using standard intake rates representative of the general population. They do not account for individuals with higher consumption rates of a specific food group (e.g., subsistence fishermen, nursing infants, and subsistence farmers--these are discussed in Section 5.5). The estimates reported here are assumed to represent typical (i.e., "central tendency") U.S. background exposures, and do not account for these types of variations in the population as a result of differences in intake rates of the various food groups. The fish concentration used to estimate background exposures, represents the average value found in fish from fresh and estuarine waters (see Section 4.5). Correspondingly, the ingestion rate used here reflects the per capita average ingestion rate of fresh/estuarine fish (U.S. EPA, 1989). Many individuals are likely to have higher ingestion rates of marine fish. However, the limited data on marine species indicates that the dioxin levels may be one to two orders of magnitude lower than fresh/estuarine water fish (also see Section 4.5).

The contact rates for ingestion of fish, soil, and water, and inhalation were derived from the Exposure Factors Handbook (U.S.EPA, 1989). For food products such as milk, dairy, eggs, beef, pork, and poultry, a different approach was taken because there is some evidence that consumption rates have changed since the data for the Exposure Factors Handbook were collected. Contact rates for these food groups were derived from commodity disappearance data from the United States Department of Agricultures's (USDA) report on Food Consumption, Prices, and Expenditures between 1970 and 1992 (USDA, 1993), and intake data from USDA's Nationwide Food Consumption Survey (NFCS) (USDA, 1992). The average of USDA disappearance and NFCS intake rates were used in this study to represent typical contact rates in the United States. USDA (1993) estimated per capita consumption rates using disappearance data (i.e., the quantity of marketable food commodities utilized in the United States over a specified time period) divided by the total population. Consumption rates were calculated for several commodities including meats, eggs, milk and dairy products. For meats, the boneless equivalent quantity was calculated by adjusting the carcass weight based on the amount of fat and bone removed at different market levels. USDA (1992) reported one-day NFCS intake data for several meat categories. These included: beef; pork; poultry; frankfurters, sausages, and luncheon meats; fish and shellfish; and mixtures containing meat, poultry, and fish. Total intake rates for beef, pork, and poultry were estimated by assuming that the rate of consumption of these meats in (1) frankfurters, sausages, and luncheon meats, and (2) meat mixtures was proportional to the intake of these meats on an individual basis. Thus, the intake in these categories was apportioned among the meat groups. In general, intake rates based on NFCS data are lower than those based on USDA disappearance data. NFCS data are believed to underestimate consumption for the general population because they may not adequately account for consumption of foods contained in mixtures (i.e., the intake rate for eggs may include eggs eaten separately or as a main ingredient in a dish, but may not be counted if they are an ingredient in a cake). Additional uncertainty is associated with the use of data for only one day during the Spring of 1988 and the use of survey data based on recall. In contrast to NFCS data, disappearance data may overestimate per capita consumption because they are based on the quantity of marketable commodity utilized, divided by the total population. Disappearance data do not account for losses from the food supply from waste or from the production of items not intended for human consumption (i.e., pet foods). Thus, the average of USDA disappearance and NFCS intake rates were believed to be representative of typical contact rates in the United States.

Background exposure levels are also presented for Germany based on data from Fürst et al. (1990; 1991). The total background TEQ exposure shown in Table 5-5 is 119 pg/day for North America. Based on Fürst et al. (1990; 1991), the estimated total TEQ background exposure for Germany is 79 pg/day (Table 5-6). However, it should be noted that the estimated background level for the United States and Germany are based on limited data, and exposure to all food groups was not considered. Also, the addition of TEQs for multiple pathways presumes that individuals are exposed by all pathways, and assumes that the fraction absorbed into the body is the same for all pathways. The following sections present observations about CDD/CDF exposures in North America, comparisons between exposure estimates from this and previous studies, and comparisons between North American and European exposures to CDD/CDFs.

Based on the data collected for this study, the total background CDD/CDF TEQ exposure for North America was estimated to be 119 pg/day, for all media combined. Exposure to 2,3,7,8-TCDD accounts for approximately 10.5 percent (12.0 pg/day) of the total TEQ exposure. Estimated exposures based on total CDD/CDF TEQs from the various exposure pathways are presented in Figure 5-1. The highest exposures were estimated to occur via ingestion of CDD/CDFs in beef (37 pg/day) which accounted for over 30 percent of the total TEQ exposure. The ingestion of foods accounted for over 97 percent of the total TEQ exposure. Exposure to CDD/CDFs via ingestion of water appears to be very low. Exposure via inhalation and soil ingestion are 2.2 and 0.8 pg/day, respectively. These exposures account for approximately 2.0 percent and <1.0 percent of the total CDD/CDF TEQ exposure in North America.

Previous studies of CDD/CDF exposures in North America were presented in Section 5.2 of this report. These studies reported CDD/CDF exposures based on the most toxic congener, 2,3,7,8-TCDD, and not on the total TEQ value for all congeners combined. For the purposes of comparison, mean background levels of 2,3,7,8-TCDD in North America from this assessment were used to calculate exposure via various pathways. Background exposures were calculated using background environmental levels of 2,3,7,8-TCDD, standard contact rates, and appropriate unit conversion factors, as described previously. Total 2,3,7,8-TCDD exposure for all pathways combined was 12.0 pg/day for the current assessment compared to 15.9 and 34.8 pg/day for the two previous studies of 2,3,7,8-TCDD exposure in North America (Henry et al., 1992; and Travis and Hattemer-Frey, 1991). Figure 5-2 depicts the comparisons of the percent contribution of various exposure pathways to total exposure to 2,3,7,8-TCDD for the current assessment and for previous North American studies. Figure 5-2 indicates that exposure via ingestion of meats accounted for a large portion of the exposure in all three studies. However, fish accounted for a higher percentage, and dairy products accounted for a lower percentage of the total 2,3,7,8-TCDD exposure in the Henry et al. (1992) study than in the Travis and Hattemer-Frey (1991) study and the current assessment. These differences reflect differences in assumptions for food ingestion rates as well as in TCDD levels. All three studies indicate that beef, dairy products, and fish comprise over 94 percent of the total exposure. Because of the data base weaknesses noted earlier, it is not known if these differences can be considered significant.

European CDD/CDF exposure studies may also be compared to the exposures estimated in U.S. reports and in the current assessment. Comparisons may be made based on the 2,3,7,8-TCDD congener or on total TEQ exposures (Table 5-7). Exposures to 2,3,7,8-TCDD in North America range from 12.0 pg/day to 34.8 pg/day based on the current assessment and two other U.S. studies. These values are comparable to the 2,3,7,8-TCDD exposures reported in Germany and the Netherlands by Fürst et al. (1991) and Theelen (1991). Fürst et al. (1991) reported an estimated 2,3,7,8-TCDD exposure of 25 pg/day based on ingestion of dairy products, meat, and fish; Theelen (1991) reported an estimate of 20 pg/day based on dairy, meat, poultry, and fish intake. Total CDD/F TEQ background exposure estimates for North America range from 119 pg/day for the current assessment to 140 to 290 pg/day based on Gilman and Newhook's (1991) Canadian study. For Europe, total TEQ exposure estimates range from 79 pg/day based on Fürst et al. (1990) to 158 pg/day based on Fürst et al. (1991). Figure 5-3 depicts the contributions of various exposure pathways to total background TEQ exposures for North America, Germany, and the Netherlands based on data from the current assessment, Fürst et al. (1990) and Theelen (1991). For all three geographic regions, over 90 percent of the exposures were attributed to ingestion of CDD/Fs in foods.

Based on the data presented in Figure 5-3, it is reasonable to expect that the CDD/CDF body burden in vegetarians would be lower than the body burden in non-vegetarians because vegetarians avoid the consumption of meat and fish and their derivative products. Welge et al. (1993) tested this hypothesis by comparing the CDD/CDF levels in the blood of 24 German vegetarians with the blood levels of 24 non-vegetarians, matched for age, sex, body weight, and height. With the exception of two individuals, all vegetarians had practiced a diet without meat and fish for at least 3 years.

The CDD/CDF levels in the vegetarian group ranged from 14.64 to 52.85 pg TEQ/g (lipid basis) with a mean of 32.60 pg TEQ/g. In the non-vegetarian group, the CDD/CDF levels ranged from 14.26 to 97.98 pg TEQ/g (lipid basis) with a mean of 34.32 pg TEQ/g. There was no significant difference (a = 0.05) between the vegetarian and non-vegetarian group in the mean levels of any of the 2,3,7,8-substituted congeners, in the total CDD levels, in the total CDF levels, in the total CDD/CDF levels, or in the total TEQ levels (each on a lipid and on a whole weight basis). Welge et al. (1993) suggested several reasons why no differences were found. First, all tested vegetarians had at one time been non-vegetarians. The higher levels of exposure during this non-vegetarian period coupled with the long biological half-life of CDD/CDFs may be responsible for the apparent similarity in body burdens using blood as the measure of body burden. Second, the vegetarians may have a higher level of consumption of dairy products than the non-vegatarians and thus have a similar CDD/CDF exposure even without consumption of fish and meat.

5.4. ASSESSMENT OF BACKGROUND EXPOSURES ON THE BASIS OF BODY BURDEN DATA

Examination of body burden data provides another, potentially more accurate, way to estimate exposures of humans to CDD/CDFs. However, these data may not represent only background exposure to CDD/CDFs as defined here because the sampled individuals may have lived in areas where dioxin sources were present. The most extensive U.S. study of CDD/F body burdens is the National Human Adipose Tissue Survey (NHATS) (EPA, 1991b). This survey analyzed for CDD/Fs in 48 human tissue samples that were composited from 865 samples. Each composite contained an average of 18 specimens. These samples were collected during 1987 from autopsied cadavers and surgical patients. The sample compositing prevents use of these data to examine the distribution of CDD/F levels in tissue among individuals. However, it did allow conclusions in the following areas:

· National Averages - The national averages for all TEQ congeners were estimated as listed in Table 5-8. Nondetects were treated as half the detection limit for averaging purposes. As shown in this table, all congeners except some of the CDFs had a very low frequency of nondetects. Thus, the overall TEQ estimate is not sensitive to how nondetects were treated in the averaging.

· Age Effects - Tissue concentrations of CDD/Fs were found to increase with age.

· Geographic Effects - In general the average CDD/F tissue concentrations appeared fairly uniform geographically. Only one TEQ congener was found to have a significant difference among geographic regions of the country. This compound, 2,3,4,7,8-PeCDF, was found at the lowest level in the West (4.49 pg/g) and the highest in the Northeast (13.7 pg/g).

· Race Effects - No significant difference in CDD/F tissue concentrations was found on the basis of race.

· Sex Effects - No significant difference in CDD/F tissue concentrations was found between males and females.

· Temporal Trends - The 1987 survey showed decreases in tissue concentrations relative to the 1982 survey for all congeners. However, it is not known whether these declines were due to improvements in the analytical methods or actual reductions in body burden levels. The percent reductions among individual congeners varied from 9 percent to 96 percent.

New information on levels of dioxin-like compounds in human tissue/blood has recently been published (Patterson et al., 1994). Human adipose from 28 individuals was collected. The individuals studied were ones that died suddenly in the Atlanta area during 1984 or 1986. Their ages ranged from 19 to 78 yr and averaged 49 yr. The tissue data are summarized in Table 5-9. This table shows that the mean PCB levels generally exceeded the mean 2,3,7,8-TCDD level and PCB-126 exceeded the 2,3,7,8-TCDD level by over an order of magnitude. The mean TEQ levels for these coplanar PCBs summed to about 17 ppt (using the toxic equivalency factors proposed by Safe, 1990). A complete CDD/F congener analysis was conducted on tissues of five of the individuals, resulting in an average of 25 ppt on a TEQ basis. These tissue samples were also analyzed for PCBs 77, 126 and 169. The TEQ levels for these coplanar PCBs summed to 8.2 ppt (using the toxic equivalency factors proposed by Safe, 1990). Thus, the PCBs contributed 24% of the total TEQs. Patterson et al. (1994) also studied serum collected by the CDC blood bank in Atlanta during 1982, 1988 and 1989. These samples were pooled from over 200 donors. The average levels for 2,3,7,8-TCDD and PCBs are summarized in Table 5-10. The serum data appears to indicate a decrease in exposure to PCBs from 1982 to 1988/1989. In general, the Patterson et al. (1994) data suggests that the coplanar PCBs can contribute significantly to body burdens of dioxin-like compounds. The data suggest that the coplanar PCBs can increase the total background body burden to over 40 ppt of TEQ. This conclusion is uncertain because the people studied by Patterson et al. (1994) may not be representative of the overall U.S. population, and the toxic equivalency factors proposed by Safe (1990) have been acknowledged to be conservative.

Schecter et al. (1993) reported on the comparison of the congener-specific measurement of CDDs, CDFs, and dioxin-like PCBs in whole blood samples of four individuals with known exposures to that of the general population. In this comparison, the analytical results of separate 450 ml blood samples collected from 50 Michigan residents, and a pooled blood sample from 5 donors at a blood bank in Missouri were used as the control group. Two of the exposed individuals were pulp and paper plant workers with potential exposure to dioxins, and the other two were Michigan residents who had elevated blood PCB levels from consuming contaminated fish. It was found that the control group and the pulp and paper mill workers who had no known exposures to PCBs had relatively high levels of coplanar, mono-ortho and di-ortho PCBs in their whole blood. On average, the Michigan and Missouri control samples showed a mean CDD/CDF TEQ concentration of 27 ppt and 25 ppt, respectively. These same samples showed PCB-TEQ (based on Safe, 1990) mean concentrations of 66 ppt for the Michigan controls, and 45 ppt for Missouri controls.

Levels of these compounds found in human tissue/blood appear similar in Europe and North America. Schecter (1991) compared levels of dioxin-like compounds found in blood among people from U.S. pooled samples from 100 subjects and Germany (85 subjects). Although mean levels of individual congeners differed by as much as a factor of two between the two populations, the total TEQ averaged 42 ppt in the German subjects and was 41 ppt in the pooled U.S. samples. In a later report, Schecter et al. (1994a) reported human blood levels for the general population from various countries. These data are presented in Table 5-11. Schecter (1991) reports adipose tissue levels in various countries, as summarized in Table 5-12. The adipose tissue data show more variation between countries but also involved much fewer samples, reducing confidence in the accuracy of the mean. Beck et al. (1994) reported on levels of CDD/CDFs in adipose tissue from 20 males (mean age-50 years) from Germany. TEQs ranged from 18 ppt to 122 ppt with a mean of 56 ppt, on a fat weight basis. Beck et al. (1994) also observed that CDD/CDF levels were found to be dependent on the age of the individual. 2,3,7,8-TCDD was found to increase at a rate of 0.12 pg/g fat per year, and TEQs increased at a rate of 0.77 pg/g fat per year. Beck et al. (1994a) also reported on CDD/CDF levels in various organs of the body. In comparison to adipose tissue, the concentrations of CDD/CDFs in brain and placental tissue were found to be low. Accumulation of CDD/CDFs was not found to occur in the thymus, spleen, and liver, based on whole weight concentrations. Schecter et al. (1994a) also reported on TEQ levels in organs of two autopsy patients from New York. The highest concentrations of CDD/CDFs were found in adipose tissue (28 ppt TEQ), adrenal tissue (14 ppt TEQ), and liver (12 ppt TEQ), on a whole weight basis. Lower concentrations were observed in spleen (4.6 ppt TEQ), muscle (2.4 ppt TEQ), and kidney (0.8 ppt TEQ). Schecter et al. (1994b) reported PCB levels for these two autopsy potients. Total PCBs in adipose tissue were 280.7 ppb on a wet weight basis and 344.2 ppb on a lipid weight basis.

In Chapter 6, the level of 2,3,7,8-TCDD found in human adipose tissue is assumed to average about 5.0 to 6.7 ppt in the United States based on data from a variety of studies. These adipose tissue data were used to estimate the associated exposure levels using a simple pharmacokinetic model that back calculates the dose needed to achieve the observed adipose tissue levels under the assumption of steady state exposure/dose. This model requires an estimate of the elimination rate constant. Based on available data, this elimination rate constant was assumed to be about 5 to 7 years which yielded a background dose rate of about 10 to 31 pg/day. This estimate agrees very well with the background exposure estimates (to 2,3,7,8-TCDD only) of 35 pg/day by Travis and Hattemer-Frey (1991), 25 pg/day by Fürst et al. (1991) and 12 pg/day from this assessment, all derived using typical media levels and contact rates. Further discussion of body burden data and associated exposures is presented in Chapter 6. Chapter 6 presents biologically-based pharmacokinetic models to estimate body burden levels. Some less sophisticated approaches have also been presented in the literature. For example, Travis and Hattemer-Frey (1988) developed linear relationships between the bioaccumulation of organic chemicals in human tissues and the octanol-water partition coefficients (K_ow) of the chemicals. The biotransfer factors (BTFs) that can be calculated using this relationship can be used to estimate adipose tissue and breast milk concentrations of organics. The BTF for human adipose tissue (B_f) is defined as the concentration of an organic in adipose tissue (mg/kg) divided by the average daily intake of that organic (mg/day). The human breast milk BTF (B_m) is defined as the concentration of an organic in breast milk (mg/kg) divided by the average daily intake of that organic (mg/day). Adipose tissue and milk concentrations are assumed to be equilibrium concentrations resulting from long-term, consistent daily intake of an organic.

Measured tissue concentrations and either measured or estimated daily chemical intakes were used to estimate the BTFs (12 chemicals for B_f and 6 chemicals for B_m). Geometric mean regression analysis was used to ascertain the correlation between K_ow values and the calculated BTFs. The results of the analyses are as follows:

B_f = 3.2 x 10-4 K_ow n = 12 r = 0.98

B_m = 6.2 x 10-4 K_ow n = 6 r = 0.94

The high correlation coefficients demonstrate that the BTFs for human adipose tissue and breast milk are strongly, positively correlated with the octanol-water coefficient. While the data upon which these correlations are based are limited both in terms of number of chemicals and the extent of measured vs. estimated intakes, the results are consistent with results reported by Travis and Hattemer-Frey (1988) for beef and dairy cattle.

Horstman and McLachlan (1994) measured CDD/F levels in human skin using an adhesive tape stripping method. Skin samples of the stratum corneum were collected from the backs of eight volunteers of varying age and sex. Two additional layers of increasing depth were collected from 5 people. All showed a decrease in CDD/F levels with depth. The concentration in the first layer ranged from 1,000 to 7,800 pg/g on a total CDD/F basis. The second layer was an average of 43 percent lower and the third layer was an average of 33 percent lower. OCDD was the dominant congener in all three layers. Also, non-2,3,7,8 substituted congeners were identified, congeners which are not normally present in human tissue.

In addition, samples of the epidermis and subcutis were analyzed. These analyses indicated that levels of the non-2,3,7,8 substituted congeners were much higher in the stratum corneum than in the epidermis and none were identified in the subcutis. The authors argue that because these congeners could not be transported from inside the body to the stratum corneum, the CDD/F in the stratum corneum must originate from external sources. Horstman and McLachlan (1994) hypothesized that textiles could be the source of skin contamination. Thirdy-five new textiles, primarily cotton products, were analyzed and found to have a total CDD/F level that was generally less than 50 ng/kg, but several colored T-shirts had high levels, with concentration up to 290,000 pg/g. The homolog patterns in the textiles were similar to the patterns found in the skin. Experiments were then conducted measuring the CDD/F levels in human skin before and after wearing T-shirts. Significant increases in CDD/F levels in the skin occurred after wearing the highly contaminated shirts for 1-2 weeks and significant decreases in CDD/F levels in the skin occurred after wearing the uncontaminated shirts for 1-2 weeks.

Thus, this work strongly suggests that dermal exposure to textiles may be contributing to background exposures to CDD/Fs. Horstman and McLachlan (1994) comment that although the levels of most CDD/F congeners in humans can be explained on the basis of diet, the origins of OCDD in humans is less clear. Since OCDD was found to be the dominant congener in textiles and skin, they speculate that the human body burden of this congener may result from dermal absorption. Horstman and McLachlan (1994) further discuss that human scale (stratum corneum) contributes to house dust and could lead to exposure via inhalation.

Certain groups of people may have higher exposures to the dioxin-like compounds than the general population. The following sections discuss the potential for higher exposures that result from dietary habits. Other population segments can be highly exposed due to occupational conditions or industrial accidents. For example, several epidemiological studies have evaluated whether elevated dioxin exposure has occurred to certain workers in the chemical industry, members of the Air Force who worked with Agent Orange, and residents of Seveso, Italy who were exposed as a result of a pesticide plant explosion. These epidemiological studies are fully discussed in the Epidemiology Chapter of the Dioxin Health Reassessment Document (EPA, 1994) and should be consulted if further details are desired.

Although certain subpopulations have the potential for high exposure to dioxin like compounds, a careful evaluation is needed to confirm this possibility. It would generally be inappropriate to compute the total background exposure for a particular group by simply adding the dioxin intake from the highly consumed food to the typical background exposure levels. Ideally the assessor should base this evaluation on the entire diet of the subpopulation and use case-specific values for food ingestion rates and concentrations of the dioxin-like compounds. The following points should be considered:

• Ingestion Rates - A subpopulation who has a high consumption rate of one particular food type is likely to eat less of other food types. For example, a subsistence fisher may ingest much more fish than the typical individual, but is also likely to ingest less meat of other types. Lacking case-specific data, it may be reasonable to assume that most people generally consume the same total amount of meats. The background exposure estimate assumes ingestion of 230 g/d of meat and eggs. Thus, as a first approximation, if a subsistence fisher consumes 140 g/d of fish, then he may consume 90 g/d of other meats and eggs.

• Concentrations in Foods - The levels of dioxin-like compounds in all major food types should be established on a case-specific basis. This is particularly important for the food groups which are consumed at unusually high rates.

• Estimating Intake - Finally the ingestion rates and concentrations are multiplied to get intake of dioxin-like compounds for each food category and then summed. This total can be compared to the typical intake of 119 pg of TEQ/d to determine if the subpopulation is truly exposed over background levels.

Schecter et al. (1992) reports that a study of 42 U.S. women found an average of 16 ppt of TEQ (3.3 ppt of 2,3,7,8-TCDD) in the lipid portion of breast milk. A much larger study in Germany (n= 526) found an average of 29 ppt of TEQ in lipid portion of breast milk (Fürst et al., 1994). Bates et al. (1994) analyzed breast milk samples from 38 women in New Zealand and reported mean lipid-based TEQs of 16.5 ppt for urban women and 18.1 ppt for rural women. The age of the mother was found to be positively correlated with the concentration of CDD/CDFs in breast milk. Beck et al. (1994) reported a mean TEQ of 30 ppt in the milk fat based on 112 human milk samples from Germany. The congeners that contributed the most to the total TEQ were 2,3,4,7,8-PeCDF (35 percent), total HxCDD (22 percent), and 1,2,3,7,8-PeCDD (21 percent). Beck et al. (1994) observed that CDD/CDFs levels decreased with the number of children and the duration of breast feeding, but increased with the age of the mother. Beck et al. (1994) also compared the adipose tissue levels of breast-fed and bottle-fed infants who had died of sudden infant death syndrome. The breast-fed infants had higher tissue levels (5.4 to 22 pg/g fat; n=4) than the bottle-fed infants (2.1 to 4.4 pg/g fat; n=2).

The levels in human breast milk can be predicted on the basis of the estimated dioxin intake by the mother. Such procedures have been developed by Smith (1987) and Sullivan et al. (1991) and are also presented in Chapter 6. The approach by Smith assumes that the concentration in breast milk fat is the same as in maternal fat and can be calculated as:

where,

C_{milk fat} = concentration in maternal milk (pg/kg of milk fat)

m = average maternal intake of dioxin (pg/kg of body weight/day)

h = half-life of dioxin in adults (days)

f₁ = proportion of ingested dioxin that is stored in fat

f₂ = proportion of mother's weight that is fat (kg maternal fat/kg total body weight)

This steady-state model assumes that the contaminant levels in maternal fat remain constant. Though not described here, Smith (1987) also presents more complex approaches that account for changes in maternal fat levels during breast feeding. The model developed by Sullivan et al. (1991) is a variation of the models proposed by Smith (1987). The Sullivan model considers changes in maternal fat levels and predicts chemical concentrations in milk fat as a function of time after breast feeding begins. The model proposed by Smith assumes that infant fat concentration at birth is zero, whereas Sullivan assumes that the infant fat concentration at birth is equal to the mother's fat concentration.

As discussed in Chapter 6, the half-life of 2,3,7,8-TCDD in humans is estimated to be 5 to 7 years. For the purpose of this preliminary analysis, it is assumed that a 7-year half-life applies to all of the dioxin-like compounds. Smith (1987) suggests values of 0.9 for f₁ and 0.3 for f₂. Using these assumptions and a background exposure level of 1 to 3 pg of TEQ/kg-d (derived from diet analysis, see Section 5.3), the concentration in breast milk fat is predicted to be about 10 to 30 ppt of TEQ, which agrees well with the measured values.

Using the estimated dioxin concentration in breast milk, the dose to the infant can be estimated as follows:

where,

ADD_infant = Average daily dose to the infant (pg/kg/d)

IR_milk = Ingestion rate of breast milk (kg/d)

ED = Exposure duration (yr)

BW_infant = Body weight of infant (kg)

AT = Averaging time (yr)

f₃ = Fraction of fat in breast milk

f₄ = Fraction of ingested contaminant that is absorbed

This approach assumes that the contaminant concentration in milk represents the average over the breast feeding time period. If the dynamic models mentioned above are used, the dose can be estimated using an integration approach to account for the changes in concentration over time.

Smith (1987) reports that a study in Britain found that the breast milk ingestion rate for 7 to 8-month old infants ranged from 677 to 922 ml/d and that a study in Houston measured the mean production of lactating women to range from 723 to 751 g/d. Smith (1987) also reports that breast milk ingestion rates remain relatively constant over an infant's life, that the milk can be assumed to have a 4 percent fat content, and that 90 percent of the ingested contaminant are absorbed. The National Center for Health Statistics (1987) reports the following mean body weights for infants:

6-11 months: 9.1 kg

1 year: 11.3 kg

2 year: 13.3 kg

Using Equation 5.2 and assuming that an infant breast feeds for 1 year, has an average weight during this period of 10 kg, ingests 0.8 kg/d of breast milk, and that the dioxin concentration in milk fat is 20 ppt of TEQ, the ADD to the infant over this period (i.e., AT = 1 yr) is predicted to be about 60 pg of TEQ/kg-d. This value is much higher than the estimated range for background exposure to adults (i.e., 1-3 pg of TEQ/kg-d). However, if a 70 year averaging time is used, then the LADD (Lifetime Average Daily Dose) is estimated to be 0.8 pg of TEQ/kg-d, which is near the lower end of the adult background exposure range. On a mass basis, the cumulative dose to the infant under this scenario is about 210 ng compared to a lifetime background dose of about 1700 to 5100 ng (suggesting that 4 to 12 percent of the lifetime dose may occur as a result of breast feeding). Traditionally, EPA has used the LADD as the basis for evaluating cancer risk and the ADD (i.e., the daily exposure per unit body weight occurring during an exposure event) as the more appropriate indicator of risk for noncancer endpoints. This issue is discussed further in Chapter 6 and in the companion document on dioxin health effects.

The simplified procedure described above contains a number of uncertainties. A tendency toward overestimates of the dose to the infant is caused by the assumption that reductions do not occur in maternal fat levels during breast feeding. Sullivan et al. (1991) estimates that the steady-state assumption may lead to overestimates of 20 percent. Uncertainty is also introduced by the assumption that the assumed half-life rate and partitioning factors apply to all the dioxin related compounds. Although these properties are likely to be similar among the various congeners, some variation is expected. It is unknown whether the net effect of these uncertainties would lead to over or under estimates of dose. However, the simple model appears to provide reasonable predictions of background levels found in breast milk and was judged adequate for purposes of a preliminary analysis. For detailed assessments, readers should consider using the more complex models and developing chemical-specific property estimates.

Travis and Hattemer-Frey (1988) presented an alternative approach to estimating breast milk contaminant levels. They proposed a biotransfer approach:

where:

C_m = contaminant concentration in breast milk (mg/kg)

B_m = biotransfer factor for breast milk (kg/d)

I = maternal intake of contaminant (mg/d)

They also argue that the biotransfer factor is primarily a function of the octanol-water partition coefficient (K_ow ) and developed the following geometric mean regression:

This regression was derived from data on 6 lipophilic compounds (log K_ow range: 5.16 to 6.5), but did not include any dioxins or furans. Assuming a log K_ow of 6.6 for 2378-TCDD, a B_m of 3700 kg/d is predicted. Combing this value with a maternal intake of 10 pg/d (or 10^-7 mg/d), a breast milk concentration of 37 ppt is predicted. This prediction is about 10 times higher than what has been measured in the U.S. Thus, this approach does not appear to work as well as the earlier approach suggested by Smith et al (1987).

The possibility of high exposure to dioxin as a result of fish consumption is most likely to occur in situations where individuals consume a large quantity of fish from one location where the dioxin level in the fish are elevated above background levels. Most people eat fish from multiple sources and even if large quantities are consumed they are not likely to have unusually high exposures. However, individuals who fish regularly for purposes of basic subsistence are likely to obtain their fish from one source and have the potential for elevated exposures. Such individuals may consume large quantities of fish. EPA (1989) presents studies that indicate that recreational anglers near large water bodies consume 30 g/day (as a mean) and 140 g/day (as an upper estimate). Wolfe and Walker (1987) found subsistence fish ingestion rates up to 300 g/day in a study conducted in Alaska.

Svensson et al. (1991) found elevated blood levels of CDDs and CDFs in high fish consumers living near the Baltic Sea in Sweden. Three groups were studied: nonconsumers (n=9), moderate consumers (n=9, 220 to 500 g/wk) and high consumers (n=11, 700-1750 g/wk). The high consumer group was composed of fishermen or workers in the fish industry who consumed primarily salmon (30 - 90 pg TEQ/g) and herring (8-18 pg TEQ/g) from the Baltic Sea. The TEQ blood level was found to average about 60 pg TEQ/g lipid among the high consumers and 20 pg TEQ/g lipid for the nonconsumers. This difference was particularly apparent for the PeCDFs.

Studies are underway to evaluate whether native Americans living on the Columbia River in Washington have high dixoin exposures as a result of fish consumption. These tribes consume large quantities of salmon from the river. A recent study (Columbia River Intertribal Fish Commission, 1993) suggests that these individuals have an average fish consumption rate of 30 g/day and a 95th percentile rate of 170 g/day. Currently studies are underway to measure dioxin levels in fish from this region.

Dewailly et al. (1994) observed elevated levels of coplanar PCBs in the blood of fishermen on the north shore of the Gulf of the St. Lawrence River who consume large amounts of seafood. Of the 185 study samples, the 10 samples with the highest total PCB levels were analyzed for coplanar PCBs. Samples from Red Cross blood donors in Ontario served as controls. Coplanar PCB levels were 20 times higher among the 10 highly exposed fishermen than among the controls. Based on these results of the 10 highest samples, Dewailly et al. (1994) estimated that for the entire fishing population studied, coplanar PCB levels would be eight to ten times higher than the control group. Dewailly et al. (1994) also observed elevated levels of coplanar PCBs in the breast milk of Inuit women of Arctic Quebec. The principal source of protein for the Inuit people is fish and sea mammal consumption. Breast milk samples were collected from 109 Inuit women within the first three days after delivery and analyzed for di-ortho-coplanar PCBs during 1989 and 1990. Subsets of 35 and 40 randomly selected samples were analyzed for mono-ortho coplanar and non-ortho coplanar PCBs, respectively. Samples from 96 caucasian women from Quebec served as controls. The levels of non-ortho coplanar PCBs for Inuit women ranged from 24.7 to 220.9 ppt. These values were 3 to 7 times higher than those observed in the control group. For mono-ortho and di-ortho coplanar PCBS, the levels among the Inuit women were three to ten times higher than in the control group.

The possibility of high exposure to dioxin as a result of consuming meat and dairy products is most likely to occur in situations where individuals consume a large quantity of these foods from one location where the dioxin level is elevated above background levels. Most people eat meat and diary products from multiple sources and even if large quantities are consumed they are not likely to have unusually high exposures. However, individuals who raise their own livestock for purposes of basic subsistence have the potential for elevated exposures. No epidemiological studies were found in the literature that evaluated this issue. However, Volume III of this document presents methods for evaluating this type of exposure on a site-specific basis.

Bates, M.N.; Hannah, D.J.; Buckland, S.J.; Taucher, J.A.; Van Maanen, T. (1994) Chlorinated organic contaminants in breast milk of New Zealand women. Environmental Health Perspectives. Vol. 102, Suppl. 1:211-217.

Beck, H.; Dross, A.; Mathar, W. (1994) PCDD and PCDF exposure and levels in humans in Germany. Environmental Health Perspectives. Vol. 102, Suppl. 1:173-185.

Columbia River Intertribal Fish Commission (1993) A fish consumption survey of the Umtilla, Nez Perce, Yakima, and Warm Springs tribes of the Columbia River basin. Peer Review Draft Report.

Dewailly, E.; Ryan, J.J.; Laliberte, C.; Bruneau, S.; Weber, J.P.; Gingras, S.; Carrier, G. (1994) Exposure of remote maritime populations to coplanar PCBs. Environmental Health Perspectives. Vol. 102, Suppl. 1:205-209.

Fürst, P.; Fürst, C.; Groebel, W. (1990) Levels of PCDDs and PCDFs in food stuffs from the Federal Republic of Germany. Chemosphere 20(7-9): 787-792.

Fürst, P.; Fürst, C.; Wilmers, K. (1991) Body burden with PCDD and PCDF from food. In: Gallo, M.; Scheuplein, R.; Van der Heijden, K. eds. Biological basis for risk assessment of dioxins and related compounds. Banbury Report #35. Plainview, NY: Cold Spring Harbor Laboratory Press.

Fürst, P.; Fürst, C.; Wilmers, K. (1994) Human milk as a bioindicator for body burden of PCDDs, PCDFs, organochlorine pesticides, and PCBs. Environmental Health Perspectives. Vol. 102, Suppl. 1:187-193.

Gilman, A.; Newhook, R. (1991) An updated assessment of the exposure of Canadians to dioxins and furans. Chemosphere 23(11-12): 1661-1667.

Henry, S.; Cramer, G.; Bolger, M.; Springer, J.; Scheuplein, R. (1992) Exposures and risks of dioxin in the U.S. food supply. Chemosphere 25(1-2):235-238.

Hites, R.A. (1991) Atmospheric transport and deposition of polychlorinated dibenzo-p-dioxins and dibenzofurans. Prepared for the U.S. Environmental Protection Agency, Methods Research Branch, Atmospheric Research and Exposure Assessment Laboratory, Office of Research and Development, Research Triangle Park, NC. EPA/600/3-91/002.

Horstmann, M.; McLachlan, M.S. (1994) Textiles as a source of polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDD/F) in human skin and sewage sludge. Environ. Sci. and Pollut. Res. 1(1):15-20.

Koester, C.J.; Hites, R.A. (1992) Wet and dry deposition of chlorinated dioxins and furans. Environ. Sci. Technol. 26:1375-1382.

McCrady, J.K.; McFarlane, C.; Gander, L.K. (1990) The transport and fate of 2,3,7,8-TCDD in soybean and corn. Chemosphere 21:359-376.

Patterson, D.G.; Todd, G.D.; Turner, W.E.; Maggio, V.; Alexander, L.R.; Needham, L.L. (1994) Levels of nonortho-substituted polychlorinated biphenyls, dibenzo-p-dioxins, and dibenzofurans in human serum and adipose tissue. Environmental Health Perspectives. Vol. 101, Suppl. 1:195-204.

Safe, S. (1990) Polychlorinated biphenyl, dibenzo-p-dioxins, dibenzofurans, and related compounds: environmental and mechanistic consideratins which support the development of toxic equivalency factors. CRC Crit. Rev. Toxicol. 21:51-88.

Schecter, A. (1991) Dioxins and related chemicals in humans and in the environment. In: Gallo, M.; Scheuplein, R.; Van der Heijden, K. eds. Biological basis for risk assessment of dioxins and related compounds. Banbury Report #35. Plainview, NY: Cold Spring Harbor Laboratory Press.

Schecter, A.; di Domenico, A.; Tirrio-Baldassarri, L.; Ryan, J. (1992) Dioxin and dibenzofuran levels in milk of women from four geographical regions in Italy as compared to levels in other countries. Presented at: Dioxin '92, 12th International symposium on Chlorinated Dioxins and Related Compounds; Tampere, Finland; August 1992.

Schecter, A.; DeVito, M.J.; Stanely, J.; Boggess, K. (1993) Dioxins, dibenzofurans and dioxin-like PCBs in blood of Americans. Presented at: Dioxin '93, 13th International Symposium on Chlorinated Dioxins and Related Compounds; Vienna, Austria; September, 1993.

Schecter, A.; Fürst, P.; Fürst, C.; Päpke, O.; Ball, M.; Ryan, J.; Cau, H.D.; Dai, L.C.; Quynh, H.T.; Cuong, H.Q.; Phuong, N.T.N.; Phiet, P.H., Biem, A.; Constable, J.; Startin, J.; Samedy, M.; Seng, Y.K. (1994a) Chlorinated dioxins and dibenzofurans in human tissue from general populations; a selective review. Environmental Health Perspectives Vol. 102, Suppl. 1:159-171.

Schecter, A.; Stanley, J.; Boggess, K.; Masuda, Y.; Mes J.; Wolff, M.; Fürst, P.; Fürst, C.; Wilson-Yang, K.; Chisholm, B. (1994b) Polychlorinated biphenyl levels in the tissues of exposed and nonexposed humans. Environmental Health Perspectives. Vol. 102, Suppl. 1:149-158.

Smith, A.H. (1987) Infant exposure assessment for breast milk dioxins and furans derived from waste incineration emissions. Risk Analysis. 7(3):347-353.

Smith, R.M.; O'Keefe P.; Briggs, R.; Hilker, D.; Connor, S. (1992) Measurement of PCDFs and PCDDs in air samples and lake sediments at several locations in upstate New York. Chemosphere 25:1-2, pp. 95-98.

Svensson, B.G.; Nelsson, A.; Hansson, M.; Rappe, C.; Akesson, B.; Skerfving, S. (1991) Exposure to dioxins and dibenzofurans through the consumption of fish. New England Journal of Medicine. 324(1):8-12.

Sullivan, M.J.; Custance, S.R.; Miller, C.J. (1991) Infant exposure to dioxin in mother's milk resulting from maternal ingestion of contaminated fish. Chemosphere 23(8-10):1387-1396.

Theelen, R.M.C. (1991) Modeling of human exposure to TCDD and I-TEQ in the Netherlands: background and occupational. In: Gallo, M.; Scheuplein, R.; Van der Heijden, K. eds. Biological basis for risk assessment of dioxins and related compounds. Banbury Report #35. Plainview, NY: Cold Spring Harbor Laboratory Press.

Theelen, R.M.C.; Liem, A.K.D.; Slob, W.; Van Wijnen, J.H. (1993) Intake of 2,3,7,8 chlorine substituted dioxins, furans, and planar PCBs from foods in the Netherlands, median and distribution. Chemosphere. 27(9):1625-1635.

Travis, C.C.; Hattemer-Frey, H.A. (1987) Human exposure to 2,3,7,8-TCDD. Chemosphere 16:2331-2342.

Travis, C.C.; Hattemer-Frey, H.A. (1988) Relationship between dietary intake of organic chemicals and their concentrations in human adipose tissue and breast milk. Arch. Environ. Contam. Toxicol. 17:473-478.

Travis, C.C.; Hattemer-Frey, H.A. (1991) Human exposure to dioxin. Sci. Total Environ. 104: 97-127.

U.S. Department of Agriculture (1992) Food and nutrient intakes by individuals in the United States, 1 day, 1987-88: Nationwide Food Consumption Survey 1987-88. Washington, DC: USDA Human Nutrition Information Service. NFCS Rpt. No. 87-I-1 in preparation.

U.S. Department of Agriculture (1993) Food consumption, prices, and expenditures, 1970-1992. Washington, DC: USDA Economic Research Service. Statistical Bulletin 867.

U.S. Environmental Protection Agency (1989) Exposure factors handbook. Washington, DC: Office of Health and Environmental Assessment. EPA/600/8-89/043.

U.S. Environmental Protection Agency (1990a) Background document to the integrated risk assessment for dioxins and furans from chlorine bleaching in pulp and paper mills. Washington, DC: Office of Toxic Substances. EPA 560/5-90-014.

U.S. Environmental Protection Agency (1990b) Chlorinated dioxins and furans in the general U.S. population: NHATS FY1987 results. Washington, DC: Office of Toxic Substances. EPA 560/5-91-003.

U.S. Environmental Protection Agency (1991a) Feasibility of environmental monitoring and exposure assessment for a municipal waste combustion: Rutland Vermont Pilot Study. Washington, DC: Office of Research and Development. EPA-600/8-91/007.

U.S. Environmental Protection Agency (1991b) Chlorinated dioxins and furans in the general U.S. population: NHATS FY87 results, Washington, DC: Office of Toxic Substances. EPA-560/5-91-003.

U.S. Environmental Protection Agency. (1994) Health Assessment for 2,3,7,8-TCDD and related compounds, Washington, DC: Office of Research and Development. EPA 600/BP-92/001.

Welge, P.; Wittsiepe, J.; Schrey, P.; Ewers, U.; Exner, M.; Sclenka, F. (1993) PCDD/F-levels in human blood of vegetarians compared to those of non-vegetarians. Presented at: Dioxin '93, 13th International Symposium on Chlorinated Dioxins and Related Compounds; Vienna, Austria; September 1993.

Wolfe, R.J.; Walker, R.J. (1987) Subsistence Economics in Alaska: Productivity, Geography and Developmental Impacts. Arctic Anthropology 24(2):56-81.

Yang, Y-Y.; Nelson, C.R. (1986) An estimation of daily food usage factors for assessing radionuclide intake in the U.S. population. Health Phys. 50(2):245-257.