Volume II Chapter 6.0 Pages 2 of 2 page    
    • 6.3.1. Pharmacokinetic Model 6-14
    • 6.3.2. Model Utilization 6-17
    • 6.3.3. Determining Liver Concentrations from Fat Levels 6-20
    • 6.4.1. Determination of Daily Intake Dose from Exposure Concentrations 6-25
    • 6.4.2. Dose Through Lactation 6-26
      • Concentration in the Milk 6-26
      • Dose to Infant 6-28

6.3.1. Pharmacokinetic Model

Assuming a linear relationship between the concentrations in the fat and the body at low exposure concentrations and the near linear elimination profile, a simpler model relating exposure, whole body elimination, and whole body concentration is developed here. As described earlier, after a prolonged exposure most of the body's organs can be assumed to have very similar kinetic profiles and can thus be lumped together.

The fat, while sharing such a similar profile, is kept separate because of its important role in storing most of the body burden and because it has been the most typically monitored tissue. The three compartments in this model are blood, fat, and a "body" representing all other tissues. The model is a flow- or perfusion-limited model with the assumption that the toxin is well stirred or uniformly distributed within each compartment. The pertinent equations for the model follow. The rate of change of concentration of toxin in the body is estimated by:

Formula V2 6-22
The rate of change of concentration of toxin in the fat is estimated by Equation 6-23 below:
Formula V2 6-23

The integral of the above differential equation (Eq. 6-23) over time gives the actual concentration.

The rate of change of concentration of toxin in blood is estimated by Equation 6-24 below:

Formula V2 6-24

It should be noted that the description of intake is somewhat different than what is typically found in most PBPK models. D here is actually a dose rate. That is, D is in terms of pg/kg/day coming into the body as a dose, not just a concentration in the food, drinking water, or air. The usual description for gastrointestinal absorption of toxin from food would be

Formula V2 6-25

In this case, because of inadequate knowledge regarding the absorption rate constants for many of the congeners, a dose rate is used instead. This does not allow for estimation of body burden directly from environmental concentrations (e.g., Fc in Equation 6-25). As more data are collected, more accurate values of the parameters and descriptions of absorption functions can be input into Equation 6-22. For the present, other approaches can be used to relate concentration in the environmental media to daily intake (see Section 6.4.1).

6.3.2. Model Utilization

Most of the model's parameters are known or can be estimated from experimental data for many of these toxins. For example, the equilibrium concentration ratios between the fat and the body and between the fat and the blood for 2,3,7,8-TCDD are approximately 10 and 100, respectively, on a tissue basis. The blood flows and compartment volumes are well known. One parameter that had to be estimated when applying this model to 2,3,7,8-TCDD was the clearance term K.

This was done by first allowing the model to simulate elimination as though exposure was suddenly terminated (D becomes zero). Initial concentrations for the tissues were taken as those typically expected in the general population (7.0 ppt in the fat). The value of K was then adjusted until the model predicted a half-life of 7 years. Values of clearance for the other compounds in this series would be determined similarly. The necessary information includes the equilibrium concentration distribution ratios, the half-lives of the compounds, and some reasonable approximation of steady state fat concentrations.

With an estimated value for the clearance, the model can now be used with various exposure inputs to establish body burdens of toxin. The model can also be used to estimate elimination profiles from the body. In addition, events such as lactation can be incorporated into the model with knowledge about the appropriate parameters.

Figure 6-2 shows the results of the model run describing the elimination of 2,3,7,8-TCDD from fat. The clearance rate was adjusted to predict a half-life of 7 years. The same clearance value was used with different starting conditions (concentration of TCDD in tissues) and the model produced a half-life of 7 years.

Next, the model was used with a constant daily intake as an input. Figure 6-3 shows the resulting profile of 2,3,7,8-TCDD in the fat. Figure 6-3 shows the results of a model run using an input of 0.44 pg/kg/day. Note that the steady state fat concentrations are approximately 7.0 ppt. The clearance rate used in this model run had been independently determined in the previous run based on reported half-life values. Thus, for these conditions this model does an adequate job of predicting tissue levels of 2,3,7,8-TCDD. Figure 6-4 shows a similar profile for a 0.30 pg/kg/day dose. The daily intakes chosen were similar to those calculated by the steady state equations in Section 6.1.2 (Table 6-1). Note that the steady state fat levels predicted by the pharmacokinetic model agree closely with those used as a starting point for the steady state calculation using seven years for the half-life.

table Figure 6-2 Model Estimates of Elimination of 2,3,7,8-TCDD from Fat. table Figure 6-3 Accumulation of TCDD in Fat with 0.44 pg/kg/day dose - Human.
expand table Figure V2 6 - 2 expand table Figure V2 6 - 3

This three compartment model appears to provide reasonable approximations of body burden, at least under the circumstances that have been tested. The necessary information for use with other substances in this series include the equilibrium distribution ratios between fat and blood and between fat and the rest of the body.

An estimate of the half-life is also needed in order to establish an appropriate value for the clearance term used in the model. Table 6-3 shows the results when the model was adjusted for other substances. The congener-specific TEQ intakes can be added together to arrive at a TEQ-based total daily intake for all the congeners of interest to a particular assessment.

Under certain circumstances, it might also be possible to use the compartmental model on a total TEQ basis. Considering, for example, a mixture of dioxins and furans whose collective biophysical properties are similar to those of 2,3,7,8-TCDD the model could be applied to approximate daily intake from steady-state fat levels. For example, Schecter (1991) reports the blood lipid levels in a sample of 85 persons from Germany to be 42 ppt TEQ (CDDs and CDFs).

Using that figure and the parameters of 2,3,7,8-TCDD, the model calculates a daily intake of 2.64 pg TEQ/d/kg. Fürst et. al (1991), based on food consumption patterns of the German population, estimated a daily intake of 2.3 pg TEQ/d/kg. Obviously great caution should be taken before using such an approach. Such estimates should only be relied upon if there is strong evidence that the mix of congeners is such that the collective properties of the mixture result in properties similar to those of 2,3,7,8-TCDD.

6.3.3. Determining Liver Concentrations from Fat Levels

As mentioned previously, higher resolution PBPK models are necessary to estimate and predict concentration at cellular and sub-cellular targets (vidae supra). At the present time, many of the data necessary to develop and apply these models to estimate target tissue or cellular dose in humans are lacking.

table Figure 6-4 Accumulation of TCDD in Fat with 0.30 pg/kg/day dose - Human. table Table 6-3 Model-Determined Daily Intakes.
expand table Figure V2 6 - 4 expand table Figure V2 6 - 3

Andersen and Greenlee (1991) provide an approach that uses the equations of a PBPK model (Leung et al., 1990) to predict liver concentrations from monitored fat concentrations. It should also be noted that blood or milk concentrations expressed on a per lipid basis could also be used. Basically the approach calculates the ratio of fat to liver concentration by dividing the tissue concentration equations of the PBPK model as follows:

Formula V2 6-26

Andersen and Greenlee (1991) further simplify Equation 6-26 for various limiting conditions. For the case of very low doses where CVL<<KB1, KB2 and their is no induction of the cytochrome P450 enzyme (microsomal binding protein), the equation simplifies to:

Formula V2 6-27
For the case where KB2<<CVL<<KB1, P450 is maximally induced but binding is less than half saturation, Equation 6-26 becomes:
Formula V2 6-28
For the case where CVL>>KB1, KB2, Equation 6-26 becomes:
Formula V2 6-29

Andersen and Greenlee (1991) provide values for each of the above conditions for experimental animals. It can be readily observed from examining Equation 6-26 that when binding levels are very small the ratio of concentration between liver and fat is influenced mostly by the ratio of partitioning coefficients.

With this approach and by knowing the necessary parameters Equation 6-26 can be used to estimate liver concentrations from fat (including blood lipid and milk lipid) concentrations. Andersen and Greenlee (1991) further suggest that most of the necessary binding parameters can be determined from in-vitro studies. Further, the pharmacokinetic model from which Equation 6-26 was derived can be used to estimate and predict cellular concentrations (both free and bound) under various exposure conditions.

The equations used by Leung et al. (1990) or other like equations provide estimates of amount bound to intracellular receptors sites, and can thus provide a relationship between multiple binding sites. Denison et al. (1991) suggest that binding to the Ah receptor is only one of several steps necessary for 2,3,7,8-TCDD to have an intracellular toxic effect.

As further knowledge becomes available about the mechanism and kinetics of each step, the model can be expanded to include these other processes such as DNA enhancing and hormonal modulation. The pharmacokinetic model will therefore become a pharmacodynamic model which will more explicitly link exposure to effect. Also, other tissues can be more specifically described in the model, if the mechanism of action data so warrant.


6.4.1. Determination of Daily Intake Dose from Exposure Concentrations

As was discussed in Section 6.3.1, it would be most advantageous to know more about the kinetics of absorption in the various animal species and the human. This is necessary for both extrapolation between species in the risk assessment and for determining body burdens from levels in the exposure media. For the time being, until more data become available regarding the kinetic absorption constants, a slightly modified approach can be used. Basically, the needed information is the concentration of the toxins in the media, the fraction of toxin absorbed from each of the media, and the amount of media coming into contact with the body. The following equation describes this in more detail.

Formula V2 6-30

This approach should be used with caution. The major assumption which impacts upon Equation 6-30 is that the fraction absorbed is constant across various concentrations and doses. As discussed in Section 6.4.2 and in U.S. EPA (1994), this assumption cannot always be considered to be sound. It most probably only applies at low doses and within any one species. It is an approach which, with care, can be used for certain conditions to give estimates until more reliable kinetic absorption data become available.

6.4.2. Dose Through Lactation

There is great concern regarding the potential dose resulting from lactation. Given the body burdens discussed in previous chapters and sections, lipid soluble substances might be expected to compartmentalize into milk and thus be transferred to nursing infants. Methods are needed to assess this potentially important route of exposure into the body. Concentration in the Milk

The first step to calculating the daily intake for infants is to determine the levels in mother's milk. There are two general methods that can be used as a basis. The first assumes that levels in maternal fat remain at steady-state and reach an equilibrium with milk fat. Under these assumptions, Equation 5-1 ( repeated here as Equation 6-31) can be used to calculate the levels in maternal milk:

Formula V2 6-31

Application of this equation to 2,3,7,8-TCDD, where an intake of 0.5 pg/kg/day, a half-life of 2555 days (7 years), 0.9 for f1 and 0.3 for f2 are assumed, results in a concentration in maternal milk fat of 5.6 ppt. This is higher than the 3.3 ppt. reported by Schecter et al., (1989).

A second and theoretically more accurate approach uses some type of physiologically based pharmacokinetic model to estimate the dynamically changing concentrations in mother's milk. One way to accomplish this is to add a mammary compartment to the compartmental model described in Section 6.3. The model is then extended to depict the toxin's transport into the milk. In the simplest form the following two equations would be added:

Formula V2 6-32
Formula V2 6-33

Other symbols as previously defined.

When the model is actually implemented, the changes in proportion of body fat in the mother that normally occur during lactation are taken into account. Observation of the two previous equations quickly shows that several new parameters are now added to the compartmental model. Many of these parameters have not been determined for most of the congeners of interest. In fact, even some of the physiologic and anatomic parameters are not readily available.

Thus, for the time being it may be best to use some type of steady-state model and Equation 6-31 or to actually use monitored data for calculating dose to a lactating infant. It should be noted that the model was applied for 2,3,7,8-TCDD and parameters were adjusted to predict levels near the 3.3 ppt value published by Schecter et al. (1989). This is not a validation of the model or its parameters; of interest, however, is that the model predicted the same ratio between milk lipids and plasma lipids as reported by Schecter et al. (1989). Dose to Infant

There are a number of measures of dose that can be used to compare the impact of exposure through lactation versus exposure through background. A common measure of dose, described in Chapter 5, is to calculate an average daily dose (ADD). Equation 5-2 (repeated here as Equation 6-34) describes such a calculation for the infant:

Formula V2 6-34

Assuming a concentration of 2,3,7,8-TCDD in milk fat of 3.3 ppt, values for f3 and f4 of 0.04 and 0.9, respectively, an average infant body weight of 10 kg, an exposure duration of one year, an ingestion rate of milk of 800 g/d, and an averaging time of one year, an ADD equal to 9.5 pg/kg/d is calculated. Using the same assumptions, except for an averaging time of 70 years (the entire assumed lifetime), an ADD equal to 0.135 pg/kg/d is calculated. Thus, depending upon the averaging time, lactation results in ADD similar to that resulting from background (0.135 pg/kg/d compared to 0.51 pg/kg/d) or an ADD over an order of magnitude higher (9.5 pg/kg/d compared to 0.51 pg/kg/d).

Little agreement exists regarding the appropriate choice of an averaging time for less than lifetime exposures. This is especially true for cases where exposure is occurring in a particularly sensitive developmental period. Lifetime averaging may be used for long time or even lifetime exposures, but the logic of applying it to short exposures is not clear.Equally unclear is the utility of the ADD calculation itself. The ADD is actually an average intake over an arbitrarily chosen period of time. The significance of the ADD is not apparent, especially in cases of compounds, such as the CDDs and CDFs, that reach steady-state levels during chronic exposure.It is recommended that other measures of dose be examined.

As advanced and new mechanistic knowledge develops, new measures of dose should be used and examined. Recent studies (Andersen and Greenlee, 1991; Andersen et al., 1993) show the importance of receptor-mediated processes. The actual toxicologically relevant dose may be related to the induced production of specific hepatic proteins. Scientific research is still needed to determine the exact role these play in the various toxic responses that may be caused by the CDDs and CDFs. As this type of mechanistic information develops, and its relevance to humans becomes better established, pharmacokinetic models will be expanded to become pharmacodynamic models and then more realistic measures of dose can be determined.

For the time being, macroscopic measures other than the ADD should also be examined for assessment purposes. One approach to estimating such measures is to use the compartmental model described in Section 6.3. For the case of infants and children, the model is modified to account for the changing body weight of the child during the first 20 years of life. The model is used here under each of the three following exposure conditions for 2,3,7,8-TCDD:

· One year lactational exposure at levels corresponding to a milk fat concentration of 3.3 ppt followed by 69 years exposure at levels corresponding to background daily intakes of 0.51 pg/kg/d.

· 70 year exposure at levels corresponding to background daily intakes of 0.51 pg/kg/d.

· One year lactation exposure at milk fat concentration of 3.3 ppt.

Results of the simulation reveals several things. Figures 6-5 through 6-7 show the fat concentration profiles after the above three exposure scenarios for 2,3,7,8-TCDD. Figure 6-5 shows the fat level to peak quickly during the lactation period and then diminish until it reaches the steady-state level of about 8.0 ppt (8000 pg/kg). This is contrasted to Figure 6-6 where only background (no lactation exposure) is simulated. Note that in both cases the steady-state fat levels are 8.0 ppt. Lactation causes a temporary peak that diminishes over the next several years. Figure 6-7 shows the results of the simulation for the lactational phase only.

This scenario assumed no further exposure after lactation ceased and was performed to investigate the impact of the lactational exposure apart from background exposure. Note that the half-life during the childhood years is shorter than that for adults (discussed in previous sections). This shorter half-life may be attributed to the changing size of the body and the compartments represented in the model. In fact, after age 20 (after which the body weight is assumed to remain constant) the model predicts the same half-life as previously discussed (7 L t1/2 3 8 years).

table Figure 6-5 Combined Exposure Adipose Tissue Concentration. table Figure 6-6 Background Exposure Adipose Tissue Concentration.
expand table Figure V2 6 - 5 expand table Figure V2 6 - 6
table Figure 6-7 Concentration After Lactational Exposure Only . Careful inspection of Figures 6-5 and 6-6 also reveals that the time to reach steady-state is the same as discussed in previous sections (> 30 years).

Other measures of dose can also be used for comparison purposes. Care should be taken when ascribing meaning to different measures of dose, especially when exposure patterns of widely different duration are compared. Additional complexity is added by the fact that the lactational exposure occurs during the early developmental period of the individual. Given, these caveats, it may still be worthwhile to look at time integral (area under the curve or AUCs) measures of mass that has resided in the body throughout the exposure period.

expand table Figure V2 6 - 7

This is not a simple calculation of intake, but rather a measure of cumulative mass within the body as calculated by the model. The time integral of the mass in the body compartment (AUCBO) is used here for discussion and comparison. Lactation exposure during the first year and background exposure during the remaining 69 years results in an AUCBO at 70 years of 3.01 X 106 pg-years. Seventy years of background exposure with no lactational exposure results in 2.95 X 106 pg-years. The one year lactation exposure results in 7,862 pg-year at the one year mark. Obviously, in terms of pg-years the contribution from lactation is minor compared to background exposure.

It is not clear, however, whether this is an appropriate measure of dose to compare exposure patterns occurring at very different times of the lifetime. This same measure of dose can be averaged over the exposure time and then compared. The results appear different when the above three AUCBOs are converted to a pg/kg/day basis. The AUCBO averaged over exposure for the combined lactation and background exposure is 1.68 pg/kg/day ([3.01 X 106 pg-years]/70 kg/[365 d X 70 years]), 1.64 pg/kg/day for the background only case, and 2.15 pg/kg/day ([7,862 pg-years]/10 kg/365 d X 1 year]) for the lactational portion of exposure.

Using this time-averaged measure, it appears that the one year lactational phase contributes slightly more, on a daily average basis, than the 69 year background phase. In addition, for some toxic end-points the first year of life may be more vulnerable than later years. How any of these measures are related to risk remains unknown and caution should be exercised before drawing conclusions. The goal here is to present different methods that might be used for comparison purposes. As mentioned previously, as the mechanisms of action become better elucidated other more appropriate measures of dose should be used to make relative risk comparisons.


  • Andersen, M.E.; Mills, J.J.; Gargas, M.L.; Kedderis, L.; Birnbaum, L.S.; Neuber, D.; Greenlee, W.F. (1993) Modelling receptor-mediated processes with dioxin: implications for pharmacokinetics and risk assessment.
  • Andersen, M.E.; Greenlee, W.F. (1991) Biological determinants of TCDD pharmacokinetics and their relation a biologically based risk assessment. In: Biological Basis for Risk Assessment of Dioxins and Related Compounds; Michael, G.; Scheuplein, R.; Van Der Heijden, K. eds.; Banbury Report 35, Cold Spring Harbor Laboratory Press.
  • Bowman, R.E.; Schantz, S.L.; Weerasinghe, N.C.A.; Gross, M.L.; Barsotti, D.A. (1989) Chronic dietary intake of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) at 5 or 25 parts per trillion in the monkey: TCDD kinetics and dose-effect estimate of reproductive toxicity. Chemosphere 18(1-6):243-252.
  • Denison, M.S.; Phelphs, C.L.; Dehoog, J.; Kim, H.J.; Bank, P.A.; Yao, E.F. (1991) Species variation in Ah receptor transformation and DNA binding. In: Biological Basis for Risk Assessment of Dioxins and Related Compounds; Gallo, M.; Scheuplein, R.; Van Der Heijden, K. eds.; Banbury Report 35, Cold Spring Harbor Laboratory Press.
  • Fürst, P.; Fürst, C.; Wilmers, K. (1991) Body burden with PCDD and PCDF from food. In: Biological Basis for Risk Assessment of Dioxins and Related Compounds; Gallo, M.; Scheuplein, R.; Van Der Heijden, K. eds.; Banbury Report 35, Cold Spring Harbor Laboratory Press.
  • Gasiewicz, T.A.; Henry, E.C. (1991) Different forms of the Ah receptor: possible role in species- and tissue-specific responses to TCDD. In: Biological Basis for Risk Assessment of Dioxins and Related Compounds; Gallo, M.; Scheuplein, R.; Van Der Heijden, K. eds.; Banbury Report 35, Cold Spring Harbor Laboratory Press.
  • King, F.G.; Dedrick. R.L.; Collins, J.M.; Matthews, H.B.; Birnbaum, L.S. (1983) Physiological model for the pharmacokinetics of 2,3,7,8-tetra-chlorodibenzofuran in several species. Toxicology and Applied Pharmacology 67:390
  • Leung, H.; Paustenbach, J. (1987) A proposed occupational exposure limit for 2,3,7,8-tetrachlorodibenzo-p-dioxin. Palo Alto, CA: Environmental Health and Safety, Syntex, U.S.A. Inc.
  • Leung, H-W.; Ku, R.H.; Paustenbach, D.J.; Andersen, M.E. (1988) A physiologically based pharmacokinetic model for 2,3,7,8-tetra-chlorodibenzo-p-dioxin in C57BL/6J and DBA/2J mice. Toxicology Lett. 42:15-28.
  • Leung, H-W.; Poland, A.; Paustenbach, D.J.; Murray, F.J. Andersen, M.E. (1990) Pharmacokinetics of [125I]-2-iodo-3,7,8-trichlorodibenzo-p-dioxin in mice: analysis with a physiological modeling approach. Toxicology and Applied Pharmacology 103:411-419.
  • McConnell, E.E.; Lucier, G.W.; Rubaugh, R.C.; et al. (1984) Dioxin in soil: bioavailability after ingestion by rats and guinea pigs. Science 223:1077-1079.
  • Nau, H.; Bab, R.; Neuber, D. (1986) Transfer of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) via placenta milk, and postnatal toxicity in mouse. Arch. Toxicol. 59:36-40.
  • Perdew, G.H.; Hollenbeck, C.E. (1990) Analysis of photoaffinity labeled aryl hydrocarbon receptor heterogeneity by two dimensional gel electrophoresis. Biochemistry 29: 6210-6214.
  • Schecter, A.; Fürst, P.; Ryan, J.J.; Fürst, C.; Meemken, H.A.; Groebel, W.; Constable, J.; Vu, D. Polychlorinated dioxin and dibenzofuran levels from human milk from several locations in the United States, Germany, and Vietnam. Chemosphere 19(1-6):979-984.
  • Schecter, A. (1991) Dioxins and related chemicals in humans and in the environment. In: Biological Basis for Risk Assessment of Dioxins and Related Compounds; Gallo, M.; Scheuplein, R.; Van Der Heijden, K. eds.; Banbury Report 35, Cold Spring Harbor Laboratory Press.
  • Schlatter, C. (1991) Data on kinetics of PCDDs and PCDFs as a prerequisite for human risk assessment. In: Biological Basis for Risk Assessment of Dioxins and Related Compounds; Gallo, M.; Scheuplein, R.; Van Der Heijden, K. eds.; Banbury Report 35, Cold Spring Harbor Laboratory Press.
  • Stanley, J.S.; Boggess, K.; Onstot, J.; Sack, T.; Remmers, J.; Breen, J.; Kutz, F.W.; Robinson, P.; Mack, G. (1986) PCDDs and PCDFs in human adipose tissues from the EPA FY82 NHATS repository. Chemosphere 15:1605-1612.
  • U.S. Environmental Protection Agency (1994) Health assessment for 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) and related compounds. Washington, DC: Office of Health and Environmental Assessment. Public review draft. EPA/600/EP-92/001.
  • Whitlock, J.P., Jr. (1991) Mechanism of dioxin action: relevance to risk assessment. In: Biological Basis for Risk Assessment of Dioxins and Related Compounds; Gallo, M.; Scheuplein, R.; Van Der Heijden, K. eds.; Banbury Report 35, Cold Spring Harbor Laboratory Press.