Document 6k3b5M7de0QDqEebL2Zx6QJR

5.0. POTENTIALLY ELEVATED EXPOSURES 5.1. INTRODUCTION Certain groups of people may have higher exposures to the dioxin-like compounds than the general population. The follow ing sections discuss higher exposures that may result from dietary habits, localized im pacts, and cigarette smoking. Other population segm ents can be highly exposed due to occupational conditions or industrial accidents. For example, several epidem iological studies have evaluated w hether elevated dioxin exposure has occurred to certain w orkers in the chemical industry, members of the Air Force w ho worked w ith A gent Orange, and residents of Seveso, Italy, w ho were exposed as a result of a pesticide plant explosion. These epidem iological studies are fully discussed in the Epidemiology Chapter of the Dioxin Health Reassessment Docum ent (U.S. EPA, 1996) and should be consulted if further details are desired. This chapter, however, does not address occupational or accidental exposure. Instead, it focuses on elevated exposures among the general population from dietary habits such as breast feeding or high rates of fish ingestion, localized sources, or cigarette smoking. 5.2. NURSING INFANTS Nursing infants may be exposed to dioxin-like com pounds via consum ption of breast milk. These com pounds are deposited in the fa tty tissues (i.e., adipose tissue, blood lipids, and breast milk) of the m other and may be transferred to the infant during nursing. Based on data from 1989, approxim ately 52 percent of U.S. m others initiate breastfeeding w ith their newborn infants, and 40 percent continue breastfeeding for 3 m onths or longer (NAS, 1991). A t 5 to 6 months of age, only about 20 percent of infants are breast-fed (NAS, 1991). This section w ill show how breast m ilk ingestion exposures, w hich are higher during breast-feeding, on a body w e ig h t basis, than during any other period in an individual's life, im pact lifetim e exposures and body burdens. First, data showing the im pact of breast milk ingestion to the infant body burden of CDD/CDF/PCBs are reviewed. Then, an estim ate of the dose (average daily dose, ADD, and lifetim e average daily dose, LADD) to the infant via breast m ilk is made. The section w ill close by developing, testing, and applying a model on the im pact of breast-feeding to body burdens for growing infants. DRAFT--DO NOT QUOTE OR CITE 5-1 December 2003 5 .2 .1 . The Impact of Breast Feeding on Infant Body Burden Abraham et al. (1994 , 1995) studied CDD/CDF/PCB levels in the blood of a breast fed and a form ula-fed infant at 11 and 25 m onths of age. Sampling of blood showed that the body burden of dioxin-like com pounds w as more than an order of m agnitude higher for the breast-fed infant than the formula-fed infant during both tim e periods, w ith CDD/CDF ranging from 3 4 .7 (11 m onths) to 4 3 .9 (25 months) ppt TEQDF-W HO98 lipid-basis in the breast-fed infant compared to 2.7 to 3.3 ppt TEQDF-W HO98 for the form ula-fed infant. Dioxin-like PCB concentrations were sim ilarly an order of m agnitude d ifferent, w ith the breast-fed infant having a concentration of 3 1 .4 ppt TEQP-W HO98, compared to 2.5 ppt TEQP-W HO98 for the form ula-fed infant at 11 m onths (PCB 126 not measured at 25 m onths, so a comparison for th a t age is not inform ative). The full congener profiles for these results are show n in Table 5-1. The increase in the lipid-based CDD/CDF TEQ concentration in the blood of the breast-fed infant at 25 m onths w as attributed to the relative decrease in body fat mass during the period between sampling and slight increases in body burden concentrations. Abraham et al. (1994) also analyzed m other's m ilk at 1 m onth and m other's blood along w ith the in fa n ts' blood at 11 m onths. They found m other's m ilk to contain 23.5 ppt TEQdf-W HO98 at 1 m onth and m other's blood to contain 14.2 ppt TEQDF-W HO98 at 10 m onths. If blood and m ilk concentrations in a m other during lactation are the same at any given tim e, then these data suggest a reduction of about 40 percent in TEQ concentration in the m other between the 1st and 10th m onth of lactation. The reduction in m other's milk concentration of dioxins during nursing is discussed further in Section 5 .2 .3 below. Kreuzer et al. (1997) developed and tested a toxicokinetic model of human lifetim e body burden of TCDD, starting w ith a model for breast-feeding. To support their model, they presented adipose tissue and liver data on 3 stillborn and 17 infants w ho had died from sudden infant death syndrom e (SIDS). Nine of the 17 infants had spent some portion of their lives breast-feeding, while the other 8 infants were formula-fed. Average congener and TEQ concentrations for these three groups are show n in Table 5-2. The highest TEQ concentrations were found in the infants w ho had some breast-feeding, w ith adipose concentrations at 15.9 ppt TEQDF-W HO98, as compared to form ula-fed infants who had concentrations at 4 .3 ppt TEQDF-W HO98. The breast-fed in fa n ts' concentrations included four infants w ho were weaned several weeks prior to their death from SIDS. DRAFT--DO NOT QUOTE OR CITE 5-2 December 2003 This may have generally led to reductions in their body burdens as their higher daily intake from breast-feeding was reduced after weaning. The average TEQ concentration for the five infants w ho died w hile still breast-feeding w as 20.1 ppt TEQDF-W HO98. The highest concentration found was for the infant w ho was breast-fed the longest at 19 weeks, and w ho died at th a t tim e; the TEQ concentration was 35 ppt TEQDF-W HO98. Other breast fed infants, how ever, did not have as much im pact - infants w ho died w hile breast-feeding at 12 and 16 weeks had concentrations of 9 and 7.5 ppt TEQDF-W HO98. W hile Kreuzer et al. (1997) concluded th a t breast-fed infants had elevated concentrations compared to form ula-fed infants, they also observed th a t breast-fed infants had adipose TEQ concentrations that were w ith in the range or lower than the values published for adults. Abraham et al. (2000) reported on a study in Germany in w hich the blood of 80 breast-fed infants between the ages of 4 and 11 m onths were analyzed for the 17 CDD/CDFs . Of these 80 infants, 27 were from a region where a copper recycling plant led to elevations in the m other's milk. The I-TEQDF blood concentration in this group of 80 children at 11 m onths ranged between 2.0 and 107 pg/g lipid-basis, w ith a median of 25 .3 ppt. Of these children, 6 had I-TEQDF concentrations greater th a t 50 ppt, and 5 of these 6 were from the region impacted by the copper recycling plant. From a control group of 21 children w ho had been form ula-fed, individual dioxin measurements were performed in 5 children. Concentrations were found to range narrow ly from 1.9 to 3.2 ppt I-TEQDF lipid-basis at 11 m onths. W ith several measurements over tim e, Abraham et al. (2000) found the blood concentrations of the breast-fed infants to increase over time. Infant or tim e-specific concentration data were not presented in Abraham et al. (2000). However, they did show the ratio between the concentration of specific com pounds, as w ell as the I-TEQDF concentrations, in the children's blood and the blood of the mothers. When the ratio exceeded 1.0, this meant that the child's concentration was higher than the m other's. Abraham et al. (2000) showed the I-TEQDF ratio to increase from 1.12 at 16-24 weeks (n = 11 meaning there were 11 paired measurements at th a t tim e) to 2.12 at 2 5 -32 weeks (n = 29) to 3 .2 0 at 3 3 -4 0 weeks (n = 33) to 3 .7 3 at 4 1 -4 8 weeks (n = 7). This was the m ost com prehensive data set found in the literature, although it was not fully described in the Abraham et al. (2000) abstract. Patandin et al. (1997) looked at the plasma levels of four polychlorinated biphenyls (PCBs) in 173 Dutch children 3.5 years of age, 91 of w hich had been breast-fed and 82 DRAFT--DO NOT QUOTE OR CITE 5-3 December 2003 of w hich had been form ula-fed. Children in the breast-fed group had significantly higher median PCB levels in plasma (p < 0 .0 0 0 1 ) than children in the form ula fed group. The four PCBs measured were 118, 138, 153, and 180. The median sums of these four PCBs in the tw o groups of children were 0 .7 5 :g /L in the breast-fed group versus 0.21 :g /L in the form ula fed group. By means of an extensive questionnaire on dietary history, combined w ith data on the concentrations of dioxin and PCBs in foods provided by the Dutch National Institute of Public Health and the Environm ent, Pantadin et al. (1997) were able to determine th a t the TEQ intake via the diet w as virtually indistinguishable in the tw o groups. They found th a t PCB levels in the breast-fed children were significantly correlated w ith the period of breast-feeding (r = 0 .6 3 ), m ilk PCB levels (r = 0 .3 9 ), and the total TEQ in breast m ilk (r = 0 .36). They concluded th a t the plasma PCB levels in Dutch children were the result of exposure through breast m ilk and in utero exposure, and th a t the influence of dietary intake of PCBs after weaning is small compared to the intake during breast feeding. 5 .2 .2 . Calculation of an Average Daily Dose from Breast-Feeding Using the estim ated dioxin concentration in breast m ilk, the administered dose to the infant can be estim ated as fo llo w s : ^milk fat * ^3 * ^milk * infant BWinfant * AT where, AD Dinfant = Average daily dose to the infant (pg/kg-d); Cmilkfat Concentration in m ilk fat (pg/g); IRmilk Ingestion rate of breast milk (kg/d); ED = Exposure duration (yr); BW infant AT = Body w eight of infant (kg); = Averaging tim e (yr); and f 3 = Fraction of fa t in breast m ilk. (Eqn. 5-1 DRAFT--DO NOT QUOTE OR CITE 5-4 December 2003 The administered dose can be converted to an absorbed dose by m ultiplying by the fraction of ingested contam inant th a t is absorbed. This approach assumes th a t all pertinent parameters, including the body w e ig h t of the infant, the infant ingestion rate of breast m ilk, and perhaps m ost im portantly, the contam inant concentration in m ilk, represent the average over the breast feeding tim e period. Sm ith (1987) reported th a t a study in Britain found th a t the breast m ilk ingestion rate for 7- to 8-m onth old infants ranged from 677 to 9 22 mL/d and th a t a study in Houston measured the mean production of lactating women to range from 723 to 751 g/d. Smith (1987) also reported th a t breast m ilk ingestion rates remain relatively constant over an in fa n t's life. For purposes of estim ating the dose to breast feeding infants, a milk ingestion rate of 8 00 mL/d w as assumed in the analysis presented in this section. Smith (1987) also assumed th a t m other's m ilk has a 4 percent fa t content, and th a t 80 percent of the ingested contam inant are absorbed. The infant w e igh t varies w ith tim e. For example, a typical infant (average of male and female data) w eighs about 3.3 kg at birth, 7.9 kg at 6 m onths, and 10.2 kg at 1.0 year (U.S. EPA, 199 7; W alker and W atkins, 1997). The concentration of dioxin in the m other's m ilk is also expected to change, since lactation provides a significant avenue of depuration. Lakind et al. (2000) cite several references where measurements of breast milk concentrations of lipophilic compounds (PCBs, DDE, DDT, CDD/CDFs) were show n to decline during the course of lactation. They fit available data on 2,3,7,8-TC D D to a curve, and their resulting relationship showed an 86 percent loss over 6 m onths. This is comparable to a modeling e ffo rt by Kreuzer et al. (1997 ), w ho modeled a 70 percent decline in TCDD concentrations after 6 months. Their model was more m echanistic, and added the loss by breast m ilk to an overall female body burden model w hich included inputs by food consum ption and outputs by m etabolic and non-m etabolic pathw ays. Patandin et al. (1999 ), in their modeling of dioxin exposures from infancy to adulthood, cited data from Germany and England to conclude th a t breast milk concentrations of TCDD decline by 20 percent every 3 months. The data described in the previous section by Abraham et al. (1994) suggest a decline of 40 percent of TEQs from 1 month to 10 months of lactation. DRAFT--DO NOT QUOTE OR CITE 5-5 December 2003 For assignment of CmNkfat in Equation 5-1, therefore, one w ould have to assign a concentration at birth and consider a decline in th a t concentration over tim e. For purposes of this discussion, a concentration of 25 ppt TEQDFP-W HO98 on a lipid basis is assumed for m other's m ilk when lactation begins (which is the average tissue concentration derived from recent studies of dioxins in blood in background settings of the US, reviewed in Chapter 4). It is then assumed to linearly drop by 50 percent after 6 m onths, w ith an additional linear drop of 50 percent by the end of 12 m onths, for a total decline of 75 percent from initial concentrations. These assignments in concentration decline are in the middle of the range reported above. They translate to concentrations of 12.5 ppt TEQDFP-W HO98 after 6 m onths and 6.3 ppt TEQDFP-W HO98 after a year, given a starting concentration of 25 ppt TEQDFP-W HO98. The proper w ay to derive an average dose to the child is to integrate Equation 5-1 over the tim e period of interest. A t birth, for example, w ith a m other's milk concentration of 25 ppt TEQdf-W HO98, an infant body w e igh t of 3.3 kg, an average ingestion rate of 8 00 g/d breast m ilk, the administered dose is predicted to be 242 pg TEQDF-W HO98/kg bw /day [(25 pg/g x 0 .0 4 x 800 g/d) / (3.3 kg) = 242 pg/kg-d]. Table 5-3 show s infant body w eights, as w ell as doses of dioxin TEQDF-W HO98 expressed in term s of pg/day for each of the first 12 m onths of life. These body w e igh t data are the averages for male and female infants (U.S. EPA, 1997; W alker and W atkins, 1997), and along w ith other data, were used in the pharm acokinetic exercise described in the next section. Calculating m onthly doses on the basis of body w e ig h t for each of the first tw e lve m onths of life and then dividing by 12, results in an average dose to the infant of 87 pg TEQDF-W HO98/kg bw/day. This value is much higher than the estim ated average background TEQDFP-W HO98 dose for adults of approxim ately 1 pg TEQDFP-WHO98/kg-d. However, if a 70 year averaging tim e is used for this one-year nursing scenario, then the LADD (Lifetim e Average Daily Dose, calculated as ADD * ED/LT where LT is lifetim e typically assumed to be 70 years) is estim ated to be 1.2 pg TEQDFP-W HO98/kg-d ([87 pg/kg-d] * 1 y r/7 0 yr). This is close to the adult background dose of 1.0 pg TEQDFP-W HO98/kg-d. However, this can be misleading because it ignores the difference in daily intake during potentially sensitive stages in development. Also, it does not consider any exposures past the first year of life. In order to calculate a true lifetim e average daily dose, one needs to incorporate the DRAFT--DO NOT QUOTE OR CITE 5-6 December 2003 changes in dose over various life stages. Using the estim ates of dose derived in Chapter 4 for various ages in children: 1-5: 3 .3 pg TEQDFP/kg-d, 6-11: 1.9 pg TEQDFP/kg-d, and 12 19: 1.1 pg TEQDFP/kg-d, the follow ing calculates the LADD for lifetim e background exposures considering one year of breast-feeding: LADD = 87 -- 1^!- + 3.3 + 1.9 + 1.1 + 1.0 70yrs 70yrs 70yrs 70yrs 70yr (Eqn. 5-2) pg TE Q TM LADD = 2 .4 5 -- -------- -- kg- day (Eqn. 5-3) On a mass basis, the cum ulative dose to the infant after a year is about 238 ng TEQDFPW HO 98 (87 pg/kg-d x 7.5 kg x 365 d x n g /1 ,0 0 0 pg; 7.5 kg is an average annual w e igh t based on the average of 12 m onthly body weights). Using the age-dependent doses derived in Chapter 4 w ith assumed body w eights, a dose from year 1 to year 70 in a 70year lifetim e is estim ated to be about 1,687 ng TEQDFP-W HO98, so th a t a total lifetim e dose is 1,925 ng TEQDFP-W HO98 (1 ,6 8 7 + 238). This suggests th a t about 12 percent of lifetim e dose (2 3 8 /1 9 2 5 * 100 percent) may occur as a result of breast feeding, if that feeding occurred for one year. This exercise describes accumulated dose over a lifetim e, given a year of breast feeding. It w as found th a t about 12 percent of lifetim e dose came from breast-feeding. It was also found that the LADD for this scenario, including average dioxin doses after the first year until year 70, was 2.45 pg TEQDFP-W HO98/kg-day. The issue of accum ulative dose is explored in more detail in the next section, where the q ua ntity, "area under the curve" is defined and used to assess overall exposure during a lifetim e, or portions of a lifetim e, under different breast-feeding scenarios as w ell as a form ula-only scenario where the infant experiences only background exposures. 5 .2 .3 . Modeling the Impact of Breast-Feeding on Infant Body Burden The previous section described an approach to estim ate doses received by an infant due to breast-feeding, w hich included a starting concentration in m other's m ilk, a decline of th a t concentration over tim e (and the resulting decline in dose delivered to the child), DRAFT--DO NOT QUOTE OR CITE 5-7 December 2003 and the child's changing body w eight. That inform ation w ill be used in this section to evaluate the im pact of breast feeding on an in fa n t's body burden of dioxins. To better evaluate the impact of nursing on infants, a one-compartment non-steady state pharmacokinetic model was used to evaluate dioxin tissue levels. As described below , this model was validated using paired m other/child data on breast m ilk and infant blood concentrations of TEQDF-W HO98. Following this validation, several breast-feeding scenarios were modeled and compared w ith a form ula-feeding only scenario. Specifically, changes in infant TEQDFP-W HO98 tissue concentration over tim e were modeled for these scenarios: formula only, 6 weeks nursing, 6 months nursing, 1 year nursing, and 2 years nursing. The section closes w ith sensitivity analyses exercises which describe the model response to changes in the key parameters describing the dose received by the infant via breast m ilk and the rate of dissipation of dioxin TEQs in the infant. 5.2 .3 .I. Description o f the Model The pharmacokinetic model was based on the follow ing differential equation describing the mass balance of dioxin in lipids (Pinsky and Lorber, 1998): da(t)/dt = f D(t) - k(t) a(t) c(t) a(t) 1000 V(t) Eqn. (5-4) Eqn. (5-5) where: a(t) = total mass of dioxins in lipid (pg) at tim e t; c(t) = concentration of dioxins in lipid (pg/g) a tim e t; D(t) = ingested dose of dioxins (pg/yr) at tim e t; V(t) = lipid w e igh t (kg) at tim e t; k(t) = elim ination rate constant (yrs-1) at tim e t; t = tim e (yrs); and f = fraction of ingested dose absorbed into lipid com partm ent (unitless). DRAFT--DO NOT QUOTE OR CITE 5-8 December 2003 The lipid w e ig h t, V (t), is calculated as the product of the percent body lipid and the full body w e ig h t of the infant, both of w hich are provided in W alker and W atkins (1997) for infant boys and girls. This section dem onstrates the approach using the average for infant boys and girls. The dose regime for tested scenarios in this section, the body w e igh t, lipid fraction, and assumed half-lives of TEQDFP-W HO98, are show n w ith other model parameters in Table 5-3. The body lipid and body w e ig h t are also show n graphically in Figure 5-1, for the 70 year life span and in more detail for the early years of life. One key assumption of this sim plistic fram ew ork is th a t dioxins are instantaneously distributed to all body lipids. This is a com m on assum ption for TCDD PK modeling in humans, adopted by the m ulti-com partm ent model of van der Molen et al. (1996 ), and the single-com partm ent models in Kreuzer et al. (1997), Campbell et al. (1996 ), and Lakind et al. (2000). The model of Carrier et al. (1995a,b) alternately has a nonlinear response to doses, w ith different partitioning to the liver and other body lipids as a function of body concentration; when the overall body concentration is high, more of the dioxin dose is partitioned to the liver, whereas at lower body concentrations, the partitioning to the liver is lower. The other key and im portant assum ption for the model is th a t the TEQDFP-W HO98 behaves as a single com pound in humans, and can be described by a single dissipation half-life. A yo tte et al. (1994) modeled TEQ body burdens from infancy to adulthood, but it was unclear w hether they modeled individual congeners or TEQs as one com pound. Campbell et al. (1996) used a fram ew ork similar to the one used here and modeled individual congeners for an industrial exposure study. The elim ination rate constant, k(t), w as developed in similar fashions by Pinsky and Lorber (1998 ), Michalek et al. (1996 ), and Flesch-Janys et al. (1996 ), for 2,3,7,8-TC D D . All three research groups derived a relationship in w hich the elim ination rate constant was a function of percent body fat. All three also curve-fit their empirical algorithm s for k(t) on data from adult individuals, whose percent body fat was about 25 percent. W ith body fa t percent increasing over tim e, particularly in older individuals, the elim ination rate constant decreased (equivalently, the half-life increased) significantly. Given a range of body fat over tim e, from about a low of 15 percent to a high over 40 percent (for elderly females), the relationship in Pinsky and Lorber (1998) results in a half-life of 2 ,3 ,7 ,8 TCDD ranging from about 6 to over 20 years. DRAFT--DO NOT QUOTE OR CITE 5-9 December 2003 None of these efforts, however, identified processes or factors critical for infants, other than percent body fat. W ith a body fat of around 15 percent at birth, the half-life is calculated to be about 6 .4 years using the relationship in Pinsky and Lorber (1998). Kreuzer et al. (1997 ), how ever, developed a procedure for modeling the elim ination half lives for 2,3,7,8-TC D D in infants w hich considerd m etabolic, t m (breakdown by enzymes), and non-m etabolic, tf (fecal elim ination) processes. Kreuzer et al. (1997) combined these tw o half-lives to solve for an overall half-life, t 1/2. Other key parameters included total body lipid mass and liver volum es, w hich change over tim e, and a reference half-life for an adult. For their "reference a d u lt" at age 4 0 , they cited an overall half-life of 5 years, based on inform ation in Geyer et al. (1986). The Kreuzer et al. (1997) model showed a rise in half-lives from a low of less than 0 .5 years at birth to a high of 5 years at the total body lipid mass of 20 kg. W ith their parameter assignm ents, perhaps m ost im portantly this assignment of a 5 year half-life for a reference adult, the half-life w ill not go far beyond 5 years (as a function of body lipid mass, it would exceed 5 years when body lipid mass exceeds 20 kg), w hich makes the model of Kreuzer et al. (1997) im portantly different than th a t of Pinsky and Lorber (1998 ), M ichalek et al. (1996 ), and Flesch-Janys et al. (1996 ), all of w hom have half-lives varying from a value of 6 to over 20 years. In short, the model of Kreuzer et al. (1997) has half-lives w hich m ostly never exceed 5 years, while the other approaches have half-lives for TCDD w hich never go below 6 years. As noted, the model of Kruezer et al. (1997) suggests relatively short half-lives for infants. For infants, the overall half-life is driven by non-m etabolic processes and the resulting half-life for new borns is calculated to be about 0 .4 years. It rises to about 2.0 years when the total body fa t w e ig h t is about 5 kg, w hich occurs around ages 8-10 years (25-35 kg overall body w e ig h t, about 20 percent body fat). Very clearly, this rapid a half life of dioxin intake w ill have an im portant im pact on the accum ulation of dioxin residues during breast-feeding as compared to a model show ing a 6 year or higher half-life. Lakind et al. (2000) adopted the Kruezer et al. (1997) approach in their evaluation of the im pacts of breast-feeding on the 2,3,7,8-TCDD body burdens of infants. For purposes of this assessment, it was assumed th a t the overall half-life for the early years of life more closely fo llo w s the trend as derived in the modeling exercises by Kruezer et al. (1997). For later years, it was fe lt th a t the empirical data upon w hich Pinsky and Lorber (1998 ), M ichalek et al. (1996 ), and Flesch-Janys et al. (1996) derived DRAFT--DO NOT QUOTE OR CITE 5-10 December 2003 the half-life relationship for 2,3,7,8-TC D D is more valid. Therefore, a hybrid of these assum ptions, as show n in Figure 5-2, was adopted for this effo rt. The half-life at birth starts at the low value of 0 .4 yr and then slow ly rises to the levels as modeled by Pinsky and Lorber (1998) by about age 20. It is noted th a t had Kreuzer et al. (1997) established reference half-lives for 40 year-olds more in the 6-20 year range, they w ould still have had very low half-lives at birth, rising to these higher half-lives w ith age. The half-life assum ptions remain an obvious uncertainty for this type of modeling approach. Not only is there a disparity in the literature w ith regard to this critical assum ption, but the literature is also only specific to 2,3,7,8-T C D D , not TEQDFP-W HO98. The im pact of this assumption is examined later in the sensitivity analysis exercises. The final assum ption for the model is the initial lipid concentration in the infant. It w as assumed to be 10 ppt TEQDFP-W HO98, w hich w as reasonably similar to the 11.9 ppt TEQDF-W HO98 found in stillborn adipose tissue in Kreuzer et al. (1997). All other assum ptions and parameter assignm ents for this modeling exercise, including the half-life change over tim e, are show n in Table 5-3, and Figures 5-1 and 5-2. 5 .2 .3 .2 . Validation o f the Model W hile dem onstrating the im pact of breast-feeding, the studies reviewed in Section 5.2 .1 . do not contain the type of inform ation needed for model validation. W hat is needed are breast m ilk concentrations th a t are taken at the same tim e infant body burden measurements are taken. The breast m ilk concentrations are used to provide the "independent" model driving term , the dose term , and the body burden measurements provide the "dependent" model prediction, the infant body lipid concentration. One study had a set of this kind of data. Abraham et al. (1998) studied CDD/CDF/PCB levels in the blood of 6 breast-fed infants as well as the breast m ilk of the m others of these infants. A portion of this data set had been reported in their earlier articles (Abraham et al., 1994; 1995). This analysis w ill focus on the TEQDF-W HO98 concentrations reported in Abraham et al. (1998 ), since PCB concentrations were not uniform ly available for m other's m ilk and infant blood for all six m other/child pairs. Tw o of the infants were the second children from mothers whose first child was also tracked by Abraham and colleagues. It w as interesting to note th a t, for these tw o second children, both the m others' m ilk and the in fa n ts' body burdens were significantly DRAFT--DO NOT QUOTE OR CITE 5-11 December 2003 lower. Specifically, the comparison of first and second children, respectively, were: 34 .7 ppt TEQdf-W HO98 lipid compared to 1 1 .9 ppt TEQDF-W HO98, and 4 4 .2 ppt TEQDF-W HO98 compared to 18.8 ppt TEQDF-W HO98. The com parison of the m others' m ilk from the first to second children w as sim ilarly disparate: the first m other had concentrations ranging from 14 to 24 ppt TEQDF-W HO98 lipid for the first child, but 13 to 14 ppt TEQDF-W HO98 lipid only for the second child. The other m other showed a range of 15 to 27 ppt TEQDFW HO 98 lipid for the first child, but only 13 to 18 ppt TEQDF-W HO98 lipid for the second. A pparently, breast-feeding of the first child resulted in a higher body burden for this infant as compared to the second infant, and a low er body burden for the m other when the second infant was breast-fed. Table 5-4 show s the observed data th a t were available in Abraham et al. (1998) for this model validation exercise. There were tw o concentrations measured in breast-m ilk for each of 5 of 6 children. The one child w hose m other had only one measurement w as only breast-fed for 7 w eeks; all other children were breast-fed for periods ranging narrow ly from 26 to 32 weeks. Concentrations w ith in the breast-feeding period were linearly extrapolated from the tw o available data points. For example, for the first m other/child pair listed in Table 5-4, m other's m ilk w as analyzed during m onth 2 and m onth 11. TEQ concentrations were 23.5 and 14.0 ppt TEQDF-W HO98 lipid, respectively. These concentrations were extrapolated backwards to give an estim ated concentration at birth of 24 .6 ppt TEQdf-W HO98. Likewise, forw ards extrapolation gave an estim ated concentration for m onth 3 of 2 2 .4 TEQDF-W HO98, assuming linear decline. The infant body burden was ascertained by blood measurements at about 1 year of age for each child. The am ount of tim e of full breast-feeding was supplied by Abraham et al. (1998 ), and this is also listed in Table 5-4. Other assum ptions for modeling the infant body burden were outlined above, and these include the intake rate of m other's m ilk (IR = 8 00 m l/d), the fraction of fa t in m other's m ilk (0.0 4), the rate of absorption of dioxins (0.8 0), and the changing infant body w e igh t and lipid fraction (and hence lipid volum e, V (t); Table 5-3). A fte r weaning, the dose to the in fan t was assumed to be 50 pg TEQDF-WHO98/d. This is the dose developed for the age range of 1-5 in Chapter 4. The dose by form ula feeding or other foods the infants may be consuming after weaning may be low er or higher, as little DRAFT--DO NOT QUOTE OR CITE 5-12 December 2003 inform ation is available on the dioxin content of baby form ula or baby food. The assumed initial body burden of the infant was 10 ppt TEQDF-W HO98, as noted above. The rate of dissipation of dioxin residues w as identified as a principal uncertainty for this model. Tw o lines of thought discussed above include the rapid dissipation (half life < 1 year) of TCDD residues in infants modeled in Kreuzer et al. (1997) and later adopted by Lakind et al. (2000), and the much longer dissipation (half-life around 7 years) of TCDD in adults described in Pinsky and Lorber (1998 ), M ichalek et al. (1996 ), and Flesch-Janys et al. (1996). The model validation exercise described in this section tested the appropriateness of the low er infancy half-life approach adopted in this model, as show n in Figure 5-2, against an assum ption of a constant 7 year half-life for TEQDF-W HO98 during the first year of life. Table 5-4 shows the final results of this exercise. As seen, the model predictions at the selected and more rapid dissipation rate were significantly nearer to observations as compared to the predictions w ith the longer half-life. The average predicted concentration for the rapid dissipation rate for the 6 infants was 26 ppt TEQDF-W HO98 lipid, compared to the average observed concentration of 23.5 ppt TEQDF-W HO98 lipid. W ith a longer 7-year half-life, the predicted concentrations were all higher, w ith an average of 39 ppt TEQDFW HO 98 lipid. The model also seemed very adequately responsive to low er or higher in fa n ts' exposures. For the infant w ho was breast-fed for only 7 weeks w ith a low concentration in the m other's m ilk, the blood concentrations measured 5.0 ppt TEQDFW HO 98 lipid at 13 m onths, compared to a predicted 10 ppt TEQDF-W HO98 lipid (w ith the rapid dissipation assumption). The infant exposed to the highest m other's milk concentration had the highest body burden measurement at 4 4 .2 ppt TEQDF-W HO98 lipid and also the highest predicted concentration at 36 ppt TEQDF-W HO98 lipid. In summary, it appeared that, even w ith the key uncertainties identified above, including the use of a simple, one-com partm ent pharm acokinetic model and the modeling of TEQDFP-WHO98s as though they were a single com pound, this approach appears to predict infant TEQDFP-WHO98 body burdens w ith in the range observed, and is adequately responsive to the d ifferent conditions of high and low exposure via breast-feeding. DRAFT--DO NOT QUOTE OR CITE 5-13 December 2003 5 .2 .3 .3 . Scenario Evaluation The scenarios evaluated include: formula only, 6 weeks of breast-feeding, 6 m onths of breast-feeding, 1 year of breast-feeding, and 2 years of breast-feeding. These scenarios encompass current trends. In com prehensive docum entation of statistics for children born between 1990 and 1993, CDC (1997) reported th a t 55 percent of all babies breastfed, w ith about half of those breastfeeding beyond 5 months. The average duration of breastfeeding was 2 8 .7 weeks. In a policy statem ent, the Am erican Academ y of Pediatrics (1997) stated th a t exclusive breastfeeding is ideal nutrition and su fficie n t to support optim al g row th and developm ent for 6 m onths after birth. They recommend that breastfeeding continue for at least 12 m onths, and thereafter for as long as m utually desired. Ryan (1997) docum ented a resurgence in breastfeeding between 1989 and 1995. In com prehensive surveys conducted in 1989 and 1995, he found a 14 percent increase in the number of m others w ho breastfed in the hospital, rising from 52 percent in 1989 to 59 percent in 1995. He also found a 19 percent increase in m others w ho continued to breastfeed at 6 m onths, rising from 18 percent in 1989 to 22 percent in 1995. The specifics of these scenarios are: Scenario #1: Formula Only: In this scenario, the dose to the infant was assumed to be 50 pg TEQDFP-W HO98/d. This is the dose developed for the age range of 1-5, as described in Chapter 4. The dose by form ula feeding may be low er or higher, as little inform ation is available on the dioxin content of baby form ula. TEQDFP-W HO98 doses among individuals from 6 to 11, 12 to 18, and greater than 18 years of age were 54, 65, and 66 pg TEQDFP- W HO 98/d, respectively, as developed for these age ranges in Chapter 4. Scenario #2: Six-W eek Nursing: The dose to the infant was assumed to be a function of the starting concentration of dioxins in the m other's m ilk, 25 pg TEQDFP-W HO98/g lipid, and other assum ptions th a t have been described in this section: 800 g/day m ilk ingestion, 4 percent lipids in m ilk, resulting in an initial dose of 8 00 pg TEQDFP-W HO98/d. This drops to 733 after 1 m onth, and then to 6 67 pg TEQDFP-W HO98/d, when nursing stops. Doses from then are as in Scenario #1. Scenario #3: Six-M onth Nursing: It was assumed th a t the dose drops linearly from 800 to an ending dose of 4 0 0 pg TEQDFP-W HO98/day at m onth 6. From there, doses are as in Scenario #1. DRAFT--DO NOT QUOTE OR CITE 5-14 December 2003 Scenario #4: It was assumed that the doses drop linearly from 800 to 400 at 6 months and then linearly again to 2 00 pg TEQDFP-W HO98/day at the end of one year. From there, doses again are as in Scenario #1. Scenario #5: The 1 year dose of 200 pg TEQDFP-W HO98/day assumes a m other's milk concentration of 6 .2 5 ppt TEQDFP-W HO98 lipid, w h ich represented a drop from an initial concentration of 25 ppt TEQDFP-W HO98 lipid. For purposes of this dem onstration, it was assumed th a t the m other's m ilk concentration stays at 6 .2 5 ppt TEQDFP-W HO98 lipid for the second year of breast-feeding. From there, doses again are as in Scenario #1. The results from this exercise are show n in Figures 5-3 and 5-4, w hich show the lipid concentrations and the body burdens from birth up to 70 years of age (on Figure 5 3), and then these tw o quantities for the narrower tim e fram e of from birth to 10 years of age (Figure 5-4). The body burden, defined as the w hole body concentration, is sim ply calculated as the lipid concentrations tim es the lipid fraction. Other results for these 5 scenarios are provided in Table 5-5, including the peak TEQDFP-W HO98 concentrations in the infant, the time when the peak occurred, the "area under the curve" (AUC), corresponding to different tim es, and the ratio of th a t AUC for the breast-feeding scenarios and the AUC for form ula feeding only. The AUC is defined as: AUC = 1 c(t) Eqn. (5-6) where: AUC = c(t) = area under the curve, ppt TEQDFP-W HO98-day lipid-based concentration in the infant each day, ppt TEQDFP-W HO98 The AUC is a measure of accum ulated exposure. For example, a year at a lipid-based concentration of 10 ppt TEQDFP-W HO98 w ould yield an AUC of 3 ,6 5 0 ppt TEQDFP-W HO98day (10 ppt * 365 days). A lifetim e at an average body lipid concentration of 10 ppt w ould yield an AUC of 2 5 5 ,5 0 0 ppt TEQDFP-W HO98-day (10 ppt * 70 yrs * 365 days/yr). The AUC provides a parameter to compare accumulated exposure for different scenarios. For th a t reason, the ratios of the AUCs of the various breast-feeding scenario and the form ula-only scenario are provided in Table 5-5. A ratio for a particular breast-feeding scenario of 6, for example, w ould mean th a t the accum ulated exposure for th a t scenario is 6 tim es th a t of the breast-feeding only scenario. This can easily be translated to a DRAFT--DO NOT QUOTE OR CITE 5-15 December 2003 corresponding measure of percent above or below the baseline scenario of form ula feeding only. For example, if the ratio is 6, this is equivalent to saying th a t the accumulated exposure for the breast-feeding scenario is 500 percent higher than the form ula-only scenario ((6 -1 )*1 0 0 percent); if the ratio is 0 .7 , this is equivalent to saying th a t the accum ulated exposure for the breast-feeding scenario is 30 percent low er than the form ula-only scenario (-(0 .7 -1 )*1 0 0 percent). The lipid concentrations are predicted to rise to about 4 4 ppt TEQDFP-W HO98 for the 6-m onth,1-year, and 2-year scenarios. The tim e th a t these peaks occur is uniform ly at 9 weeks. For the 6 week breast-feeding scenario, the peak is at 34 ppt and it occurs at the 6 w eek mark. Body burdens fo llo w a similar trend, rising to about 9 ppt TEQDFP-W HO98 for both the 6-m onth, 1-year, and 2-year scenarios at 9 weeks. The body burdens decline for these breast-feeding scenarios, but the decline is slow er as the duration of tim e for breast feeding increases. For the 2-year scenario, the body lipid concentration stays near 40 ppt past 2 years of age (Figure 5-4). The six-w eek scenario show s a rise in infant body burden to above 30 ppt TEQDFP-W HO98, but then show s a rapid decline, tracking the form ula-only scenario fairly well after about age 2. From Figure 5-3, it appears th a t all four scenarios begin to merge at about age 10 years. The rise in concentrations seen in the later years in Figure 5-3 is due to the rise in body fa t percent and the subsequent rise in half-life as predicted by elim ination rate model of Pinsky and Lorber (1998). The AUC results in Table 5-5 show how the accum ulated exposure is higher for each of the breast-feeding scenarios as compared to the form ula-only scenarios. This exceedance after one year is about a facto r of 6 for breast-feeding for 6 m onths or more. Even for the first 10 years of life, the accum ulated exposure is about 2 tim es higher for these breast-feeding scenarios as compared to form ula feeding only. A fte r a lifetim e, the ratios suggest th a t breast feeding results in a lifetim e exposure only 1 .0 3 -1 .1 8 tim es higher than form ula feeding, or expressed in term s of a percentage, from 3 to 18 percent higher than formula feeding only. 5 .2 .3 .4 . Sensitivity Analysis A brief sensitivity analysis was conducted to study the impact of key parameters and assum ptions on this exercise. Sensitivity analysis exercises typically focus on the parameters w ith these tw o characteristics: those th a t are the m ost uncertain and those DRAFT--DO NOT QUOTE OR CITE 5-16 December 2003 th a t have an im portant im pact on the results. It was ascertained th a t the parameters: absorption, body w e ig h t, and lipid fra c tio n , while im portant, are reasonably w ell know n and assigned appropriate mid-range values fo r this exercise. The initial concentration in the infant at birth of 10 ppt TEQDFP-W HO98 lipid is an unknow n, but testing showed that infant concentrations rapidly declined to about 5 ppt regardless of the initial concentration. Therefore, this parameter does not appear to influence results like peak concentrations or accumulated exposures. The tw o most important parameters, or groups of parameters, were those used to determine dose to the infant - m other's m ilk concentration initially and over tim e, and the dissipation rate of dioxin-like com pounds in the infant. To test the influence that these parameters have on results, six sensitivity analyses were devised. These were all variations on a baseline scenario selected to be the 6-m onth breast-feeding scenario. The six scenarios were: Scenario #1: Use of the Pinsky and Lorber (1 9 9 8 ) lipid-based function for dissipation rate instead of the hybrid function used. The Pinsky and Lorber (1998) lipid-based function is show n in Figure 5-2. This is expected to result in an increase in the predictions of infant and childhood body impacts. Scenario #2: Use of the Kreuzer et al. (1997) modeled dissipation rate throughout life instead of the hybrid function used. The Kreuzer et al. (1997) function is show n in Figure 5-2. This is expected to result in a decrease in predictions of infant, childhood, and adult im p a cts. Scenario #3: Use of the assum ption th a t m others' m ilk concentrations do not decline from 25 ppt TEQDFP-WHO98 lipid during the 6 m onths of breast-feeding. This w ill obviously result in an increase in the im pact to the infant. Scenario #4: Use of the assum ption th a t m others' m ilk concentrations begin at 15 ppt TEQdfp-W HO98 and decline to 7.5 ppt TEQDFP-W HO98 after 6 m onths (instead of using concentrations th a t begin at 25 ppt TEQDFP-W HO98 and decline to 12.5 ppt TEQDFP-W HO98 after 6 months). The assignm ent of an initial concentration of 25 ppt TEQDFP-W HO98 is based on the finding that this represents a reasonable average of adult body concentrations of the sum of dioxin, furan, and dioxin-like PCB TEQs, described in Chapter 4. However, this average includes a w ide age range of populations, including older individuals. A s w ill be discussed in the next chapter on trends, higher exposures in the past have resulted in higher concentrations in older adults compared to to d a y's younger DRAFT--DO NOT QUOTE OR CITE 5-17 December 2003 adults. It may be more reasonable to assume a wom an of child-bearing age today would have concentrations closer to 15 ppt TEQDFP-W HO98 rather than the full adult population average of 25 ppt TEQDFP-W HO98. Also, for children after the first-born, reduction in a w om an's body burden of dioxin-like compounds could occur from prior breast-feeding. Scenario #5: For this scenario, both assum ptions th a t w ould lead to a higher im pact use of the Pinsky and Lorber (1998) lipid-based function for dissipation rate and the assum ption of m other's m ilk concentration not declining from 25 ppt TEQDFP-W HO98 lipid were used. Scenario #6: In contrast to Scenario #5, both assum ptions th a t w ould lead to a lower im pact - the use of the Kreuzer et al. (1997) modeled dissipation rate and the assumption of m others' m ilk concentrations beginning at 15 ppt TEQDFP-W HO98 and declining to 7.5 ppt TEQdfp-W HO98 after 6 m onths - were used. Results from this sensitivity analysis exercise are show n in Table 5-6 and Figure 5-5. It is seen th a t all results range from a reduction of 40 percent (AUC ratio of 0.6) from baseline or an increase of 160 percent (AUC ratio of 2.6). The peak concentration can rise as high as 54 ppt TEQDF-W HO98, as seen in Table 5-6 for Scenario #5 or as low as 28.5 ppt TEQdf-W HO98 for Scenario #6. Perhaps the biggest im pact is seen in assuming the higher dissipation rate (lower half-life) for the childhood years. This is seen in Scenarios #4 and #6, w ith the low est peak concentration and the steepest reduction from baseline: 30-40 percent. W hile the results displayed on Table 5-6 focus on the difference between the baseline 6-m onth scenario and sensitivity analysis scenarios, also of note is the difference between the tw o extremes as modeled by Scenarios #5 and #6. The peak concentration predicted for Scenario #5 (infant im pact maximized) is about tw ice th a t of Scenario #6 (infant im pact minimized): 5 4 .0 versus 28.5 ppt TEQDFP-W HO98. More im portantly, the accum ulated exposure from Scenario #5 is significantly higher than from Scenario #6. This can be seen In Figure 5-5, where the body concentrations are higher in Scenario #5 than #6 throughout the modeled lifetim e, particularly for the first 10 years of life. A fte r 10 years, the accumulated exposure, as measured by AUC, is about 4 tim es higher for Scenario #5 as compared to #6 (this results from the 10-year ratios of 2.6 and 0.7 displayed on Table 5-6, divided by each other; 2 .6 /0 .7 = 3.7). In short, when using this modeling approach to evaluate exposure-related im pacts to infants from breast feeding, the assessor needs to be aware th a t assum ptions relating DRAFT--DO NOT QUOTE OR CITE 5-18 December 2003 to the dissipation of dioxin-like com pounds from infants, and the dose they receive through breast m ilk, are the tw o m ost im portant inputs. Data should be sought to further validate the selected modeling procedure and parameters, if possible. 5.3. SPORT AND SUBSISTENCE FISHERS The possibility of high exposure to dioxin as a result of fish consum ption is most likely to occur in situations where individuals consume a large qu a n tity of fish from one location where the dioxin level in the fish is elevated above background levels. M ost people eat fish from m ultiple 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. U.S. EPA (1997) presents studies th a t indicate th a t Native American subsistence fisherm en consume 59 g fish/day (as a mean) and 170 g fish/day (as an upper estim ate). W olfe and W alker (1987) found Native Am erican subsistence fish ingestion rates up to 770 g/day in a study conducted in Alaska. Assum ing th a t subsistence fishermen consume 59 to 170 g of freshw ater fish per day as their primary source of protein (i.e., no meat or eggs are consumed) adult daily intake of CDD/CDFs/PCBs w ould be 2.2 to 5.7 pg/kg-day (Table 5-7). This estim ate is based on the same CDD/CDF/PCB media concentrations exposure assum ptions, and exposure algorithm s as those presented in Chapter 4. The estim ated values for subsistence fisherm en are tw o to six tim es higher than the adult general population mean daily intake from all food sources of 0 .9 4 pg/kgday, as estim ated in Chapter 4. It should be noted th a t fish ingestion rate data for subsistence fishermen are limited. The ingestion rate values used here pertain only to Native Am erican subsistence populations (U.S. EPA, 1997), but are used to dem onstrate the potential for elevated exposures among groups of individuals whose diets are know n to consist of higher proportions of fish than the general population. Fish ingestion rates for sports fisherm en w ould generally be low er than for the Native Am erican subsistence population, but higher than for the general population. Studies are underway to evaluate w hether Native Am ericans living on the Columbia River in W ashington have high dioxin exposures as a result of fish consum ption. These Tribes consume large quantities of salmon from the river. As cited in U.S. EPA (1997 ), a DRAFT--DO NOT QUOTE OR CITE 5-19 December 2003 study conducted by the Columbia River Intertribal Fish Commission (1994) suggested that these individuals have an average fish consum ption rate of 59 g/day and a 95 th percentile rate of 170 g/day. These data were used in the estim ated dietary intake calculations for subsistence fisherm en, as show n above. C urrently, studies are underway to measure dioxin levels in fish from this region. 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: nonconsum ers (n = 9 ), moderate consumers (n = 9, 220 to 500 g /w k), and high consumers (n = 11, 700 to 1,750 g/w k). The high consumer group was composed of fisherm en or w orkers in the fish industry w ho consumed prim arily salmon (30 to 90 pg I-TEQDF/g) and herring (8 to 18 pg I-TEQDF/g) from the Baltic Sea. The I-TEQDF blood level w as found to average about 60 pg I-TEQDF/g lipid among the high consumers and 20 pg I-TEQDF/g lipid for the nonconsumers. This difference was particularly apparent for the PeCDFs. Asplund et al. (1994) also found elevated plasma levels of dioxin-like PCBs in Swedish fisherm en w ho consumed large am ounts of fish. A tota l of 37 individuals w ith varying intake rates of fish from the Baltic Sea was studied. These individuals were categorized as high-fish eaters, moderate fish-eaters, and nonfish-eaters. The estim ated w eekly intake of fish correlated positively w ith plasma PCB levels among this group (Table 5-8). Cole et al. (1995) reported on CDD/CDFs and PCBs in 132 serum samples (pooled to 14) from Ontario Great Lakes anglers and control populations. Based on a prelim inary survey, anglers from the com m unities of Cornwall and M ississauga, Canada, were categorized based on the numbers, species, and locations of fish caught and kept for consum ption, and on data reflecting the contam inant levels for the fish in these areas. Individuals categorized as having the highest and low est potential for having elevated body burdens of CDD/CDFs and PCBs were selected for biological sampling. Individuals w ho did not consume fish served as controls. Study participants were further categorized by age (i.e., < 3 8 years, 3 8 -50 years, and > 5 0 years). The results, how ever, indicated th a t mean I-TEQDF levels were similar for both eaters and noneaters of Great Lakes' fish in these com m unities. I-TEQDFs ranged from 20 .8 to 4 1 .2 ppt for fish eaters and 2 4 .7 to 3 6 .8 ppt for noneaters. In general, mean I-TEQDFs increased w ith age (Table 5-9). PCBs 77, 126, and 169 were also evaluated in the serum samples collected from Cornwall DRAFT--DO NOT QUOTE OR CITE 5-20 December 2003 residents. Mean TEQP-W HO98s ranged from 2.6 to 17.3 ppt for fish eaters and noneaters combined. Again, significant differences between the tw o groups were not observed and the serum CDD/CDF and PCB levels are w ith in the range of values observed for the general population, as presented in Chapter 4. Health departments of five Great Lakes states: W isconsin, Michigan, Ohio, Illinois, and Indiana, form ed a consortium to study blood levels of chemical residues in fish consumers of three Great Lakes: M ichigan, Huron, and Erie. Anderson et al. (1998) reported on a feasibility study to determine w h ich com pounds m ight be found in very frequent Great Lakes sport fish consumers. Anderson et al. (1998) selected 32 angling enthusiasts who reported eating at least one sport fish meal per week from one of three Great Lakes (i.e., 11 Lake Huron anglers, 11 Lake Erie anglers and 10 Lake Michigan anglers). The analysis included examination of serum levels of 7 CDDs, 10 CDFs, 4 coplanar PCBs (i.e., 77, 8 1 ,1 2 6 , and 169), and 32 other PCB congeners. One individual was excluded from the data summary due to unusually high occupational/environmental exposures. The blood CDD/CDF/PCB levels for these anglers were compared to CDD/CDF/PCB blood levels for a comparison group (n = 70) from Jacksonville, Arkansas. Data for this Arkansas population are discussed in Chapter 4. The comparison groups represented the general population w ith no know n exposure to the contam inants of concern (Anderson et al., 1998). The mean CDD/CDF lipid adjusted serum concentrations for both the sport fishing populations and the com parison group used by Anderson et al. (1998) are show n in Table 5-10. The mean coplanar and other PCB lipid adjusted serum concentrations are reported in Table 5-11. The average lipid-based I-TEQDFP concentration for Great Lakes fish consumers w as calculated at 5 6.8 ppt, w ith the breakdown as follow s: I-TEQd = 27.5 ppt; I-TEQF = 11.9 ppt, and TEQP-W HO94 = 17.4 ppt. Anderson et al (1998) suggested th a t these values were higher than the background population used in the comparison. The Anderson et al. (1998) study led to a larger study, in w hich the blood of 100 additional sport fishers were sampled, and a com parison population of 100 other individuals were sampled. Falk et al. (1999) reported on the results of the blood sampling from 96 (of the 100) additional sport fishers. Results for the 100 comparison population were not provided. Falk et al. (1999) presented results in term s of I-TEQDF and TEQ PW HO94 (congener-specific data were not provided), and also examined relationships DRAFT--DO NOT QUOTE OR CITE 5-21 December 2003 between the CDD/CDF/PCB measurements in blood and factors such as: age, gender, w hich Great Lakes the fish came from , the type of sport fish consumed, and am ount of sport fish consumed, as reported by the participants. The median lipid-based TEQDFPW HO 94 from the 96 participants w as 2 1.3 ppt, w ith the breakdown as follow s: I-TEQD = 9.6 ppt; I-TEQf = 7.4 ppt, and TEQP-W HO94 = 4 .3 ppt. This finding of 21 .3 ppt TEQDFPW HO 94 appears significantly low er than the original finding of 56 .7 TEQDFP-W HO98. One reason for this is th a t the low er finding from the 96 participants w as a m edian, w hile the finding from the 31 individuals in the pilot study was a mean. It also appears th a t the smaller population had a fe w individuals w ith very high levels of dioxins w hich resulted in a higher mean concentration. Other differences th a t could be identified from Anderson et al. (1998) and Falk et al. (1999) include: 1) the tim e of sampling of the tw o studies; 2) the age of the participants; and 3) the sport fish consum ption rates. Anderson et al. (1998) reported th a t sampling of the initial 31 individuals in the pilot study occurred in 1993. A lthough Falk et al. (1999) did not identify the date at w hich the follow u p study occurred, it appears likely to have been in 1995 or 1996. A lthough not expected to be a large facto r explaining the differences in the populations, it is possible th a t average body burdens w ith in the population decreased during this tim e period. The mean age of the 31 participants in the pilot study w as 52 years (range 36 to 76), w hile the mean age in the second population of 96 participants w as 46 years (range 27-67). A clear age relationship has been dem onstrated in other studies, show ing th a t older individuals have higher body burdens of dioxins. The follow up study population of 96 individuals clearly showed lower consum ption of Great Lakes fish as compared to the pilot population of 31 individuals, as evidenced by questionnaire response data. The fo llow u p study population reported an average of 52 fish meals consumed per year, w hile the pilot group reported an average of 77 fish meals per year. Likewise, consumption of Great Lakes fish was lower for the fo llow u p group than the pilot group: 43 Great Lakes fish meals per year and 49 Great Lakes fish meals per year, respectively. Also, the follow up group reported a lower number of years consuming Great Lakes sports fish (26 years) than the pilot group (33 years). In sum m ary, w hile the larger population of 96 sport fishers in the full survey appeared to show a much low er body burden of dioxin-like com pounds as compared to pilot population of 31 sport fishers, the differences could be explained by factors of data description (median vs. mean), year of sam pling, age of participants, and exposure to dioxins in fish. DRAFT--DO NOT QUOTE OR CITE 5-22 December 2003 Another observation from the Anderson et al. (1998) study w as th a t Lake Erie sport fish consum ers had consistently low er CDD/CDF/PCB serum concentrations than consum ers of sport fish from Lakes M ichigan and Huron. Serum levels observed for the Lake M ichigan and Lake Huron fish consum ers were similar and higher than those observed in consum ers of Lake Erie sport fish. These interlake differences parallel the pattern observed in previously reported EPA sport fish tissue m onitoring data from the respective lakes (Anderson et al., 1998) and indicate th a t serum concentrations may also be affected by variations in fish concentrations among the lakes. Kolic et al. (2000) reported on sampling of fish for dioxin-like CDD/F/PCBs conducted by the Ontario M inistry of the Environment between 1996 and 1998 in the Ontario Great Lakes region. Table 5-12 presents the data on this sampling e ffo rt as it was reported in Kolic et al. (2000). A total of 193 samples are reported on in Table 5-12. TEQ concentrations were calculated using the WHO 1998 TEF scheme, w ith zero values used for non-detects. W hen no congeners were detected in the sample, the value was reported as zero. Fish were sampled for the Ontario Sports Fish Program. This program m onitors sport fish in the Ontario fresh w ater lakes and issues consum ption advisories through its biannual guide. Because of this purpose, the fish sampled are typically larger and from suspect areas in order to obtain positive results and therefore be able to set consum ption advisories (Reiner et al., 1995). An initial set of data on CDD/CDFs from this program were reported on in Reiner et al. (1995). Sampling reported there occurred between 1991 and 1994. Kolic et al. (2000) examined the data taken between 1996 and 1998 for purposes of studying the relationship between dioxin-like PCBs and CDD/CDFs in the fish. They observed PCB TEQ concentrations substantially greater th a t CDD/CDF TEQ concentrations in m ost locations, as is evident from Table 5-12. Over all lakes, they observe an average ratio of 6.5 (om itting the large value of 86 from the Welland River, as well as all circum stances when CDD/CDF TEQ concentrations are zero). Also no te w o rth y is th a t the concentrations reported for these fish are substantially higher than the fish concentrations used for background exposure calculations in Chapter 5. CDD/CDF concentrations for marine and freshw ater sources were 1.0 and 0 .2 6 pg/g w hole w e ig h t TEQDF-W HO98, respectively. PCB concentrations were 1.2 and 0 .2 5 pg/g w hole w e igh t TEQP-WHO98, respectively. In contrast, the overall concentrations for this Great Lakes data set were 5.4 and 26 .6 TEQDF-W HO98/g w hole w e ig h t for CDD/CDFs and DRAFT--DO NOT QUOTE OR CITE 5-23 December 2003 PCBs, respectively. As observed by Kolic et al. (2000 ), CDD/CDF concentrations appeared to be declining since their earlier reporting in 1995. A t th a t tim e, Reiner et al. (1995) reported on 198 samples, and the average I-TEQDF concentration was 11.5 pg/g whole weight. Hong et al. (1994) analyzed PCBs in human m ilk from M ohaw k and control wom en to evaluate the potential effe ct that relatively high levels of environmental contam ination may have had on the body burdens of lactating M ohaw k wom en in N ew York. PCBs were found to be present in fish and w ild life in the vicin ity of the M ohaw k Reservation, and the M ohaw k people form erly depended on local fish and w ild life for food. However, no significant differences were observed between the mean total dioxin-like PCB levels in milk from 30 M ohaw k wom en and the 20 control wom en. The mean PCB concentrations for these wom en were 49 ppb and 55 ppb, respectively. The age of the m other, the length of the nursing period, and the number of breastfed children were found to influence PCB levels in human milk. Older w om en, m others of first born children, and smokers had higher levels of PCBs. PCB levels were also higher at the onset of lactation and in earlier samples during a breastfeeding session. Dewailly et al. (1994) observed elevated levels of dioxin-like PCBs in the blood of fisherm en on the north shore of the G ulf of the St. Lawrence River w ho consume large amounts of seafood. Of the 185 study samples, the 10 samples w ith the highest total PCB levels were analyzed for dioxin-like PCBs. Samples from Red Cross blood donors in Ontario served as controls. Dioxin-like PCB levels were 20 tim es higher among the 10 highly exposed fisherm en than among the controls (Table 5-13). Based on these results of the 10 highest samples, Dewailly et al. (1994) estim ated that for the entire fishing population studied, dioxin-like PCB levels w ould be eight to ten tim es higher than the control group. Dewailly et al. (1994) also observed elevated levels of dioxin-like PCBs in the breast milk of Inuit women of A rctic Quebec. The principal source of protein for the Inuit people is fish and sea mammal consum ption. Breast m ilk samples were collected from 109 Inuit wom en w ith in the first 3 days after delivery and analyzed for di-ortho dioxin-like PCBs during 1989 and 1990. Subsets of 35 and 40 randomly selected samples were analyzed for m ono-ortho dioxin-like and non-ortho dioxin-like PCBs, respectively. Samples from 96 Caucasian wom en from Quebec served as controls. The levels of non-ortho dioxin-like PCBs for Inuit wom en ranged from 24 .7 to 2 2 0 .9 ppt. DRAFT--DO NOT QUOTE OR CITE 5-24 December 2003 These values were three to seven tim es higher than those observed in the control group. For m ono-ortho and di-ortho dioxin-like PCBS, the levels among the Inuit wom en were three to ten tim es higher than in the control group. Humphrey et al. (2000) reported on an elevation in the PCB concentrations in serum of humans consuming Great Lakes fish. They described a careful identification of "sport fish eaters" and "non-sport fish eaters," or controls, from 11 Lake Michigan shoreline com m unities during the years 1979 and 1982. Sport fish eaters were defined as those individuals who consumed 26 or more pounds of sport-caught fish annually, while non-sport fish eaters consumed less than 6 pounds of sport-caught fish annually. This cohort was revisited a second tim e in 1992 for a study of individuals over the age of 50. Blood sampling occurred in 1 9 9 3 -1 9 9 5 for 101 fish-eaters and 78 controls. These samples were measured to 90 PCB congeners, including dioxin-like PCB congeners 77, 105, 1 18, 123, 157, 169, and 180. They found th a t sport fishers had significantly higher PCB concentrations as compared to controls. The mean concentration of total PCBs (sum of the 90 measured) in the fish-eaters was 14.26 ppb w hole w e igh t, w hile for the controls, the concentration was 4.56 ppb. They found that 22 of the congeners explained m ost of the concentration, and dioxin-like PCBs 105, 118, and 180 were among those 22. The w hole w e igh t concentrations (ppb) of these three PCBs in sport and control fishers were 2.0 0 (fishers) vs. 0 .7 9 (control) for PCB 180; 0 .2 6 vs. 0 .0 2 ppb for PCB 105; and 0 .8 3 vs. 0 .0 6 for PCB 118. 5.4. LOCALIZED IMPACTS Data have been collected that demonstrate that localized im pacts may occur from emissions of dioxins from incinerators and other potential sources. "Localized im pacts" are defined as measurements of CDD/CDFs in environm ental (air, soil) or biotic (vegetation, animal tissue) samples near incinerators or other sources that show elevation above typical background levels for the area being studied. Therefore, "im p a cts," as used below , refer to elevation above background. These localized im pacts may result in elevated exposure among some members of the population. Most of the data on localized impacts originate from studies conducted outside the United States, specifically from the European countries of England, Switzerland, Germany, Austria, The Netherlands, Belgium, and France. Data collected include concentrations of dioxins in air and soil, biota including DRAFT--DO NOT QUOTE OR CITE 5-25 December 2003 grass and c o w 's m ilk, as w ell as human blood and hair samples. This section reviews several o f these studies, prim arily discussing results in term s o f TEQs from CDD/CDFs only. Following a review o f the studies, the principal findings w ith regard to localized im pacts are summarized. Goldman et al. (2000) compared serum concentrations of CDD/CDFs among residents of 2 homes where contam inated home-produced eggs and beef were consumed to residents of a similar rural area th a t did not consume home-produced eggs and beef. The contam inated eggs and beef originated from a residence located near the site of a 1987 fire at a w ood preservative plant. The chicken eggs had an I-TEQDF mean concentration of 10 pg I-TEQDF/g; beef fa t contained 27 pg I-TEQDF/g. These concentrations are 10- to 100-fold higher than the results observed in samples of commercial foods. The soil near the home where contam inated eggs were observed contained CDD/CDF levels ranging from 30 to 40 pg I-TEQDF/g, and the soil CDD/CDF profile w as similar to th a t observed in the eggs. Serum samples were collected from 9 individuals residing in homes where contam inated eggs and beef were consumed. I-TEQDF concentrations in serum were 2 6 .7 ppt for the 4 individuals w ho had consumed eggs from the contam inated site over a 2-year period and 6 3 .7 ppt for the 5 individuals w ho had consumed both eggs and beef from the contam inated site for up to 15 years, compared to 17.0 ppt for the comparison group. Beck et al. (1990) sampled m ilk from a rural and an industrial area in Germany, and from dairies near a metals reclam ation plant in Austria. Beck et al. (1990) observed average lipid-based concentration of 0.9 pg I-TEQDF/g in rural, background m ilk, 2.5 pg I-TEQDF/g in "industrial m ilk," and 9.6 pg I-TEQDF/g in the m ilk obtained from dairies near the metals reclamation plant. The dairy nearest the metals reclamation plant was located about a kilom eter in the dow nw ind direction and had the highest m ilk concentration (i.e., 14 pg I-TEQDF/g). The Austrian metals reclam ation plant described above has also been studied for im pacts to air, soil, vegetation, and human blood by another research team (Riss et al., 1990; Riss, 1993). The plant was located in a rural Alpine river valley in Tyrol, A ustria, in a m ostly agriculture area. A lthough em issions data were unavailable, air concentrations measured near the incinerator were 1.2 to 2.3 pg I-TEQDF/m 3 (Riss, 1993). These data suggest very high emissions, because typical urban air concentrations are approxim ately DRAFT--DO NOT QUOTE OR CITE 5-26 December 2003 0 .1 0 pg/m 3 in Europe as well as in the United States, and rural air concentrations are typically less than 0 .0 5 pg I-TEQDF/m 3. (See Chapter 3.) Soil concentrations averaged 4 2 0 pg I-TEQDF/g at the site of the incinerator, 170 pg/g w ith in 200 meters of plant, and 46 ppt about 2 km in the dow nw ind direction (Riss et al., 1990). This compares w ith typical urban soil concentrations of approxim ately 10 to 20 pg I-TEQDF/g in both Europe and the United States and rural soil concentrations of less than 5 pg I-TEQDF/g. (See Chapter 3.) A dairy farm was located between 1,400 and 2 ,1 0 0 meters from the same metals reclamation site in the dow nw ind direction, and members of th a t farm ing fam ily consumed m ilk from their ow n cow s. Samples of the cow s m ilk ranged from 20.1 to 69 .5 pg I-TEQDF/g on a lipid basis. Given a general background level of m ilk in the low to sub ppt level on a lipid basis, it is clear th a t the m ilk showed elevated dioxin levels. (See Chapter 3.) Samples in the grass and hay from th a t farm were also elevated at 13 to 36 pg I-TEQDF/g dry w eight. This compares to typical grass samples found in rural areas at the low to sub ppt levels (Reed et al., 1990; Kjeller et al., 1991; 1996). Blood samples from tw o farm ers w ho consumed this m ilk were also elevated. Their blood CDD/CDF concentrations were 152 and 9 46 pg I-TEQDF/g on a lipid basis. Subsequent samples from three additional fam ily members were also slightly elevated above typical levels at 4 1 ,6 6 , and 77 pg TEQP-W HO94/g lipid. The Austrian samples described above were taken in the late 1980s, before emission controls and other practices (i.e, removal of some plastics) were undertaken to reduce em issions from these plants. Riss (1993) reported on reductions in both c o w 's m ilk and fodder from this nearby farm in the early 1990s and speculated th a t they resulted from reductions in incinerator emissions. CDD/CDF concentrations in c o w s ' milk dropped steadily from a high in 1 9 8 7 /8 8 sam plings, averaging 49 pg I-TEQDF/g fa t, to an average of 5 pg I-TEQDF/g fa t in the 1 9 9 2 /9 3 sampling. Grass concentrations sim ilarly dropped from 33 pg I-TEQDF/g dry w e igh t to 4 pg I-TEQDF/g dry w e ig h t between the tw o sample dates. This trend dem onstrates an im portant expectation w ith regard to environm ental responses to reductions in emissions from tall industrial stacks. Specifically, vegetation appears to respond im m ediately to reduced air concentrations, and if dairy cow s are being fed w ith vegetation th a t has reduced concentrations, c o w 's milk should sim ilarly respond in a rapid manner. Fries and Paustenbach (1990) stated th a t a DRAFT--DO NOT QUOTE OR CITE 5-27 December 2003 steady state is reached in c o w 's m ilk w ith a constant dietary input of dioxins after about 30 to 60 days. Therefore, reductions in em issions w ill result in both a reduction in vegetation and c o w 's m ilk concentrations alm ost sim ultaneously. Another study was conducted in Austria by Moche and Thanner (1997). The study evaluated am bient air patterns and CDD/CDF concentrations in a vicin ity of steel production plants in Leoben/Donawitz. Samples were collected from sites in the immediate vicin ity of the production plants, in an area th a t was expected to be impacted by the production plants, and in an area th a t was shielded by m ountains in the northw est. Sampling occurred over four periods to address the potential influence of the summer and w in te r fluctuation s in CDD/CDF concentration. The CDD/CDF concentrations in these samples were compared to previous data collected in the three Austrian conurbations Graz, Linz, and Wien. The previous data suggested average summer levels of CDD/CDFs in the range of 20 to 40 fg I-TEQDF/N m 3 and w in te r levels in the range of 50 to 220 fg ITEQDF/N m 3. The data collected at Leoben/Donawitz indicated higher am bient air levels of CDD/CDF concentrations. Only the levels in the area shielded by m ountains fall w ith in the levels of the previously reported data. In addition, the CDD/CDF profiles of the Leoben/Donawitz sites indicated a high contribution of the lower chlorinated CDFs (tetrathrough hexachlorinated CDFs as the m ost abundant). The patterns were in good agreement w ith emission profiles of m etallurgical processes reported by Hagenmaier et al. (1994) (Moche and Thanner, 1997). Liem et al. (1991) reported on the analysis of over 200 samples of c o w 's m ilk that were taken in various regions in The Netherlands, including some th a t were near municipal solid w aste incinerators and metals reclam ations plants, and some identified as background sites. Background levels ranged from 0 .7 to 2.5 pg I-TEQDF/g lipid. The highest levels were found approximately 2 km from the largest municipal solid waste incinerator identified at the tim e, w ith concentrations ranging from 2.8 to 12.6 pg ITEQDF/g lipid. Higher than background levels were also found in samplings near other incinerators. The researchers did a principal com ponent analysis on congener profiles in the milk samples to determine if there were any discernable differences among groupings of samples. Liem et al. (1991) observed a d istin ct pattern for samples around the metals reclamation plant compared to samples around municipal solid waste facilities. A higher CDF/CDD ratio was found around the metals reclamation plant (i.e., higher furan DRAFT--DO NOT QUOTE OR CITE 5-28 December 2003 concentrations were in the m ilk near the metals reclamation plant than near the municipal solid w aste incinerator). Liem et al. (1991) speculated th a t metals reclam ation plants process cables th a t contain PVC, and according to Christmann et al. (1989), furans are predom inantly form ed in the com bustion of PVC. Subsequently, the higher levels of furans w ould be taken up into vegetation and then into c o w 's milk. Liem et al. (1991) also found d istin ct patterns in samples associated w ith other facilities, as characterized by the relative am ounts of low er and higher chlorinated congeners. Tw o of the incinerators were closed in April of 1990, and a marked decrease in sample concentrations associated w ith these tw o incinerators w as noted between the February and A ugust 1990 sampling. This supports the expectation described above regarding the response of vegetation and m ilk to changes in nearby source emissions. A limited sample from six cow s in Switzerland showed sim ilarly elevated CDD/CDFs in association w ith incinerators or m anufacturing sites. Higher CDD/CDF concentrations were observed in m ilk samples th a t were w ith in 1 ,000 meters of an incinerator (tw o samples) and those th a t were w ith in 1,000 meters of a production site for various chlorinated samples (one sample) than samples from a background farm (one sample) and from local dairies that pooled milk from several farms (tw o samples) (Rappe et al., 1987). Insufficient inform ation was available in this report to calculate I-TEQDF concentrations. De Fre and W evers (1998) evaluated paired CDD/CDF deposition and c o w 's milk data from several locations in Belgium to evaluate the relationship between deposition rates and m ilk levels, and the potential im pact th a t elevated deposition rates may have on local m ilk supplies. CDD/CDF deposition ranged from approxim ately 2 ng I-TEQDF/m 2/year to 45 ng I-TEQDF/m 2/year, and CDD/CDF concentrations in m ilk fa t ranged from approxim ately 1 pg I-TEQDF/g to 19 pg I-TEQDF/g. The correlation co efficie nt (R) for CDD/CDF deposition rates and m ilk fat concentrations w as 0 .6 9 . The results of a regression analysis using these data indicated th a t m ilk fat concentrations of I-TEQDFs could be predicted from deposition rates using the equation y = 0 .3 3 3 2 x , where y is the m ilk fa t concentration of CDD/CDFs in units of pg TEQDF/g and x is the CDD/CDF deposition rate in units of ng TEQDF/m2/y. In France, the M inistry of A griculture and Fisheries investigated CDD/CDF concentrations in c o w 's m ilk sampled from farm s in a dow nw ind direction w ith in 11 km, but m ostly w ithin 5 km, of 26 industrial facilities (Defour et al., 1998). These industries DRAFT--DO NOT QUOTE OR CITE 5-29 December 2003 included: steel m anufacturing, secondary lead and aluminum sm elting, copper refining, chemical and oil refining industries, electricity production, and municipal w aste incinerators. Of the 49 m ilk samples analyzed, 46 samples had CDD/CDF concentrations th a t were less than 3 pg I-TEQDF/g fa t w ith an average of 1.53 pg I-TEQDF/g on m ilk fat basis. One milk sample collected from a site near a chem istry industry was found to contain 3 to 5 pg I-TEQDF/g fa t, and tw o m ilk samples collected 250 m and 1 km dow nw ind of incinerators had concentrations higher than 5 pg I-TEQDF/g fat. The average concentration in m ilk and diary products in France assessed through a 1996 survey conducted by the M inistry of A griculture and Fisheries w as 1.33 pg I-TEQDF/g fa t (Defour et al., 1998). Abraham et al. (1998) reported on the levels of CDD/CDFs in the human m ilk of 10 mothers who lived w ithin a radius of 8 km of Ilsenburg, Germany. The to w n was identified as an area highly contam inated w ith CDD/CDFs reportedly resulting from emissions from a copper plant. A t the tim e of sample collection (i.e., 1997) the plant had been closed for approxim ately 6 years, Abraham et al. (1998) compared the findings to the results of a previous study of human m ilk levels conducted in 1990/1991 when the plant w as still in operation. The 1990/1991 human m ilk samples contained a mean ITEQdf of 59 ppt, lipid based (n = 9). The 1997 human m ilk samples contained a mean ITEQdf of 41 ppt, lipid based (n = 10). Abraham et al. (1998) docum ents th a t this decrease in CDD/CDFs is low er than the decline reported in general background concentrations in human m ilk from W estern Germany in recent years. These values are som ew hat higher than the values reported in Chapter 4 for the general population of the United States. An extensive study was undertaken in the Pontypool environm ent of South Wales (Ball et al., 1993; Ball et al., 1994a; Ball et al., 1994b; Ball et al., 1995). Evidence of the im pact of em issions from w aste incineration at Rechem International Ltd. (a chemical company) prom pted extensive investigations into im pacts from emissions of PCBs and CDD/CDFs to nearby and regional media including soil, grass, w ater, air, fruit/vegetables, c o w 's m ilk, duck m eat, and eggs from chicken and ducks. The region has a com bination of residential and industrial uses, w ith very little agricultural uses. The greatest im pact was found at a residence adjacent to the site, located only about 100 meters away. The soil at Rechem International averaged 8 10 pg I-TEQDF/g (n = 4 ), w hile at this nearby DRAFT--DO NOT QUOTE OR CITE 5-30 December 2003 residence, the concentration averaged 66 pg I-TEQDF/g (n = 8). Other areas evaluated ranged from 4 to 24 pg I-TEQDF/g. Data were not available on CDD/CDF em issions, but air measurements at this residence suggested high emission rates. For five air samples taken at the residence, air concentrations ranged from 1.6 to 14.8 pg I-TEQDF/m 3. This compares to air concentrations ranging from 0 .0 2 to 0 .6 8 pg I-TEQDF/m 3 taken from a site about 2 ,5 0 0 meters away in the same direction from the Rechem site. The researchers also compared these air concentrations to average air concentrations ranging from 0.21 to 0 .6 7 pg I-TEQDF/m 3 in four other UK urban areas. C oncentrations of CDDs/CDFs in grass were found to be elevated at the same residence, but described as more typical for other grass sampling sites. Perhaps m ost im portantly, samples of duck and bantam eggs from this residence showed concentrations th a t exceeded other duck and bantam egg samples in the area by a facto r of 10. Duck and bantam egg concentrations in the area, but not at this residence, were described as typical of background. Duck m eat at the impacted residence was not described as elevated compared to duck meat from nearby settings. Sampling of sediments in a nearby reservoir did not indicate elevated concentrations of dioxins or PCBs. There w as no sampling of human blood or tissue. However, a simple exposure exercise showed that consumption of duck eggs, duck meat, apples, inhalation, and incidental soil ingestion at this im pacted residence w ould result a daily intake of 165 pg I-TEQDF/day, compared to a background intake from these pathw ays of 4 3 .2 pg I-TEQDF/day (consum ption rates described as typical derived from consum ption data from the M inistry of Food and Fisheries in the UK). Foxall et al. (1997) also reported geographical variations in environm ental levels and human exposure to CDD/CDFs and PCBs of the above study. The data indicated a particular im pact in a 200-m eter w ide strip of land around the boundary of the incineration plant owned by Rechem International Ltd. This location is predom inantly dow nw ind from the incinerator and there had been evidence suggesting th a t fu gitive emissions from the plant contributed to the environmental impacts. Marked differences were noted between the CDD/CDF and PCB content of samples (i.e., air, soil, and foods) collected at the impacted site and those collected at rural background locations. Intakes of CDD/CDFs (pg I-TEQDF/day) and PCBs (vg/day) were estim ated using mean daily food consum ption rates, inhalation and soil ingestion rates of 20 m3/day and 100 m g/day, respectively, and the median concentrations of CDD/CDFs and PCBs found in the samples. These estim ates DRAFT--DO NOT QUOTE OR CITE 5-31 December 2003 indicated th a t exposure to CDD/CDFs and PCBs at the im pacted site was much higher than for background levels and the main contributors to these higher levels were residues in bantam and duck eggs. The estim ated intake of CDD/CDFs from ingestion of bantam or duck eggs at the im pacted site were 2 04 pg I-TEQDF/day and 103 pg I-TEQDF/day, respectively; levels th a t are substantially higher than the average UK dietary intake of 88 pg I-TEQDF/day from all food sources. Based on a body mass of 60 kg, these egg intakes (i.e., 3 .4 and 1.7 pg I-TEQDF/kg body mass/day) w ould represent 34 and 17 percent of the WHO (World Health Organization) TDI (Total Dietary Intake) value of 10 pg I-TEQDF/kg body mass. Sim ilarly, the corresponding PCB intake of 7.3 and 6.3 jvg/day would represent 73 and 63 percent, respectively, of an average dietary intake (10 jvg/day) of PCBs. Lovett et al. (1998) performed additional analysis of chicken, bantam, and duck eggs; and also duck meat collected from the vicin ity of the Rechem incinerator and compared the results to PCB and CDD/F levels of comparable fo o d stu ffs collected from rural areas in the same Welsh district. Poultry produced at the impacted residence displayed a congener profile w ith noticeable variations compared to those collected from nearby rural sites. A prominence of higher chlorinated congeners in the egg and duck meat samples for the residence located near the incinerator was observed. Analysis of 46 PCB congeners resulted in a median fresh mass total PCB concentration in duck eggs of 191 yvg/kg (n = 2), 341 yvg/kg in bantam eggs (n = 2), and 43 yvg/kg in duck meat (n = 2) from samples collected in the impacted area. O bservations from rural areas showed fresh mass based total PCB concentrations of 14 yvg/kg for duck eggs (n = 6), 22 yvg/kg for bantam eggs(n = 4), and 25 yvg/kg for duck meat (n = 6). A second location in the United Kingdom, the Derbyshire area in central England, has show n elevations in c o w 's m ilk and other animal tissues. Initially, samples of c o w 's m ilk were taken by the M inistry of A griculture, Fisheries, and Food (MAFF) from individual farm tanks on 11 farm s in 1990. W hen 2 of the samples showed high concentrations of 40 and 42 ng I-TEQDF/kg fa t (the other 9 showed more typical concentrations in the 1.1 to 7.1 ng I-TEQDF/kg fa t), the sampling was expanded to 30 farms. These original tw o farm s, plus an additional farm , continued to show high concentrations in the milk. Testing continued in m ilk through 1994. M ilk concentrations dropped at one farm , but overall concentrations appeared to remain high (i.e., 29 ng I-TEQDF/kg) for the m ost recently DRAFT--DO NOT QUOTE OR CITE 5-32 December 2003 reported sampling in July of 1994 (Harrison et al., 1996). As a result of these findings, MAFF tested animal tissue from the three farms. Calves from one of the three farm s had extrem ely elevated levels of dioxins and furans, w ith concentrations ranging from 2.5 to 6.9 ng I-TEQDF/kg w hole w e ig h t (i.e., not lipid basis) in muscle tissue (MAFF, 1992a). This compares, for example, w ith I-TEQDF concentrations approxim ately 0 .2 0 ng ITEQDF/kg in the United States beef supply (from the national study on beef back fa t, assuming 19 percent fa t in whole beef (W inters et al., 1996). Egg samples were taken from one of the three farm s and a second "free range" supplier. The concentrations found were reported as 2.2 and 2.1 ng I-TEQDF/kg in w hole eggs. A second sample from one of the farm s taken a year later showed a low er concentration of 0 .8 ng I-TEQDF/kg whole w eight. These w ould appear to be elevated, considering th a t eggs found in a background setting in M ississippi (Cooper et al., 1995), had concentrations less than 0 .1 0 ng ITEQDF/kg w hole w eight. MAFF (1992b) also sampled leafy herbage (grass, hay, etc.) from the three farms described above. Concentrations ranged from about 2 to 14 ng I-TEQDF/kg dry w eight. This appears elevated, considering th a t samplings of background grass in England showed 0 .8 9 ng I-TEQDF/kg dry w e ig h t, as reported in 1991 (Kjeller et al., 1991) and 0 .5 7 ng I-TEQ/kg dry w e igh t in 1996 (Kjeller et al., 1996). The evidence suggests th a t the im pacts were due to nearby industrial em issions from a fuel plant and chemical w aste incineration. Her M ajesty's Inspectorate of Pollution (HMIP) conducted additional studies to evaluate this possibility. These included stack testing of the Coalite Fuels, Ltd. and the Coalite Chemicals, Ltd. (which were adjacent to one of the farm s and near the other tw o ), air dispersion modeling, and soil m onitoring. No results were available to evaluate the stack testing, but the air dispersion modeling predicted th a t the three impacted farm s w ould be in the sectors having the highest air concentrations. The soil sampling on the three farm s showed concentrations ranging from 10 to 90 ppt. This can be considered elevated above typical rural background and in the range or even higher than typical urban concentrations. Specifically, this compares w ith typical urban soil concentrations of 10 to 20 pg I-TEQDF/g in both Europe and the United States, and rural soil concentrations typically less than 5 pg I-TEQDF/g. (See Chapter 3.) Also, these concentrations are similar to concentrations found near incinerators em itting very high concentrations of dioxins in Tyrol, A ustria, as described above, and in Columbus, Ohio, as described below. The DRAFT--DO NOT QUOTE OR CITE 5-33 December 2003 National Rivers A u th o rity sampled sediment upstream and dow nstream of the effluent discharge pipe of the Coalite Chemicals, Ltd site. Samples collected about 1.5 km dow nstream of the discharge site had CDD/CDF levels th a t were 1 ,000 tim es greater than background samples collected 1.5 km upstream of the discharge point (these studies reported in MAFF, 1992a). Another interesting finding associated w ith the samplings of foods and environm ental media in the Derbyshire area were the congener profiles. Compared to background m ilk samples, the samples from the three im pacted farm s had proportionally higher concentrations of low er chlorinated dioxins, particularly 2,3,7,8-TC D D and 1,2,3,6,7,8-H xC D D . The background samples of m ilk tended to be dom inated by OCDD (MAFF, 1992b). Blood samples were also collected from residents of these three impacted farm s and analyzed for CDD/CDFs (Startin et al., 1994). The I-TEQDF concentrations for the 10 individual blood samples ranged from 49 pg/g to 291 pg/g on a lipid basis. In contrast, the tw o control samples had I-TEQDF concentrations of 16 pg/g and 26 pg/g on a lipid basis. The Derbyshire samples were dom inated by OCDD, follow ed by 1,2 ,3 ,4 ,6 ,7 ,8 -H p C D D , 2,3,7,8-T C D D , and 1,2,3,6,7,8-H xC D D . Sandalls et al. (1998) analyzed soil concentrations around the site of a chemical waste incinerator near Bolsover, Derbyshire, United Kingdom. A t each of 46 sites, five surface soil samples were collected, at varying depths up to a depth of 5 centim eters, every 1 square meter. All 46 sample sites had total TCDD concentrations exceeding background concentrations. Higher concentrations of TCDD were observed at locations closer to the incinerator and there was a strong correlation between the TCDD concentration in a given quadrant and the am ount of tim e th a t the w ind w as blowing in that direction. A t four quadrants around the site, TCDD soil concentrations were reported as 603 ppt for the northeast (approxim ately 42 percent of the total deposition); 315 ppt for the southeast (22 percent of the total); 269 ppt for the southw est (19 percent of the to ta l); and 244 ppt (17 percent of the total) for the northw est. The results of the spatial distribution of CDD/CDFs im plicated the incinerator as the likely source, and the correlation between deposition and w ind direction suggested th a t these com pounds reached the ground via the atmosphere. Also, 42 of the 46 sample sites showed similar CDD/CDF congener ratios to the flue gas of the w aste incinerator. Soil concentrations were well in excess of background concentrations, up to 5 kilom eters around the site. DRAFT--DO NOT QUOTE OR CITE 5-34 December 2003 Ohta et al. (1997) studied levels of CDD/CDF and non-ortho coplanar PCBs in soil at a high cancer-rate area close to a batch-type municipal solid w aste (MSW) incinerator in Japan. Sixty-one soil samples were collected around the MSW incinerator. Among them, 52 samples were radially collected w ith in 2 km from the center of the MSW incinerator, and 9 samples were collected across the high cancer-rate area. High concentrations of CDD/CDFs and coplanar PCBs were observed in all the soil samples from the leeward side of the MSW incinerator. Total concentrations ranged from 5,303 to 3 2,167 pg/g; mean = 1 3 ,9 3 4 pg/g. On the other hand, all but one sample on the w indw ard site showed high contam ination. Am ong the 61 samples analyzed, the total concentration w as greater than 2 .0 0 0 pg/g in 45 of the 61 samples and the TEQDFP-W HO94 concentration was over 10 pg/g in 39 of the 61 samples. In addition, the levels of CDD/CDFs and coplanar PCBs at a distance of 0 to 1.1 km from the MSW incinerator was compared w ith th a t of the area 1.1 to 2.0 km from the MSW incinerator. The area closer to the incinerator contained a higher ratio of samples w ith contam ination over 5 ,0 0 0 pg/g (63.2 percent) than in the area further away from the MSW incinerator (38.5 percent). Similarly, the area closer to the incinerator had a higher percentage of samples w ith CDD/CDF/PCB levels over 30 pg TEQDFP-W HO94/g (26.3 percent) than the area further away from the incinerator (15.3 percent). M iyata et al. (1998) collected blood samples from residents living w ith in 2 km from a batch-type municipal solid w aste incinerator in Japan where soil concentrations of CDD/CDFs and PCBs were show n to be elevated. Eighteen blood samples were collected from 13 men, aged 23 to 63 years old (average age = 45 years), and 5 w om en, aged 30 to 72 years old (average age = 46 years) in March 1996. The results indicated th a t the average lipid-based TEQDFP-W HO94 concentrations found in blood samples ranged from 34 pg/g to 200 pg/g w ith a mean of 81 pg/g for men, and from 22 pg/g to 4 6 3 pg/g w ith a mean of 149 pg/g for wom en. These mean TEQDFP-W HO94 values are higher than those reported for the general population of various countries (the mean value estim ated in Chapter 4 of this docum ent is 55 pg/g). Local im pacts around a w aste-to-energy municipal solid w aste incinerator in Columbus, Ohio, were undertaken by the Ohio Environmental Protection Agency, and the U.S. Environmental Protection Agency (Lorber et al., 1998). This incinerator operated between 1983 and 1994. A stack te st w as taken in 1992, and when the results were DRAFT--DO NOT QUOTE OR CITE 5-35 December 2003 extrapolated to typical operation of the incinerator, annual emissions were calculated at 985 g I-TEQ/yr. This is a very high emission rate, and compares to total em issions from several European countries. It is about one-tenth of the national emissions estim ated for all United States sources. (See Volum e 1.) Process m odifications were undertaken in the w inter of 1993/94. A stack test was conducted w hich indicated that annual emissions were reduced to 267 g I-TEQDF/yr. The U.S. EPA undertook a soil testing program in December of 1995. Results showed a definite im pact to soils at the site of the incinerator, w ith an average concentration of 3 56 pg I-TEQDF/g (n = 4). Of the four samples collected, three of the samples averaged 4 5 8 pg I-TEQDF/g and the fourth was much low er at 50 g I-TEQDF/g. Just o ffsite in the dow nw ind direction, a cluster of four samples w ith in 1,000 meters also showed some elevation of CDD/CDFs w ith an average concentration of 49 pg I-TEQDF/g. Fourteen additional samples, generally w ith in 2 miles of the site, averaged 10 pg I-TEQDF/g. Three soil samples at a background site 28 miles away in the upw ind direction averaged 1 pg I-TEQDF/g. These latter tw o clusters of urban and background samples have concentrations th a t are typical of urban and background situations. The urban results suggests th a t, despite large emissions from this source, soil im pacts above typical levels appeared to be restricted to w ith in 1,000 meters of the incinerator. The Ohio EPA conducted air m onitoring in 1994 and 1995 (OEPA, 1994; Lorber et al., 1998). M onitoring in 1994 occurred after process m odifications were undertaken to reduce dioxin emissions. A stack te st conducted just prior to the air sampling showed reductions of 75 percent from the levels measured in 1992. W ind rose data were taken on an hourly basis during the 1994 sampling. This showed th a t tw o samples from a sampler located about 2 miles away were in the dow nw ind direction during the 48-hour sampling period. Eight other samples from four samplers (w hich were between 1 and 2 miles from the incinerator) were clearly not in the dow nw ind direction. The tw o dow nw ind samples averaged 0 .2 6 pg I-TEQ/m3 CDD/CDFs, w hile the eight upwind samples averaged 0 .0 5 pg I-TEQDF/m 3. The incinerator shut dow n in December 1994. Five samples taken in 1995 showed an average of 0 .0 5 pg I-TEQDF/m 3 CDD/CDFs. This air sampling suggests the follow ing: (1) the typical background urban air CDD/CDF concentration in Columbus is probably around 0 .0 5 pg I-TEQDF/m 3; and (2) when the incinerator is em itting dioxins typical of the rate measured in the 1994 stack te st (not the DRAFT--DO NOT QUOTE OR CITE 5-36 December 2003 1992 stack test), air concentrations about 2 miles away are higher, perhaps by a facto r of 5. Other media, including vegetation, agricultural products, or human blood were not sampled. Schecter and Papke (1998) examined CDD/CDFs and PCBs 77, 126, and 169 in blood sampled from 10 residents living near a PCB m anufacturing fa cility in Alabama in 1997, and compared these concentrations to a pooled sample from a control group representing 100 adults. The results showed that, for the 10 residents, total lipid-based CDD/CDF concentrations ranged from 825 ppt to 6,422 ppt. The corresponding total CDD/CDF concentrations from the pooled control sample was 1,112 ppt. Total PCB concentrations ranged from 240 ppt to 5,216 ppt for the 10 nearby residents, compared to 1,112 ppt for the pooled control sample. PCB-77 was only detected in the blood of 5 of the 10 residents, ranging from 41 ppt to 713 ppt. PCB-126 concentrations ranged from 104 ppt to 4 ,0 5 0 ppt in residents, compared to 48 ppt found in the control sample. PCB-169 concentrations ranged from 136 ppt to 2,807 ppt for the 10 residents, compared to 35 ppt for the control sample. In term s of TEQ, the total I-TEQD and I-TEQF concentrations ranged from 16.3 ppt to 38 .9 ppt, and from 6.7 ppt to 131 ppt, respectively, compared w ith 18.5 ppt and 8.3 ppt from the control blood. TEQP-W HO94s ranged from 34 ppt to 360 ppt compared to 32 ppt for the control blood. There were also studies conducted in Asia to address localized impacts. Luksemburg et al. (1997) reported, in a prelim inary study, th a t high levels of CDD/CDFs were observed in soil and sediment samples collected inside and outside a sodium pentachlorophenate plant in Tianjin, China. The plant is situated in a w etland w ith rivers em ptying into the nearby Pacific Ocean and close to several large housing developments. Human hair samples collected from barber shops in the housing developm ents near the plant were also collected and CDD/CDFs were detected in these samples. The I-TEQDF concentrations in soil ranged from 15 ppt at a site upstream of the plant to 7 4 0 ,0 0 0 ppt w ith in the plant, and was 1,800 to 2 ,2 0 0 ppt at sites outside the plant. The I-TEQDF concentrations in sediment ranged from 150 ppt at a site 50 km aw ay from the site to 1 1 0 ,0 0 0 ppt in a drainage canal located ju st southw est of the plant. Hair samples contained I-TEQDF concentrations ranging from 12 to 120 ppt. According to Luksemburg et al. (1997), "the isomer profiles of all the samples were consistent w ith the DRAFT--DO NOT QUOTE OR CITE 5-37 December 2003 pentachlorophenol sources." However, it should be noted th a t no background inform ation on the te st subjects (time of residence, health records, etc.) were collected in this study. The major findings and conclusions based on this review of localized sources include: Localized impacts, meaning elevated concentrations of CDD/CDFs above background, have been found in the vicin ity of some CDD/CDF sources. Localized im pacts appear to be limited to an area w ith in 5 km of an incinerator source, perhaps only w ith in 2 to 3 km of the source, and in some cases, only w ithin a few hundred meters of the source. One study noted elevations in grass, c o w 's m ilk, and human blood on a farm located 2 km from an incinerator presumed to be em itting high am ounts of CDD/CDFs. Not all of the studies described in the literature discussed distance from the source, as the surveyed areas were sim ply identified as "in d u stria l." Several studies continued environmental samplings after effo rts were made to reduce em issions or after the sources were shut dow n. In these cases, reductions in CDD/CDF concentrations were noted for various media including, c o w 's m ilk, vegetation, and air. As discussed below , vegetation has been found to respond rapidly to reductions in air concentrations, and the tim e to reach steady state in c o w 's m ilk given a steady input of CDD/CDFs is also relatively short. In other cases, such as in the accum ulation of CDD/CDFs in soils or in body fa t, the benefits may not be as immediate. Soil and body fat are reservoirs in w hich the residence tim e of these com pounds are measured on the order of years. The available data reviewed above suggest that measurable im pacts near incinerators only occur if the incinerator em its very high am ounts of dioxins, in contrast to emissions th a t are know n to be w ith in regulatory lim its. However, tw o key descriptors here, including "measurable im pacts" and "very high am ounts of em issions" cannot be rigorously defined. The data suggest th a t "m easurable" im pacts can be defined as DRAFT--DO NOT QUOTE OR CITE 5-38 December 2003 elevations in dioxin concentrations in environm ental or biotic media on the order of 5-10 tim es higher than typical background. "V ery high am ounts" of releases is less well defined. The Columbus incinerator is the only incinerator reviewed above for w hich emission data were available, and the stack test of 1992 showed emissions th a t were about 200 tim es higher than the 1995 proposed regulatory lim it for solid waste incinerators of 30 ng of total dioxins per m3. By comparison of environm ental media sam pling, one could surmise th a t the metals reclamation plant in Tyrol, A ustria, and the incinerators in the Derbyshire area of Central England, if not others noted above, were also em itting unusually high amounts of dioxins. Also, it is im portant to understand th a t elevations in air, soil, vegetation, and animal products do not autom atically translate to higher exposure levels. This document (and several other e ffo rts w orldw ide) have concluded th a t the bulk of exposure to dioxins occurs via the diet, and specifically animal fats. Higher exposure to an individual would only result if an individual subsisted on animal food products from animals raised near incinerators (meaning also that the animal's diet was comprised of vegetation grown where the animal is raised), and perhaps only incinerators em itting high am ounts of dioxins. As described above, there are limited human tissue data for individuals where localized environm ental contam ination has been dem onstrated. In one study in Austria of a farming fam ily consuming home-grown milk near the metals reclamation plant, both the m ilk and the blood of the fam ily were show n to have elevated levels of dioxins. In the vicin ity of a municipal solid w aste incinerator in Japan where CDD/CDF and PCB concentrations in soil were elevated, nearby residents also had elevated blood levels of CDD/CDF/PCBs. In the U.S., there have been no studies to evaluate the prevalence of subsistence behavior near sources, in general, and near high em itting incinerators, in particular. Even if there is a sparsity of subsistence behaviors near sources in the U.S., it is reasonable to assume th a t animal fats produced near high em itting incinerators would likely have elevated CDD/CDF levels and be consumed. If incinerators are meeting regulation lim its, they would not be high em itters, and the likelihood of localized im pacts w ould be small. In addition to the localized contam ination resulting from incinerator em issions, as described above, there have been several incidents involving contam ination of commercial food supplies from natural and accidental sources in various parts of the w orld. For DRAFT--DO NOT QUOTE OR CITE 5-39 December 2003 example, ball clay was found to be the source of elevated CDD/CDF levels in animal food products in the United States (Ferrario et al., 2000). Ball clay from a mine in M ississippi, w hich w as used as an anti-caking agent in soy-based animal feeds, w as found to be the source of contam ination after poultry and catfish samples collected in the same region of the United States were found to have concentrations of CDD/CDFs that were significantly elevated above background concentrations. The incident, w h ich occurred in 1998, involved less than 5 percent of the national poultry production. Subsequently, the use of ball clay in animal feeds was discontinued. In Germany, dairy products were found to be contam inated in 1998 (Malisch, 1998). The concentration of CDD/CDFs in m ilk was 1.38 pg I-TEQDF/g fa t (N = 4 3 ) in 1998, compared to 0 .6 2 pg I-TEQDF/g fa t (N = 76) in 1997. A fter intense investigation, contam inated citrus pulp obtained from Brazil, w hich was used as a feed ingredient for dairy cow s in certain regions of Germany in 1998, w as found to be the source of contam ination of the food supply. In 1999, similar contam ination of the com mercial food supply occurred in Belgium when 500 tons of animal feed was inadvertently contam inated w ith approxim ately 50 kg of PCBs and 1 g of dioxins in transform er oil (Van Larebeke et al., 2001). The feed w as delivered to poultry farm s (and to a lesser extent rabbit, co w , and pig breeding facilities), prim arily in Belgium. This contam inated feed represented a "lim ited percentage" of the feed produced in Belgium, but was delivered to hundreds of farm s in the country. The mean TEQDFP-W HO94 concentrations were 170 pg/g fa t in poultry, 2.3 pg/g fa t in eggs, and 2 ,3 2 0 pg/g fa t in animal feed. Once discovered, animal products w ith excessive levels were destroyed, including approximately 2 million chickens. DRAFT--DO NOT QUOTE OR CITE 5-40 December 2003 5.5. CIGARETTE SMOKERS As discussed in Volum e 1, cigarette smoking has been found to be a source of CDD/CDFs. As a result, individuals w h o smoke cigarettes, and nonsmokers w h o are exposed to second-hand smoke, may experience higher levels of exposure to dioxin-like com pounds than the general population. Matsueda et al. (1994) reported th a t the mean I-TEQdf content of a pack of U.S. cigarettes was 8.6 pg. This estim ate is based on analytical data from seven brands of U.S. cigarettes. Assuming that a pack of cigarettes contains 20 cigarettes, the I-TEQDF content of a single cigarette w ould be 0 .4 3 pg. This value represents about half of the I-TEQDF value reported for a mainstream cigarette smoke from a Swedish brand of cigarettes (Lofroth and Zebuhr, 1992) and is about five tim es higher than the I-TEQDF level in m ainstream smoke from German cigarettes (Ball et al., 1990). The daily intake of CDD/CDFs by smokers can be estim ated by m ultiplying the CDD/CDF content of a single cigarette by the mean number of cigarettes smoked per day by current smokers. According to U.S. EPA (1992 ), 25.5 percent of the adult U.S. population were smokers in 1990. The average daily number of cigarettes smoked by this population was 19.1. Thus, mean CDD/CDF exposures via cigarette smoking are estim ated to be 8 .2 pg I-TEQDF/day for smokers. This level of exposure represents over 10 percent of the average daily background dose of CDD/CDFs from soil, air, w ater, and foods, as described in Chapter 4. The use of data on the total I-TEQDF content of a cigarette from Matsueda et al. (1994) results in uncertainties as to the estim ate of exposure to smokers because the approach assumes th a t all of the dioxin in the unburned cigarette is inhaled. It is likely th a t some of the dioxins are released w ith the sidestream smoke rather than being inhaled. It is also possible th a t dioxins are destroyed and/or form ed during the com bustion process. Thus, it is unclear if these factors w ould lead to a net increase or decrease in the am ount of dioxins inhaled. However, as described above, the I-TEQdf value reported by Matsueda et al. (1994) is less than th a t in m ainstream (i.e., inhaled) smoke reported by Lofroth and Zebuhr (1992) and greater than th a t of Ball et al. (1990) and provides the best estim ate of CDD/CDFs in cigarettes to w hich smokers may be exposed. Nonsmokers may also be exposed to CDD/CDFs from environm ental tobacco smoke. A lthough the data on the frequency, m agnitude, and duration of exposure to environm ental tobacco smoke are lim ited, an idea of the m agnitude of exposure to DRAFT--DO NOT QUOTE OR CITE 5-41 December 2003 CDD/CDFs can be gained by assuming th a t nonsmokers receive a fraction of the CDD/CDF TEQ received by smokers. Based on data for nicotine, the dose to nonsmokers exposed to environm ental tobacco smoke is estim ated to be 0.1 to 0 .7 percent th a t of smokers (U.S. EPA, 1992). For 4-am inobiphenyl, nonsmokers exposed to environm ental tobacco smoke were estimated to receive a dose that was 10 to 20 percent that of smokers (U.S. EPA, 1992). Assuming that nonsmokers receive 0.1 to 20 percent of the dose of CDD/CDFs from second-hand smoke that smokers receive, the estimated daily dose of CDD/CDFs for nonsmokers w ould range from 0 .0 0 8 pg I-TEQDF/day to 1.6 pg I-TEQDF/day. It should be noted, how ever, th a t individual exposure to sidestream smoke is highly variable, depending on a person's proxim ity to smokers, how often they are near smokers, and the ventilation rate in these areas. DRAFT--DO NOT QUOTE OR CITE 5-42 December 2003 REFERENCES Abraham , K.; Papke, O.; Ball, M .; Lis, A .; Helge, H. (1994) Concentrations of PCDDs, PCDFs and coplanar PCBs in blood fa t of a breast-fed and a form ula-fed infant. Organohalogen Compounds 21:163 -1 65. Abraham , K.; Papke, O.; Ball, M .; Lis, A .; Helge, H. (1995) Changes in blood lipid concentration of PCDDs, PCDFs, and coplanar PCBs in a breast-fed and a form ulafed infant in the second year of life. Organohalogen Compounds. 2 6 :2 2 3 -2 2 5 . Abraham , K.; Palavinskas, R.; W ahn, U.; Helge, H.; M athar, W. (1998) Still relatively high PCDD/PCDF concentrations in human m ilk of m others living in a contam inated area in Germany. Organohalogen Compounds (3 8):33 -3 6. Abraham , K.; Papke, O; W ahn, U.; Helge, H. (2000) POP accum ulation in infants during breast-feeding. Organohalogen Compounds 48, 25-26. Am erican Academ y of Pediatrics. (1997) Breastfeeding and the use of human milk (RE9729) Pediatrics. 100 (6 ):1035 -10 39. Anderson, H.A.; Falk, C.; Hanrahan, L.; Olson, J.; Burse, V .W .; Needham, L.; Paschal, D.; Patterson, D.; Hill, R.H.; Boddy, J.; Budd, M.; Burkett, M ..; Fiore, B.; Humphrey, H.E.B.; Johnson, R.; Kanarek, M ..; Lee, G.; Monaghan, S.; Reed, D.; Shelley, T.; Sonzogni, W .; Steele, G.; W right, D. (1998) Profiles of Great Lakes critical pollutants: A sentinel analysis of human blood and urine. Environmental Health Perspectives. 106(5):279-289. Asplund, L.; Svensson, B.G.; Nilson, A .; Ericksson, U.; Jansson, B.; Jensen, S.; W idegvist, U.; Skerfving, S. (1994) Polychlorinated biphenyls, 1,1 ,1 -trich lo ro -2 ,2 bis(p-chlorophenyl) ethane (p,p'-DDT) and 1,1-dichloro-2,2-bis(p-chlorophenyl)ethylene(p,p'-DDE) in human plasma related to fish consum ption. Archives of Environmental Health. 4 9 :4 7 7 -4 8 6 . A yo tte , P.; Carrier, G.; Dewailly, E. (1994) Health risk assessment for new borns exposed to organochlorine compounds through breast feeding. Organohalogen Compounds. 21: 45-50. Ball, D.J.; Creaser, C.S.; Foxall, C.; Lovett, A .A .; MacGuire, F.A.S.; and Patteson, M.T. (1993) Polychlorinated Biphenyls, Dioxins, and Furans in the Pontypool Environment. The Panteg M onitoring Project. Final Report. Centre for Environmental and Risk M anagement, School of Environmental Sciences, University of East Anglia Norw ich NR4 7TJ, UK. Subm itted to the Welsh O ffice of the W orld Health Organization, April, 1993. Ball, D.J.; Creaser, C.S.; Foxall, C.; Lovett, A .A .; MacGuire, F.A.S.; and Patteson, M.T. (1994a) Polychlorinated Biphenyls, Dioxins, and Furans in the Pontypool Environment. Research Report No. 21. Centre for Environmental and Risk DRAFT--DO NOT QUOTE OR CITE 5-43 December 2003 Management, School of Environmental Sciences, University of East Anglia Norwich NR4 7TJ, UK. Subm itted to the W elsh O ffice of the W orld Health Organization, January, 1994. Ball, D.J.; Creaser, C.S.; Foxall, C.; Lovett, A .A .; MacGuire, F.A.S.; and Patteson, M.T. (1994b) Polychlorinated Biphenyls, Dioxins, and Furans in the Pontypool Environment. Research Report No. 22. Centre for Environmental and Risk Management, School of Environmental Sciences, University of East Anglia Norwich NR4 7TJ, UK. Subm itted to the W elsh O ffice of the W orld Health Organization, October, 1994. Ball, D.J.; Creaser, C.S.; Foxall, C.; and Lovett, A .A . (1995) Polychlorinated Biphenyls, Dioxins, and Furans in the Pontypool Environment. Research Report No. 25. Centre for Environmental and Risk Management, School of Environmental Sciences, University of East Anglia Norw ich NR4 7TJ, UK. Subm itted to the Welsh O ffice of the World Health Organization, June, 1995. Ball, M .; Ppke, O.; Lis, A. (1990) Polychloridibenzodioxine und polychlordibenzofurane in cigaretterauch. Beitrge zur Tabakforschung International. 14 (6 ):3 9 3 -4 0 2 . Beck, H.; Dross A .; M athar, W .; and W ittk o w s i, R. (1990) Influence of different regional emissions and cardboard containers on levels of PCDD, PCDF, and related com pounds in cow milk. Chemosphere. 21: 789 -7 9 8 . Campbell, C .A .;, Teschke, K.; Bert, J.; Quintana, P.J.E.; Herzman C. (1996) Pharmacokinetic model of dioxin and furan levels in adipose tissue from sawm ill w ork involving chlorophenate fungicides. Chemosphere 33:2373-2381. Carrier, G.; Brunet, R.C.; Broduer, J. (1995a) Modeling of the toxicokinetics of polychlorinated dibenzo-p-dioxins and dibenzofurans in mammalians, including humans. I. Nonlinear distribution of PCDD/PCDF body burden between liver and adipose tissues. Toxicology and Applied Pharmacology. 131: 253 -2 66. Carrier, G.; Brunet, R.C.; Broduer, J. (1995b) Modeling of the toxicokinetics of polychlorinated dibenzo-p-dioxins and dibenzofurans in mammalians, including humans. II. Kinetics of absorption and disposition of PCDDs/PCDFs. Toxicology and Applied Pharmacology. 131: 253 -2 6 6 . CDC (1997) Vital and health statistics. Fertility, fam ily planning, and w o m en 's health: new data from the 1995 National Survey of Family Growth. National Center forH ealth S tatistics, Centers for Disease Control and Prevention, U.S. Department of Health and Human Serivces. Series 23, No. 19. May 1997. Christmann, W .; Kasiske, D.; Kloppel, K.D.; Partscht, H.; Rotard, W. (1989) Combustion of polyvinylchloride - an im portant source for the form ation of PCDD/PCDF. Chemosphere. 19:387-392. DRAFT--DO NOT QUOTE OR CITE 5-44 December 2003 Cole, D.C.; Kearney, J.; Gilman, A.P.; Ryan, J.J. (1995) Serum PCB, dioxin, and furan levels in Ontario Great Lake anglers. Organohalogen Compounds. 2 6 :1 9 3 -1 9 6 . Columbia River Intertribal Fish Commission (1994) A fish consum ption survey of the Um tilla, Nez Perce, Yakima, and W arm Springs Tribes of the Columbia River basin. Technical Report 94-3. Cooper, K.R.; Fiedler, H.; Bergek, S.; Anderson, R.; Hjelt, M .; Rappe, C. (1995) Polychlorinated dibenzo-p-dioxins (PCDD) and polychlorinated dibenzofurans (PCDF) in food samples collected in Southern M ississippi. Organohalogen Compounds. 26:51-57. Defour, S.; Fraisse, D.; Scherrer, M .C .; Schnepp, B.; Le Querrec, F. (1998) Analysis of polychlorodibenzo-dioxins (PCDDs) and polychlorodibenzo-furans (PCDFs) in milk around industrial sites in France. Organohalogen Compounds. 3 8 :8 5 -8 8 . De Fre, R.; W evers, (1998) Measurements of dioxins in deposition and in c o w 's m ilk in Belgium. Organohalogen Compounds. 3 8 :1 1 3 -1 1 6 . Dewailly, E.; Ryan, J .J.; Laliberte, C.; Bruneau, S.; W eber, J.P.; Gingras, S.; Carrier, G. (1994) Exposure of remote m aritime populations to coplanar PCBs. Environmental Health Perspectives. Vol. 102, Suppl. 1:205-209. Falk, C.; Hanrahan, L.; Anderson, H.; Kanarek, M .S.; Draheim, L.; Needham, L.; Patterson, D.; and the Great Lakes Consortium (1999) Body burden levels of dioxin, furans, and PCBs among frequent consumers of Great Lakes sport fish. Environmental Research Section A. 80:S19-S25. Ferrario, J.; Byrne, C.; Cleverly, D. (2000). Summary of evidence for the possible natural form ation of dioxins in mined clay products. Organohalogen Compounds. 4 6 :2 3 26. Flesch-Janys, D.; Becher, H.; Gurn, P.; Jung, D.; Konietzko, J.; Manz, A .; Papke, O. (1996) Elimination of polychlorinated dibenzo-p-dioxins in occupationally exposed persons. Journal of Toxicology and Environmental Health 4 7 :3 6 3 -3 7 8 . Foxall, C.D.; Lovett, A .A .; Creaser, C.S.; Ball, D.J. (1997) Environmental levels of PCDD/DFs and PCBs in the Panteg D istrict of South W ales, UK and their implications for human exposure. Organohalogen Compounds. 3 1 :342 -3 47. Fries, G.F., Paustenbach, D.J. (1990) Evaluation of potential transm ission of 2 ,3 ,7 ,8 tetrachlorodibenzo-p-dioxin-contam inated incinerator emissions to humans via foods. Journal of Toxicology and Environmental Health. 2 9 :1-43. Geyer, H.J.; Scheunert, I.; Filser, J.G .; Korte, F. (1986) Bioconcentration potential of (BCP) of 2,3,7,8-tetrachlorodibenzo-p-dioxin (2,3,7,8-TC D D ) in terrestrial organisms including humans. Chemosphere 15: 1495-1502. DRAFT--DO NOT QUOTE OR CITE 5-45 December 2003 Goldman, L.R.; Harnly, M .; Flattery, J.; Patterson, D.G.; Needham. L.L. (2000) Serum polychlorinated dibenzo-p-dioxin and polychlorinated dibenzofurans among people eating contam inated home-produced eggs and beef. Environm ental Health Perspectives. 108:13-19. Hagenmaier, H.; Lindig, C.; She, J. (1994) Correlation of environmental occurrence of polychlorinated dibenzo-p-dioxins and dibenzofurans w ith possible sources. Chemosphere. 29:2163-2174. Harrison, N.; de M. Gem, M .G .; Startin, J.R.; W right, C.; Kelly, M .; Rose, M. (1996) PCDDs and PCDFs in m ilk from farm s in Derbyshire, U.K. Chemosphere. 3 2 :4 5 3 460. Hong, C.S.; Kiao, J.; Casey, A.C .; Bush, B.; Fitzgerald, E.F.; Huang, S.A. (1994) M ono ortho and non-ortho-substituted polychlorinated biphenyls in human m ilk from M ohaw k and control wom en: e ffe cts of maternal factors and previous lactation. Archives of Environmental Contam ination and Toxicology. 2 7 :4 3 1 -4 3 7 . Humphrey,H.E.B.; Gardiner,J.C.; Pandya, J.R.; Sweeney, A .M .; Gasior, D.M .; M cCaffrey, R.J.; Schantz, S.L. (2000) PCB congener profile in the serum ofhum ans consuming Great Lakes fish. Environmental Health Perspectives. 108(2):167-172. Kjeller, L-O; Jones, K.C.; Johnston, A.E.; Rappe, C. (1991) Increases in the polychlorinated dibenzo-p-dioxin and -furan content of soils and vegetation since the 1840s. Environmental Science and Technology. 2 5 :1 6 1 9 -1 6 2 7 . Kjeller, L-O; Jones, K.C.; Johnston, A.E.; Rappe, C. (1996) Evidence for a decline in atm ospheric emissions of PCDD/Fs in the UK. Environmental Science and Technology. 30:1398-1403. Kolic, T .M .; MacPherson, K .A.; Reiner, E.J.; Gobran, T.; Hayton, A. (2000) A comparison of TEQ contributions from chlorinated dioxins, furans, and dioxin-like PCBs in Great Lakes fish. Organohalogen Compounds 46:5 6 2 -5 6 5 . Kreuzer, P.E.; Csanady, G y.A .; Baur, C.; Kessler, W .; Papke, O.; Greim, H.; Filser, J.G. (1997) 2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD) and congeners in infants. A toxicokinetic model of human lifetim e body burden by TCDD w ith special emphasis on its uptake and nutrition. Archives of Toxicology 7 1 :3 8 3 -4 0 0 . Lakind, J.S.; Berlin, C.M .; Park, C.N.; Naiman, D.Q.; Gudka, N.J. (2000). Methodology for characterizing distributions of incremental body burdens of 2,3,7,8-TC D D and DDE from breast m ilk in north american nursing infants. Journal of Toxicology and Environmental Health 59:605 -6 39. Liem, A .K .D .; Hoogerbrugge, R,; Kootstra, P.R.; van der Velde, E.G.; de Jong, A .P.J.M . (1991) Occurrence of dioxins in c o w 's m ilk in the vicin ity of municipal w aste incinerators and metal reclam ation plant in The Netherlands. Chemosphere. 21:1675-1684. DRAFT--DO NOT QUOTE OR CITE 5-46 December 2003 Lofroth, G.; Zebhr, Y. (1992) Polychlorinated dibenzo-p-dioxins (PCDD) and dibenzofurans (PCDFs) in m ainstream and sidestream cigarette smoke. Bulletin of Environmental Contam ination and Toxicology. 4 8 :7 8 9 -7 9 4 . Lorber, M .; Pinsky, P.; Gehring, P.; Braverman, C.; W inters, D.; Sovocool, W. (1998) Relationships between dioxins in soil, air, ash, and emissions from a municipal solid w aste incinerator em itting large am ounts of dioxin. Chemosphere. 3 7 (9 -1 2 ):2 1 7 3 2197. Lovett, A .A .; Foxall, C.D.; Creaser, C.S.; Chewe, D. (1998) PCB and PCDD/F concentrations in egg and poultry meat samples from know n urban and rural locations in Wales and England. Chemosphere. 3 7 (9 -1 2 ):1 6 7 1 -1 6 8 5 . Luksemburg, W .J.; Mitzel, R.S.; Hedin, J.M .; Silverbush, B.B.; W ong, A.S. (1997) Polychlorinated and dibenzofurans (PCDDs/PCDFs) levels in environm ental and human hair samples around a pentachlorophenol plant in China. Organohalogen Compounds. 32:38-40. M alisch, R. (1998) Increase of PCDD/F - Contam ination of m ilk and butter in Germany by use of contam inated citrus pulps as com ponent in feed. Organohologen Compunds. 38:65-70. Matsueda, T.; Kurokawa, Y.; Nakamura, M .; Takada, S.; Fukamachi, K. (1994) Concentrations of PCDDs, PCDFs and coplanar PCBs in cigarettes from various countries. Organohalogen Compounds. 2 0 :331 -3 34. M ichalek, J.; Pirkle, J.; Caudill, S.; Tripathi, R.; Patterson, D.G.; Needham, L.L. (1996) Pharmacokinetics of TCDD in veterans of operation Ranch hand: 10 year follow -up. Journal of Toxicology and Environmental Epidemiology 47: 209 -2 20. M inistry of A griculture, Fisheries, and Food (MAFF) (1992a) Report of Studies on Dioxins in Derbyshire Carried Out by the M inistry of A griculture, Fisheries, and Food. M inistry of A griculture, Fisheries, and Food (MAFF) (1992b) Third Report of Studies on Dioxins in Derbyshire Carried Out by the M inistry of A griculture, Fisheries, and Food. M iyata, H.; Kuriyama, S.; Nakao, T.; Aozasa, O.; Ohta, S. (1998) Contam ination levels of PCDDs, PCDFs and non-ortho coplanar PCBs in blood samples collected from residents in high cancer-causing area close to batch-type municipal solid w aste incinerator in Japan. Organohalogen Compounds. 3 8 :1 4 3 -1 4 6 . M oche, W .; Thanner, G. (1997) Am bient air patterns and concentrations of PCDD/F in the vicin ity of steelw orks in Austria. Organohalogen Compounds. 3 2 :3 8 0 -3 8 3 . National Academ y of Sciences (NAS) (1991) Nutrition during lactation. W ashington, DC: National Academ y Press. DRAFT--DO NOT QUOTE OR CITE 5-47 December 2003 Ohio Environmental Protection Agency (OEPA) (1994) Risk assessment of potential health effects of dioxins and dibenzofurans em itted from the Columbus solid waste authority's reduction fa cility. The Ohio Environmental Protection Agency, Division of Air Pollution Control. February 28, 1994. Ohta, S.; Kuriyama, S.; Nakao, T.; Aozasa, O.; M iyata, H.; Tanahaski, M. (1997) Levels of PCDDs, PCDFs, and non-ortho coplanar PCBs in soil collected from high cancercausing area close to batch-type municipal solid w aste incinerator in Japan. Organohalogen Compounds. 3 2:155 -1 60. Patandin, S.; Dagnelie, P.C.; M ulder, P.G.H.; de Coul, E.O.; van der Veen, J.E.; WeisglasKupersu, N.; Sauer, P.J.J. (1999) Dietary exposure to polychlorinated biphenyls and dioxins from infancy until adulthood: a comparison between breast-feeding, toddler, and long-term exposure. Environmental Health Perspectives 1 07 :45 -5 1. Patandin, S.; Weisglas-Kuperus, N.; de Ridder, M .A .J.; Koopman-Esseboom, C.; van Staveren, W .A ., van der Paauw, C.G.; Sauer, P.J.J. (1997) Plasma polychlorinated biphenyl levels in Dutch preschool children either breast-fed or form ula-fed during infancy. American Journal of Public Health 8 7 :1 7 1 1 -1 7 1 4 . Pinsky, P.F.; Lorber, M.N. (1998) A model to evaluate past exposure to 2,3,7,8-TCDD. Journal of Exposure Analysis and Environmental Epidemiology 8: 187-206. Rappe, C.; Nygren, M .; Lindstrom , G.; Buser, H.R.; Blaser, O.; W uthrich, C. (1987) Polychlorinated dibenzodioxins and dibenzo-p-dioxins and other chlorinated contam inants in cow m ilk from various locations in Switzerland. Environmental Science and Technology. 2 1 :9 6 4 -9 7 0 . Reed, L.W .; Hunt, G.T.; Maisel, B.E.; Hoyt, M .; Keefe, D.; and Hackney, P. (1990) Baseline assessment of PCDDs/PCDFs in the vicin ity of the Elk River, M innesota generating station. Chemosphere. 21:159-171. Reiner, E.J.; MacPherson, K.A.; Kolic, T .M .; Hayton, A .; Clement, R.E. (1995) Characteristic levels of chlorinated dibenzo-p-dioxins and dibenzofurans in fish from Ontario Great Lakes. Organohalogen Compounds 24:379-383. Riss, A. (1993) Impact of PCDD/PCDF em issions from a copper reclam ation plant: five years of experience w ith environm ental m onitoring. Proceedings of Dioxin 93, held in Vienna, A ustria, September, 1993. Organohalogen Compounds. 14:23-26 . Riss, A .; Hagenmaier, H.; W eberruss, U.; Schlatter, C.; W acker, R. (1990) Comparison of PCDD/PCDF levels in soil, grass, c o w 's m ilk, human blood and spruce needles in an area of PCDD/PCDF contam ination through emissions from a metal reclamation plant. Chemosphere. 21:1451-1456. Ryan, A. (1997) The resurgence of breastfeeding in the United States. Pediatrics. 99(4):12-16. DRAFT--DO NOT QUOTE OR CITE 5-48 December 2003 Sandalls, J.; Berryman, B.; Bennett, L.; Newstead, S.; Fox, A. (1998) PCDD/PCDF on land around an industrial site and identification of source. Chemosphere. 37(912):2225-2237. Schecter, A .; Papke, O. (1998) Dioxins, dibenzofurans, and coplanar and marker PCBs in blood of persons residing near a PCB m anufacturing fa cility in the United States. Organohalogen Compounds. 3 8:175 -1 78. Sm ith, A.H. (1987) Infant exposure assessment for breast m ilk dioxins and furans derived from waste incineration emissions. Risk Analysis. 7(3):34 7-353. S tartin, J.R.; W right, C.; Kelly, M .; C harlesw orth, E.A. (1994) Dioxin concentrations in the blood of individuals resident on farms near Bolsover U.K. Organohalogen Compounds. 21:117-120. Svensson, B.G.; Nelsson, A .; Hansson, M .; Rappe, C.; Akesson, B.; Skerfving, S. (1991) Exposure to dioxins and dibenzofurans through the consum ption of fish. New England Journal of Medicine. 3 2 4 (1 ):8 -1 2 . U.S. Departm ent of Agriculture (1995) Food and nutrient intakes by individuals in the United States, 1 day, 1989-91. W ashington, DC: Agricultural Research Service. NFS Report No. 91-2. U.S. Environmental Protection Agency (1992) Respiratory health effects of passive smoking: lung cancer and other disorders. W ashington, DC: O ffice of Research and Development. E P A /600/6-90/006F. U.S. Environmental Protection Agency (1996) Health Assessm ent for 2,3,7,8-TC D D and related com pounds, W ashington, DC: O ffice of Research and Development. EPA/600/BP-92/001. U.S. Environmental Protection Agency (1997) Exposure Factors Handbook. W ashington, DC: O ffice of Research and Development. EPA/600/P-95/002B. U.S. Environmental Protection Agency (1999) Risk assessment guidance for Superfund Volum e I: Human health evaluation manual (Part E, supplem ental guidance for dermal risk assessment) interim guidance, draft. W ashington, D.C. Office of Emergency and Remedial Response. Van der Molen, G .W .; Kooijman, B .A .L.M .; W ittsiepe, J.; Schrey. P.; Flesch-Janys, D.; Slob, W. (2000) Estim ation of dioxin and furan elim ination rates w ith a pharm acokinetic model. Journal of Exposure Analysis and Environmental Epidemiology. 10:579-585. Van Larebeke, N.; Hens, L.; Schepens, P.; Covaci, A .; Baeyens, J.; Everaert, K.; Bernheim, J.L.; V lietinck, R.; DePoorter, G. (2001) The Belgian PCB and dioxin incident of January - June 1999: Exposure data and potential im pact on health. Environmental Health Perspectives. 109: 265-273. DRAFT--DO NOT QUOTE OR CITE 5-49 December 2003 W alker, W .A .; W atkins, J.B ., eds. 1996. N utrition in pediatrics. Basic Science and Clinical Applications. Second Edition. 8 6 4 pp. B.C. Decker Inc., Ham ilton, Ontario. W inters, D.; Cleverly, D.; Meier, K.; Dupuy, A .; Byrne, C.; Deyrup, C.; Ellis, R.; Ferrario, J.; Harless, R.; Leese, W .; Lorber, M .; M cDaniel, D.; Schaum, J.; W a lco tt, J. (1996) A statistical survey of dioxin-like com pounds in United States beef: A progress report. Chemosphere. 32:469-478. W olfe, R.J.; W alker, R.J. (1987) Subsistence econom ics in Alaska: p rodu ctivity, geography and developmental impacts. A rctic Anthropology. 24 (2 ):5 6 -8 1 . DRAFT--DO NOT QUOTE OR CITE 5-50 December 2003 Table 5-1. Concentrations of CDDs, CDFs, and Dioxin-Like PCBs in Blood (lipid based) of a Breast-Fed and a Formula-Fed Infant at the Age of 11 and 25 Months Age (Months) Compound (conc. in pg/g fat) 2,3,7,8-T4CDF 2,3,7,8-T4CDD 1,2,3,7,8-P5CDF 2,3,4,7,8-P5CDF 1,2,3,7,8-P5CDD 1,2,3,4,7,8H6CDF 1,2,3,6,7,8-H6CDF 2,3,4,6,7,8-H6CDF 1,2,3,4,7,8-H6CDD 1,2,3,6,7,8-H6CDD 1,2,3,7,8,9-H6CDD 1,2,3,4,6,7,8-H7CDF 1,2,3,4,6,7,8-H7CDD OCDF OCDD TEQdf-WH098 (<LD = 0.5*LD) PCB 77 PCB 126 Breast-Fed Infant 11 25 < 2 .7 * 3.7 < 1 .2 23.1 11.1 9.8 8.1 < 3 .4 7.8 43.0 7.1 13.1 24.3 < 5 .0 148.7 29.2 < 2 .5 4.1 n.d. (1.4) 29.7 15.2 12.2 10.2 < 3 .0 9.1 51.7 8.1 n.a. 29.7 n.a. 204.0 36.8 23 (m) 287 20 (m) n.a. Formula-Fed Infant 11 25 < 3 .0 < 1.0 < 1 .2 1.5 < 1.0 < 2 .2 < 1.0 < 2 .3 n.d (1.1) 2.5 n.d. (1.2) < 5 .8 8.8 < 5 .0 79.3 2.4 < 2 .5 < 1.0 n.d. (2.5) < 2 .5 n.d. (1.8) < 2 .5 < 2 .5 < 2 .5 n.d. (2.8) < 5 .4 < 4 .5 < 6 .0 <10.0 n.a. 70.0 2.3 26 (m) 24 20 (m) n.a. PCB 169 270 183 7 11 t e q p-w h o 98 31.4 1.8 7 0.1 * for values reported as " < " a value (2.7, e.g.), % the concentration was used for TEQ calculations. n.a. = not analyzed n.d. = not detected (limit of detection) (m) = maximum value, due to possible contribution of a contaminant Source: Abraham et al. (1995). DRAFT--DO NOT QUOTE OR CITE 5-51 December 2003 Table 5-2. Concentrations of CDDs and CDFs in Adipose Tissue (lipid based) of Stillborn, Formula-Fed, and Breast-Fed Infants Compound (conc. in pg/g fat) 2,3,7,8-TCDD 1,2,3,7,8-PCDD 1,2,3,4,7,8-HxCDD 1,2,3,6,7,8-HxCDD 1,2,3,7,8,9-HxCDD 1,2,3,4,6,7,8-HpCDD OCDD 2,3,7,8-TCDF 1,2,3,7,8,-PCDF 2,3,4,7,8-PCDF 1,2,3,4,7,8-HxCDF 1,2,3,6,7,8-HxCDF 2,3,4,6,7,8-HxCDF 1,2,3,7,8,9-HxCDF 1,2,3,4,6,7,8-HpCDF 1,2,3,4,7,8,9-HpCDF OCDF TEQdf-WH098 Stillborn (n = 3) 1.6 3.4 2.5 8.8 1.3 12.9 51.2 1.4 0.2 9.2 3.7 2.4 1.0 0.1 3.6 0.4 2.1 11.9 Formula-Fed (n = 8) 0.4 1.1 1.0 4.0 0.7 5.0 29.1 1.9 1.0 3.1 1.7 1.0 0.2 0.1 1.6 0.1 1.8 4.3 Breast-Fed (n = 9) 1.7 4.9 4.0 19.9 3.7 25.2 91.6 1.1 0.5 10.6 3.5 2.8 1.1 0.1 3.8 0.1 1.6 15.9 Note: Average congener concentrations calculated assuming non-detects equal to % detection limit. Source: Kreuzer et al. (1997). DRAFT--DO NOT QUOTE OR CITE 5-52 December 2003 Table 5-3. Parameters Used for Modeling the Impact of Nursing on Body Burden and Body Lipid Concentrations of TEQs from Infancy to Adulthood Time After Birth A t birth 1 month 2 months 3 months 4 months 5 months 6 months 7 months 8 months 9 months 10 months 11 months 1 year 2 years 5 years 11 years 18 years 34 years 55 years TEQd f p -W H 0 9 s Concentration in Breast Milk Fat (pg/g) 25 22.9 20.8 1S.S 16.7 14.6 12.5 11.5 10.4 9.4 8.3 7.3 6.3 ------- Body Weight (kg) 3.3 4.3 4.6 6.0 6.7 7.4 7.9 8.4 8.8 9.2 9.4 9.8 11.3 13.3 19.7 41.1 65.1 71.5 73.8 Lipid Fraction 0.14 0.16 0.18 0.20 0.22 0.23 0.25 0.25 0.24 0.24 0.23 0.23 0.23 0.20 0.15 0.15 0.13 0.21 0.27 Half-life (yrs) 0.40 0.50 0.60 0.70 0.75 0.80 1.00 1.00 1.05 1.08 1.10 1.12 1.14 1.39 2.12 3.60 5.33 7.70 9.76 Administered Dose of Dioxin TEQDFP- WHO9S, pg/daya Formula Only 50 50 50 50 50 50 50 50 50 50 50 50 50 50 54 65 66 66 66 6-week BF 800 733 667 / 50 50 50 50 50 50 50 50 50 50 50 50 54 65 66 66 66 6 Months BF 800 733 667 600 533 467 400 50 50 50 50 50 50 50 54 65 66 66 66 1 Year BF 800 733 667 600 533 467 400 367 333 300 267 233 200 50 54 65 66 66 66 2 Year BF 800 733 667 600 533 467 400 367 333 300 267 233 200 200 54 65 66 66 66 a Dose = TEQDFP-WHO9S concentration in milk fat (pg/g) x lipid fraction in milk (0.04) x ingestion rate of milk (800 g/day). DRAFT--DO NOT QUOTE OR CITE 5-53 December 2003 Table 5-4. Model Validation Data and Results (all concentrations in ppt TEQDF-W HO98 lipid basis) Description 1st Mother, 1st child 1st Mother, 2nd child 2nd Mother, 1st child 2nd Mother, 2nd child 3rd Mother, only child 4th Mother, only child Average, pg/g TE Q df-W H Q 98 Observed Data Milk TEQdfW H O 98/m o n th 23.5 / 2 13.7 / 2 26.5 / 2 18.3 / 2 13.7 / 2 12.7 / 2 Milk TEQdfW H O 98/m o n th 14.0 / 11 12.7 / 5 15.2 / 11 13.1 / 6 NA 13.0 / 6 Child TEQdfW H O 98/m o n th 34.7 / 11 11.9 / 11 44.2 / 12 18.8 / 12 5.0 / 13 26.5 / 12 23.5 Weeks Breast-Fed 26 29 30 32 7 30 Model Predictions Child TEQdfw h o 98 @ Selected k(t)* 34 27 36 27 10 21 Child TEQdfw h o 98 @ Long k(t) 51 39 56 41 16 32 26 39 * Predicted child TEQ DF-W H O 98 at the time of measurement (between 11 and 13 months) using the more rapid dissipation rate, k(t), selected for this model in comparison to the slower k(t) (longer half-life of 7 years) shown in the last column. DRAFT--DO NOT QUOTE OR CITE 5-54 December 2003 Table 5-5. Results of PK Modeling for Formula Feeding and 4 Breast Feeding Scenarios Description of Model Output Peak concentration, pg TEQDFp-WHO98/g lipid Time after birth of peak AUC1 after 1 year AUC after 10 years AUC after 70 years AUC (bf) / AUC (formula)2 - 1 yr AUC (bf) / AUC (formula) - 10 yr AUC (bf) / AUC (formula) - 70 yr Formula 13.0 6-wk BF 34.1 6-mo BF 44.3 1-yr BF 44.3 2-yr BF 44.3 9 yr 2,168 39,433 27,5419 ____ ____ -- 6 wk 5,989 46,516 282,654 2.8 1.2 1.03 9 wk 12,129 62,696 299,304 5.6 1.6 1.09 9 wk 13,645 73,183 310,210 6.3 1.9 1.13 9 wk 13,645 86,370 324,202 6.3 2.2 1.18 1 AUC = measure of accumulated exposure defined as, "area under the curve", equal to lipidconcentration, ppt TEQDFP-WHO98 * days. For example, a lifetime at 10 ppt TEQDFP-WHO98 lipid would yield an AUC of 10 ppt TEQDFP-WHO98 * 70 years * 365 d/yr = 255,500 ppt TEQDFP-WHO98day. 2 AUC (bf) / AUC (formula) = ratio comparing the accumulated exposure difference between the breast-feeding scenario and the formula. A result of 6.0 means that the accumulated exposure for the breast-feeding scenario being evaluated is 6 times more than a formula only scenario. DRAFT--DO NOT QUOTE OR CITE 5-55 December 2003 Table 5-6. Sensitivity Analysis Testing of PK Model for Breast-Milk Im pacts3 Description of Model Output Peak concentration, pg TEDFp-WHO98/g lipid Time after birth of peak AUC1 (sa) / AUC (6-mo)2 - 1 yr AUC (sa) / AUC (6-mo) - 10 yr AUC (sa) / AUC (6-mo) - 70 yr 6-mo baseline 44.3 Dissipation Tests Sc Sc #1 #2 49.8 44.3 Dose Tests Sc Sc #3 #4 48.1 28.5 Extremes Sc #5 54.0 Sc #6 28.5 9 wk _ ----- 9 wk 1.3 2.2 1.4 9 wk 1.0 0.8 0.7 9 wk 1.2 1.1 1.03 9 wk 0.6 0.8 0.96 10 wk 1.6 2.6 1.5 9 wk 0.6 0.7 0.6 1 AUC = measure of accumulated exposure defined as, "area under the curve", equal to lipidconcentration, ppt TEQDFP-WHO98 * days. For example, a lifetime at 10 ppt TEQDFP-WHO98 lipid would yield an AUC of 10 ppt TEQDFP-WHO98 * 70 years * 365 d/yr = 255,500 ppt TEQDFP-WHO98day. 2 AUC (sa) / AUC (6-mo) = ratio comparing the accumulated exposure difference between the sensitivity analysis scenario and the 6-month breast-feeding baseline scenario. A result of 6.0 means that the accumulated exposure for the sensitivity analysis scenario being evaluated is 6 times more than the 6-month baseline scenario. 3 Scenario definitions: Sc #1: Use of the Pinsky and Lorber (1998) lipid-based function for dissipation rate Sc #2: Use of the Kreuzer, et al. (1997) modeled dissipation rate Sc #3: Assumption of mother's milk concentration not declining from 25 ppt TEQDFP-WHO98 lipid Sc #4: Assumption of mother's milk concentration beginning at 15 ppt TEQDFP-WHO98, declining to 7.5 ppt TEQDFP-WHO98 after 6 months. Sc #5: Use of Pinsky and Lorber (1998) lipid-based function for dissipation rate and the assumption of mother's milk concentration not declining from 25 ppt TEQDFP-WHO98 lipid. Sc #6: Use of Kreuzer, et al. (1997) modeled dissipation rate and the assumption of the mother's milk concentration beginning at 15 ppt TEQDFP-WHO98, declining to 7.5 ppt TEQDFPWHO98 after 6 months. DRAFT--DO NOT QUOTE OR CITE 5-56 December 2003 Table 5-7. Estimated CDD/CDF/PCB Exposures for Adult Subsistence Fishermen Media Soil ingestion Soil dermal contact Freshwater fish ingestion Marine fish ingestion Inhalation Water ingestion Milk ingestion Dairy ingestion Vegetable fat ingestion Conc. TEQDFP-WHO98a Contact Rateb 11.6 ppte 50 mg/day 11.6 ppt 12 g/dayf 2.2 pptg 59 to 170 g/day 0.51 pptg 12.5 g/day 0.12 pg/m3 13.3 m3/day 0.00056 ppq 1.4 L/day 0.027 ppt 175 g/day 0.18 ppt 55 g/day 0.093 ppte 17 g/day Total Daily Intake (pg/kg-day)c 8.2 x 10-3 1.9 x 10-3 1.9 to 5.3 x 10+0 9.1 x 10-2 2.3 x 10-2 1.1 x 10-5 6.8 x 10-2 1.4 x 10-1 2.2 x 10-2 2.2 to 5.7 x 10+0 d a Values from Table 3-64. b Values for adult soil ingestion, inhalation, water ingestion, and subsistence fish ingestion from Exposure Factors Handbook (U.S. EPA, 1997). Contact rates for milk, dairy, and vegetable fats are based on data from USDA (1995). c Daily intake (mg/kg-day) = [Contact rate (g/day; m3/day; L/day; mg/day) x Conc. TEQ x Unit Conversion (soil unit conversion = 10-3, all other media no unit conversion needed)/Body Weight (kg)] or Contact rate (g/kg-day) x Conc. TEQ x Unit Conversion. d Approximately equivalent to 77 to 186 pg/day, assuming an adult body weight of 70 kg. e Calculated by setting nondetects to zero. f Calculated as the surface area of the body that contacts the soil (5,700 cm2/day) x the rate that soil adheres to the skin (0.07 mg/cm2) x the fraction of CDD/CDFs absorbed through the skin (0.03); exposure factors based on recommendations in U.S. EPA (1999) for an adult resident, which assumes that the lower legs, forearms, hands, and head are exposed to the soil. g This concentration is a species-specific ingestion-weighted average value. DRAFT--DO NOT QUOTE OR CITE 5-57 December 2003 Table 5-8 . Levels o f D iffe re n t PCB Congeners in Blood Sam ples fro m Three G roups o f Men w ith D iffe re n t Fish C onsum ption Habits Congener (UIPAC) 77 (pg/g)a 126 (pg/g) 169 (pg/g) 105 (ng/g) 118 (ng/g) 156 (ng/g)d 157 (ng/g)d 180 (ng/g)e None Fish Intake M od e ra te Plasma (n = 9) Lipid (n = 8 ) Plasma (n = 14) Lipid (n = 7 0.04 (0 .0 1 -0 .0 9 ) 0.73 (0.3-1.2) 0.65 (1.3-1.5) 0.02 (0-0.03) 0.12 (0 .0 5 -0 .2 1 ) 0.13 (0 .0 5 -0 .3 4 ) 0.02 (0 .0 1 -0 .0 5 1 15 (3-38) Non-ortho-PCBs 0 .1 b (9.03-0.2) 41b (26-62) 220 (100-450) 1.05 (0.6-2.4) 400b (210-650) 200 (100-340) 0.86 (0.4-1.7) 250 (170-360) Mono-ortho-PCBs 5 (0-13) 0.04 (0 .0 2 -0 .0 7 ) 14b (9-20) 41 (17-92) 0.21 (0 .1 2 -0 .4 3 ) 76 (45-120) 40 (19-68) 0 .1 4 (0 .0 7 -0 .2 8 ) 44 (30-64) 6.6 (2.8-11) 0 .0 2 (0 .0 1 -0 .0 5 ) 7.8 (5.4-11) Di-ortho- and other PCBs 400 1 400 Notes: M eans and ranges in d icated on plasma and lipid basis. a Near the detection lim it. b p < .0 5 , compared w ith group "none." c p < .0 5 , compared w ith group "m oderate." d Q uantified, using single-response factors. e CB-180 quantified from tw o fractions, concentrations thus estimated. Source: A splund et al. (19 9 4 ). High Plasma (n = 14) Lipid (n = 11) 0 .2 b,c (0.1-0.5) 2 .8 bc (1.2-4.9) 1 .8 0 bc (0.3-3.6) 50b (15-140) 7 9 0 bc (380-1400) 5 7 0 bc (210-1200) 0 .1 4 bc (0 .0 4 -0 .3 ) 0 .5 8 b,c (0 .2 1 -1 .0 0 ) 0 .3 b,c (0 .0 5 -0 .7 ) 0 .0 6 bc (0 .0 1 -0 .1 4 ) 3 9 bc (18-77) 1 6 0 bc (84-300) 9 0 b,c (36-180) 1 8 bc (7.4-39) 2 600 DRAFT--DO NOT QUOTE OR CITE 5-58 December 2003 Table 5-9. Mean TEQ Levels in Pooled Serum Samples Cornwall Sports Fishers < 38 years, lower higher 38 years, lower higher > 50 years, higher Nonfish Eaters < 38 years 38-50 years > 50 years Mississauga Sports Fishers < 38 years 38-50 years > 50 years Nonfish Eaters < 38 years 38-50 years > 50 years Source: Adapted from Cole et al. (1995). I-TEQdf (ppt, lipid basis) t e q p-w h o 94 (ppt, lipid basis) 20.8 22.2 28.4 31.4 33.5 24.7 29.8 36.8 32.4 40.1 41.2 34.0 29.1 34.3 -3.6 3.1 9.5 17.3 2.6 6.8 9.7 ---- ---- DRAFT--DO NOT QUOTE OR CITE 5-59 December 2003 Table 5-10. Mean CDD/CDF Levels in Serum of Consumers of Great Lakes Sport Fish (ppt, lipid adjusted) 2,3,7,8-TCDD 1,2,3,7,8-PeCDD 1,2,3,4,7,8-HxCDD 1,2,3,6,7,8-HxCDD 1,2,3,7,8,9-HxCDD 1,2,3,4,6,7,8-HpCDD 1,2,3,4,6,7,9-HpCDD OCDD Total Total I-TEQd 2,3,7,8-TeCDF 1,2,3,7,8-PeCDF 2,3,4,7,8-PeCDF 1,2,3,4,7,8-HxCDF 1,2,3,6,7,8-HxCDF 1,2,3,7,8,9-HxCDF 2,3,4,6,7,8-HxCDF 1,2,3,4,6,7,8-HpCDF 1,2,3,4,7,8,9-HpCDF OCDF Total Total I-TEQf All Sport Fish Consumer Subjects (ppt) (n = 31)a 5.6 10.4 8.4 126 7.0 134 777 1,062 27.5 2.2 2.0 17.7 12.7 9.0 ND 5.1 20.0 ND 58.2 11.9 Lake Michigan Participants (ppt) (n = 9) Lake Huron Participants (ppt) (n-11) CDD Congeners 4.7 10.5 9.8 16 11.4 8.4 120 142 8.7 5.5 144 153 ND ND 783 918 1,087 1,258 25.8 36 CDF Congeners 2.4 2.1 ND 1.7 20.4 22.8 11.6 16.0 8.0 10.5 ND ND 6.0 4.8 22.1 22.9 ND ND ND 70.8 79.3 13.2 14.8 Lake Erie Participants (ppt) (n = 11) 4.9 5.8 5.5 115 5.8 95.9 623 844 20.7 ND 10.4 10.2 7.7 ND 8.0 15.2 ND ND 48.3 7.8 Comparison Group3 (ppt) (n = 70) 2.8 5.5 9.0 70.8 8.4 124 4.4 971 1,188 15.5 2.1 1.6 5.5 8.0 5.3 1.8 3.8 21.3 NA 6.9 87.3 4.9 a One individual was excluded from the data summary due to unusually high occupational/environmental exposures. b Comparison group is from a 1991 unpublished NCEH/CDC data set of a Jacksonville, Arkansas, population of 70 individuals. Source: Anderson et al. (1998). DRAFT--DO NOT QUOTE OR CITE 5-60 December 2003 Table 5-11. Mean PCB Levels in Serum of Consumers of Great Lakes Sport Fish (ppt, lipid adjusted) 77 81 126 169 Coplanar PCB T otal I-TEQP-W H O 94 28 52 56 58 74 99 101 1.5 118 130 138 146 153 156 157 187 170 172 177 178 180 183 187 189 193 194 195 201 2 0 3 /1 9 6 206 209 Total A ll S p o rt Fish Consumer Subjects8 (ppt) (n = 31) 14.6 13.5 148 80.8 228 17.4 0.08 0.01 0.02 0.04 0.3 0.4 ND 0.1 0.4 0.1 0.8 0.2 1.1 0.02 0.1 0.03 0.1 0.02 0.04 0.07 0.4 0.1 0.3 ND 0.03 0.09 0.04 0.2 0.09 0.07 0.02 5.2 Lake Michigan Participants (ppt) ( n = 9) Lake Huron Participants (ppt) (n-11) Coplanar PCB Congeners 16.5 14.2 17.4 13.2 261 187 113 84.2 340 282 26 23 Congener-Specific PCBs 0.08 0.1 ND 0.01 0.06 ND 0.08 0.04 0.6 0.4 0.7 0.5 0.01 ND 0.2 0.1 0.8 0.5 0.2 0.1 1.3 0.8 0.3 0.2 1.7 1.1 0.04 0.02 0.2 0.1 0.07 0.03 0.2 0.2 0.05 0.03 0.09 0.06 0.13 0.08 0.8 0.4 0.2 0.1 0.4 0.3 ND ND 0.08 0.02 0.1 0.1 0.07 0.06 0.3 0.2 0.2 0.11 0.08 0.08 0.03 0.04 8.6 5.7 Lake Erie Participants (ppt) (n = 11) 13.3 28 48.4 75.4 4.8 0.08 ND ND 0.01 0.2 0.1 ND 0.02 0.08 0.02 0.4 0.04 0.6 ND 0.08 ND 0.07 ND ND 0.03 0.2 0.03 0.08 ND ND 0.05 ND 0.09 0.03 0.04 ND 2.2 C o m pa riso n Groupb (ppt) (n = 70) 12.6 8.6 18.4 17.9 57.4 1.8 ND ND ND ND 0 .0 0 9 ND ND 0.4 0.03 NA 0.4 ND 0.4 NA NA ND ND ND ND ND 0.4 ND 0.04 NA NA 0 .0 0 4 ND 0.04 0 .0 0 7 ND NA 1.2 a One individual w as excluded from the data sum m ary due to unusually high occupational/environm ental exposures. b Com parison group fro m a 1 9 9 6 unpublished data set o f 41 non-G reat Lake sp o rt fish consum ers analyzed by the W isconsin State Laboratory of Hygiene. Source: Anderson et al. (1 9 9 8 ). DRAFT--DO NOT QUOTE OR CITE 5-61 December 2003 Table 5.12. Average PCB and CDD/F TEQ-WHO98 Concentrations (all concentrations in pg/g whole weight) Main Water Body Lake Superior Lake Huron Lake Erie Niagara River Lake Ontario Northern Areas Location Black Bay Jackfish Bay Peninsula Harbour Algoma Area Algoma-Agawa Bay Goulais Bay Manitoulin Island Nottawassaga River Nottawasaga River Tobermory Oliphant/Fishing Is. Oliphant/Fishing Is. Bruce Cty Grand Bend Western Basin Niagara River Bar Niagara River Bar Niagara River Bar Niagara River Niagara River Niagara River Welland River Hamilton Harbour Hamilton Harbour Bronte Creek Port Credit Credit River Credit River Don River W hitby Harbour Whitby/Pickering Cobourg Trent River Trent River - 4 Upper Bay of Quinte L. Bay of Quinte Cataraqui River Mattagami River Cochrane Species Lake Trout (5) Lake Trout (5) Lake Trout (5) Lake Trout (5) L. W hitefish (5) L. W hitefish (5) L. W hitefish (5) Chinook (5) Rainbow Trout (5) L. Trout (5) L. Trout (5) L. W hitefish (5) Carp (5) L. Trout (5) Ch. Catfish (5) Lake Trout (5) Chinook (5) Brown Trout (5) White Perch (5) Rainbow Trout (5) White Bass (4) Carp (5) Carp (5) Ch. Catfish (5) Brown Trout (5) Lake Trout (4) Brown Trout (5) Chinook (5) White Sucker (5) Carp (5) Chinook (5) L. Trout (5) Chinook (5) L. W hitefish (5) W hitefish (5) L. Trout (5) Carp (5) White Sucker (5) TEQp W H O g s 9.4 13 9.5 83 4.5 4.1 1.4 14 6.9 21 12 4.4 31 22 55 79 52 21 6.2 22 1.6 11 25 58 56 72 41 39 16 27 29 67 182 52 8.7 110 57 21 White Sucker (5) 21 TEQ WHO98 0.68 5.3 2.8 23 5.4 1.8 0 0.94 0 3.5 0.75 2.7 2.1 0.76 8.1 22 12 4.6 3.2 4.8 0 0.13 1.9 6.7 8.3 19 10 8.7 4.4 40 6.8 26 59 5.8 7.0 29 4.7 0 0 TEQD p F -W H O 9 8 14 2.5 3.3 3.7 0.83 2.3 > 1 .4 15 > 6 .9 5.9 17 1.7 14 29 6.8 3.5 4.4 4.5 1.9 4.6 > 1 .6 86 13 8.7 6.7 3.8 4.1 4.5 3.6 0.66 4.3 2.6 3.1 9.0 1.2 3.8 12 >21 >21 Source: Kolic et al. (2000). DRAFT--DO NOT QUOTE OR CITE 5-62 December 2003 Table 5-13. Comparison Between Mean PCB Levels in Fish-eating Populations and Controls PCBs 126 169 118 170 180 TOTAL TEQP-WHO94 Fishermen Mean Concentration (ppt, lipid basis) TEQp-WHO94 (ppt, lipid basis) 1540 1010 568 539 1776 -- 154 10.1 56.8 53.9 17.76 292.6 Controls Mean Concentration (ppt, lipid basis) TEQp-WHO94 (ppt, lipid basis) 48 29 25.4 27.7 48.2 -- 4.8 0.29 2.54 2.77 0.48 10.9 Source: Adapted from Dewailly et al. (1994). DRAFT--DO NOT QUOTE OR CITE 5-63 December 2003 DRAFT--DO NOT QUOTE OR CITE 5-64 December 2003 Half-life, yrs --------Selected half-life ---------Lipid-based half-life, Lorber & Pinsky (1998) --------Modeled half-life, Kreuzer, et al (1997) Figure 5-2. Comparison of the selected half-life of TEQs in the body with two options that were available in the literature for 2,3,7,8-TCDD. DRAFT--DO NOT QUOTE OR CITE 5-65 December 2003 (A) (B) Figure 5-3. Demonstration of the model for evaluating impacts on lipid concentrations (A) and body burdens (B) of infants resulting from various nursing scenarios during a lifetime. DRAFT--DO NOT QUOTE OR CITE 5-66 December 2003 (A) (B) Figure 5-4. Demonstration of the model for evaluating impacts on lipid concentrations (A) and body burdens (B) of infants resulting from various nursing scenarios during the first 10 years of life. DRAFT--DO NOT QUOTE OR CITE 5-67 December 2003 (A) (B) Figure 5-5. Results of sensitivity analysis showing the difference when making modeling assumptions that lead to a high impact to the infant (high impact scenario) and to a low impact to the infant (low impact scenario) as compared to the baseline scenario for a 6-month breast-feeding scenario (6-month scenario). DRAFT--DO NOT QUOTE OR CITE 5-68 December 2003