ToxSci Advance Access originally published online on August 6, 2007
Toxicological Sciences 2007 100(1):224-237; doi:10.1093/toxsci/kfm199
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
An Adaptable Internal Dose Model for Risk Assessment of Dietary and Soil Dioxin Exposures in Young Children
* Health Science Resource Integration, Tallahassee, Florida 32309
Private Consultant, Philadelphia, Pennsylvania 19103
ChemRisk, Pittsburgh, Pennsylvania 15222
ChemRisk, San Francisco, California 94105
1 To whom correspondence should be addressed at 2976 Wellington Circle West, Tallahassee, FL 32309. Fax: (850) 906-9777. E-mail: brentkerger{at}att.net.
Received March 8, 2007; accepted July 25, 2007
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMODEL DESCRIPTIONMODEL APPLICATIONMODEL VALIDATIONDISCUSSIONFUNDINGSUPPLEMENTARY DATAREFERENCES |
|---|
An adaptable model is presented for assessing the blood lipid concentrations of polychlorodibenzodioxins and polychlorodibenzofurans (PCDD/Fs) from dietary (breast milk, formula, milk, and other foods) and soil pathway exposures (soil ingestion and dermal contact) utilizing age-specific exposure and intake estimates for young children. The approach includes a simple one-compartment (adipose volume) toxicokinetic model that incorporates empirical data on age-dependent half-lives and bioavailability of PCDD/F congeners, child body size and intake rates, and recent data on breast milk and food dioxin levels. Users can enter site-specific soil concentration data on 2,3,7,8-chlorinated PCDD/F congeners for specific assessment of body burden changes from soil pathways in combination with background dietary exposures from birth through age 7 years. The model produces a profile of the estimated PCDD/F concentration in blood lipid (in World Health Organization 1998 dioxin toxic equivalents) versus time for a child from birth through age 7 years. The peak and time-weighted average (TWA) internal dose (defined as blood lipid dioxin toxic equivalents) for a variety of specific child exposure assumptions can then be compared to safe internal dose benchmarks for risk assessment purposes, similar to an approach taken by United States Environmental Protection Agency for assessing child lead exposures. We conclude that this adaptable toxicokinetic model can provide a more comprehensive assessment of potential health risks of PCDD/Fs to children because it integrates recent empirical findings on PCDD/F kinetics in humans and allows users to assess contributions from varied dietary and site-specific environmental exposure assumptions.
Key Words: dioxins; furans; childhood; diet; soil; toxicokinetic model.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMODEL DESCRIPTIONMODEL APPLICATIONMODEL VALIDATIONDISCUSSIONFUNDINGSUPPLEMENTARY DATAREFERENCES |
|---|
Exposures to infants and young children have been a prominent concern for the human health risk assessment of dioxins (Greene et al., 2003
; USEPA, 2003). These highly lipophilic and persistent compounds are transferred from mother to infant during breast-feeding and are known to be present in the fats of dairy products, meats, and other foods commonly consumed by young children (ATSDR, 1998). The higher fat intake per body weight in younger children (e.g., ages 0–2 years) reinforces concerns that they receive considerably higher intake of dioxins per body weight compared to adults. Moreover, the greater hand-to-mouth contact rates among younger children may lead to intake of dioxins present in environmental sources such as contaminated soils (Greene et al., 2003; Paustenbach et al., 2006; USEPA, 2002a). Accordingly, the combined impact of these higher potential exposures to dietary and environmental sources in young children and the associated health risks must be properly characterized and understood.
It is commonly assumed that children may be more sensitive to potential toxic effects of chemicals in general based on conservative choices for childhood exposure parameters in risk assessment (USEPA, 1989; Williams et al., 2003) and, especially, to chemicals with potential neurodevelopmental effects like lead and dioxins that are known to bioaccumulate in the human body (Greene et al., 2003; USEPA, 2002b, 2003). Bioaccumulation leads to potential hazards that are not well predicted by daily dose, but rather must be examined with respect to body burden or target organ/tissue concentration accumulated over time.
Fortunately, the general population body burdens of dioxins in developed countries have decreased over the past two decades (Aylward and Hays, 2002; Hays and Aylward, 2003; Lorber, 2002). Indeed, general population body burden estimates for polychlorodibenzodioxins and polychlorodibenzofurans (PCDD/Fs), expressed in World Health Organization (WHO) toxic equivalents (TEQs), have been lowered more than an order of magnitude in the last 20 years. This downward trend is associated with the recognition of dioxin hazards in the 1970's and the extensive interventions to limit emissions and exposures through research that characterized the prominent sources and through regulatory interventions and application of risk assessment principles (USEPA, 2003). This trend is also important to consider in the assessment of total exposures and health risks of dioxins in young children since the body burdens present in reproductive-age women can create the baseline dioxin body burden in their offspring via in utero transfer and breast-feeding (Richter et al., 2006). Further, it is expected that background dietary exposures to dioxins have been lowered concurrently with the observed decrease in general population body burdens.
Measurements of adipose and liver concentrations of 2,3,7,8-tetrachlorodibenzodioxin (TCDD) in infants have suggested that the relatively high daily TCDD doses associated with breast-feeding may not lead to the high body burdens predicted by toxicokinetic models using a constant half-life for all ages (Kreuzer et al., 1997). Using a simple model that accommodated an age-dependent adipose tissue volume, Kreuzer et al. (1997) estimated that the half-life for TCDD in infants was only about 0.4 years, much shorter than the 7–11 years estimated for adults. Lorber and Phillips (2002) reported a modeling exercise of PCDD/F body burdens during breast-feeding of infants using a similar one-compartment model and including TCDD TEQ as TCDD. Leung et al. (2006) recently verified the half-life of TCDD in infants and identified similar short half-lives for other PCDD/F congeners (e.g., 0.2–0.5 years). The reasons for the shorter half-lives in young children are not fully understood, but might be due to a combination of factors, including the effect of dilution from the rapid growth of the adipose mass, higher fecal lipid excretion relative to adults, and increased metabolism (Kreuzer et al., 1997). Recent observations that children and adolescents (under age 18 years) reduced their total TCDD burdens more readily than adults exposed during the Seveso incident, beyond those accounted for by growth-related dilution, suggest that excretion probably plays an important role (Kerger et al., 2006). Our preliminary model (paustenbach et al., 2004) adapted the Kreuzer et al. (1997) model for TCDD to examine child body burdens of all relevant PCDD/Fs, and some of the limitations and needed refinements of these earlier models were analyzed by Kerger et al. (2007). Like the Integrated Environmental Uptake Biokinetic Model (USEPA, 2002b) developed for evaluating childhood lead exposures, the age-dependent half-life model for childhood dioxin body burdens provides more accurate predictions of both peak and long-term average blood lipid concentrations which are almost certainly the best dose metric for predicting the adverse effects of the PCDD/Fs.
The purpose of this study is to describe an adaptable internal dose model for evaluating dietary and soil exposures to PCDD/Fs in young children. Internal dose is defined here as being equivalent to the blood lipid concentration of dioxin TEQ. The model algorithms, data sets, parameters, and assumptions are fully described, and the implications of the model outputs and further refinements and applications are explored.
|
MODEL DESCRIPTION |
|---|
|
TOPABSTRACTINTRODUCTIONMODEL DESCRIPTIONMODEL APPLICATIONMODEL VALIDATIONDISCUSSIONFUNDINGSUPPLEMENTARY DATAREFERENCES |
|---|
Kerger et al. (2007)
described our preliminary age-dependent half-life model for all 2,3,7,8-chlorinated PCDD/Fs based on a modification of the toxicokinetic model for TCDD by Kreuzer et al. (1997). Our preliminary model developed in 2004–2005 (Kerger et al., 2007) used a similar framework to the current model but it lacked flexibility and much of the detailed assessment of empirical congener-specific pharmacokinetic parameters, monthly intake, and dose estimates corresponding to body size parameters for ages 0–2 and congener-specific relative bioavailability and doses from soil ingestion and dermal contact based on the probabilistic assessment of Paustenbach et al. (2006). In the basic one-compartment model used here and by Kreuzer et al. (1997), Lorber (2002), and others, the distribution of dioxins in humans is assumed to be exclusively in the volume of total body fat. This volume expands substantially during childhood and tends to reduce the effective half-life of dioxins at early ages. An equilibrium is proposed to exist between ingested/excreted dietary fats and the body burden of dioxins, particularly for infants and young children. Since breast-fed infants take in more dietary fat, and a greater proportion of dietary fat is excreted in the feces of these infants, the net body burden in infants may be attenuated compared to adults who do not experience the rapid growth rate and fat excretion seen in young children.
The model is comprised of four basic modules that characterize all the data inputs, assumptions, and algorithms for calculating two outputs: (1) the temporal pattern of blood lipid TEQ concentrations from birth through age 7 years and (2) the peak and TWA TEQ concentration. To simplify the calculations, congeners with the same WHO TEF value were combined (i.e., all 2,3,7,8-hexaCDD/Fs, all 2,3,7,8-heptaCDD/Fs, and OCDD/F). The two outputs can be incorporated into risk assessments and compared to appropriate "safe" internal dose benchmarks to be determined by scientific consensus as was the case for the Integrated Environmental Uptake Biokinetic Model for lead (USEPA, 2002b).
Module 1: Child Body Size Parameters
Child body weight and body fat mass from birth through age 7 years are two necessary parameters for this toxicokinetic model. Recently updated 50th percentile body weights and body mass index (BMI) for children in the United States are utilized in the model (CDC, 2000). Body fat percentages are estimated based on BMI by applying the age- and sex-specific equations derived by Deurenberg et al. (1991). Data for male children are used in the model due to the leaner body fat percentage in males compared to females (CDC, 2000; ICRP, 1975); leaner body fat percentage corresponds mathematically to higher lipid PCDD/F concentrations in the model and hence is more conservative. Figure 1 illustrates the rapid rise in BMI and body fat percentage in male children during the first 7 months of life, followed by a plateau between ages 8 and 14 months and a decline to a leaner state that continues through about age 60 months (5 years). Early childhood obesity could have competing effects of both diluting dioxin body burden (by increasing volume of distribution in fat) and lengthening elimination half-life (by sequestering dioxins in adipose), but the current model did not attempt to evaluate this issue.
|
Body fat mass was derived by multiplying the estimated age-specific male body fat percentage (Deurenberg et al., 1991
) times the 50th percentile age-specific male child body weight (CDC, 2000). Figure 2 illustrates the pattern of body fat mass (i.e., the estimated volume of distribution for PCDD/Fs) for an average male child in the United States through age 84 months (7 years). Consistent with the rapid increase in body fat percentage during the first year of life, body fat mass shows a steep rise during this age span, followed by a transition plateau between ages 1 and 2, and a more gradual and relatively linear rate of increase from about age 2.5–7 years (Fig. 2). These age-dependent body fat mass parameters, based on recent data (CDC, 2000), have important implications for the toxicokinetic model predictions.
|
The model default estimates for central tendency body size parameters for ages 0–2 years (monthly) and 2–7 years (at 6-month intervals) are tabulated in the Supplementary Table A-1.
Module 2: Child Dietary PCDD/F Intake Parameters
Average daily intake rates for breast milk, formula, and other dietary items commonly ingested by children ages 0–7 years are integrated into the model. The model adopts an approach that integrates empirical measurements of average daily intake of breast milk (Dewey and Lonnerdal, 1983; Neville et al., 1988) expressed as grams of milk fat per kilogram body weight per day. A similar approach is applied to data reported for baby formula intake (Koletzko et al., 2000). Figure 3 illustrates the pattern of fat intake from breast milk and formula during the first year of life when expressed as total intake (grams fat per day) and body weight–adjusted intake or dose (grams fat per kilogram body weight per day). Note that the degree of variability in estimated intake in fat grams per day during this rapid growth period is attenuated by normalizing the intake to age-specific body weight using the metric of fat grams per kilogram body weight per day (Fig. 3).
|
The model also takes into account the nutritional recommendation by international experts on pediatric nutrition that about 50% of infant calorie intake should be from fats (Aggett et al., 1991
, 1994; European Union, 1996; Koletzko et al., 2000; WHO/FAO, 1994). The model assumes that infants ages 0–1 year require 110 kcal/kg body weight for normal growth, based on the range of 100–120 kcal/kg reported by Tigges (1997). The age-specific requirement for calories from fats is estimated by multiplying 110 kcal/kg times 50th percentile male body weight, and the amount of calories from breast milk or formula fats is approximated at 8.99 kcal/g fat (Tigges, 1997), assuming 3.5% breast milk fat content based on measurements by Butte et al. (1984). Figure 4 illustrates that full-time breast-feeding provides the recommended amount of dietary fat intake to satisfy the child's caloric needs through age 3 months, then additional sources of dietary fats are needed. This is in part due to reductions in breast milk fat content and milk fat intake per body weight as full-time breast-feeding continues (Dewey and Lonnerdal, 1983). In contrast, Figure 5 illustrates that fat intake rates (from vegetable fats) reported by Koletzko et al. (2000) for commercial baby formulas are more consistent over the first year of life. However, baby formulas contain a lower percentage of fats compared to breast milk and therefore technically require supplementation even at age 1 month in order to meet the recommended fat calorie intakes for children. The underlying data and calculation methods for Figures 4 and 5 are presented in the Supplementary Tables A-2 and A-3.
|
|
The model integrates as a default assumption that 50% of the infant caloric intake during the first year of life occurs as breast milk intake and/or vegetable fats from baby formula or other fat sources other than animal fats. Current nutritional recommendations indicate that fats from cow's milk and other animal fats are not recommended during the first year of life (Koletzko et al., 2000)
. Children ages 6–12 months are reported to have a relatively high per capita intake rate of 111 grams per kilogram body weight per day for dairy products (USEPA, 2002a), but such intake from infant formulas is likely to represent skim milk if typical label directions are followed.
The default PCDD/F concentrations in breast milk fat are the mean values reported by Wittsiepe et al. (2004), which correspond to a relatively large population sample taken between the years 2000 and 2003 in urban Germany (Supplementary Table A-4). Little data are available concerning PCDD/Fs in commercial baby formulas; however, fats added to baby formulas are those of vegetable origin which contain far lower PCDD/F content compared to breast milk, cow's milk, or other animal fats (Tigges, 1997). The default PCDD/F concentrations for formula fats (Supplementary Table A-5) are based on the 0.21 pg TEQ/g weighted mean estimate of JECFA (2001) for all fats and oils, and due to lack of published congener profile data on baby formula fats, this total TEQ estimate was assumed to follow the congener pattern for the mean value trends in breast milk reported by Wittsiepe et al. (2004). These default assumptions for PCDD/Fs are consistent with the limited data on baby formulas reported in the preliminary National Health and Nutrition Examination Survey (NHANES) data set (CDC, 2005) and by Schecter et al. (2002) and with estimates for child intake of PCDD/Fs in fats and oils based on the Total Diet Study results from 2001 to 2002 (USFDA, 2004).
Table 1 presents the congener-specific TEQ intake rates for the first 12 months of life under assumptions of either full-time breast-feeding or full-time formula-feeding. These estimates represent the relevant source intake rates for fats (grams fat per kg body weight per day) times the default PCDD/F measurements expressed as TEQ (pg TEQ/g fat), which equals the estimated daily dose in pg TEQ per kilogram body weight per day. For the first year of life, the model conservatively assumes 100% bioavailability of dioxins present in breast milk and formula fats.
|
For children between ages 1 and 7 years, the per capita dietary intake rates for the eight standard food categories reported for children (USEPA, 2002a
) are utilized as default data for the model. The 50th percentile age-specific intake rates are estimated for fruits, vegetables, grains, meats, eggs, dairy, fish, and fats or oils as summarized by Williams et al. (2003) and USEPA (2002a). The estimated PCDD/F TEQ intakes for North America as estimated by JECFA (2001) are utilized, with total TEQ distributed across congener TEQ groups based on the patterns presented by the USEPA (2003) for meat, poultry, eggs, dairy products, fats and oils, and fish. Table 2 presents the absorbed daily dose estimates in pg TEQ per kilogram body weight per day derived by multiplying the food category intake rates times the corresponding TEQ content and applying a bioavailability assumption of 50%. This bioavailability is based on the mass balance study of PCDD/F absorption in human volunteers by Schlummer et al. (1998), who showed that net absorption in younger individuals (ages 24–36 years) ranged from 42 to 62% for TCDD (average 49.3%). Schlummer et al. (1998) showed similar net absorption trends for the pentachlorinated PCDD/Fs that comprise most of the background TEQ in blood lipid and lower net absorption for hepta-and octa-chlorinated PCDD/Fs.
|
Module 3: Child Soil Pathway Intake Parameters
The model is designed to allow users to enter their own site-specific PCDD/F congener TEQ data that represent average and/or upper bound soil concentrations applicable to child resident exposures. Environmental PCDD/F intake parameters in the model are based on the detailed literature review and probabilistic risk analysis regarding PCDD/F intakes from urban residential scenario recently reported by Paustenbach et al. (2006)
and are summarized in the Supplementary Table A-6. The probabilistic analysis provided 95th percentile values for PCDD/F intake via incidental soil ingestion and dermal contact as represented by age-specific dose constants that approximate absorbed doses after being multiplied by the assumed soil TEQ concentration and a relative bioavailability factor according to Equation 1 and is subsequently summed across all congener TEQ groups:
 | (1) |
, and RBsoil is congener-specific relative bioaccessibility compared to the baseline assumption of 50% bioavailability for TCDD (a factor of 1.0 for all tetra and penta congeners, 0.6 for HxCDD/Fs, 0.4 for HpCDD/Fs, and 0.3 for OCDD/F). The relative bioavailability factors (relative to TCDD) adjust for the known tendency of the higher chlorinated PCDD/Fs to adhere more tightly to soils and to be more poorly absorbed by oral and dermal routes relative to TCDD (Birnbaum and Couture, 1988; Norback et al., 1975; Ruby et al., 2002; reviewed in Paustenbach et al., 2006). These relative bioavailability factors are listed in the Supplementary Table A-6.
Three hypothetical data sets for soil PCDD/Fs used for model illustration are contained in the Supplementary Table A-7. The first data set contains a soil PCDD/F pattern consistent with contamination from a 2,4,5-T manufacturing site (Wenning et al., 1993) (proportions by congener) dominated by tetra-, penta-, and hexa-CDD/F TEQ concentrations and having a total TEQ of 100 ppt. The second soil data set contains a higher magnitude of total TEQ concentration (1000 ppt) and a PCDD/F pattern consistent with municipal soil waste incineration that is dominated by penta- to hepta-CDD/F TEQ concentrations (Wenning et al., 1993). The third soil data set contains a total TEQ concentration of 1000 ppt TEQ that is dominated by the hepta- and octa-chlorinated congeners which exhibit considerably lower bioavailability compared to the lower chlorinated PCDD/Fs. This pattern is consistent with PCDD/F contamination of pentachlorophenol (Wenning et al., 1993).
The age-specific dose constant based on the 95th percentile of exposure from the probabilistic risk assessment was estimated using the exposure parameter distributions presented in the Supplementary Table A-6 and the methods for estimating soil ingestion and dermal contact for an urban residential scenario as presented in Paustenbach et al. (2006).
The model calculates the internal dose (i.e., the absorbed blood lipid TEQ concentration) by multiplying the 95th percentile soil exposure factor times the congener-specific soil TEQ concentration and the relative bioavailability factor. The absorbed doses estimated for each of the three example soil data sets are summarized in Table 3.
|
Module 4: Child Toxicokinetic Parameters and Cumulative Internal Dose Algorithms
Table 4 provides a summary of the default toxicokinetic parameters not described above that are integrated into the model. These include the estimated average congener-specific lipid TEQ concentration at birth, empirical estimates of congener-specific elimination rates during age 0–1 year, and the empirical rate of increase in congener half-life through age 7 years. Table 4 also provides the calculated elimination rate constants that are integrated into the model for each 12-month period through age 7 years.
|
An estimate of average blood lipid TEQ for each congener at birth is required in order to properly cumulate the PCDD/F internal doses in the model. Three reports examined either cord blood lipid TEQ concentrations at birth (Abraham et al., 1998
; Schecter et al., 1998) or adipose TEQ concentrations from stillborn infants (Kreuzer et al., 1997), of which the highest average was that of Kreuzer et al. (1997). The average congener-specific TEQ concentrations across the three stillborn infants and an infant that died 3 days after birth were adjusted (multiplied by 0.4) to account for the reduction in background adult internal doses for reproductive-age females between the collection time of the Kreuzer infant data (1991–1992) and the mid-2000's. This correction factor represents the reduction in average breast milk TEQ for German women sampled in 1986–1991 (Furst et al., 1994) and in 2000–2003 (Wittsiepe et al., 2004). Table 4 provides the mean lipid TEQ estimates by congener that serve as the initial internal dose at birth in the modeled infant.
Half-life values for various ages were estimated based on measurements of PCDD/F half-lives reported for human infants (Kreuzer et al., 1997; Leung et al., 2006) and for young children and adolescents (Kerger et al., 2006; Leung et al., 2005). These empirical data indicate that the half-lives of PCDD/Fs increase at a rate of about 0.12–0.18 years for every year of advancing age (Kerger et al., 2006; Leung et al., 2005). The congener-specific rate of increase in half-life with age was combined with the estimated PCDD/F congener half-lives determined in infants during the first 12 months of life (Leung et al. 2006) to provide annual average elimination rate estimates (expressed as first-order rate in days–1) for each of the years examined in the model (Table 4).
The model algorithms are designed to calculate age-specific PCDD/F TEQ internal doses by incrementally adding the absorbed PCDD/F TEQ doses from each substantial source (dietary and environmental) to the existing internal dose, subtracting the amount eliminated, and adjusting for growth-related expansion of adipose volume. Inhalation exposures to PCDD/Fs in vapor or house dust forms are assumed to be negligible compared to dietary and soil ingestion/dermal contact routes (Paustenbach et al., 2006). In addition, we did not include dioxin-like Polychlorinated biphenyls in our assessment due to a paucity of pharmacokinetic data in children.
The age-specific internal dose (i.e., cumulative blood lipid TEQ concentration) across all congeners at time t (months) through the age of 84 months can be calculated from:
 | (2) |
At t = 0, Ci,0 = concentration of the specific congener at birth, Ki,t is the first-order elimination rate constant of the specific congener at the specific age (per day), Xi,t is the TEQ intake of the specific congener from breast milk at the specific age which ranges from 0 to 12 months (pg/kg/day), Yi,t is the TEQ intake of the specific congener from formula at the specific age which ranges from 0 to 12 months (pg/kg/day), Si,t is the TEQ intake of the specific congener from soil (oral and dermal) at the specific age which ranges from 7 to 84 months (pg/kg/day), Di,t is the TEQ intake of the specific congener from food at the specific age which ranges from 13 to 84 months (pg/kg/day), Wi,t is the body weight of child at the specific age which ranges from 0 to 84 months (kg), and Fi,t is the fat mass of child at the specific age which ranges from 0 to 84 months (g).
|
MODEL APPLICATION |
|---|
|
TOPABSTRACTINTRODUCTIONMODEL DESCRIPTIONMODEL APPLICATIONMODEL VALIDATIONDISCUSSIONFUNDINGSUPPLEMENTARY DATAREFERENCES |
|---|
Figure 6 illustrates the model output for varied breast-feeding durations, ranging from 0 (full-time formula-feeding) to 12 months of full-time breast-feeding for children subsequently exposed to environmental soil TEQ concentrations of 100 ppt (data set #1). For breast-feeding durations of 3, 6, and 12 months, the peak internal dose occurs within the first 4 months and ranges from about 35 ppt TEQ (3 months duration) to 40 ppt TEQ (6 or 12 months). A relative plateau in internal dose is achieved after about 24 months of age for all three breast-feeding durations, although a continued gradual rate of depuration is apparent between 24 and 84 months under the 12 months full-time breast-feeding assumption. The TWA internal dose from birth through age 7 years (84 months) increases at a linear rate of 0.5 ppt TEQ per month of full-time breast-feeding in this scenario.
|
Figure 6 also shows that the modeled internal doses among full-time formula-fed infants are lowered during the rapid growth phase of the first 12 months of life and, subsequently, begin accumulating higher TEQ internal doses from background dietary sources (primarily dairy) in the second 12 months of life. A slower rate of accumulation occurs after 24 months of age, reaching an approximate plateau in TEQ internal dose after about 60 months of age. The TWA internal dose for strictly formula-fed infants is about fourfold lower than that for 12 months of full-time breast-feeding, differing by about 7 ppt TEQ in blood lipid according to the modeled assumptions for this scenario. After 24 months of age, the formula-fed and 3- or 6-month breast-feeding duration groups show similar average internal doses (4.5, 5.3, and 6.2 ppt, respectively), while the average for 12 months breast-feeding is considerably higher at 10.2 ppt (Fig. 6).
Figure 7 illustrates the model outcome for a constant breast-feeding duration (3 months full time) and varied ingestion and dermal contact exposures to dioxins in soils. Data set #1 represents exposures to soils with 100 ppt TEQ in a pattern dominated by TEQ from lower chlorinated congeners (mainly TCDD) and corresponds to a TWA internal dose of 7.4 ppt and a plateau after 24 months of age averaging 5.3 ppt. Data set #2 represents exposures to soils with 1000 ppt TEQ in a pattern dominated by TEQ from penta-, hexa-, and hepta-chlorinated congeners and corresponds to a TWA internal dose of 8.5 ppt and a plateau after 24 months of age averaging 6.5 ppt. Data set #3 represents exposures to soils with 1000 ppt TEQ in a pattern dominated by TEQ from higher chlorinated congeners (hepta-, and octa-chlorinated) and corresponds to a TWA internal dose of 7.9 ppt and a plateau after 24 months averaging 5.7 ppt. Note that the difference between data sets #2 and #3, which have the same soil TEQ concentration, relates to the lower relative bioavailability of the higher chlorinated congeners (e.g., hepta- and octaCDD/Fs) that are more prevalent in data set #3.
|
|
MODEL VALIDATION |
|---|
|
TOPABSTRACTINTRODUCTIONMODEL DESCRIPTIONMODEL APPLICATIONMODEL VALIDATIONDISCUSSIONFUNDINGSUPPLEMENTARY DATAREFERENCES |
|---|
Two reports by German investigators provide individual infant PCDD/F internal dose data following specified breast-feeding durations. Kreuzer et al. (1997)
reported data for nine infants who died of sudden infant death syndrome in the mid-1990's who were breast-fed 0.2–4.4 months before sampling. Abraham et al. (1998) reported data for six infants born in 1992–1995 who were breast-fed 1.6–7.4 months, four of which were first born and two second born to the same mother. Since Kreuzer et al. (1997) did not collect breast milk PCDD/F concentration data, our validation exercise was initially done by applying the model algorithms for predicting internal doses from the reported breast-feeding durations and sampling times using the average breast milk PCDD/F concentrations for 526 German mothers sampled in 1986–1991 by Furst et al. (1994). Abraham et al. (1998) reported breast milk concentrations for each of the six infants studied, showing PCDD/F TEQ concentrations (11–22 pg TEQ/g lipid) which were lower than the average value of 34 pg TEQ/g lipid reported by Furst et al. (1994). Thus, separate model runs were also made using the specific early breast milk concentrations for each of these six infants.
Table 5 provides a validation exercise for the model internal dose predictions relating to varied durations of breast-feeding based on published internal dose data for 15 German infants sampled in the early to mid-1990's. The assumed breast milk PCDD/F concentration and congener distribution used in the model was based on the average (34 pg TEQ/g lipid) reported for 526 German mothers sampled in 1986–1991 by Furst et al. (1994). The modeled infant TEQ values for the nine Kreuzer infants, the six Abraham infants, and the 15 combined (respective averages 48.3, 35.3, and 43.1 pg/g) were 2.1- to 3.5-fold higher than the measured TEQ values (respective averages 13.7, 16.7, and 14.9 pg/g). Birth order was not reported for the nine breast-fed infants of Kreuzer et al. (1997), but four of the six infants examined by Abraham et al. (1998) were first-born and two were second-born infants (B1-2 and B2-2). A greater margin of difference was observed for the shortest breast-feeding durations (e.g., less than 2 months) and for the two second-born infants. Since individual-specific breast milk PCDD/F concentrations were reported by Abraham et al. (1998), the model was run for each infant using the specific early breast milk TEQ concentration. Table 5 shows that modeled infant TEQ values (averaging 17.2 pg/g) are quite consistent with the measured TEQ values (averaging 16.7 pg/g) across the six infants studied by Abraham et al. (1998). The predicted values for individual infants varied from the measured values by only 19–41%.
|
Although data for model validation in this case are quite limited, Table 5 illustrates that the modeled internal doses for 15 infants with different breast-feeding durations are generally about two- to threefold higher than the measured internal doses for children up to about 1 year old. To some degree, this tendency toward overestimation may be related to the use of breast milk PCDD/F concentrations in an earlier era (1986–1991) than that for the infant internal dose sampling (about 1992–1995). Concurrent breast milk concentrations were apparently somewhat lower for the six Abraham infants, but we chose to use breast milk data from the most robust survey of German mothers preceding the infant sampling dates. Shorter durations of breast-feeding (e.g., < 2 months) were associated with greater margins of internal dose overstatement by the model for unknown reasons. Also, the second-born infants reported by Abraham et al. (1998)
showed a greater margin of internal dose overstatement due in part to the fact that the modeled breast milk concentration of 34 pg TEQ/g lipid from Furst et al. (1994) overstated the measured breast milk levels by about two- to threefold for infants B1-2 and B2-2 (Abraham et al., 1996, 1998). Individual-specific model runs for the six infants with early breast milk TEQ concentrations showed predicted values remarkably close to the measured infant internal doses (within 19–41%, for predicted vs. measured, Table 5). It should be noted that individual-specific factors that likely affect infant internal doses were not included in the model, such as actual body weight and growth pattern, gender, breast-feeding pattern, and specific infant diet from weaning until blood sampling at age 11.3–12.1 months. Since breast milk exposures generally dominate the peak and TWA internal dose in young children, this limited validation exercise would reasonably support the use of this model for risk assessment purposes, mindful of the fact that the model is an inexact tool, the accuracy of which largely depends on proper concordance of the input data with the characteristics of the population of interest.
We were unable to locate data on background internal doses in young children after infancy (e.g., 12–84 months old) that provided adequate information on breast-feeding and other dietary exposures to allow for testing of the current model. However, the TCDD half-life information utilized in the current model was based on data from a cohort of children and adolescents exposed during the trichlorophenol reactor explosion incident in Seveso, most of whom were under age 10 years at the time of exposure (Kerger et al., 2006). Also, the predicted child internal doses at age 7 are less than twofold higher than those for Australian children averaging 10 years old (Harden et al., 2004). The other PCDD/F congener half-lives from Leung et al. (2006) were based on the unique mass balance data set for two German infants that included specific assessments of breast milk and dietary intakes along with the internal dose measurements (Abraham et al., 1996, 1998). However, similar PCDD/F mass balance data sets have not been published for older children to our knowledge. We recommend further validation testing of the model as appropriate childhood internal dose data become available.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMODEL DESCRIPTIONMODEL APPLICATIONMODEL VALIDATIONDISCUSSIONFUNDINGSUPPLEMENTARY DATAREFERENCES |
|---|
This report expands on a preliminary age-dependent half-life model for PCDD/Fs in infants and young children (Kerger et al., 2007
) by incorporating more detailed empirical data on model parameters and adapting the algorithms into a user-friendly spreadsheet program that allows for input of site-specific PCDD/F congener data in environmental media including soils, foods, and breast milk. The current model includes empirical estimates of infant half-lives and rates of increase in half-lives with age for PCDD/F congeners (Kerger et al., 2006; Leung et al., 2006), more current and detailed data on monthly child body weight/body fat mass (CDC, 2000) and body weight–scaled dietary fat intakes during the period of peak internal dose shifts (ages 0–2 years), and robust recent data on background breast milk dioxin concentrations (Wittsiepe et al., 2004). Data on relative bioavailability of higher chlorinated PCDD/Fs compared to TCDD were also incorporated (Birnbaum and Couture, 1988; Norback et al., 1975; Ruby et al., 2002). The model also incorporates recent food dioxin concentrations for North America estimated by JECFA (2001) and empirical age-specific intake rates for breast milk, other foods (USEPA, 2002a; Williams et al., 2003), and for age-specific dermal contact and ingestion exposures to PCDD/F contaminated soils (Paustenbach et al., 2006).
The model produces a profile of blood lipid TEQ concentration versus age through 84 months and allows for calculation of the TWA child internal dose during the first 7 years of life (Figs. 6 and 7). Estimates of peak internal dose and TWA internal dose are considered the most important dose metrics for determining the human health risks of dioxins (USEPA, 2003). Thus, comparison of the output from this simple model to benchmark doses based on toxicity studies and/or background exposures may be considered more scientifically relevant than the currently used or suggested USEPA approaches based on daily dose estimates (USEPA, 1989, 2003).
A somewhat similar model was developed by USEPA (2002b) for evaluation of child internal doses of lead, another persistent pollutant with health risks that are difficult to assess based on daily dose estimates. The proposed dioxin child internal dose model similarly integrates the chemical uptake from important sources and utilizes empirical human toxicokinetic data to model the internal dose versus time profile in young children. Similar to the safe internal dose benchmarks identified through scientific consensus among experts in the toxicology of lead compounds (USEPA, 2002b), we recommend that similar benchmarks be identified for evaluating the health risk implications of the dioxin child internal dose model. Benchmarks based on human clinical studies and perhaps also on the range of current background blood lipid concentrations may be reasonably considered for comparison to the model-predicted results for children through age 7 years.
Possible benchmarks for comparison of the model-estimated peak internal dose might rely on the low-effect levels for TCDD-induced chloracne in children of 828 ppt identified by Mocarelli et al. (1991) and the approximate threshold of 800–1000 ppt for mild chloracne and for liver enzyme induction among Viennese laboratory workers accidentally exposed to TCDD (reviewed in Guzelian et al., 2006). Dividing these human low-effect/no effect levels of approximately 800 ppt by a safety factor of 10 results in a benchmark internal dose of 80 ppt in blood lipid for peak exposures that are unlikely to pose an acute health hazard in human infants. Accordingly, breast milk exposures associated with peak blood lipid levels exceeding 80 ppt might provide a rationale for actions to limit breast-feeding exposures.
The TWA internal dose estimates from the model may be compared to estimates of current background population internal doses of PCDD/Fs among younger reproductive-age females. Ferriby et al. (2007) summarized the NHANES study findings for younger adult females (ages 20–34) with a median TEQ in blood lipid at about 10 ppt and the 95th percentile at 23 ppt. Accordingly, model-estimated TWA internal doses that exceed 23 ppt TEQ might define a benchmark for taking corrective actions to avoid sources of excessive exposure to younger children. Further research on health-based benchmarks based on TWA internal dose estimates in human children and young adults is needed.
These modeling exercises illustrate that the much shorter half-life of dioxins in young children (e.g., 0.27–0.46 years) are capable of rapidly moderating internal doses from background dietary and environmental intakes, translating to average internal doses from age 0 to 7 that are well below those for adults who exhibit a much longer half-life (e.g., 7–11 years). An earlier modeling analysis of infant dioxin internal doses by Lorber and Phillips (2002) came to similar conclusions using TCDD as a surrogate for TEQ from all PCDD/Fs in breast milk. Breast-feeding is the dominant factor for dioxin internal dose in young children (Fig. 6). The longest assumed duration of full-time breast-feeding (12 months) showed a predicted TWA internal dose of 13.9 ppt, which is about twofold lower than background PCDD/F internal doses around 19–30 ppt TEQ in the blood lipid of adults in the United States and Europe (Aylward and Hays, 2002; Ferriby et al., 2007; Hays and Aylward, 2003; Lorber, 2002; USEPA, 2003) and is about fourfold higher than the background PCDD/F internal doses of 3.9 ppt TEQ in the blood lipid of children with an average age of 10 years in Australia (Harden et al., 2004). Under more typical durations of breast-feeding (i.e., 3–6 months), the TWA internal doses are lower (7.3–9.7 ppt) and approach a similar plateau as for formula-fed infants after about 24 months. These modeled TEQ internal doses are approximately two- to fourfold below estimated current adult internal dose estimates (Aylward and Hays, 2002; Ferriby et al., 2007; Hays and Aylward, 2003; Lorber, 2002; USEPA, 2003) and less than twofold higher than current internal dose estimates for children averaging 10 years old (Harden et al., 2004).
The model also allows evaluation of dioxin uptake from soils via dermal contact and soil ingestion. The algorithms for the internal dose via soil ingestion and dermal contact were adapted from a detailed probabilistic risk assessment approach recently reported by Paustenbach et al. (2006). This probabilistic risk assessment methodology is based on comprehensive assessment of central tendency data and uncertainties for the key exposure factors in an urban residential setting (see Supplementary Table A-6); it does not incorporate data or assumptions regarding child pica or possible indirect exposure pathways (e.g., site-related impacts on fish or livestock) other than soil ingestion and dermal contact (Paustenbach et al., 2006). However, the 95th percentile estimates for PCDD/F uptake via soil ingestion and dermal contact used in the current model are considered appropriate to capture the most likely upper bound soil exposures for young children in an urban residential setting.
Three different soil PCDD/F patterns are used to illustrate the impact of 95th percentile soil exposures on the TEQ internal dose profile in young children (Fig. 7). Despite a 10-fold difference in soil TEQ between data sets #1 and #2, the difference in TWA internal dose was only 1.2 ppt and during the peak soil ingestion timeframe (ages 3–4 years) differed by less than 2 ppt TEQ. A smaller difference was apparent between data set #1 and #3, despite the same 10-fold difference in soil TEQ concentrations. The muted difference between data set #1 and either #2 or #3 is due to the increasingly prominent presence of higher chlorinated PCDD/F congeners in each data set (Supplementary Table A-7); the model accounts for the lower relative bioavailability of hepta- and octa-chlorinated congeners, and this equates with lower TEQ internal dose in the child.
The current model is not comprehensive of all PCDD/F sources and has certain weaknesses. First, it does not include algorithms for all plausible direct and indirect PCDD/F exposure pathways. As discussed by Paustenbach et al. (2006), most of the additional exposure pathways are either negligible compared to uptake via dietary and soil pathways (e.g., inhalation or homegrown vegetable pathways) or they are unique to certain populations whose consideration would greatly skew the central tendency uptake estimates (e.g., subsistence farmer or fisher groups). Thus, one must cautiously evaluate possible contributions from other relevant exposure pathways when evaluating the outputs from the current model. Second, the current model does not attempt to include dioxin-like polychlorinated biphenyls due to the paucity of toxicokinetic data for these congeners in humans. Third, the current model does not attempt to account for cumulative impacts of soil exposures on adult female/breast milk dioxin concentrations. However, application of the age-dependent half-life model suggests that shorter PCDD/F half-lives during early childhood and adolescence may greatly attenuate contributions to the accumulated adult internal dose (and adult breast milk concentration) from soil-related exposures in reproductive-age females (Richter et al., 2006). Fourth, the estimated half-lives included in the current model are empirical estimates and surrogate values that require further validation based on human studies (Kerger et al., 2006; Leung et al., 2006). And fifth, the current model does not include an algorithm to account for the depuration of breast milk PCDD/F concentrations with longer breast-feeding durations due to the paucity of available data and the uncertainties on this topic. As additional data become available to resolve these issues, adjustments should be made to increase the accuracy of the model.
In conclusion, the proposed child dioxin internal dose model is an important step toward identifying a useful risk assessment approach for noncancer effects of PCDD/Fs relating to exposures during early childhood. It is hoped that the current approach can be further developed and refined to become a useful risk assessment tool as was the case for the Integrated Environmental Uptake Biokinetic Model adopted by USEPA (2002b). Further, it is important to assess the potential health risks of PCDD/Fs by taking into account the most recent empirical findings in humans. The age-dependent half-life findings provide some very important scientific caveats regarding past concerns about exposures from breast-feeding and soil exposures.
|
FUNDING |
|---|
|
TOPABSTRACTINTRODUCTIONMODEL DESCRIPTIONMODEL APPLICATIONMODEL VALIDATIONDISCUSSIONFUNDINGSUPPLEMENTARY DATAREFERENCES |
|---|
Partial funding was provided by the Dow Chemical Company, which is currently involved in litigation related to PCDD/F contaminated soils.
|
SUPPLEMENTARY DATA |
|---|
|
TOPABSTRACTINTRODUCTIONMODEL DESCRIPTIONMODEL APPLICATIONMODEL VALIDATIONDISCUSSIONFUNDINGSUPPLEMENTARY DATAREFERENCES |
|---|
Supplementary Tables A1–A7 are available online at http://toxsci.oxfordjournals.org/.
|
ACKNOWLEDGMENTS |
|---|
Each of the authors has conducted research on the toxicology and risk assessment of dioxins and has consulted with various academic, industrial, nongovernmental, legal, regulatory, and governmental organizations, including expert witness testimony.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMODEL DESCRIPTIONMODEL APPLICATIONMODEL VALIDATIONDISCUSSIONFUNDINGSUPPLEMENTARY DATAREFERENCES |
|---|
Abraham K, Knoll A, Ende M, Papke O, Helge H. Intake, fecal excretion, and body burden of polychlorinated dibenzo-p-dioxins and dibenzofurans in breast-fed and formula-fed infants. Pediatr. Res. (1996) 40:671–679.[Web of Science][Medline]
Abraham K, Papke O, Gross A, Kordonouri O, Wiegand S, Wahn U, Helge H. Time course of PCDD/PCDF/PCB concentrations in breast-feeding and their infants. Chemosphere (1998) 37:1731–1741.[Medline]
Agency for Toxic Substances and Disease Registry (ATSDR). Toxicological Profile for Chlorinated Dibenzo-p-dioxins (1998) Washington, DC: U.S. Dept. of Health and Human Services.
Aggett PJ, Haschke F, Heine W. Comment on the content and composition of lipids in infant formulas. ESPGAN Committee on Nutrition. Acta Paediatr. Scand. (1991) 80:887–896.[Web of Science][Medline]
Aggett PJ, Haschke F, Heine W. Childhood diet and prevention of coronary heart disease. ESPGAN Committee on Nutrition. J. Pediatr. Gastroenterol. Nutr. (1994) 19:251–269.[Medline]
Aylward LL, Hays SM. Temporal trends in human TCDD body burden: Decreases over three decades and implications for exposure levels. J. Expo. Anal. Environ. Epidemiol. (2002) 12:319–328.[CrossRef][Web of Science][Medline]
Birnbaum LS, Couture LA. Disposition of octachlorodibenzo-p-dioxin (OCDD) in male rats. Toxicol. Appl. Pharmacol. (1988) 93:22–30.[CrossRef][Web of Science][Medline]
Butte NF, Garza C, Stuff JE, O'Brian Smith E, Nichols BL. Effect of maternal diet and body composition on lactational performance. Am. J. Clinical Nutr. (1984) 39:296–306.
Centers for Disease Control and Prevention (CDC). CDC Growth Charts: United States. Advance Data from Vital and Health Statistics of the Centers for Disease Control and Prevention, National Center for Health Statistics (2000) Atlanta, GA: Centers for Disease Control and Prevention. Number 314.
Centers for Disease Control and Prevention (CDC). NHANES Baby Formula Data from Online Database (2005) Atlanta, GA: Centers for Disease Control and Prevention.
Deurenberg P, Weststrate JA, Seidell JC. Body mass index as a measure of body fatness: Age- and sex-specific prediction formulas. Br. J. Nutr. (1991) 65:105–114.[CrossRef][Web of Science][Medline]
Dewey KG, Lonnerdal B. Milk and nutrient intake of breast-fed infants from 1 to 6 months: Relation to growth and fatness. J. Pediatr. Gastroenterol. Nutr. (1983) 2:497–506.[Web of Science][Medline]
European Union. Commission of the European Communities. Commisssion Directive 96/4/EU amending Directive 91/32/EEC on infant formulae and follow on formulae. Off. J. Eur. Comm. (1996) L94/12.
Ferriby LL, Knutsen JS, Harris M, Unice KM, Scott P, Nony P, Haws LC, Paustenbach D. Evaluation of PCDD/F and dioxin-like PCB serum concentration data from the 2001–2002 National Health and Nutrition Examination Survey of the United States population. J. Expo. Sci Environ. Epidemiol. (2007) 17:358–371.[CrossRef][Web of Science][Medline]
Furst P, Furst C, Wilmers K. Human milk as a bioindicator for body burdens of PCDDs, PCDFs, organochlorine pesticides, and PCBs. Environ. Health Perspect. (1994) 102:187–193.[CrossRef]
Greene JF, Hays S, Paustenbach DJ. Basis for a proposed reference dose (RfD) for dioxin of 1-10 pg/kg-day. A weight of evidence evaluation of the human and animal studies. J. Toxicol. Environ. Health Part B (2003) 6:115–159.[CrossRef]
Guzelian P, Quattrochi I, Karch N, Aylward L, Kaley R. Does dioxin exert toxic effects in humans at or near current background body levels? An evidence-based conclusion. Hum. Exp. Toxicol. (2006) 25:99–105.
Harden F, Muller J, Toms L. National Dioxins Program. Technical Report No. 9, Dioxins in the Australian Population: Levels in Blood (2004) Canberra, Australia: Prepared for Australian Government Department of the Environment and Heritage.
Hays S, Aylward L. Dioxin risks in perspective: Past, present, and future. Regul. Toxicol. Pharmacol. (2003) 37:202–217.[CrossRef][Web of Science][Medline]
International Commission for Radiological Protection (ICRP). International Commission for Radiological Protection Report of the Task Group on Reference Man (1975) New York: Pergamon Press.
Joint FAO/WHO Expert Committee on Food Additives (JECFA). Joint FAO/WHO Expert Committee on Food Additives, Fifty-seventh Meeting (2001) Rome, Italy. 24–40. 5–14 June 2001: Rome, Italy, pp. 24–40. Available at: http://www.who.int/ipcs/food/jecfa/summaries/en/summary_57.pdf. Accessed January 14, 2007.
Kerger BD, Leung H-W, Scott P, Paustenbach DJ. Refinements on the age-dependent half-life model for estimating child body burdens of polychlorodibenzodioxins and dibenzofurans. Chemosphere (2007) 67:S272–S278.[Medline]
Kerger BD, Leung H-W, Scott P, Paustenbach DJ, Needham LL, Patterson DG, Gerthous PM, Mocarelli P. Age- and concentration-dependent elimination half-life of 2,3,7,8-tetrachlorodibenzo-p-dioxin in Seveso children. Environ. Health Perspect. (2006) 114:1596–1602.[Web of Science][Medline]
Koletzko B, Dokoupil K, Reitmayr S, Weimert-Harendza B, Keller E. Dietary fat intakes in infants and primary school children in Germany. Am. J. Clin. Nutr. (2000) 72:1392S–1398S.
Kreuzer PF, Csanady GA, Baur C, Kessler W, Papke O, Greim H, Filser JG. 2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD) and congeners in infants. A toxicokinetic model of human lifetime body burden by TCDD with special emphasis on its uptake by nutrition. Arch. Toxicol. (1997) 71:383–400.[CrossRef][Web of Science][Medline]
Leung H-W, Kerger BD, Paustenbach DJ. Elimination half lives of selected polychlorinated dibenzodioxins and dibenzofurans in breast-fed human infants. J. Toxicol. Environ. Health Part A (2006) 69:437–443.[CrossRef][Web of Science][Medline]
Leung H-W, Kerger BD, Paustenbach DJ, Ryan JJ, Masuda Y. Age- and concentration-dependent elimination half-lives of chlorinated dibenzofurans in Yusho and Yucheng patients. Organohalogen Compd. (2005) 67:1726–1729.
Lorber M. A pharmacokinetic model for estimating exposure of Americans to dioxin-like compounds in the past, present, and future. Sci. Total Environ. (2002) 288:81–95.[CrossRef][Medline]
Lorber M, Phillips L. Infant exposure to dioxin-like compounds in breast milk. Environ. Health Perspect. (2002) 110:A325–A332.[Web of Science][Medline]
Mocarelli P, Needham L, Marocchi A, Patterson DG, Brambilla P, Gerthoux PM. Serum correlation of 2,3,7,8-tetra-chlorodibenzo(p)dioxin and test results from selected residents of Seveso, Italy. J. Toxicol. Environ. Health (1991) 32:357–364.[Web of Science][Medline]
Neville MC, Keller R, Seacat J, Lutes V, Neifert M, Casey C, Allen J, Archer P. Studies in human lactation: Milk volumes in lactating women during the onset of lactation and full lactation. Am. J. Clin. Nutr. (1988) 48:1375–1386.
Norback DH, Engblom IE, Allen JR. Tissue distribution and excretion of octachlorodibenzeo-para-dioxin in the rat. Toxicol. Appl. Pharmacol. (1975) 32:330–338.[CrossRef][Web of Science][Medline]
Paustenbach DJ, Leung H -W, Scott P, Kerger BD. An approach to calculating childhood body burdens of dibenzodioxins and dibenzofurans which accounts for age-dependent biological half-lives. Organohalogen Compounds (2004) 66:2714–2721.
Paustenbach DJ, Fehling K, Scott P, Harris M, Kerger BD. Identifying soil cleanup criteria for dioxins in urban residential soils: How have 20 years of research and risk assessment experience affected the analysis? J. Toxicol. Environ. Health Part B (2006) 9:1–59.[CrossRef]
Richter RO, Kerger BD, Leung H-W, Paustenbach DJ. Implications of age-dependent half-lives of dioxins on assessment of breast milk dose and body burden. Toxicol. Sci. (2006) 90. 177(Abstract 1561).
Ruby MV, Fehling KA, Paustenbach DJ, Landenberger BD, Holsapple MP. Oral bioaccessibility of dioxins/furans at low concentrations (50–350 ppt toxicity equivalents) in soil. Environ. Sci. Technol. (2002) 36:4905–4911.[Medline]
Schecter A, Kassis I, Papke O. Partitioning of dioxins, dibenzofurans, and coplanar PCBs in blood, milk, adipose tissue, placenta and cord blood from five American women. Chemosphere (1998) 37:1817–1823.[Medline]
Schecter A, Wallace D, Pavuk M, Piskac A, Papke O. Dioxins in commercial United States baby food. J. Toxicol. Environ. Health Part A (2002) 65:1937–1943.[CrossRef][Web of Science][Medline]
Schlummer M, Moser A, McLachlan MS. Digestive tract absorption of PCDD/Fs, PCBs, and HCB in humans: Mass balances and mechanistic considerations. Toxicol. Appl. Pharmacol. (1998) 152:128–137.[CrossRef][Web of Science][Medline]
Tigges BB. Infant formulas: Practical answers to common questions. Nurse Pract. (1997) 22:70–87.[Medline]
United States Environmental Protection Agency (USEPA). Risk Assessment Guidance for Superfund, Vol. 1, Human Health Evaluation Manual (1989) Washington, DC: Office of Emergency and Remedial Response, United States Environmental Protection Agency. EPA/540/1–89/002.
United States Environmental Protection Agency (USEPA). U.S. Environmental Protection Agency Child-Specific Exposure Factors Handbook (2002a) Washington, DC: NCEA. 3–59. EPA-600-P-00-002B.
United States Environmental Protection Agency (USEPA). Short Sheet: Overview of the IEUBK Model for Lead in Children. EPA #PB 99-9635-8, OSWER #9285-7-31 (2002b) Washington, DC: Office of Solid Waste and Emergency Response, United States Environmental Protection Agency.
United States Environmental Protection Agency (USEPA). Exposure and Human Health Reassessment of 2,3,7,8-Tetrachlorodibenzo-p-Dioxin (TCDD) and Related Compounds (2003) United States Environmental Protection Agency, Office of Research and Development, National Center for Environmental Assessment. NAS Review Draft, December 2003. EPA/600/P-00/001C(af). Available at: http://epa.gov/ncea/dioxin.htm. Accessed January 14, 2007.
United States Food and Drug Administration (USFDA). PCDD/PCDF Exposure Estimates, Center for Food Safety and Applied Nutrition, Office of Plant and Dairy Foods (2004) Washington, DC: U.S. Food and Drug Administration. Available at: http://www.cfsan.fda.gov/
lrd/dioxee.html. Accessed January 14, 2007.
Wenning RJ, Paustenbach DJ, Harris M, Bedbury H. Principal components analysis of potential sources of polychlorinated dibenzo-p-dioxins and dibenzofurans in surficial sediments from Newark Bay, New Jersey. Arch. Environ. Contam. Toxicol. (1993) 24:271–289.[CrossRef][Web of Science]
WHO/FAO. Fats and Oils in Human Nutrition. Report of a Joint Expert Consultation (1994) Rome, Italy: Food and Agriculture Organization of the United Nations.
Williams PRD, Holicky KC, Paustenbach DJ. Current methods for evaluating children's exposures for use in health risk assessment. J. Child. Health (2003) 1:41–98.
Wittsiepe JP, Furst P, Schrey P, Lemm F, Kraft M, Eberwein G, Winneke G, Wilhelm M. PCDD/F and dioxin-like PCB in human blood and milk from German mothers. Organohalogen Compd (2004) 66:2865–2872.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  H. Leung, B. Kerger, D. Paustenbach, J. Ryan, and Y Masuda Concentration and age-dependent elimination kinetics of polychlorinated dibenzofurans in Yucheng and Yusho patients Toxicology and Industrial Health, September 1, 2007; 23(8): 493 - 501. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||