INHIBITION OF THE SODIUM-IODIDE SYMPORTER BY PERCHLORATE:
AN EVALUATION OF LIFESTAGE SENSITIVITY USING
PHYSIOLOGICALLY-BASED PHARMACOKINETIC (PBPK) MODELING
NOTICE
THIS DOCUMENT IS A PRELIMINARY DRAFT. It is undergoing peer review and public
comment and should not at this stage be construed to represent any final Agency policy. It is
being circulated for comment on its technical accuracy and policy implications.
U.S. Environmental Protection Agency
Office of Research and Development
EPA/600/R-08/106A
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	.	DRAFT

-------
Disclaimer
This document is intended to support EPA's preliminary regulatory determination for
perchlorate; it is, however, still undergoing peer review under applicable information quality
guidelines and public comment. As a result, this document may change. This document docs
not represent and should not be construed to represent any final Agency determination or policy.
It is being circulated for review of its technical accuracy and science policy implications.
Mention of trade names or commercial products does not constitute endorsement or
recommendation for use.
Contributing Authors
National Center for Computational Toxicology
Rory B. Conally
National Center for Environmental Assessment
Lynn Flowers
Eva D. McLanahan
Jacqueline Moya
Paul M. Schhsser
Paul White
National Health and Environmental Effects Research Laboratory
Mary E. Gilbert
Office of Science Policy
Danielle C. Tillman
This document is a draft for review and public comment purposes only and does not constitute fin at Agency policy.
October 2,2008	~	DRAFT

-------
TABLE OF CONTENTS
EXECUTIVE SUMMARY 											4
1.	INTRODUCTION....................							.........6
2.	EXAMINATION OF PBPK MODEL COMPUTER CODE				8
3.	EVALUATION OF PBPK MODEL TECHNICAL APPROACH - MODEL
DEVELOPMENT AND MODEL PARAMETERIZATION						10
3.1.	URINARY CLEARANCE 							 10
3.2.	PARAMETER SCALING							13
3.3.	POST-NATAL PBPK MODELING...........					14
3.3.1.	BREAST-FED INFANT SUCKLING RATE					 15
3.3.2.	BOTTLE-FED INFANT MODEL SIMULATION APPROACH		 16
4.	EPA MODIFIED PBPK MODEL RESULTS AND LIFESTAGE ANALYSIS	18
4.1.	EPA MODIFIED PBPK MODEL RESULTS							18
4.2.	LIFESTAGE RELATIVE SENSITIVITY ANALYSIS 			 21
4.3.	LIFESTAGE COMPARISON FOR THREE DRINKING WATER
CONCENTRATIONS......................									.........24
5.	SUMMARY AND CONCLUSIONS 								29
REFERENCES............................							......30
APPENDIX A: DESCRIPTION OF PBPK MODEL CODE ISSUES AND RESOLUTION... 34
APPENDIX B: EVALUATION OF URINARY CLEARANCE PARAMETERS 			 39
APPENDIX C: MODEL REVIEW FINAL REPORT FROM EPA CONTRACTOR	51
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	t	DRAFT

-------
EXECUTIVE SUMMARY
Perchlorate competitively inhibits uptake of iodide by the sodium-iodide symporter (NIS)
in laboratory animals and humans. NIS is found in many tissues, but is primarily responsible for
sequestering iodide from the bloodstream into the thyroid, enabl ing biosynthesis of thyroid
hormones. The National Research Council (NRC, 2005) concluded that hypothyroidism is the
first adverse effect in the continuum of effects that could result from perchlorate exposure.
However, NRC advised that hypothyroidism not be used as the basis of the perchlorate RID,
recommending that the most health protective and scientifically valid approach was to base the
perchlorate RfD on the inhibition of iodide uptake by the thyroid. In this analysis, the
physiologically-based pharmacokinetic (PBPK) models of perchlorate and radioiodide, which
were developed to describe thyroidal radioactive iodide uptake (RAIU) inhibition by perchlorate
for the average adult (Merrill et aL 2005), pregnant woman and fetus, lactating woman and
neonate, and the young child (Clewell et al., 2007), were evaluated based on their ability to
provide additional information about this critical effect for potentially sensitive subgroups.
EPA evaluated the PBPK model code provided by the model authors and found minor
errors in mathematical equations and computer code, as well as some inconsistencies between
model code files. ORD scientists made corrections to the code, with agreement from model
authors that the corrections should be made, in order to harmonize the models and more
adequately reflect the biology.
EPA determined that model parameters describing urinary excretion of perchlorate and
iodide were particularly important in prediction of RAIU inhibition in all subgroups; therefore, a
range of biologically plausible values available in peer-reviewed literature was evaluated in
depth using the PBPK models. EPA also determined that exposure rates were critical for
estimation of RAIU inhibition by the models and thus evaluated exposure rates further.
EPA's analysis identified the near-term fetus (only gestation week 40 fetus could be
adequately modeled) as the most sensitive subgroup with respect to percent RAIU inhibition at a
perchlorate dose equal to the point of departure (7 /ig/kg-day). Specifically, at a perchlorate dose
of 7 /ig/kg-day, the percent RAIU inhibition predicted by the model for the near-term fetus is 5-
fold greater than the average adult. After correcting the model for reduced urinary clearance in
infants, the same analysis predicts percent RAIU inhibition approximately 1- to 2-fold higher for
the breast-fed and bottle-fed infant (7-60 days) than for the average adult (differing from Clewell
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	A	DRAFT

-------
et ai. (2007)), and predicts percent RAIL! inhibition slightly lower for the 1-2 year old child than
for the average adult. Clewell et al. (2007) predicted percent RAIU inhibition in the older child
to be about one-half that of the adult; ORD's results are closer to, but still less than, the adult.
Overall, detailed examination of Clewell et al. (2007) and Merrill et al. (2005) reflected
that the model structures were appropriate for predicting percent inhibi tion of RAIU by
perchlorate in most lifestages. Unfortunately, the lack of biological information and data that
might be used to validate model predictions, particularly for early fetal development, limits
EPA's confidence on predictions for fetal endpoints. Therefore the EPA simply chose not to use
model predictions for the early- or mid-term fetus. However, because many of the physiological
and iodide/perchlorate-specific parameters in the late-term fetus are expected to be quite close to
those of the newborn, and there are much more data available for validation of the model in the
newborn, our higher confidence in model predictions for the newborn is then partially extended
to the late-term fetus (although there is still lower confidence in the late-term fetal predictions
than in those for the newborn). Quantitative outputs of the PBPK models as updated by the
EPA differ by up to 3-fold from published values, though many of the outputs are within 20% of
the published values. Nevertheless, the EPA evaluation determined ihat, with those
modifications as described herein, the Clewell et al. (2007) and Merrill et al. (2005) models are
acceptable to calcu late the lifestage differences in the degree of thyroidal NIS RAIU inhibition at
a given level of perchlorate exposure.
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	«-	DRAFT

-------
1. INTRODUCTION
The sodium-iodide symporter (NTS) transports iodide from blood into the thyroid gland,
enabling biosynthesis of thyroid hormones. Perchlorate is a potent competitive inhibitor of the
NIS. Perchlorate has been shown to cause thyroid-related pathologies and neurodevelopmcntal
effects in rodents by disrupting the thyroid axis homeostasis (e.g., NRC, 2005; York et al., 2005
a, 2005b; Gilbert and Sui, 2008). The National Research Council (NRC, 2005) evaluated the
human health implications of perchlorate and stated that inhibition of iodide uptake has been
unequivocally demonstrated in humans exposed to perchlorate, and it is the key event that
precedes all thyroid-mediated effects of perchlorate exposure. NRC concluded that
hypothyroidism is the first adverse effect in the continuum of effects that could result from
perchlorate exposure. However, NRC advised that hypothyroidism not be used as the basis of
the perchlorate RiD, recommending that the most health protective and scientifically valid
approach was to base the perchlorate RfD on the inhibition of iodide uptake by the thyroid. NRC
further concluded that iodide uptake inhibition, although not adverse, would precede any adverse
health effects of perchlorate exposure. The lowest dose (7 (ig/kg-day) administered in the Greer
et al. (2002) study was considered a no-observed effect level (NOEL) because iodide uptake
inhibition was considered to not be an adverse effect. The NRC also recommended that EPA use
this dose as the point of departure and apply an intraspecics uncertainty factor of 10 to account
for differences in sensitivity between the healthy adults in the Greer et al. (2002) study and the
most sensitive population, fetuses of pregnant women who might have hypothyroidism or iodine
deficiency. EPA's Integrated Risk Information System (IRIS) adopted the NRC's
recommendations (U.S. EPA, 2005).
Merrill et al. (2005) described a deterministic, physiologically based pharmacokinetic
(PBPK) model for radioiodide and perchlorate and the competitive interaction of perchlorate and
radioiodide at the NIS in adult humans. Clewell et al. (2007) extended this work and previous
lifestage models in the rodent (Clewell et al., 2003a, 2003b) to predict inhibition of the NIS for
pregnant and lactating women, nursing infants, and for the subsequent stages of childhood.
The following report provides a quantitative analysis of perchlorate-mediatcd inhibition
of the NIS in humans using the Merri ll et al. (2005) and Clewell et al. (2007) PBPK models,
focusing on the variability in the degree ofNIS inhibition as a function of lifestage. This set of
models, largely completed after the NRC (2005) report, provides new information that may be
This document is a draft for review ami public comment purposes only and does not constitute final Agency policy.
October 2,2008	.	DRAFT

-------
useful to EPA for addressing differences in human responses to percMorate across lifestages.
The analysis presented here was conducted to inform EPA's regulatory determination for
perchlorate under the Safe Drinking Water Act, Specifically, the Office of Water requested that
staff in the Office of Research and Development's (QRD) National Health and Environmental
Effects Research Laboratory (NHFERL), National Center for Environmental Assessment
(NCEA) and National Center for Computational Toxicology (NCCT) assist in a thorough
evaluation of the Merrill et al. (2005) and Clcwcll et al (200?) PBPK models to address their
scientific soundness and to determine whether or not they are suitable to provide quantitative
predictions to the Agency on the lifestage variability of perchlorate NIS inhibition of thyroidal
iodide uptake.
The evaluation of the PBPK models was conducted in two stages. First, the scientific
credibility of the models was evaluated based on (1) the information presented in Merrill et al,
(2005) and Clewell et at (2007), (2) limited inspection of the computer codes of the models, and
(3) limited execution of the computer model codes as supplied by the model authors. In
concluding this first stage of model evaluation, EPA decided that the PBPK models were
potentially suitable for regulatory use by the Agency, but a more detailed and thorough
evaluation of the models was necessary. Thus, the second stage of model evaluation involved a
more complete inspection of the computer codes and examination of the technical approach used
to develop the model structures and parameter values. In addition, the published models were
modified by EPA to fix errors and incorporate new data, particularly data on lifestage variability
in the urinary clearance of perchlorate, to which NIS inhibition is sensitive.
The first-stage evaluation included staff1 with expertise in PBPK modeling,
developmental neurotoxicology, perchlorate toxicology, and risk assessment. The second-stage,
-------
first-stage evaluation. The second-stage group consisted of NHEERL, NCEA and NCCT
personnel with expertise in PBPK modeling and in the application of PBPK models to risk
assessment, although expertise in developmental neurotoxicology and in risk assessment was
retained. The overall approach taken for both the first and the second stages of this evaluation
followed the recommendations for evaluation of PBPK models provided by Clark et al. (2004)
and Chiu et al. (2007).
PBPK model-predicted inhibition of"thyroidal NIS radioiodide uptake by perchlorate was
evaluated for several lifestages. The lifestages evaluated by EPA included the pregnant woman,
fetus, lactating woman, breast-fed infant, bottle-fed infant, 1 year old and 2 year old child,
"average" adult, and non-pregnant woman of child-bearing age. Ciewel! et al. (2007) developed
separate PBPK model codes for the pregnant woman/fetus and for the lactating woman/breast-
fed infant. These model codes were provided to the EPA by the authors of Clewell et al. (2007).
EPA obtained results for the "bottle-fed" neonate by altering the dose specification in the model
for the breast-fed infant. Simulation results for bottle-fed infants were compared to information
contained in a consultative letter transmitted from the US Air Force Research Laboratory
(AFRL) to the EPA (Mattie, 2006). The PBPK model code for the average adult was obtained
from the authors of M errill et al. (2005), while the code for the non-pregnant woman of child-
bearing age was modified by EPA from the pregnant woman code by removing the placental and
fetal compartments, but retaining the mammary compartment.
2. EXAMINATION OF PBPK MODEL COMPUTER CODE
A number of coding errors were found in each model version/file provided to EPA by the
authors and several inconsistencies between the various code files were identified. Except for
those instances noted below and in Appendix A, correction of these errors resulted in only minor
changes in model outputs, and the model codes were still able to reproduce human datasets
shown in Clewell et al. (2007).
An example of a coding error with little quantitative impact relates to NIS inhibition in
tissues other than the thyroid. Inhibition of NIS radioiodide transport by perchlorate was
described for the thyroid in all model codes, but other NIS-containing tissues (e.g.
gastrointestinal tract, skin, mammary gland, placenta, and excretion into breast-milk) were found
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	g	DRAFT
-------
to inconsistently include perchlorate inhibition of radioiodide transport acxoss model codes. The
model code obtained from the authors of Clewell et al. (2007) for human pregnancy included
inhibition of NIS radioiodide transport by perchlorate in the skin and gastrointestinal tract, but
this inhibition was not described mathematically in the model code for the lactating woman,
breast-fed neonate, or young child. Addition by EPA staff of inhibition of NIS radioiodide
transport by perchlorate in the skin and gastrointestinal tract into the code for the pregnant
woman did not significantly impact the kinetics or predictions of percent inhibition of thyroidal
uptake of radioiodide.
In contrast, the radioiodide excretion into breast-milk by NIS was not described in the
lactating woman code as being inhibited by perchlorate, but inclusion of this inhibition markedly
increased the predicted percent inhibition of thyroidal radioiodide uptake in the breast-fed infant
(about 2-fold al lower perchlorate doses and less at higher doses of perchlorate). Discussion with
the Clewell et al (2007) model authors concluded that the competitive inhibition of NIS
radioiodide transport by perchlorate should have been described for all NlS-containing tissues.
Thus, EPA staff added inhibition of radioiodide transport hy perchlorate when it was absent in
the model codes obtained from the authors. These and other model code modifications made by
EPA are described in detail in Appendix A.
EPA has not identified any coding errors that invalidate the use of the overall model
structure of Clewell et al. (2007) for quantitative prediction of perchlorate-mediated competitive
inhibition of thyroidal NIS uptake of radioiodide, although EPA predictions with the corrected
code differ to some extent from those described in Clewell et al. (2007) as a result of those
corrections. EPA has conducted a complete audit of model codes for potential errors. This effort
was in support of EPA5s ln-house analysis of the model code, and a report of the analysis is
attached as Appendix C.
NOTE; A PDF document of the model code modified by EPA and used in this analysis is
available upon request. Please contact Eva D. McLanahan at McLanahan.Eva@epa.aov or 919-
541-1396 to request a copy of the code. Please include your name, affiliation, e-mail address,
phone number, and reason for requesting the code.
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	Q	DRAFT
-------
3. EVALUATION OF PBPK MODEL TECHNICAL APPROACH - DEVELOPMENT
AND MODEL PARAMETERIZATION
3.1. URINARY CLEARANCE
The urinary clearance values for perchlorate and iodide across all lifestages were
determined to be sensitive parameters for prediction of N1S thyroidal iodide uptake inhibition by
perchlorate. Thus, urinary clearance was examined further to determine if the approach used in
Clewell et al. (2007) appropriately represents the available peer-reviewed literature data on
urinary clearance. Details of this evaluation axe found in Appendix B and a brief summary
follows.
For parameters based on human data, the following issues were identified. Clewell et al,
(2007) scaled urinary clearance of perchlorate and iodide by body weight (BW) as a function of
overall metabolism and clearance (BW0"75). This scaling causes urinary clearance per unit of
BW to increase as BW decreases. EPA determined, however, that this relationship does not
accurately describe the reported rate of urinary clearance in neonates. In fact, several indices of
renal function indicate that urinary clearance of perchlorate and iodide in neonates is
considerably slower than is indicated by BW0'75. For example, glomerular filtration rate (GFR;
normalized to surface area) in 1 -week old neonates is 11.0 ± 5.4 mUmin/1.73 m2 wMle in infants
aged 9-12 months GFR is 86.9 ± 8,4 mb'min/1.73 m2 (Goutcz and Norwood, 2005). (Note: A
convention in literature reporting these data is to calculate GFR per unit of body surface area
¦y
(m ) for the tested individuals, but to express the results normalized to a standard adult surface
area (1.73 irf) regardless of the tested individuals age.) Data on urinary elimination of a number
of compounds including drugs and drag metabolites also indicate that renal clearance is slower
per unit of body weight in neonates (Clewell et al., 2002; Dome et al., 2004). Modification of
the PBPK models to describe slower clearance of perchlorate and iodide in neonates
(approximately 50% of adult values when normalized to BW) versus that described in Clewell et
al. (2007) (approximately 200% of adult values when normalized to BW) resulted in an increase
in predicted levels of NIS inhibition in infants at a perchlorate dose-rate of 7 ng/kg-day (amount
ingested by the infant, equal to the point of departure for the RfD). For example, the 7-day-old
bottle-fed infant model RAIU inhibition predictions increased from 1.5% (after other corrections
were made) to 4.4% (Table 2) as urinary clearance decreased.
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	, „	DRAFT
-------
An analysis of urinary excretion data in children (2 -12 years) showed that the default
scaling used by Clewell et al. (2007) fell within the range of the data for cimetidine, whose
primary clearance is renal excretion (Lloyd et al., 1985), but the average for those data was better
described as scaling by BW1, resulting in somewhat lower average predicted clearance (see
Appendix B for details). Therefore, EPA chose to estimate perchlorate-induced inhibition using
scaling of urinary clearance proportional to BW for children at 1 year of age and older, which
results in somewhat higher estimates of iodide uptake inhibition than reported by Clewell et al.
(2007), though still slightly less than predicted for the average adult exposed at the same dose.
(See Appendix B, including Figure B-4, for details.) EPA's estimates of urinary clearance in
infants and children are lower than those used in Clewell et al. (2007), but arc values EPA judges
to be best scientific estimates, not bounds. In particular, the GFR values used for infants were
obtained based on the best estimates of statistical fits to experimental data (Dewoskin and
Thompson, 2008; Guignard et al., 1975).
Data indicating that urinary clearance of iodide by the mother during pregnancy and the
first few months postnatal may be as much as two times higher than in non-pregnant woman
were also identified (Aboul-IChair et al., 1964). These data include radioiodide uptake during
pregnancy and lactation. Details of the PBPK model analyses with data from the Aboul-Khair
study are discussed in Appendix B. Model fits using these data also required the assumption of
higher maternal NIS levels over this period, which may be consistent with the information
indicating higher thyroid activity during pregnancy (Fantz et al., 1999). However, the review
paper by Delange (2004) indicates that observations on urinary clearance changes during
pregnancy are mixed and should not be considered as generally occurring (and not explanatory
for the increased likelihood of maternal iodide deficiency during pregnancy). The existing
model of Clewell et al. (2007) applies maternal urinary clearance constants for iodide and
perchlorate that are about half of the average adult, based on pregnant :non-pregnant comparisons
in rats, which is the opposite of what the data in Aboul-Khair et al. (1964) suggest (but also
inconsistent with conclusions by Delange 2004; see Tables 1 and 2).
Because of the differences in urinary clearance values reported in the literature, EPA
considered three alternatives for pregnancy using: 1) urinary clearance values reported by
Clewell et al. (2007); 2) increased urinary clearance based on the pregnant:non-pregnant ratio
reported, by Aboul-Khair et al. (1964); and 3) urinary clearance constants assumed to be
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	, .	DRAFT
-------
unchanged in pregnancy from the average adult consistent with the observations of Delange
(2004) (see Table 2). In each case, the urinary clearance of perchlorate and iodide were assumed
to vary proportionately; i.e., the ratio of perchlorateriodide clearance constants as estimated in
average adults is maintained (from Merrill et al., 2005). Unfortunately, perchlorate clearance has
not been measured in other human lifestages, so there are no human data with which to validate
this assumed constant proportionality.
Since there are no conclusive human pregnancy data to distinguish among these
alternatives as to which is more likely, EPA selected the lower clearance values reported in the
peer-reviewed, published paper by Clewell et al. (2007) for relative response estimation
(lifestage sensitivity analysis). These lower clearance values were used in producing Tables 1, 3
and 4 below. While this analysis uses the lowest urinary clearance value among the alternatives
evaluated, it does not provide an overall upper-bound effect estimate because the impact of
uncertainty and variability in parameters other than those examined here (e.g., uncertainty in
thyroid NIS parameters and inter-individual variability in urinary excretion) was not evaluated.
For lactation, Clewell et al. (2007) used a clearance rate for iodide equal to the average
adult, but a clearance rate for perchlorate about 40% of the average adult value, again based on
lifestage comparisons in rats. The data of Aboul-Khair et al. (1964) show an iodide clearance
rate that is close to the late-pregnancy value immediately following birth, but that falls to within
control range at postnatal week 12. Thus, EPA again considered three possibilities in a
sensitivity analysis: 1) clearance parameters as used by Clewell et al. (2007); 2) clearance (for
both iodide and perchlorate) higher than the average adult, based on Aboul-Khair el al. (1964);
and 3) clearance equal to non-pregnant "average" values. Simulation results with all three
options are shown below in Table 2. Since it does not seem biologically realistic for perchlorate
clearance to be reduced while iodine clearance is not reduced, EPA decided not to use the lower
perchlorate clearance of Clewell et al. (2007) (option 1). But given the biological uncertainty
between setting both clearance values equal to the average adult (option 3) and the higher
clearance indicated by Aboul-Khair et al. (1964) (option 2), EPA chose to base the model
predictions on the middle of these three possibilities, option 2. The resulting estimates of
perchlorate effects are not "upper bound" values, as this analysis did not address a range of other
uncertainties in the modeling, such as uncertainty thyroid MS parameters and inter-individual
variability in urinary clearance.
This document is a draft for review and public comment purposes only and does not constitute final Agency policy
October 2,2008	,0	DRAFT
-------
3*2. PARAMETER SCALING
For parameters scaled up from rodents, Qewell et al. (2007) scaled permeability-area
cross products (PA) by BW0'75. The use of such scaling is common practice for PBPK. model
parameters describing metabolic clearance, and for such applications has been tested through
applications with a number of chemicals. EPA tested the impact of scaling PA values by BW0 5
or BW1'0 for most tissues and found that this variation had little impact on predictions of NIS
inhibition.
During pregnancy and lactation, the placenta and mammary gland tissues undergo size
changes that are disproportionate to overall BW, In particular, the placenta volume increases
hundreds of times in size over the full course of pregnancy and 16-fold from the end of the first
trimester to the end of pregnancy, while total BW increases only about 10%, So even if one
accepts the scaling power of 0.75 as being correct, assuming that the transport parameters in
these tissues change with total BW, rather than tissue weight, may be viewed as biologically
inappropriate because this assumption leads to the prediction that total NIS levels in the placenta
remain approximately constant as the size of the placenta increases hundreds of times.
Therefore, EPA tested the effect of alternate scaling of the chemical-specific placental constants.
Specifically, EPA tested scaling chemical-specific placental constants by tissue weight rather
than total BW, but this scaling approach was found to only result in minimal quantitative
changes. However, with the model code "as is" from the authors of Ciewell et al (2007), high-
frequency oscillations in fetal levels of perchloratc were predicted at the time when the fetal
thyroid begins to develop, which was presumed to be biologically unrealistic. Changing the
scaling for these constants to depend on tissue weight rather than total BW removed these
oscillations, which were then presumed to result from a numerical instability in the model.
Nevertheless, in keeping with EPA's intent to change only those model components that are
either clear coding errors or have significant quantitative impact, EPA left these quantities as
described in the original code for its reported simulations.
In addition, while the adult woman PBPK code was adjusted by the model authors to
reflect changes in body fat and mammary size during lactation, the equation used to adjust Wood
flow to the mammary in proportion to its size appeared to be in error, since the flow rate set by it
at birth only reflected the mammary volume change (increase) during pregnancy, and not the
This document is a draft for review and public comment purposes only and does not constitute final Agency policy,
October 2,2008	<,	DRAFT
-------
initial (pre-pregnancy) volume. The result is that the mammary blood flow at birth was less than
the pre-birth blood flow, even though the tissue volume was greater. The equation was corrected
to reflect the total tissue volume at the end of pregnancy/birth.
Finally, the equation and scaling used for binding of perchlorate and iodide to blood
proteins are believed to not appropriately reflect the biology. In particular, if the concentration
of the binding protein (Cbp) is constant across age, as stated by the authors, and the blood volume
(Vq) is a constant fraction of BW (i.e., V0 = VCb'BW), as is assumed in the model, then the total
amount of binding protein should scale as blood (i.e., Abp = Cpb*VB - Cpb'VCo'BW1), Since the
maximum rate of blood binding would be expected to be proportional to the total amount of
protein, it then follows that this maximal rate should scale as BW1, not BW0 75, as currently
modeled. However, the impact of changing the scaling coefficient from 0.75 to 1.0 was found to
be minimal, so the scaling was left in the original form for EPA's subsequent analysis.
Blood binding of perchlorate and iodide is described using a Michaelis-Menten equation:
rb,nd = Vmax,b*C/(TCm.b + C), where C is the blood concentration of perchlorate or iodide. In the
case where C is approximately constant over a long period of time, as is expected for dietary
iodide, this equation would result in a constant rate of binding. However, as the amount of
bound material increases, assuming that the total concentration of binding protein is constant, the
amount offree binding protein available would be expected to decrease, which in turn would
cause a decrease in the rate of binding. The Michaelis-Menten equat ion used in the model is
qualitatively inconsistent with that mechanistic expectation. Since changing the scaling of this
rate had minimal effect on EPA's predictions for tracer radioiodide uptake (where blood
concentration is not: constant and declines to near-zero levels over a few days), this matter was
not pursued further and the model was considered adequate for the evaluation of such tracer
kinetics. However, if the models were to describe dietary iodide, rather than the current trace
amounts of radioiodide, describing blood binding using a Michaelis-Menten equation would not
allow for a variable rate of binding, thus limiting EP A's confidence in using these models for
predictions of various intake rates of dietary iodide
3.3. POST-NATAL PBPK MODELING
Procedurally, to account for exposure to perchlorate throughout pregnancy, and hence
that the newborn and mother would carry a level of perchlorate from the moment of birth,
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2»2008	^	DRAFT
-------
simulations for both the mother and fetus should be first run using the pregnancy model to
estimate body levels at birth. However, EPA found that by post-natal day (PND) 7, the
predicted fetal and maternal levels had almost no dependence on pre-birth levels versus
exposures that began after birth. Rapid changes in infant thyroid function in the first few days
immediately following birth also make the model parameters, and hence model predictions of
RAHI inhibition, quite uncertain for those days. Therefore, EPA chose to simulate postnatal
exposure beginning at birth, and to use PND 7 as EPA's earliest prediction.
3.3,1. BREAST-FED INFANT SUCKLING RATE
The suckling rate used by Clcwell et al. (2007) was determined to be an inadequate
description based on data currently available in peer-reviewed literature. The original model
used a table function to describe the baby's suckling rate, which is a volumetric transport rate
(L/h) between the breast milk and baby's stomach, as the route of exposure for perchlorate.
Fairly recent data on breast-milk ingestion rates (Arcus-Arth et al., 2005) indicate that in the first
couple weeks of life, suckling rates are higher than were set by the table function implemented
by Clewell et al., but then fall below that table between 2.5 weeks and several months of age
(Figure 1). Therefore, to improve PBPK model predictions, the suckling rate was altered from
the original approach used by the authors. A smooth function of infant body weight was fit to
the mean ingestion rate data from Arcus-Arth et al. (2005) and implemented in the description of
infant breast-milk ingestion:
Milk ingestion rate = KTRANS = 28.3*(BW - 3.375)0'175 (mL/h).
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	, .	DRAFT
-------
£ 35
E 30 -
r 25 ¦
o
40 r	
oc ' A 7 days - model BVV
a15Y
£
j Cieweffetai
i CieweffetaL
(200?) table
| {unction
~ 10
; 7 clays :
1 fereaa-'w;k data'
= 5 - '
IS ol--
3.36
3.86
4.36
4.86
BW (kg)
Figure I. Breast-milk consumption values used by EPA, Data from Arcus-Arth et al. (2005),
It should also be noted that for breast-fed infant simulations, the intravenous (IV) dose of
radioiodide was treated as being given to the mother, a portion of which passed to the infant
through breast milk. The portion passing to the infant in the absence of percMorale sewed as the
control value. So, in addition to the inhibition of iodide uptake by the infant's thyroid predicted
by the model, the ingestion of perchloratc by the mother inhibited iodide transport into the breast
milk, thus accounting for a perchlorate-induced alteration in nutritional iodide. The reduction of
iodide transport in milk was small in the sense that the predicted reduction was only 1.3-1.5%
when the infant was receiving 7 jjtg/kg-day of perc hi orate (for infants between 7 and 60 days old,
with maternal perchlorate clearance at the average adult value). But this reduction had a close to
additive effect on the predicted reduction of iodide uptake by the infant's thyroid at this dose rate,
which was 2.5% in the 60-day-old bottle-fed infant (see below) but 3.9 % for the breast-fed
infant of the same age (both using the tow infant clearance, based on glomerular filtration, but
adult-average maternal clearance),
3.3,2. BOTTLE-FED INFANT MODEL SIMULATION APPROACH
The model code for the breast-fed infant was modified to allow for exposure to
perchlorate via ingestion of a water-based formula, rather than breast milk. This provided for
direct comparison of a bottle-fed infant with a breast-fed infant using the same PBPK model
structure and parameter set but with, an alternate exposure scenario. Briefly, the lactating
mother's perchlorate dose rate was set equal to zero, which allowed the infant's perchlorate
exposure to be controlled independent of the mother's. The model was coded such that a fixed
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	,.	DRAFT
-------
dose of perchlorate could be administered to the infant or a water concentration could be
multiplied by a formula ingestion rate. Also, a direct IV dose of radioiodide to the bottle-fed
infant was used in the determination of perchlorate inhibition of iodide uptake. The presence of
perchlorate in formula is assumed not to decrease the iodide available to the infant in the
formula. More details of this approach are included in Appendix A.
In addition, while iodide uptake inhibition in the infant was estimated by Clewell et al.
(2007) using a simulated radioiodide injection directly to the infant, that approach would not
account for the effect of maternal perchlorate exposure on iodide ingestion by the infant in breast
milk, as noted above. The model code already described iodide transport to breast milk;
however, EPA added perchlorate inhibition of that transport. Additionally, the model was
extended to allow for transfer of the breast-milk iodide to the breast-fed neonate's stomach
(using the same suckling rate as for perchlorate), and estimated total infant thyroid iodide at 24
hr after simulated IV injection in the mother as the measure of effect, (No such change is needed
for the bottle-fed infant, since in that case the amount of iodide in the formula is presumed to be
unaffected by the presence of perchlorate.) Predicted radioiodide kinetics in infant blood is
much different under this scenario, with a slower rise and fall, and a peak around 12 hours after
maternal injection. Thus, some of the dissimilarity between bottle-fed and breast-fed infant
predictions can be attributed to this difference in radioiodide kinetics.
Finally, to account for the fact that water ingestion will vary with age and BW in the
bottle-fed infant, a smooth function of age was fit to the results ofKahn and Stralka (2008),
which had been plotted against the mid-point for each age range, as shown in Figure 2 (upper
panel, solid line, quadratic equation). However, to account for the minimal ingestion occurring
in the first couple days of life, the equation was multiplied by a rising exponential function:
I - e"day. Note that the function describes the BW-specific ingestion rate (L/kg-day), so it is
multiplied by BW to obtain the total ingestion (L/day) which is then a continuously increasing
function of age, as shown in the lower panel o f Fi gure 2. (The function is only used for
predic tions up to 60 days of age.)
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	.	DRAFT
-------
250
I >>
m
5 200
as
! JC
o
E 150
c
0
ys
g
an
c
1
!|
100
50
Water ingestion
,, ... ^ (direct + indirect)
Ingestion-- exp;,-oay)} >
(increasing from zero o\er:
first few day s after birth} i
-0.0015x': - 0.036SX + 235.9
R2= 1
40
120 160 200
Age (days)
240
_ 1200
j a
5 1000
BOO
\E
c
I 600
oi
¦E 400
200
0
/
• - - y,;'i - exp(C£
T otai water
ingestion
14 21 28 35 42
Age (days)
49 56 63!
Figure 2. 90th Percentile water consumption values used by EPA for the bottle-fed
infant. Upper panel: fit of body-weight-spccific function to 90L" percentile water
ingestion data from Kalin and Stralka (2008); lower panel: total ingestion, after
multiplication by body weight.
A detailed description of issues with parameterization and coding errors in the PBPK.
models and the resolution of these issues is provi ded in Appendix A,
4. EPA-MODIF1ED PBPK MODEL RESULTS AND LIFESTAGE ANALYSIS
4.1. EPA-MODIPIED PBPK MODEL RESULTS
Model predictions obtained with ihe model as modified by EPA are compared to
published values by Cleweli et al. (2007} in Tabic 1. The two sets of values are generally close
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2.2008	, Q	DRAFT
-------
with the exception of the lactating mother, in which model predictions are approximately one
third of the values previously published in the lower dose range. This change is primarily
attributed to the corrections in model code whereby inhibition of iodide transport into maternal
skin and particularly breast tissue/milk by perchlorate was added to the code. Addition of these
terms leads to the prediction that with perchlorate exposure, more iodide is kept in the circulating
maternal blood rather than being transported to skin or transferred to the infant. Hence, more is
available for uptake by the thyroid, reducing the impact of perchlorate inhibition of thyroid
uptake.
Table I: Comparison of Clewell et aL (2007) published model predicted percent inhibition
of thyroidal radioiodide uptake across llfestages with EPA modified versions.*
External
Dose
(mg/kg-day)
Fetus*
{% Inhibition)
Breast-fed
Neonate*
(% Inhibition)
Child c
(% Inhibition)
Pregnant
Woman*
(•/• Inhibition)
Lactating
Woman
{% Inhibition)

Clewell
EPA
Clewell
EPA
Clewell
EPA
Clewell
EPA
Clewell
EPA
0.001
11
1.3
0.9
1.3
0.3
0.3
1
0.95
LI
0.3
o.o i
10
12
8
12
3
2.9
9
8.9
10
2.9
0.1
49
52
34
56
21
23
50
50
54
25
1
84
86
63
92
72
75
91
90
92
78
* For this comparison, infant and maternal urinary clearance rates were set as in Clewell et aL (2007).
" Fetus and pregnant woman shown at gestation week (GW) 38 using clearance values as published in Clewell et al.
(2007) that are equal to about half of the average adult value.
!> Breast-fed neonate shown at post-natal month 1.5; suckling rate was set to the ingestion rate-function fit to the
data of Arcus-Aith et al. (2005); external dose is that ingested by the mother; neonate ingestion (mg/kg-day) is
2.2, 2.1, 1.6, and 0.56 times maternal at external doses of 0.001, 0.01,0.1, and 1 mg/kg-day, respectively, due to
saturation of NIS-mediatcd transport of perchlorate into breast tissue and milk at higher doses.
!' Child shown at 7 years of age and EPA prediction nses "medium" estimate for urinary clearance,
d Lactating woman shown at post-natal day (PND) 7.
Table 2 compares the effects of alternate urinary clearance parameters on EPA-modified
PBPK model predictions of RA1U inhibition for the different lifestages. A decrease in the PBPK
model urinary clearance rate of iodide and perchlorate resulted in increases ofRAIU inhibition
predictions for all lifestages. The largest effect was seen for the near-term fetus (GW40), such
that as the prediction of fetal RAIU inhibition increased from 3.3% inhibition at the highest rate
of clearance to 11% inhibition at the lowest rate of clearance (Table 2). The detailed effects on
RAIU inhibition resulting from EPA model modifications, as described in section 3, are provided
in Appendix A.
This document is a draft for review and public comment purposes only and does not constitutefinal Agency policy,
October 2,2008	, Q	DRAFT
-------
Table 2: Effect of urinary clearance on model predicted percent inhibition of thyroidal radioiodide uptake at the POD (7
|ig/kg-day) for various lifestages using the EPA version of model code.
Urinary
Clearance
Rate
Gestation Model b
Lactation Model
Bottle-Fed Infant^
Older Child"
Pregnant
Woman
Fetus
Lactating Woman'
Breast-Fed Infant''
10 kg
14 kg
GW 40
GW 40
7d
30d
60d
7d
30d
60d
7d
30d
60d
0.97yr
2yr
High
1.7%
3.3%
1.4%
1.6%
1.8%
3.0%
2.9%
2,8%
1,5%
1.5%
1.6%
1.4%
1.4%
Medium "
3.0%
5.3%
2.1%
2.0%
1.9%
3.3%
2.7%
2.6%
2.0%
1.5%
1.3%
1,9%
1.9%
Low
6.3%
1 i%
43%
4.0%
3.9%
6.5%
4.8%
3.5%
4.4%
3,0%
2.5%
2.3%
2.3%
™ Average adult value for urinary clearance was used as a "medium" estimate for the pregnant and lactating woman and the older child, where clue i = 0.11 and
cluc j = 0,125.
h "High" value for urinary clearance was determined from Aboul-Khair et al, (1964) and cluc i = {1 + 0,0703 x GW - 0.0012 x GW2) x 0.11 and clue_p ¦= {1 +
0.0703 x GW-0.0012 x GW2) x 0,125, where GW=40. The Vmax for thyroidal uptake of iodide and perchlorate were adjusted (VmaxTc x 1.8631) to fit
Aboul-Kfaair et al. (1964) maternal thyroid uptake data. Fetus percent inhibition of RAIU was affected by maternal urinary clearance and thus included in the
table; however, no fetal parameters were altered. Maternal MS Vmax values were not readjusted for the "medium" clearance. "Low" value for urinary
clearance was used as published in Clewell et al. (2007) where cluc i = 0.06 and cluc p -= 0.05, determined from the parallelogram parameterization approach.
c The dose of 7 jig/kg-day was provided to the lactating woman, and the breast-fed infant received a concentration in breast milk that corresponded to maternal
intake of 7 jig/kg-day. High clearance rate for the lactating woman was set equal to that of the average adult values from Merrill et al. (2005), but the Vmax
for thyroid perchlorate and iodide uptake was adjusted to fit literature data. Central estimate did not include Vmax adjustment. The low estimate was as
published by Clewell et al. (2007) where the clearance for iodide was equal to the average adult, but the clearance rate for perchlorate was about half (cluc p ~
0.05).
* The breast-fed infant received perchlorate dose from maternal milk that was estimated by the model following maternal ingestion of 7 jig/kg-day. The bottle-
fed infant was simulated using a constant 7 /ig/kg-day perchlorate dose rate. "High" is estimated using BW#',S scaling as published in the Clewell et al. (2007)
model. "Medium" assumes that clearance is equal to GFR; i.e. cluc i = cluc p = (7.5 L/h)/(702'3), with a scaling coefficient of 2/3, "Low" assumes that
clearancc/GFR is the same as in the adult (-40%).
e "High" value was estimated as published in Clewell et al. (2007) using iodide and perchlorate urinary clearance constants equal to the average adult and the
constants were scaled by BW" 7'\ "Medium" used the same constants but scaled by DW1; this sealing described the average renal excretion of cimetidine
(Lloyd et al., 1985) better. The "low1" clearance estimate was estimated by scaling by BW1 and multiplying by the ratio (0.76) of the lower 95% confidence
bound to the mean for the Lloyd ct al. (1985) data.
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2.2008	DRAFT
-------
4.2. LIFESTAGE RELATIVE RESPONSE ANALYSIS
For this document and analysis, sensitivity is defined as the predicted response in percent
RAIU inhibition 24 hours after iodide intravenous injection for an average individual within a
specific subgroup (e.g., bottle-fed infants) relative to the predicted response in percent RAIU
inhibition for an average, non-pregnant adult, where response is the percent RAIU inhibition 24
hours after iodide IV injection.
The PBPK models published by Merrill et al. (2005) and Clewcll et al. (2007) were
modified as described above and in the appendices, and used to estimate the predicted percent
RAIU inhibition for the average adult and different subgroups, including potentially sensitive
subgroups. These estimates were made assuming a dose equal to the point of departure (POD) of
7 /xg/kg-day, which was identified by the National Research Council (NRC, 2005) as a no-
observed-effect- level (NOEL) for the derivation of the RfD and adopted by EP A. Table 3,
column 3 shows the PBPK model predictions of percent inhibition of R AIU at the 7 /ig/kg-day
dose rate. The relative sensitivity of different subgroups was determined by comparing the
percent RAIU inhibition of each subgroup to the percent RAIU inhibition for an average adult at
a dose equal to the POD (Table 3, column 4).
EPA's model predictions may generally be considered central estimates for each
subgroup (at the consumption levels modeled) that account for PK differences, and do not take
into account within-group variability in pharmacokinetics, uncertainty in model parameters and
predictions, or population differences in PD. Fetal simulations are only reported for the end of
gestation (GW 40) as key fetal parameters are considered to be too uncertain for reliable use
earlier in gestation. However, maternal parameters are considered to be more reliable, so
maternal predictions arc also shown for GW 13 and 20.
This document is a draft for review and public comment purposes only and does not constitute final Agency policy,
October 2, 2008	21	DRAFT
-------
Table 3. Model-predicted radioactive iodide uptake (RAIL1) inhibition and relative
sensitivity of different subgroups compared to the average adult at a dose equal to the
point-of-departure (POD) based on EPA's modified PBPK models.
1
Population or lifestage
Body weight
(kgr
RAIU inhibition at the
POD (7 jttg/kg-day)
Relative sensitivity vs.
average adult @ the POD
Average Adult
70
2.1%
1
Woman (child-bearing age)
68
3.1%4
1.5
GW13
Mom: 69
6.6%'"
3.1
GW20
Mom: 71
6.5%r
3.1
GW 40
Mom: 78
Fetus: 3.5
6.3%r
ll%f
3.0
5.3
Mother and
Mom: 74
2.1%rf
0.99
breast-fed infant (7 d)
Infant: 3.6
53%d-e-f
2.9
Mother and
Mom: 73
2.0% a
0.95
breast-fed infant (30 d)
Infant: 4.2
4.3%"'eJ
2.1
Mother and
Mom: 72
1.9%*
0.93
breast-fed infant (60 d)
Infant: 5.0
3,9%da/
1.9
Bottle-fed infant (7 d)
Infant: 3.6
4.4%'
2.1
Bottle-fed infant (30 d)
Infant: 4.2
3.0%''
1.4
Bottle-fed infant (60 d)
Infant: 5.0
2.5%'
1.2
Child (0.97 yif
Child: 10
1.9%*
0.9
Child (2 yr)
Child: 14
1.9%4
0.9
" The body weight (70 kg) for the average adult is the default weight used by the Office of Water (OW). All other
body weights are generated by the model.
h Results were obtained using modified code, in which fetal and placental compartments were removed from the
code for pregnancy. Maternal body weight was held at the value defmed at the start of pregnancy (BW 67.77
kg), and the "average adult' urinary clearance values as published by Merrill et al. (2005) were used.
c Results are based on using the maternal urinary clearance as published in Clewell et al. (2007), which equal to
about half of the average adult clearance.
d Results are based on setting the maternal clearance rates of both perch! orate and iodide during lactation equal to
that of the average adult. Clewell et al. (2007 ) used an iodide clearance rate equal to that of an average adult, but
a perchlorate rate only half that of the average adult.
*	%RAIU inhibition given for the infant is provided based upon a value of urinary clearance scaled from the adult
by BW2" to approximate surface-area scaling, and then multiplied by a rising fraction vs. age based on data
( DeWoskin and Thompson 2008) to reflect the reduction in glomerular filtration rates (see bullet in text for further
details). Clewell et al. (2007) scaled urinary clearance by BW0"75, rather than adjusting based on GFR.
*	These %RAIU inhibition values are based on an internal dose to the breast-fed infant of 7 /xg/kg-day, the same as
for the other subgroups. Maternal dose rates lower than the POD are needed to provide 7 jig/kg-day to the infant
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2.2008	22	DRAFT
-------
(see Table 2 notes), as follows: 7 and 30 days - 3.1 ftg/kg-day; 60 day - 3.4 |ig/kg-day. These doses differ due to
changes in body weights and other PK. factors with age.
s Because OW typically uses a 10 kg child as a default assumption for its health advisories, the model was run for a
child at 0.97 yr, the age at which the model-simulated body weight for a child is 10 kg.
h Results obtained by setting urinary clearance constants for the older child equal to the average adult (Merrill et al.,
2005) and scaling by BW1.
In this analysis, urinary clearance was identified as a key parameter (i.e., model
predictions were highly sensitive to the value used for this variable). Given the range of
uncertainty about urinary clearance during pregnancy and early infancy, the most conservative
value was selected from a range of potential values that were identified, while during lactation
(breast-feeding woman), the middle option (# 2) was selected. (See section 3.1, above, for
details.) However, a full population analysis of urinary clearance was not conducted, and given
that variability in other PK parameters was not addressed, these estimates should not be
considered a true upper confidence bound on RAIU inhibition.
When compared to the average adult, the fetus was identified by EPA's analysis as
the most sensitive subgroup with respect to percent RAIU inhibition at a dose equal to the
POD. This finding is consistent with prior PBFK modeling analyses by Clewell et al. (2007).
The predicted percent RAIU inhibition is approximately 5-fold higher for the fetus at gestational
week 40 than for the average adult. (Simulations at earlier gestation weeks indicate that the fetus
is more sensitive than the adult throughout pregnancy, but are considered too quantitatively
uncertain to assign exact relative sensitivities.)
The same analysis shows that the predicted percent RAIU inhibition is approximately
one- to two-fold higher for the breast-fed and bottle-fed infant (7-60 days) than for the average
adult, and is slightly lower for the 1-2 year old child compared to the average adult.
To the extent that predictions of percent RAIU inhibition for the different subgroups are
close to the average adult, this provides greater confidence in applying the existing RID for these
subgroups, which is based on an uncertainty factor of 10 to account for intra-species variability.
While the difference estimated for fetuses is larger than for other groups, EPA's analyses
indicate that, due to differences in exposure, fetuses whose mothers drink water containing
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	23	DRAFT
-------
perchloratc would be predicted to have somewhat lower predicted RAID inhibition than would
other sensitive subgroups.
4.3. LIFESTAGE COMPARISON FOR THREE DRINKING W ATER
CONCENTRATIONS
EPA evaluated the percent of RARJ inhibition at various water concentrations, with and
without perchlorate intake from food sources, for the different lifestages. The EPA-adjusted
models based on Clewell et al. (2007), as described earlier, were used to simulate three (15,20,
24.5 ppb) drinking water concentrations (Table 4). Available literature was used to estimate
water intake rates for the different lifestages, as well as the dietary contribution to the average
daily dose of perchlorate, as described below.
The water intake rates used for the average adult, non-pregnant woman, and pregnant
woman are based on normalized 90th percentile values for total (direct and indirect) consumers-
only water intake multiplied by the age- or gestation-week-dependent BW. The water intake
rates used to estimate daily perchlorate exposure from drinking water were 0.032 LAg/day
(Kahn and Stralka, 2008; U.S. EPA, 2004) for the average adult and non-pregnant woman and
0.033 LAg/day for the pregnant woman (U.S. EPA, 2004). However, a constant water intake
rate (2.96 L/day, 90th percentile, consumers-only (U.S. EPA, 2004)) for the lactating mother was
used since her BW is expected to decrease during the weeks following pregnancy, while
demands of breast-feeding increase. For the 6- to 12-month and 1- to 2-year-old children, the
water intake rates of 0.971 L/kg-day and 0.674 L/kg-day, respectively, were set based on 90th
percentile values for direct and indirect water consumers-only intake (Kahn and Stralka, 2008).
Additionally, to calculate L/day for these age groups, the corresponding age group mean body
weights obtained from NHANES 1999-2006 were used: 9.2 kg for 6- to 12-month and 11.4 kg
for 1- to 2-year-old children. Using the PBPK model-predicted BW from growth equations, this
approach resulted in model predictions for a 9.6-month old child and a 1.3-year old child. A
different approach was used to estimate the breast- and bottle-fed infant breast milk and formula
intake rates, respectively. Refer to sections 3.3.1 and 3.3.2 for detail regarding these intake rates.
This document is a draftfor review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	24	DRAFT
-------
The dietary closes of perchlorate used correspond to the midpoint of the range of lower -
and upper-bound average perchlorate dietary intakes for each subgroup, as identified from the
FDA TDS (Murray et al., 2001), except for the breast- and bottle-fed infants. The breast-fed
infants arc assumed to have no direct exposure via food or water. The estimates for breast-fed
infants in Table 4 result from the combined food and water dose to the mother providing breast
milk to the infant
EPA used perchlorate concentrations in infant formula based on perchlorate data from the
FDA TDS, The data gathered for 2005-2006 resulted in detection (level of detection (LOD) 1,0
ppb) of perchlorate in 8 of the 12 samples (soy- and- milk based formulas) with a detected
concentration mean of 1.875 ppb	Using H the
LOD for the samples in which perchlorate was not detected, the average is 1.42 ppb. Each of the
12 values represents a composite sample, based on samples collected 4 times a year in 4
geographical locations for S week period and in 3 cities in each region. In addition to the FDA
TDS data, EPA also considered the results of a study by Pearce et al. (2007), Samples of 17
brands of prepared liquid formula analyzed by Pearce et al, (2007) averaged 1.45 ppb
perchlorate, consistent with the FDA TDS information.
Assuming a 90th percentile water ingestion rate of 0.033 L'kg-day and perchlorate intake
from food consumption of 0.1 jxg/kg-day and using the Clewell el al. (2007) PBPK model-fitted
body weight, the pregnant woman's dose of perchlorate was estimated to not exceed the
reference dose if she consumed less than 15 iig/L of perchlorate in drinking water.
There are uncertainties associated with this modeling, as there are for any modeling
effort. For example, this analysis does not take into account within-group variability in PK,
uncertainty in model parameters and predictions, or population differences in PD. Also, the
NRC identified fetuses of pregnant women that are hypothyroid or iodine deficient as the most
sensitive subpopulation. These models were not designed to account for whether the pregnant
woman is hypothyroid or iodine deficient. Model predictions of doses in the various subgroups
apply to a subgroup average for typical, healthy individuals, and effectively describe the RA1U
inhibition relative to that same individual as his/her own control. Some members of a group
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	25	DRAFT
-------
would be expected to have RAIU inhibition greater than indicated in Table 4 for a particular
perchlorate concentration, while others would have lesser inhibition. This would be expected for
fetuses as well as for other subgroups. Likewise, the model does not allow for predictions of how
RAIU inhibition, or the impact of that inhibition, might change with dietary iodide status (i.e., in
an iodide deficient individual, or one with more than sufficient dietary iodide).
There is also some uncertainty regarding the water intake rates, particularly for infants.
EPA described water intake by infants as a smooth function fit to the 90th percentile community
water-consumers intake-rate data (intake per unit BW) of Kahn and Stralka (2008), which is then
multiplied by the age-dependent BW to account for the changes occurring over the first weeks of
life. This resulted in an estimated 90th percentile water intake rate of 0.84 L/day for the 7-day
bottle fed infant and used by EPA in PBPK model simulations. General information on water
and formula intake for 7-day old infants is also available in guidelines for healthy growth and
nutrition of the American Academy of Pediatrics (AAP, 2008). The values estimated using the
guidelines from the AAP (0.126 L/kg-day assuming 80% is the percent water used in preparation
of formula) for 7-day-old infants are close to the mean consumers-only intake rate for the 1-30
day-old infants from Kahn and Stralka (2008; 0.137 L/kg-day N =40).
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	26	DRAFT
-------
Table 4, Predicted percent radioactive Iodide uptake (RAIL) inhibition and corresponding perchlorate intake (pg/kg-day) at
three different water concentrations with and without food intake.

Body
weight
(kg)"
90th
Percentile
% RAID inhibition
TDS
% RAW Inhibition

Water
Intake
(L/dayf
ISftg/L
Water only
20 fJg/L
24,5 pg/L
Food
fHg/kg-da>f
Food ~i-
15 pg/L
Food + Water ,
Food + Food + 1
20 ug/L 24.5 Ug/L
Average adult
('¦: LR ••
70
2.24
0.15
0.20
0,24
0.1
0.IE
0,23
0.27
i >..,
IS on-pregnant
woman
66
2.11
0.21
1.	? \
0,28
0.35
V
0.1
0.26
¦ , ¦.
0.33
0.39
¦
Pregnant woman
Mom -- GW 13
69
2.18
0.49
'¦
0.65
f - ;
0,80
i.i
0.59
0.75
r-
'
0.90
:
Mora - GW 20
i' *i
71
234
0,49
; t -i 1;
0.65
D.l.f.
0.79
0 V
0.1
0.59
(i,. i
0.75
I ! •
0.89
i ,)¦;
Mom — GW 40
78
257
0,47
0.63
0.77
n
0,1
0.57
0.72
0.86
	•:*:	j
Fetus - GW 40
3,5

0.90
1.2
1.5

1.1
1.4
1.6 1
•••
Breast-fed infant
Mom - 7 d
74
236
©.18
-
§,24
<« Kfj
0.29
u II
0.1
0.21
¦ ¦ >
0.27
.i.
			
0.32
"
Infant - 7 d
' / -MX
M
mf
LI
1.5
1.8
d
1.3
1.6
s
2.0
Mom - 6# d
71
136
0,17
0.23
0.28
o.i
0.20

i
Infant — 68 d
S
0.74"
0.73
0.97
1.2
„d
0.84
1.1

Bottle-fed infant
Infant — 7 d
.i,,
3,6
0.84'*
2.0
2.7
f
3.3
1.42 pg/L
2.2
2 J
3.5
Infant - 60 d
W,;1;' k ' -«ll\
5
1.14*
1,3
1.7
2,0
1.42 fig/L
1.4
1.8
2,2
This document is a draft for review and public comment purposes only and does not constitute final Agency policy,
October 2,2008	27	DRAFT
-------
Older child
6-12 mof	9,2 1.03 0,46 0.61	0.75	0,275 0.53 0.68 0.82
	kr_ n 	 '•	¦ ': !	' ! ¦ c	7^1 _ *'1'
" ~""l-2yrf	1L4 0.64 0.23 03 f	'"'0.38	 03?	0.33	 0.41""""
-,sa, (» • : ;	s :?'¦	i •
Calculations for a 70 kg "average" adult are shown, while the tody weight (BW) for the non-pregnant woman is from I :.S, EPA 2004 (based on CSPII 94-
96.98) and BWs for the child are mean values from Kikhn and Stralka (2008). BWs in italics are predicted weights (functions of age or gestation week) using
growth equations from Gentry et al. (2002) as implemented in the PBPK models ((Tew ell ef »L 2007; rum-pirgnant value is BW ai day 0 of gestation),
Water intake levels for adults other than the Isolating mother are based on normalized 90"" percentile values for total waif? intake (direct and indirect)
multiplied by the age- or gestation week dependent BW, as follows. 32 mL/kg day for average adult and non-picgnant woman; 33 ml. kg-day for the pregnant
woman, A fixed ingestion rate was used for the lactating mother because, while her BW is expected to drop during the weeks following the end of pregnancy,
the demands of breast-feeding will be increasing. Values are from Kahn and Stralka (2008 K with the exception of the values for women, which come from
U.S. EPA (2004).
The dietary values used correspond to the midpoint of the range of lower- and upper-bound average perchloratc levels for each subgroup, as identified from the
FDA TDS in Murray ct al. (2008). except for the buttle fed infant. Kl'A used 1.42 fig/L as the concentration of petchloratc in infant formula. This is based on
an average of available FDA TDS data, with lA 1 ()T> included in the average for the samples in which pcrchlnrate was not detected
The breast-fed infants are assumed to have no direct exposure via food or water. The prediction for breast-fed infants in this table results from the dose from
both food and water to the mother providing breast milk to the infant. Breast-fed infant "water intake" is the breast milk ingestion rate obtained by fitting an
age-dependent function to the breast-milk ingestion data (L kg-day) from Arcus-Arth et al. (2005). Urinary clearance rates for the lactating woman equal to
that of the average adult were used, consistent with data presented in Delangc (2004).
Tor the bottle-fed infant, normalized totai water intake {direct and indirect, L/kg-day) was described as a smooth function of infant age fit to the results from
Kahn and Stralka (2008). and multiplied by BW(age). FDA has suggested an alternate approach, using the caloric intake requirement of a 7-day old infant as
the basis for calculating consumption (FDA, 2008). This would likely yield a lower estimate of intake than the 0.14 L/Jay HP A has used in the model.
For the 6- to 12-month and 1- to 2-year-old children, I-PA set the water ingestion based on published exposure tables and .selected the age at which the model-
predicted BW matched the exposure-table mean. This approach resulted in model predictions for 0 7%-yr olds (to represent {>- to 12-month-old children) and
1.285-yr olds (to represent 1- to 2-vear-old children).
This document is a drafi for review and public comment purposes only ami does not constitute final Agency policy.
October 2» 2008	28	DRAFT
-------
5. SUMMARY AND CONCLUSIONS
Detailed examination of CIewell et al. (2007) determined that the model structure is
appropriate for predicting percent inhibition of thyroidal RAIU following perchlorate exposure.
While some coding errors were found, correction of these led to only minor changes in the NIS
inhibition prediction in most cases. Beyond the issue of coding errors, a number of concerns and
questions were raised regarding choices of parameter values in the models for each lifestage.
However, discussions with the authors of Clewell et al. (2007) about the technical basis for some
of the parameter values clarified most such issues. Several issues with model choices made by
the authors have also been noted and tested, but found to not have marked impacts on model
predictions (e.g., the equation and scaling used for binding of perchlorate and iodide to blood
proteins, as described above). In cases that represent differences in scientific judgment rather
than coding errors (some noted in Appendix A) and that result in minimal changes to the results,
the model code has been left as-is. The existing model structure has previously been peer-
reviewed; thus, it was that the most expedient course was not to make these changes because of
their minimal impact.
A few adjustments to the model components and procedures were made if it was
determined that they could have more substantive impacts on the results. For example, the
arrangement of terms in the equations describing transport in blood (mixing of venous blood
streams and arterial blood compartment) was giving rise to numerical instabilities in the
computer implementation. While the software used to solve the model appeared to be robust
enough to handle this instability, such that there was minimal changes in model predictions when
the equations were adjusted to remove them, it was deemed appropriate and better to use the
modified code, without instabilities.
Use of more accurate values for the rate of urinary clearance in neonates led to the
greatest changes in predicted levels ofNIS inhibition relative to those provided by Clewell et al.
(2007). A few adjustments were also made to the lactation/breast-feeding model components
and procedures, including the neonate's rate of milk ingestion. Overall, however, while the
quantitative outputs of the PBPK model as modified by EPA differ from those published in
Clewell et al. (2007) (Table 1), the EPA evaluation determined that, with modifications as
described herein, Clewell et al. (2007) is acceptable to calculate the lifestage differences in the
degree ofNIS inhibition of thyroidal radioiodide uptake at a given level of perchlorate exposure.
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	29	DRAFT
-------
REFERENCES
Aboul-Khair, SA; Crooks, J; TurnbulU AC; Hytlcn, FE. (1964) The physiological changes in
thyroid function during pregnancy. Clin Sci 27:195-207,
Arcus-Arth, A; Krowech, G; Zeise, L. (2005) Breast milk and lipid intake distributions for
assessing cumulative exposure and risk. I Expo Anal Environ Epidemiol l5(4):357-365.
Chin, TW; MacLeod, SM; Fenje, P; Baltodano, A; Edmonds, JF; Soldin, SJ. (1982)
Pharmacokinetics of cimetidine in critically ill children. Pediatr Pharmacol 2:285-92.
Chiu, WA; Barton, HA; DcWoskin, RS; Schlosser, P; Thompson, CM; Sonawane, B; et al.
(2007) Evaluation of physiologically based pharmacokinetic models for use in risk
assessment. J Appl Toxicol 27:218-237.
Clark, LH; Setzer, RW; Barton, HA. (2004) Framework for evaluation of physiologically-based
pharmacokinetic models for use in safety or risk assessment. Risk Anal 24:1697-1717,
Clewell, HJ; Teegiiarden, J; McDonald, T; Sarangapani, R; Lawrence, G; Covington, T; el al.
(2002) Review and evaluation of the potential impact of age- and gender-specific
pharmacokinetic differences on tissue dosimetry. Crit Rev Toxicol 32:329-89.
Clewell, RA; Merrill, EA; Yu, KO; Mafale, DA; Stemer, TR; Mattie, DR; et al. (2003a)
Predicting fetal perchlorate dose and inhibition of iodide kinetics during gestation: A
physiologically-based pharmacokinetic analysis of perchlorate and iodide kinetics in the
rat. Toxicol Sci 73:235-255.
Clewell, RA; Merrill, EA; Yu, KO; Mahle, DA; Stemer, TR; Fisher, JW; et al (2003b)
Predicting neonatal perchlorate dose and inhibition of iodide uptake in the rat during
lactation using pbysio 1 ogical 1 y-based pharmacokinetic modeling. Toxicol Sci 74:416-
436.
Clewell, RA; Merrill, EA; Gcarhart, JM; Robinson, PJ; Stemer, TR; Mattie, DR; et al. (2007)
Perchlorate and radioiodide kinetics across lifestages in the human: Using PBPK models
to predict dosimetry and thyroid inhibition and sensitive subpopulations based on
developmental stage. I Toxicol Environ Health Part A 70:408-428.
Delange. F. (2004) Optimal iodine nutrition during pregnancy, lactation and the neonatal period,
kit J Endocrinol Metabol 2:1-12.
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	30	DRAFT
-------
DeWoskin, RS; Thompson, CM. (2008) Renal clearance parameters for PBPK model analysis of
early lifestage differences in the disposition of environmental toxicants. Regul Toxicol
Pharmacol 51:66-86.
Dome, IL; Walton, K; Renwick, AG. (2004) Human variability in the renal elimination of
foreign compounds and renal excretion-related uncertainty factors for risk assessment.
Food Chem Toxicol 42:275-98.
Fantz, CR; Dagogo-Jack, S; Ladenson, JH; Gronowski, AM, (1999) Thyroid function during
pregnancy. Clin Chem 45:2250-2258.
Gardner, DF; Centor, RM; Utiger, RD. (1988) Effects of low dose oral iodide supplementation
on thyroid function in normal men. Clin Endocrinol (Oxf) 28:283-288.
Gentry, P.R.; Covington, T.R.; Andersen, M.E.; and Clewell, H.J. 2002. Application of a
physiologically-based pharmacokinetic model for isopropanol in the derivation of an
RfD/RfC. Regul Toxicol Pharmacol 36:51-68.
Gilbert, ME; Sui, L. (2008) Developmental Exposure to Perchlorate Alters Synaptic
Transmission in Hippocampus of the Adult Rat. Environ Health Perspect 116:752-760.
Gomez, RA; Norwood, VF. (2005) The kidney in infants and children. In: Greenberg, A;
Cheung, AK: Coffman, TM; Falk, RJ; Jennette, JC; eds. Primer on Kidney Diseases. 4th
edition. Philadelphia, PA: WB Saunders; pp. 420-424.
Greer, MA; Goodman, G; Pleus, RC; Greer, SE. (2002) Health effects assessment for
environmental perchlorate contamination: The dose response for inhibition of thyroidal
radioiodine uptake in humans. Environ Health Perspect 110:927-937.
Guignard, JP; Tornado, A; DaCunha, O; Gautier, E. (1975) Glomerular filtration rate in the first
three weeks of life. J Pediatr 87:268 272.
Kahn, H; Stralka, K. (2008) Estimated daily average per capita water ingestion by child and
adult age categories based on USDA's 1994-96 and 1998 continuing survey of food
intakes by individuals. J Expo Sci Environ Epidemiol: in press. Advance online
publication. May 14, 2008; doi:10.1038/jes.2008.29
Lloyd, CW; Martin, WJ; Taylor, BD; Hauser, AR. (1985) Pharmacokinetics and
pharmacodynamics of cimetidine and metabolites in critically ill children. J Pediatr
107:295-300.
This document is a draft for review and public comment purposes only and does not constitute final Agency policy,
October 2,2008	31	DRAFT
-------
Lorber, M. (2008) Use of a simple pharmacokinetic model to characterize exposure to
perchlorate. J Expo Sci Environ Epidemiol: in press. Advance online publication, April
16, 2008; doi:10.1038/jes.2008.8.
Mattie, DR. (2006) Memorandum from David R. Mattie, AFRL/HEPB Bldg 837, 2729 R Street,
Wright Patterson AFB, OH 45433-5707 to Bruce D. Rodan, Assistant Center
Director/Medical Officer (Research), US EPA/ORD/NCEA (8601D), 4th floor, 808 17th
St., NW, Washington D.C. 2006, April 18, 2006.
Merrill, EA; Clewell, RA; Gearhart, JM; Robinson, PJ; Sterner TR; Yu, KO; et al. (2003) PBPK.
predictions of perchlorate distribution and its effect on thyroid uptake of radioiodide in
the male rat. Toxicol Sci 73:256-269.
Merrill, EA; Clewell, RA; Robinson, PJ; Jarabek, AM; Sterner, TR; Fisher, JW. (2005) PBPK
model for radioactive iodide and perchlorate kinetics and perchlorate-induced inhibition
of iodide uptake in humans. Toxicol Sci 83:25-43.
Murray, CW; Egan, SK; Kim, H; Bern, N; Bolger, PM. (2008) US Food and Drug
Administration's Total Diet Study: Dietary intake of perchlorate and iodine. I Exp
Science Environ Epidemiol: in press. Advance online publication, January 2,2008;
doi:10.1038/sj.jes.7500648
NRC (National Research Council). (2005) Health Implications of Perchlorate Ingestion.
National Research Council of the National Academies. National Academies Press,
Washington, D.C. Available from: dur.y >>> \\t&i'.euu	>.
Pearce, EN; Leung, AM; Blount, BC; Bazraishan, HR; He, X; Pino, S; et al. (2007) Breast milk
iodine and perchlorate concentrations in lactating Boston-area women. J Clin Endocrinol
Metab 92:1673-1677.
Soleimani, M; Xu, J. (2006) SLC26 chloride/base exchangers in the kidney in health and
disease. Semin Nephrol 26:375-385.
U.S. EPA (Environmental Protection Agency). (2004) Estimated Per Capita Water Ingestion and
Body Weight in the United States— an Update: Based on Data Collected by the United
States Department of Agriculture's 1994-96 and 1998 Continuing Survey of Food Intakes
by Individuals. U.S. Environmental Protection Agency, Office of Water, Washington,
D.C., EPA-822-R-00-001. October 2004.
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	3 2	DRAFT
-------
U.S. EPA. (2005) Integrated Risk Information System (IRIS), Perchlorate and Perchlorate Salts.
National Center for Environmental Asssessment, Office of Research and Development,
Washington, DC. February 2005. Available from:
<. v.-,,.. .in,''.-n*";,i ' j'i" .
U.S. FDA (Food and Drug Administration), Center for Food Safety and Applied Nutrition.
(2008). Surv ey Data on Perchlorate in Food: 2005/2006 Total Diet Study Results. Available
from: v.r	sd	•. io4»..al rr.^». Accessed on September 15.2008.
U.S. FDA. (2008) Volume of feeds for infants. Memorandum from Benson M. Silverman, M.D.,
Staff Director, Infant Formula/Medical Foods Staff, Center for Food Safety and Applied
Nutrition, to P. Michael Bolger.
York, RG; Bamett, J; Girard, MF; Mattie, DR; Bekkedal, MV; Garman, RH; et al. (2005a)
Refining the effects observed in a developmental neurobehavioral study of ammonium
perchlorate administered orally in drinking water to rats. II. Behavioral and
neurodevelopment effects, kit J Toxicol 24:451-467.
York, RG; Lewis, E; Brown, WR; Girard, MF; Mattie, DR; Funk, KA; et al. (2005b) Refining
the effects observed in a developmental neurobehavioral study of ammonium perchlorate
administered orally in drinking water to rats. I. Thyroid and reproductive effects. Int J
Toxicol 24:403-418.
This document is a draft far review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	33	DRAFT
-------
APPENDIX A
DESCRIPTION OF MODEL CODE ISSUES AND RESOLUTION
Computer code (acsl language .csl and xmd files) were provided by the authors (R.A.
Clewell and E.A. Merrill) for the average adult (human 10.csl; Merrill et al., 2005), pregnant
woman and fetus (HPregF.csl; Clewell et al., 2007), lactating woman and breast-fed infant
(HLactF.csl; Clewell et al., 2007), and older child (HKidF.csl; Clewell et al., 2007).
Descriptions of specific issues and discrepancies identified in the code or between the code and
model descriptions in the published papers follow, along with the resolutions. At the end of this
appendix is a brief "impact of changes" listing showing the impact of each change or set of
changes on model predictions of inhibition of radio-iodide uptake given exposure at the point-of-
departure (7 ^g/kg-day), to illustrate the quantitative results of these corrections.
Perchlorate Inhibition of Iodide Transport
While reviewing the acsl CSL code provided by the authors (R.A. Clewell and E.A.
Merrill) for the average adult (human lO.csl; Merrill et al., 2005), pregnant woman and fetus
(HPregF.csl; Clewell et al., 2007), lactating woman and breast-fed infant (HLactF.csl; Clewell et
al., 2007), and older child (HKidF.csl; Clewell et al., 2007) several apparent discrepancies
between the manuscript and the code were noted that were related to NIS and Pendrin inhibition
of iodide transport by perchlorate. Understanding of the biology of NIS, along with the
statement, "Inhibition of iodide uptake was included in the maternal, neonatal , and fetal thyroid
follicle and colloid, GI contents, and skin, as well as the maternal placenta, mammary gland, and
milk, based on various literature sources showing inhibition in these tissues in laboratory animals
and humans" in Clewell et al. (2007) lead to the following concerns and actions
(1) Transport of iodide via NTS into GI contents was described with perchlorate
inhibition in the human l O.csl and HPregF.csl code for the average adult and pregnant
woman and fetus, but was not included in the code for lactating woman/infant and
older child.
Action: GI inhibition of NIS iodide transport by perchlorate was added to the
model code for the older child and lactating woman/infant.
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	34	DRAFT
-------
(2)	Transport of iodide via N1S into skin was described with perchlorate inhibition in the
model code for the pregnant woman and fetus, but was not included in the model
code for the other lifestages.
Action: Skin inhibition ofNIS iodide transport was added to model code for
the average adult, older child, and lactating woman/infant.
(3)	(a) Perchlorate inhibition of mammary iodide active transport was not included in the
model code for both the pregnant woman and lactating woman. Also, iodide
transport from mammary tissue to milk was not inhibited by perchlorate in the
lactating woman model code.
(b) Of particular concern is that, qualitatively, one would expect a dual impact on the
breast-fed infant, due to reductions in iodide it receives in breast milk, which the
model code obtained from the authors did not predict due to the lack of an inhibition
in the lactation/milk compartment.
Action: (a) Inhibition of iodide transport by perchlorate into the mammary
tissue was added to the model code for the pregnant woman and lactating
woman, and inhibition o f iodide transport to milk was added in the
lactating woman model code.
(b) The value of Km_Mkp of le6 ng/L for the lactating woman in breast
milk was used. Iodide transfer to the infant was then simulated as the
iodide concentration in breast milk times the suckling rate a clearance
term in the existing maternal model, by adding the term to the infant
*	125
gastric juice compartment. This revised code was used to simulate I
levels in the infant thyroid at 24 hr after maternal I25I-dosing (along with
lactational transfer of perchlorate) to obtain the percent inhibition in the
breast-fed infant.
Tissue and Blood Flow during Pregnancy/Lactation
While the adult woman PBPK code was adjusted by the model authors to reflect changes
in body fat and mammary size during lactation, the equation used to adjust blood flow to the
mammary in proportion to its size appeared to be in error, since the flow rate set by it at birth
only reflected the portion of mammary vo lume resulting from the change (increase) during
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	35	DRAFT
-------
pregnancy, and not the initial (pre-pregnancy) volume to which that increase is added. The result
is that the mammary blood flow at birth was less than the pre-birth blood flow, even though the
tissue volume was greater. The equation was corrected to reflect the total tissue volume at the
end of pregnancy/birth.
Modification to Lactating Woman/Breast-Fed Infant Model to Provide for Bottle-Fed
Infant Simulations
The model code for the lactating woman/ breast-fed infant was modified slightly to
provide model simulations of the bottle-fed infant. Specifically, for bottle-fed infant simulations,
the perchlorate dose to the mother was set equal to zero, then the existing direct-dose rate, either
set PDOSE_N was set > 0 for fixed /ig/kg-day rates and/or added to a fixed concentration
multiplied by the suckling rate (K TRANS; change in code) for fixed water concentrations. In
those simulations the maternal code still ran, but contributed nothing since the maternal dose was
set to zero. An IV dose of iodide to the infant was simulated in order to calculate percent of
thyroidal radioiodide uptake (%RAIU) inhibition in the bottle-fed infant, in contrast to the iodide
dose being received via breast-milk for the calculation if %RAIU inhibition in the breast-fed
infant.
Model Code Errors, Little to No Quantitative Impact on Model Predictions
The binding equation for iodide in the blood for the pregnant woman, lactating woman,
and older child models had the term for concentration of iodide in the arterial blood (Ca_i or
Ca ni) twice in the denominator. This effectively reduced the Km and Vmax for blood binding
by half, which does not affect model predictions at the concentrations tested (within linear range
of blood binding). The extra Ca_i term was removed in these models.
Mass balance was corrected for perchlorate by adding the amount of perchlorate in the
"deep" thyroid, ADT_p, to the total mass in tissue (TM_lp) equation. The iodide mass balance
(BAL i) was corrected by adding the amount bound in blood (ABND_i) to the total mass in
blood (TM_2i) equation and changing QS to QF in the equation for the rate of change in fat
tissue (RAFi).
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	36	DRAFT
-------
In the model code equation for the concentration of iodide in venous blood (CV_i) in the
average adult model, blood binding of iodide had an addition sign (+) instead of multiplication
(*) between the Km and the perchlorate inhibition term; ie,
['Vmax_Bi*CA_i)/(CA_i+Km_Bi+(l_+ Ca_p/Km_Bp))]
instead of
iVmax Bi"CA i)/(CA i+Km Bi*( 1 _+ Ca_p/Km_Bp))].
(The equation for binRAbnd i equation has the correct operator. EPA changed the + to a * and
this had minimal impact on model predictions.
Additionally, blood binding of iodide and perchlorate are described slightly differently in
the average adult model code (human lO.csl) arid in the maternal and neonatal code (IIlactF.csl).
For perchlorate the rate of change of bound perchlorate is subtracted from the equation for the
rate of change in the arterial plasma (RPLAS_p), which determines the concentration in arterial
blood (Ca p), and Ca_p is subsequently used in the Michaelis-Menten (MM) binding equation;
however, for iodide, the rate of change of bound iodide is subtracted from the venous blood
concentration (CV_i) equation and Ca_i is still used in the MM binding equation. EPA notes that
Ca i and Cv i are typically very close if not equal to one another, so this was expected to have
minimal impact on model predictions. That was found to be the case when the term was moved
to the arterial equation. However, making this change did seem to improve computation speed
and stability.
Impacts of Various Changes on Model Predictions
Described below are differences between the percent RA1U inhibition predictions of the
PBPK models as originally published/described by the authors and the percent RAIL" inhibition
predictions now obtained with the models. The differences are illustrated with model predictions
at the POD of 7 /xg/kg-day. The predictions of the models originally published/described are in
the first sub-bullet in each category, and the effects of EPA's changes are noted in the
subsequent sub-bullets. In most cases, only those adjustments that resulted in relatively larger
changes were noted. For each of these lifestages/populations, other technical corrections that arc
not described here and had only minimal effects were also made.
This document & a draft for review and public comm en t purposes only and does not constitute final Agency policy.
October 2,2008	37	DRAFT
-------
Changes in % RAIU inhibition predicted bv the model with changes in code and parameters
•	Average (non-pregnant) adult original model predictions assumed perchlorate in 4 equal
doses at 4-hour intervals, corresponding to the Greer et al. study protocol:
>	3.3% - original value from Table 2 (1" one, page # 8 at bottom), Mattie (2006)
>	3.1 % - assuming continuous/steady-state perchlorate exposure
>	2.1 % adding inhibition of iodide transport in skin (correction)
•	Pregnant woman:
>	6.4% - original value from Table 2 (Mattie, 2006), GW 38
>	5,7% - multiple, small corrections (e.g., + inhibition in mammary), continuous exposure
•	Fetus (exposure is to mother, per total maternal BW):
>	8.6% - original value from Table 2 (Mattie, 2006), GW 38
>	7.6% • multiple, small corrections, as above, GW 38
>	9.9% ^ inhibition at GW 40
•	Lactatrng woman - original model used 4 doses at 4-hr intervals:
>	6,9% - original value, postnatal 30
>	3.8% - adding inhibition of iodide transport in skin and breast milk (corrections)
>	4.1% assume continuous dosing/ingestion; other small changes (e.g., blood binding)
•	Breast-fed neonate (exposure is to mother, per maternal BW):
>	~ 7% - original value (interpolated from Table 1, Mattie (2006)), age 1 month
>	11.6% - including inhibition of iodide transfer in breast milk & maternal skin, other fixes
>	18.5% - revised (reduced) urinary clearance to scale as GFR vs. adult (Appendix B)
>	17,6% - revised lactation expression (Appendix C)
>	20.2% - age 7 days
•	Bottle-fed neonate (since dose-rate fixed,, milk ingestion rate does not impact these):
>	1.3% - original value (interpolated from Table 1, Mattie (2006)), 1 month-old
>	1.2% - small corrections
>	3,0% - revised (reduced) urinary clearance to scale as GFR vs. adult (Appendix B)
>	4.1% - age 7 days
•	Older child:
>	2.1% original value Table 5 (Clewell et al. 2007); 7 year-old
>	1.7% •• w/ inhibition in G1 tract and skin, steady-state exposure simulation
>	2.0% - revised clearance for perchlorate (scale as BW1, Appendix B), 7 year-old
>	1.9% - revised clearance, 2 year-old
>	1,9% - revised clearance, 1 year-old
This document is a draft for review and public comment purposes only and does not constitute final Agency policy,
October 2,2008	38	DRAFT
-------
APPENDIX B
EVALUATION OF URINARY CLEARANCE PARAMETERS
In the PBPK models of Merrill et al. (2005) and Clewell et al. (2007) the urinary
clearance of perchlorate and iodide are described using a common form of alfometric scaling by
body-weight (BW) raised to the 3A power. The actual urinary clearance constant for an
individual of a given BW is given by:
CLUi = CLUQ x BW0 75,	(Eq. Bl)
where CLU* has units of L/fa and "k" is either "P" for perchlorate or "I" for iodide.
Note 1: The tables in the papers identify the units of the CLUCs as L/h/kg, but clearly this
should be I/h/kg0 75 to be consistent with this mathematical formulation, which is how the
CLU values are calculated in the computer code. Moreover the tables list CLUCP = 0.13 (±
0.05 in Merrill et al.) and CLUQ = 0.11 (Merrill et al.) or 0.1 (Clewell et al.). The values as
set in the computational command (,cmd) file were 0.125 and 0.11 for perchlorate and iodide
respectively, so these values will be used below.
These values for CLUQ were determined from PK data in adult humans, so henceforth
they will be identified as the "adult" values.
Note 2: The similar values of CLUQ for perchlorate and iodide suggest that these are handled
similarly by the kidney, as would be expected given their similar charge and diameter. As
will be discussed below, there is evidence for re-uptake activity by the pendrin transporter for
iodide in the kidney, though only significant at lower concentrations, and for perchlorate-
iodide interactions in renal clearance. While this transporter appears to operate on both
iodide and perchlorate, the Vmax for iodide is significantly higher than for perchlorate, and
other tissues where it is explicitly described in the models capture this differential activity.
This difference, and the fact that it appears to have small impact at higher (test) iodide
concentrations corresponds nicely with the small difference in adult CLUQ values: if- 10%
of iodide is actively resorbed, but a much smaller faction of perchlorate, such a difference
would be predicted.
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	3 9	DRAFT
-------
The current model does not include active renal transport per se, but takes renal excretion
to be:
rCLUt = CLUCi x CVKfe	(Eq. B2)
where CVK* is the concentration of k in the venous blood exiting the kidney. Inclusion of active
(saturable) transport would lead to a nonlinear formulation. The error from not including the
active transport is considered to be within the realm of pharmacokinetic (PK) uncertainty and
variability that is not included in the current model applications. So, a revision o f the model to
include it is not proposed . But in application of these results, one should be mindful of the fact
that not all of the inter-individual variabil ity and u ncertain ty in the perchlorate and iodide PK has
been quantified.
Given that this linear formulation is accepted, and the implicit suggestion that renal
clearance is largely controlled by glomerular filtration and non-specific fluid resorption, the
expectation is that the relative clearance for iodide and perchlorate, i.e., CLUj/CLUp, should be
constant across ages, body weights, and lifestages. In EPA's evaluation for the child and
"average" (non-pregnant, non-lactating) adult, this proportionality has been maintained.
Note 3: In Clewell et al. (2007), "The maternal urinary clearance value (Cluci) was set at 60%
of the value in the non-pregnant human based on observed difference in the pregnant and
male rat models (Clewell et al., 2003b; Merrill et al., 2003)." In fact both CLUCi and
CLUCp were set to 0.05 in the pregnant woman (both reduced by ahout the same proportion),
but in the lactating woman, only CLUCP was so reduced while CLUCi was not. These
maternal lactation values go against the argument given just above that the proportionality
should be maintained, but EP A chose to use the maternal values as so set. It is likely
worthwhile to evaluate these maternal values in light of the generally higher urinary
excretion seen in pregnant/lactating women, but alteration of these clearance constants would
require refitting o f other parameters, and so EPA chose not to conduct that specific
evaluation.
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	40	DRAFT
-------
Urinary Clearance in Adults (outside of pregnancy/lactation)
Using the calculations as indicated above, the clearance rate from the Merrill et al. (2005)
model for "average" adults is CLUp = CLUCp x 70°'75 - 3,025 L/fa = 50.4 mL/roin, The average
glomerular filtration rate (GFR) in adults is 125 mL/min - 7.5 L/h, so CLUp/GFR = 40%. For
iodide the values are 2.662 I/h or 443 mL/min, 35% of GFR,
For comparison, Gardner et al. (1988) examined the effects of iodide supplementation in
men, and a plot and regression of their data (urinary clearance vs. blood concentrations) is shown
in Figure B-l. The slope of the regression line, 49.9 mL/min, is quite close to the clearance
value used by Merrill et al, (2005), and if the intercept is forced to zero, the slope reduces to 40.5
mL/min, bracketing that value. However, as indicated by the dashed line, drawn for illustration,
the clearance must be considerably reduced at lower concentrations, assuming that clearance
does not become zero until the blood concentration becomes zero. (The slope of the dashed line
is about 10 mL/min.)
4000
3500 -
E
I 3000
1- 2500 -
2000 -
•2 1500
5 ' 000 -
1 500 -
0 -¦
0	20	40	60	80
lodlcto blood concn (ng/ml)
Figure B-l: Iodide clearance vs. blood levels in men, from Gardner et al. (1988)
What these data and regression indicate is that there is a non-linearity in clearance at low
levels, which could well be due to active re-absorption that becomes saturated at higher
concentrations. The presence of the pendrin transporter in the kidney is noted in the review of
Soleimani and Xu (2006).
While EPA does not propose changing the model to explicitly include active transport in
the kidney and thereby describe this nonlinearity in excretion, it is noted here as a source of
y = 49,87% - 450.
R => 0.
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	41	DRAFT
-------
uncertainty or variability in model predictions: clearance values obtained at high concentrations
(clinical experiments) might not be completely predictive of values at lower concentrations.
The presence of a transporter, (more) active towards iodide, could explain the slightly
lower clearance of iodide vs. perchlorate in the adult It is assumed that this transporter acts at a
similar proportion of activity in all Mfestages, and hence that the ratio of perchlorateriodide
transport is approximately constant. Likewise, it is assumed (as one means of estimation) that
the clearance rates remain at about 40 and 35% of GFR for perchlorate and iodide, respectively.
Urinary Clearance la tie Neonate
Note 4: The analysis here focuses on clearance of perchlorate, but as indicated above, iodide
clearance was always changed in parallel to maintain the ratio of 0.11/0.125. Further, for
each alternate value of CLUi evaluated, the Vmax for NlS-mediated uptake of both iodide
and perchlorate in the follicle (from blood to follicle tissue) was adjusted to maintain model
fits to radio-iodide uptake (RAIll) data (available for infants and other life-stages in the
absence of perchlorate).
Data from Guignard et al. (1975) on GFR m infants (age 1-25 days) with a linear
regression is shown in Figure B-l. As a basis for comparison, EPA will consider the clearance
of a 3.6-kg child, the (average) weight predicted by the model to occur at 7 days of age. Based
on the regression shown below, GFR at that age/body weight is 3.557 mL/min or 0.21 L/h. As
implemented in Clewell et al (2007), the clearance of perchlorate is predicted to be CLL> =
0.125 x 3.6°75 = 0.33 L/h. Clearly this value of CLUp does not fit with the assumptions on
clearance/GFR stated above; the only way in which clearance could be higher than GFR is if
there is active excretion of iodide with no or substantially reduced resorption. EPA is aware of
no data on renal transporters during infancy to suggest the level and pattern of expression
changes required to bring about such an effect. If instead an assumption is made that perchlorate
clearance was 40% of GFR, as it is in the adult, the value one would obtain at 3.6 kg BW is
0.085 L/h, almost 4 times lower than the default extrapolation.
Since the intent is to account for BW changes in a convenient way, it should be noted that
GFR is typically normalized to body surface area, as this scaling has been found to explain much
of inter-individual variability. Even with such normalization GFR is below adult levels near
litis document is a draft for review and public comment purposes only and does no! constitute final Agency policy
October 2,2008	42	DRAFT
-------
birth, rising toward adult values over the first few months (DeWoskin and Thompson, 2008),
Therefore, consideration was given to scaling of renal excretion for infants by BW*', as an
approximation of surface area normalization, and using the average values compiled by
DeWoskin and Thompson (2008) for different age ranges. Values for the ratio of normalized
GFR in infants vs. adults (ratio mean value to adult, plotted vs. mid-point of each age range) are
shown in Figure B-3, along with a simple power-function curve fit, SGR = 0.2087 x day0 ,
where "day" is the child's age in days. This function was used, together with BW2/3 scaling to
estimate urinary clearance for perchlorate in the infant as:
CLUpchild = SGR x CLUpaduSt x
BW,
child
BW.
%
adult -
0.23333
= 0.2087 x day0'23333 x 3.025 x, 		,
'	I 70 i
Clearance for iodide is similarly calculated from the adult clearance value.
fBWI1MY^
(Eq. B3)
C
I
E,
at
©
10
9
8
7
6
5
4
3
2
1
0
A Guignard etal. date
~ GFR-based scaling
—Linear {Guignard etal, date)
3	4
BW (kg)
y = 0.7S84x + 0.7908
R2 = 0.3704
Figure B-2: Glomerular filtration rate (GFR) vs. body weight (BW) in infants (Gnignard el
ai, 1975). GFR-based scaling uses equation (B1 with total adult GFR of 125 mlAiin vs. 3.025
L/h perchlorate clearance) with individual age and BW values of Guignard et al. (1975).
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	43	DRAFT
-------
_ 0.8
= 0.2087x0'2333
R2 = 0.9865
0
50
100
150
Average ag# (days}
Figure B-3: Ratio of surface-area normalized glomerular filtration rate (GFR, ral/rain/SA)
in infants vs. adults, as a function of age. Data from DeWoskin and Thompson (2008).
The result of applying equation (Bl) to the individual data of Guignard et al. (1975), but
using the total adult GFR of 125 mL/mim (7.5 L/h) instead of the iodide clearance of 3.025 L/h,
is shown in Figure B-2. The result is not a smooth function of BW because of the variation in
BW vs. age in that data set. However, one can sec that the resulting predictions arc generally
higher than the observations in that particular data set. This is not surprising since the average
clearance values of Guignard et al. (1975) are lower than many of the other results included in
the calculation of geometric means for each age range by DeWoskin and Thompson (2008), on
which the multiplicative function shown in Figure B-3 and used in equation B3 is based.
Likewise, applying equation B3 tor a 7-dav-old, 3.6-kg infant, one obtains:
Compared to the GFR of 0.21 L/h, this seems reasonable, although it is 67% of GFR rather than
40%. At 60 days, when an average child is 5 kg, equation B3 yields 0.28 L/h or 4.7 mL/min,
which again appears reasonable in comparison to the data in Figure B-l, again noting that this is
iodide clearance rather than total GFR. While the data shown above are for children below 25
days, EPA therefore extends its extrapolation to a 60-day-old, 5 kg child, though with greater
uncertainty at that age (since renal clearance does rise rapidly during that time). However, the
estimates obtained up until 30 days are expected to be fairly sound.
CUJem - 0.2087 x 7°'23333 x 3.025 x
(1A
i 70)
- 0.14 L/h
(Eq. B4)
This docitment is a draft for review and public comment purpose? only and drns not constitute final Agency policy.
October 2,2008	44	DRAFT
-------
Urinary Clearance in the Older Child
For older children, consideration was given to the data of Chin et al. (1982) and Lloyd et
al. (1985) for cimetidinc, which is primarily cleared by urinary excretion, The subjects were
children being treated primarily for close-head injuries ("secondary to motor-vehicle accident"),
and EPA restricted the Lloyd et al. data (larger data-set) analysis to only that injury category
(excluding a few cases of sepsis, for example) and ages less than 12 yeaxs (youngest was 4.1
years). The Chin et al, (1982) data are included because it includes children as young as 1 year
old, though fewer subjects. Unlike neonates, from these data it appears that either direct BW
scaling, or normalization, or scaling by BWr"75, may be appropriate. A plot of clearance/BW (y-
axis) vs, BW is shown in Figure B-4, with lines indicating BW1 (constant, dash-dot horizontal
line with this normalization), BW0'75 and BW2''3 scaling from adult values.
25
_ 20
o»
« !
f 15 ~
t »
s 10
m
1 5
O
0
3W-\2,3)
-- ..	
Dalai average —	—	- —-—~»-~—
-p. -j
Adult average-¦
Lower 9	stai ^
O
10
20	30
BW (kg)
40
Figure B-4: Ciraetidine clearance data (Chin et al., 1982 [red squares! and Lloyd et aL,
1985 [yellow diamonds]) and possible scaling relationships. BW075 and BW3"'3 curves are
normalized to BW after applying this scaling.
Lloyd et al. (1985) state that the range of clearance rates in adults is 9-103 ml/mia'kg,
and the average of these two values is shown for comparison (with no trend vs. BW) as well as
the average of the data shown, and the results of using BW0,75 and BW2''3 to scale from that
approximate adult average. It can be seen that the data of Lloyd et al {normalized to BW1) show
little residual trend vs. BW, although the more limited data of Chin ct al. (1982) show a
This document is a draft far review and public comment purposes only and does not constitute final Agency polity.
October 2,2008	45	DRAFT
-------
downward trend similar to the results of scaling by BW0'75 and BW2'3. (Regression of the Lloyd
et al. (1985) data yields a slope that is negative, but is not significantly different from zero.)
While the allometrically scaled relationships are clearly within the range of the data and
may be considered a reasonable estimation, the closeness of the normalized data average to that
for adults and the fact that the BW0Jy scaling falls above the data average suggest that simple
scaling of clearance (Clu) by BW1 better describes the data over much of the range. In the face
of the variability shown by these data and lack of clear fit by any of these functions, EPA chose
to represent the average clearance in "older" (>1 year of age) children by scaling adult clearance
values by BW1, although this relationship may be low for younger children. The results of using
BW0"73 scaling, as in the original publication of Clewell et al. (2007) were also shown in Table 2
as representing a "high" clearance values, and the results of scaling by BW1 but multiplying by
the ratio ( 0.76) of the lower 95% confidence bound to the mean for the Lloyd et al. (1985) data
are shown for the "low" clearance values.
Urinary Clearance in Pregnancy and Lactation
Clewell et al. (2007) estimated clearance of perchlorate and iodide during pregnancy and
lactation based on parallel changes in the rat (vs. average adult), obtaining clearance for both
compounds of about half the average adult values during pregnancy. During lactation this
approach lead to half of theaverage adult for perchlorate, but a value for iodide equal to adults.
The data for iodide clearance in humans of Aboul-Khair et al. (1964) during pregnancy and early
postnatal times, shown below, was also considered.
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	46	DRAFT
-------
RENAL CLEARANCE OF IODINE
IN PHfCC,NAHCV
100,
POST - TA f(T U M
VKEGNftMCY
CONTROL MKAN t ? S. K
WKKKS
Figure B-5: Renal clearance of iodine (mean ± 2 S.E.) in pregnancy and post-par turn period
compared with non-pregnant values. From Aboul-Khair et al. (1964) Figure 2,
These data show that renal (urinary) clearance for Iodide is elevated to as much as 2-
timcs control (non-pregnant) values during pregnancy, and while this declines fairly rapidly
towards control after birth, it is still elevated in the first couple of months, where EPA's analysis
on neonatal clearance has focused attention. Keeping with the assumed proportionality between
perchlorate and iodide, based on these data the same relationship would be expected to hold:
higher clearance rather than reduced. A dilemma occurs in considering the data of Aboul-Khair
et al. (1964); however, in that the control iodine clearance as measured by them is 31.05 ± 3.66
mL/min (mean ± SE), while the value determined by Merrill et al. (2005) for non-pregnant adults
is 44.3 mL/min. Likewise Aboul-Khair ct al. (1964) report thyroid iodide uptake at 2.5 hr post-
injection as 21 4 ± 1.4 % of the administered dose, but the amount predicted by the Merrill et al.
(2005) model (in the absence of perchlorate) is 7.78%. Therefore, the data of Aboul-KJhair et al
(1964) was normalized to their own controls for both urinary clearance and iodide uptake, and
then use that relative change as a model input (for clearance, multiplying the non-pregnant
clearance rate constant by the pregnant :control ratio from Aboul-Khair et al. (1964)) or in
estimating changes in thyroid NIS (to fit relative increases in thyroid uptake).
This document is a draft for review and public comment purposes only ami does not constitute final Agency policy.
October 2.2008	47	DRAFT
-------
The urinary iodine clearance data from Aboul-Khair ct al. (1964) for pregnancy and 1
week post-partum. with a quadratic interpolation function, are shown in Figure B-6. A quadratic
function was likewise fit to the data for the early postnatal period (along with the last gestational
data point) as shown in Figure B-7. The latter function was only used up to 60 days (8.6 weeks)
of infant age, since the data indicate a decline toward control values after that point.
y = -0.0012x* + 0.0703x<-1
R2 = 0.5379
27
38
9
18
45
0
Gestation week
Figure B-6. Relative pregnant:non-pregnaat iodide clearance values from Aboul
Khair et al. (1964), with quadratic interpolation function. Points are mean ± SE.
y = 0.OO82X1' -0.1645x + 19186
R2 = 0.9948
GW 39
(Limit of fwicriau
3	B
Postnatal week
•* S>.
Figure B-7: Maternal iodine urinary clearance and approximation function for early
postnatal period; data from Aboul-Khair et al. (1%4).
This document k a draft for ri%4m> ami public comment purptses only and does not constitute final Agency policy
October 2,2008	48	DRAFT
-------
For pregnancy, the Vmax values for maternal thyroid NlS-mediated uptake of perchlorate
and iodide were adjusted specifically to fit iodide uptake data of Aboul-Khair et aL (1964)
collected at the same pregnancy time-points as the urinary clearance data, A multiplier function,
R Vmax (pregnane y) - 0.0009-GW2 - 0.054-GW + 2.6,	(Eq, B5)
was used to adjust both the perchlorate and iodide values. The fit to these iodide uptake data,
given the increased urinary clearance as shown in Figure B-6 and the fitted quadratic for
increased maternal NIS is shown in Figure B-8.
Gestation w««k
Figure B-8: Thyroid iodide uptake at 2.5-hr post IV Injection relative to control. Data
(points) are from Aboul-Khair et aJ. (1964) (mean ± SE). Line is model simulation. Note that
while values drop from GW 12 to 32, they are consistently greater than one.
During the post-part urn period the urinary clearance and iodide uptake data of Aboul-
Khair et al. (1964) are both falling towards control values, but again there is the situation that the
control uptake measured, 21,4% of the IV dose at 2.5 hr post-injection, is well above the value
estimated by the PBPK model for an average adult: 7.8%. Further, a number of the
physiological parameters differ in the 1 acta ting woman model vs. the average adult, as well as
over time. Therefore, EPA first ran the lactating woman model, using the average adult
clearance constants (that scale by BW0 71) at postnatal week 50 to estimate the model-control
maternal uptake at 2.5 hr post iodide injection: 6.57%. Then to evaluate the impact of using
these observations, the NIS levels during this period were adjusted by assuming that as
clearance falls from about 2 times control values at GW 39 (1 week prior to birth) towards
control, the NIS Vmax values follow suit, dropping to the values of Merrill et al. (2005) for the
Th-is document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	49	DRAFT
-------
adult. In particular, if RCLF is the fold increase of urinary clearance over control, then the NIS
Vmax was multiplied by:
RVmax(lactation) = (RCLF - l)2 + 1,	(Eq, B6)
Thus, for RCLF = 2.09 (from the equation in Figure B-7) at GW 39 (postnatal week "-1"),
RVmax equals 2 and as RCLF falls towards 1.0 (i.e., clearance approaches control values),
RVmax also falls to 1.0, so Vmax values will approach controls. A plot of the RAIU uptake
measured by Aboul-Khair et al. (1964) and the simulated values resulting from use of this
function is shown in Figure B-9, where both have been normalized by their respective control
values.
1.6
s 1'5
c 1.4
1
jg 1.3
M
a
~
1
S5
1.2
1.1
1
0.9
~
t '
GW 39
-2
4 6 8
Postnatal week
10
12
14
Figure B-9: Radio-iodide uptake in late pregnancy and early postnatal period. Data are
from Aboul-Khair et al. (1964). PBPK simulations are with adjusted lactation model (see text
above).
This document is a draft for review and public comment purposes only and does not constitute final Agency policy,
October 2,2008	50	DRAFT
-------
APPENDIX C
MODEL REVIEW FINAL REPORT FROM EPA CONTRACTOR
Center for Biological Monitoring & Modeling
September 25, 2008
Report for Work Assignment 4-5
Evaluation of Perchlorate PBPK Model
Submitted To:
Robert DeWoskin, PhD, DABT
US EPA/NCBA
Research Triangle Park, NC 27711
Principal Investigator:
Paul M. Hinderliter, Ph.D.
ph: 509-376-3907
fax: 509-376-9064
e-mail: [/ii-j.Mi'.j"'.
Address:
Battelie, Pacific Northwest Division
902 Battelie Blvd.
P.O, Box 999, MS P7-59
Richland, WA 99352
This document is a draft for review and public comment purposes only ami does not constitute final Agency policy.
October 2,2008	51	DRAFT
-------
Overview
Battelle received PBPK models from EPA that were revised from Cleweli et al. (based on the
2007 manuscript Pcrchloraie and Radioiodide Kinetics Across Life Stages in the Human: Usimr
PBPK Models to Predict Dosimetry and Thyroid Inhibition and Sensitive Suboooulations Based
on Developmental Stave) and performed the following tasks:
•	Evaluated model code for internal consistency
•	Digitized Figures from published manuscripts
•	Compared manuscript figures to current ACSL model outputs
Results
The check of the model code found no outstanding coding discrepancies beyond those corrected
by EPA staff (as noted in the code/comments of the model files). Additionally, the EPA staff
corrections (as identified by comments in the code) all appear to appropriately result in code
equations which now reflect the model as described in the manuscript.
Model Checked;
Lactational Model HlactFrev.csl
Pregnancy Model HPregF_Yjpms2.csl
The model code has been significantly revised by EPA staff to correct mistakes (typos) in
equations, harmonize model code with statements in the manuscript, clean model code for
readability, and reduce model run irregularities (i.e. long simulation times). Extensive model
checking by Battelle was conducted on. prior versions of the csl files. The current csl files were
also checked to verify that corrections/additions were properly implemented. The m-files for
producing the figures are attached in the Appendix.
From the 2007 Cleweli paper, the following figures were analyzed:
Figure 5 Thyroid of newborns;
Figure 6 - Maternal Concentrations;
Figure 7 - Total fetal burden;
Figure 8 - Lactatmg women; and
Figure 9 - Neonatal urine.
EPA staff also provided additional parameter values and some adjusted m-files for some of the
simulations as noted m the discussion of each figure.
In addition to the output of the current model, simulation lines presented in the original
manuscript were digitized (using Digitizelt, share-it! - Digital River, Eden Prairie, MN) and axe
presented in most of the figures below.
In general, the model simulations were similar to, but occasionally not identical to, the published
model results. This may be due to modifications made by Cleweli et al. after publication or
EPA's corrections to the submitted model.
This document is a draft for review ami public comment purposes only and does nor constitute final Agency policy.
October 2,2008	52	DRAFT
-------
Figure 5 - Thyroid of newborns;
Produced using the pregnancy model and the Fig5_GFR.m file,
o The figure presents model-simulated thyroid radioiodide uptake 24 fa postdosing
in the newborn infant,
o The 3D nature of the manuscript figure makes it difficult to determine the exact
values of the points.
o Since the experimental data exists only at 30 hour after dosing (for each dose
group), only this tiiriupoiiit was simulated,
o EPA staff provided additional constants not documented in the original model
code (see attached Fig5_GFR.m file),
o The simulation closely represents the published simulations and the experimental
data.
Figure 5 Clewsll 2007
100
80
I SO -
Q>
"O
=f40
20
~experimental data
< model output
3	4
Neonate Age (days)
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	53	DRAFT
-------
Flgure 6 Maternal Oonccntrstions
• Produced using the pregnancy model and the Fig6.m file.
o The figure presents predicted radioiodide concentration in maternal (A) thyroid,
(B) urine, (C) whole blood, and (D) placenta,
o The experimental data in the m files matches that presented in the manuscript
with a few minor differences for both the mean and max Vmax values.
Figure 6a Cleweil 2007
70
60

Manuscript scan - mean
Vmax
Manuscript scan - ma*
Vmax
* exper daia

Model Mean Vmax
5
00^
I
Mcoei Max Vmax
*¦1/ •

0
5
10
15
20
25
30
Time After Dose (hr)
This document is a draft for review and public comment purposes only and does no! constitute final Agency policy.
October 2.2008	54	DRAFT
-------
Figure 6b Clewetl 2007
/

10
—t ^	
20	30
Time After Dose (hr)
40
50
Figure 0c CSeweli 200?
- - Manuscript scan - mean
Vmax
	Manuscript scan - ma*
Vmax
• exper data
• Model Mean Vmax
• Model Max Vmax
1
0 01
20
30
40
50
0
10
Time After Dose (hr)
—	Manuscript scan - mean
Vmax
	Manuscript scan - max
Vmax
• exper data
— Mode! Mean Vmax
• !v*xJ«l Max Vmax
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	55	DRAFT
-------
Figure 6d Clewel! 2007
10
01
0 01 +-
0	m	20	30
Tme AfSf,- Dose «nn
¦ • y,anu'-cf!W scan mean
v'max
	Manuscript scan- max
Vtnax
~ expe? caf.i
! — Mooet f"eafi Vma<
i
; • Moae! Ma* Vmax
This document is a draft for review and public comment purpose* otdy and does not constitute final Agency policy,
October 2,2008	56	DRAFT
-------
Figure 7 Total fetal burden
• Produced using the pregnancy model and the Fig7„m file,
o The figure presents Total fetal 131F burden,
o Experimental data appears to match (except for an anomalous value in the
manuscript day 15 panel, at 20 hours and 90+ percent iodide in the fetus),
c The simulation and publication differ slightly in the rate of elimination with the
acsiXtreme code showing slightly faster elimination.
Figure ? - GW 13 Cteweli 2007
0 25
• experimental uaia
— Model Output
02
¦' Scanned Wanuscnpt Simulation
015
01
0 05
0
• 60
0
m
40
m
80
too
140
Fime alter dose (hours)
This document a a draft for review and public comment purposes only and does not constitute final Agmey policy,
October 2.2008	5 7	DRAFT
-------
FifUW 7 <• GW 14, Ctewnil 2007
# 015
i • Experimental Data
—— Model Output
	Scanned Manuscript Simulation
80	BO
Time after dose {hours}
120
140
OS
£
m
Sni
c
*
1
02
I •
Figure 7. GW 15. Cteweli 2007
08 ¦

Experimental Data
• Model Output
Scanned Menusaipt Simulalicn
.T.1• * * ff i r^Wiifluii
20	so	m »
Trttte after toe (hours)
J 00
f20
M0
i§e
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	58	DRAFT
-------
Figure 8 - Lactating women
• Produced using the lactational model and the FIg7,m file.
o The figure presents predicted radioiodide concentration in the (A) thyroid, (B)
urine, and (C) breast milk of lactating women.,
o The experimental data in the m files matches that presented in the manuscript
(although there is data at longer post-dosing times in the m file not presented in
- the paper). EPA staff also supplied a modified m-file containing additional
parameters.
o The model simulations are close but it appears that some parameters have slightly
changed.
Figure 8a Cleweii 2007
8
©
TJ
£
-o
t
SO
50
40
30
20
10

Manuscript scan - mean
Vmax
Manuscript scan - max
Vmax
Mode! Mean Vmax
Model Max Vmax
exper. data
10
20	30
Time After Dose (hr)
40
50
This document is a draft for review and public comment purposes only anil does not constitute final Agency policy.
October 2,2008	59	DRAFT
-------
Figure ib Ctewel 200?
80
70
80
150
-------
Figure 9 - Neonatal Urine
• Produced using the lactational model and the Fig9_GFR.ni file,
o The figure presents Predicted radioiodidc in neonatal urine after a direct oral n'l
dose.
o The experimental data in the m files matches that presented in the manuscript
o The model simulations closely match the simulations presented in the manuscript.
Figure S Clews; 200?
2 Days	3 Days	i Year
Additional Items
•	Originally the pregnancy model had to run its init routine twice to get acceptable values,
EPA staff fixed this issue by correcting the order of setting constant values in the m-files.
The model currently does not use the init routine supplied by the original authors; rather
constants arc set in the csl and m-files.
•	The m-files should be fully converted away from the old command language for
consistency,
•	Many of the figures are now easily produced by running m-files provided by the EPA
staff.
•	Given the evolution of the model, simulation of figures from the author's older papers on
this model was attempted but the number of changes made this a difficult comparison.
This document a a draft for re*4ew and public comment purposes only and does not constitute final Agency policy,
October 2,2008	61	DRAFT
-------
Appendix: m-files for figures
_______
WESITG-O; WEDITG=0;
output eclear
prepare ©Clear DAYS CANI CTTOT_NI
nio_upt= [] } CINT=,1;
ACLU-0.75; RU.1.0; VCHNG-0.0; AKT=1; NDRNX=1;
cluc_i-C . 11; clu,c_p«0 .12 5 ; %adu.lt values
CLOC_NI = c luc_ i * RU * (70a (0 .75 - ACUJ) ) ;
CLUC_STP»cluc_p*RU*(70"[0,75 - ACLU));
PDOSE = 0 r IVD0SE_I = 0; IVDOSE__NI = 5 0 ; PPB-C ; PDOSE N = 0 ; CONC-O; DOSE RI	0 ;
for day=[1:5,7]
IVSTART I* (day-IS *24 ; TSTQP=IVSTART_l+24 ; IVSTRTNI = I VSTART_I ;
start -Snocallback
nio_upt = [oioupt; [day, BW, BW_K, 10 0*ATTOT _NI/1VDCSE_NI] j
end
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	62	DRAFT
-------
tuateraal iodide vs. contained literature data
~corresponds to Figure 6 ir. JTBH, 200S
IDAT* Miodide (t, «u i, attot_i, cacot_i, cpl_i)
MiodideD^
= [3699
NaN
HaN
16
3599
NaS
NaN
1.8
««
3699
Nas-
STaM
1.3
7
3700
sau
11
BaU
MaH
3700
HaJS
12
NaN
HaH
3700
NaN
18
NaN
NaN
3700
Half
10
SaH
HaH
3700
NaN
18
Hall
KaN
3702
NaU
Natl
0.9
0.4
3702
HaS
SaH
,08
0.5
3702
NaM
tfa»
1.3
0.7
3102
HaN
HaS
0.5
0.3
3702

SaH
1.0
0 . s
3708
NaN
Mas
0 .10
0.2
3708
NaN
NdN
0.19 0
,22

3 712
53
24
Mail

0 .IS
3714
NaN'
HaS
0.27 0
.32

3714
NaN*
MaS
0.39 0
.328

3714
NaV
NaN
NaH

0
3714
NaN
Natl
HaK

0..
3720
58
€6
0.1
0


3720
28
30
0.13
0.
14

3720
61
22
0.11
£
I

3720
55
22.?
0.07
0.
03

3720
59
26
0.0
HaN

3720
S6
26.B
SaH

MaN
3720
13
29
Has

SaM
3720
56
30
HaH

N*N
3720
MaS
26.4
HaN


Hall
3720
NaN
30.8
HaN


NaH
3720
Naff
29 .8
MaU


Saw
3720
NsS
35
NaN


BaS
3720
NaN
41
NaH


NaM
3720
NaN
19
SaH


NaM
3744
65
NaH
0.2

0.
06
3744
62
NaH
0 .14

0.
06
3744
56
NaH
0 .04

0.
12
3744
62
HaK
0.31

0.
07
3744
57
HaN
0.14

0

3744
21
HaB
HaN


0
3744
62
HaM
MaB


0
3792
NaS
NaN
0.2 Nail


3792
NaN
HaN
0.22 NaJM


3840
NaS
NaH
0.54 MaH


3840
HaK
NaH
0.03 NaH'


%Fig6.m
%Plot Figure 8
DCSi_:-10 0; TSTART= 30 00;1VSTART_I=£9 £;TSTOP=6 0 0;
VMAXC_TI=1.22e5;
%VCHNG=1.7; Sf»Lft=0;
prepare ATTOT_I AU_I CATOT I CPL T T
a t ar t <*noc a 11 back
via._irrot_i, v2a- au i, v3a=_catot_i, v4a=_cpl_i, rl- r.
VKRXC_TI=S.S2eS;	Imaximura vnax used in m«rrill ee al.
startsnocallback;
vlh»_ a'.r.ot v2b=__au._i, v3b=_catot_i, v4b«_cpl i, r.2- t
plot	{tl,via,t2,vlb,MiodideD(
plot	{tl,v2a,t2,v2b,MiodideD(
plot	(tl,v3a,t2,v3b,MiodideD(
plot	
,15,MiodideD( : , 5} , ' 'figSd.aps*)
This document is a draft for ra-iew and public comment purposes only and doss not constitute final Agency policy.
October 2,2008	63	DRAFT
-------
preg .r - initialization file for pregnancy
HES rTG=0; HE0I*re=0 ;
%dam cle>4 (In 45-52?, fetus C104 (Ln. S4-C0}, dam 1125 {Ln S2-69), £etus 1125 (Ln 71-76)
Sis ps_p-0,2l, pr_p=0.5S, p£_p«Q . OS, pJc_p-0 . 99, pl_p=0 - 56, P3_P"i • 25 . pgj_js=1.76
ns pt_p=0,13, pdtjp=7.Q, psk_p=X.32, prfacj-0,8, ppl_p=0.56, praa«_p=0.6S
lis vmaxc_tp>6e3, wiaxc_dtp=l.S?e4, vmaxe_sp=l.2eS, vmaxc_gp=3¦2e?
lis vmaxc_pp=6e4 , vjiinc -p-2 . 2^4 , Km_Ip=1.6eS, Km_DTp=l, OeS , Xm Qp=2.0e5
lis K;r. 3p-2 . 0e&, kmjpp=2, £te5» kni_rtip=2 , De5, pagc_p=0.6, pagjcjpsl.O
Us paskc_p>«3», 25, patc_p»l,0«-4, padtc_p=0 , 01, parbcc_p«10 .0, pipe p 0,1
lis pamc_p=0.04, cluc_p=0.05
IIS v-raxc:_bp« 588 , kmjbp=l,64e4, kunbc_pB(J. 03
lie ktrana2c=0.12» ktranslc»0.12
Ha vtuxc dcfp-167««
! Is padcc fp»0.01, patc_fp=0.01
! !s Ki»_TPp-i.eBS, Kx GFp-5 . C«5, Kn_SFp»2 .OeS
!!s V«wC_SFp=8,0e5, paskc_fp=l.25
! ! S Vmxc_<5Fp-4 . C«i€ , pagjc_fp=l, 0 , pijc fp-0 . 6S
!!s wwc_bfp*50C, km_lj£p=l .8e4, kur.bc£p 0 . 03
!!s ps_i=0.21, pr_i=0.4, pf_i=0,OS, pX_i*l.09, pl_i-0.44, pg_i=l.Q, pgj_l=2.0
!!s p'._ i-0 -15, pdt_i=7.0, p."k i C . 7 , prbc_i=l.Q, ppl_i=0.4,	. £ £
US vmaxc_ti=l .22e5, vmaxc_dt i = l, OeS, vnu»KC_si=8.4e4, vmaxc_gi=4 . 5e5
lis •.'Taxc,-_pi»s«4, vmaxc_mi=4 . 0e4, K®_Ti=»4. OeS, KniJ3Ti=1.0e9, Kr*_3i«4 . 3e6
lis Em_5i=4.OeS, km_pi=4.0e6, ktn_mi=4 . OeS, pagc_i=Q.IS, pagjc_i=12.0
lis pasJcc_i=Q. 06, patc_i=l. Qe-4 , padtc_i=l, 5e-5, parbcc_i=10.0, papc_i=0,005
! !s paaic_i=i0.ai, cluc_i=Q.06, khcrwc_ 1.0 . 03 . ksecrc_i»3 .le-7
lie vmaxc_bi»36Cr Xw bi-'.Beb, kdeiodc i-0.021
lis Xtranalc j. 0 . 12 . ktratis2c_i=0 ,12
IIS vtruce_cit f i 6 . Ce"
lis padto_£i»l,Oe-4, patc_fi»0.01
I i a Kr. TF'i-4 , C®6, Kat_GFi=4, OeS, Kx. SFi-4 . Ce6
!is VmxC_SFi=3.OeS, paskc_fi=0.02
lis Vmxc_(JFi«2 . OeS, pagj c_f 1=0,3, pagc__f 1=0.1
TSTO?«lrCINT-1;
start ©HoCallBack
%Fig7,n
%Figure 7 plots
Dat»_Hf>r«igF
ioit_preg
prepare aclear TIME attot_t cfet_I CTTCT fi
DOSE_I=10 .0; T5TAHT=210 0; .IV START1-84: TSTOP=IVSTART_I+150;
SPIA-O; start minocailfaacic
plot {_time-IVSTART_I,_c£et_i, . , .
ALL!3d!:,1 f-TSTART -IVSTART_I, RtL13d(;,3),'O', 'Fig7_GW13.aps1}
TSIART=2300; IVSTART_1-52; TSTOP=IVSTRRT_I+150;
SPIA=1; start enocallback
plot(_time-lVSTART_l,_cfet_i, ...
ALL14<3 (:, i: - 73TART - TV START	I. A14.14 d (: , 3 j , 'o' , 1 Fig?_GM14. Hps « )
ALL15d=
[252$
135
2532
190
P.013
2532
64
0.012
2538
31.4
0.139
2538
35.2
0.0948
2538
140 .3
0 .0948
2538
225 .2
0.139
2544
30
0 .017
2544
221.
NaN
2544
73 .3
HaN
2544
173. &
NaN
2544
38.9
NafT
2544
43 .1
NaN
2544
55,2
HaNJ ;
TSTART-24DC-; IVSTAP.T__I»12C r TSTOi>=lVSTART_I + 150 ?
RI.L15
-------
: in* - 7 VSTART I;pt t>0,
plot It Cpt) , cf«c _i (pt) , ALiiSd (:, 1) , ALL15<1{ :,3), 'o' , 1 Fig?_GHlS.aps ')
This document is a draft for renew and public comment purposes only and does not constitute final Agency policy,
October 2,2008	65	DRAFT
-------
%Figure 8 - RCIewell et aX, JTEH 2006
%Maternal iodide
%t, crr.k_i, attot_i, au_i
MATID-[3242 45.80 NaN NaN
3242
3242
3243
3243
3243
3243
3243 ,
3246
3246
324 6
3246
3246
324S
3249
3250
3252
3252
3252
3252
3252
3252
3252
3252
3252
3252
3258
3258
3258
32 5 8
3258
3259
3260
3263
3264
3264
32S4
3 2 64
3264
3264
3264
32 54
3268
3268
3268
3270
3272
3272
3272
3272
3275
3276
3275
3276
3276
3280
3280
3281
3284
3287
3288
3288
3288
3268
3288
3288
3288
3288
3.00
3.30
45 .80
NaN
HaN
NaN
NaN
NaN
35.00
24

NaN
NaN

NaN

NaN

NaN

NaN

NaN

30 .81
23

NaN
40 .72
29

NaN
35 .00
NaN

NaN
67 .10
NaN

NaN
NaN
NaN

Half
NaN

57.21
NaN

NaN
33 . 00
36

NaN
24 ,50
NaN

NaN
31.74
3S

NaN
28.17
NaN

NaN
28.00
NaN

NaN
23 .92
NaN

NaN
46 .70
NaN"

NaN
10.30
HaN

NaN
34.10
NaN

NaN
13.80
NaN

NaN
3.00
NetN

NaN
16.50
NaN

NaN
27.92
HaN

NaN
29.30
NaN

HaN
9.30
NaN

NaN
34.SO
NaN

NaN
20.60
NaN

NaN
12.90
NaN

NaN
11.51
NaN

MaH
11 .20
NaN

NaN
8 .14
36

46
19.22
18

NaN
2 .00
34

NaN
1,30
26

HaN
9.00
NaN

HaN
16 .50
NaN

NaN
17 , 30
NaN

NaN
18 .70
NaN

NaN
10 ,40
NaN

NaN
13 .20
NaN

NaN
13 .60
NaN

NaN
2 .19
NaN

NaN
3 .30
NaN

NaN
5.20
NaN

NaN
7.00
NaN

NaN
0.40
KaN

NaN
6.SO
NaN"

NaN
10.66
Naif

NaN
8.36
NaN

NaN
6.40
NaN

NaN
14.4 0
NaN

NaN
3 .50
NaN

NaN
3 .10
NaN

NaN
NaN
21

65 ,8
2.40
Nan

NaN
1.90
NaN

NaN
1.74
32

50
5.05
20

74
0.80
25

64
1.90
23

53
3.40
37

52
NaN
39

NaN
NaN
NaN

NaN
HaN
NaN

NaN
NaN
Man
NaN
HaN
NaN
This document is a draft for review arc! public comment purposes only and does not constitute final Agency policy.
October 2,2008	66	DRAFT
-------
3288
N«M
3286
Ha N
3294
3 .23
3300
3 .95
33X2
1 .76
3312
0 .40
3313
1. 58
3330
0 .34
3336
0 .10
3336
Ma»
33S4
D.D5
3360
0.55
3360
0.50
3360
0.03
3378
0.01
3384
0.01
3408
0.43
3408

3432
0.51
3480
0-49
3504
0.44
3528
0 .38
3552
0 .33
3672
0.18
3720
0 .13
3792
0 .13
396 0
0.0?
4152
0.01
naU
KaB
NuN
KaM
PaU
fiaJK
NaS
SaJS
29
HaS
24
Nail
NaN
NaN
NaN
NaN
26
NaN
23
HaH
NaN
NaN
24
NaN
21
HaN
NaN
SaM
NaH
HaH
20
HaS
20
Has
21
NaH
NaM
NaM
15
Pall
NaN
NaM
NaN
NaN
NaN
HoN
NaN
NaN
NaN
HaU
NaN
NaN
NaN
NaS
NdN
HaJB
%Fig8.»
IFigure 8 - RClcwell et al, JTEH 2006
WESITG=0; MiD;iG»C;
output WCleAC
prepare T CKK I ATTCT I MJ_I
ACLU-0.75; R0=1, 0; VCHNCJ-O . 0 ; AKT«0.0; HORNK=l;
cluc_i-0.Hi cluc_p=0.125; %adult v»lu«a
CLUC__NI=cluc_i*RU* (70 A (0. 75 - ftCLUi ) ;
CLUC_MP=cliic_p*P.U* {70" (0 , 75 - ACI-Ui ) ;
PD0SB.0; XVDOSE_I=lQQ; IVDOSE_NI-0; PPBeQ; PDOSE_H=0; COHC=Or POSE_8.I = 0; ITOOSE_P=0;
IVSTRRT_I=324 0; T8TQF-33SS;
!I set TIMEO-3240.0
VHAXC_T1 - 1.33e5;
starc~nocalIback
cwJtl=_cmk_i; act.ocl=_attot_i; aul=_au_i;
VMAXC_TI=7 . 4«5 ;
startSaocaliback
omk2=_cmk_i; attot2=_attot_i; au2=_«u i;
VKWvC_TI = 1.39e5;
I I set TIMBOmQ
plot (_t-TIMED,attotl,_t-TIME0,attOt2, MAUD(;,1)-TiMEQ,MAT ID(i,3), »0!,'Fi»8«.apa';
ploc (_t-TIMED,aul,_t-TIME0,au2,MATID(r,l) - TIMED,MRTID(;,4! , »o» r 'FigBb.ape")
plot (_t-TIME0.	TIMED, rank!, KATID f;»1] -T1ME0 »MATID { : , 2) , >o 1 , ' FigSc , apa 1J
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	67	DRAFT
-------
% Pig9_GFR.rti
WESITG=0; WEDITGnsO;
output @clear
prepare ©clear DAYS CA_NI CTTOT_NI
nio_urine= [] ; CINT=.l;
ACLU=0 . 75 ; R'J«1. 0; VCHNO0 . 0; AKT=1; NDR.NX-1 ;
ciuci	0.11; cluc_p=0,125; %adult values
CLDC_NI=cluc_i*RU*(70*(0.75 - ACLO));
CLUC_NP=cluc_p*RU*£7Q~ £0.75 - ACLO));
PDOSE=0; IVDOSE1-0; IVDOSENI-100; PPB=0; PDOSE_N=0; CCNC-0; DOSE_RI=0;
I I set TIMEO	0
for day	[2 3 366]
IVSTART_I= (day-1) *24 ; TSTOP=IVSTART_J>24 ; IVSTRT_NI»IVSTART_I;
if day>3S4
I I sec TIHE0-876 0
TSTOP"IVSTART'*48 % in .cmd file, TSTOP « 48 Hr after injection for ¦Proced
yeari'
end
start (BnocallbacJc
nio_urine - [nio_urin,e; [day, BW, BW_N, A" HI] j
end
! I set TIME0=0
This document is a draft for review and public comment purposes only and does not constitute final Agency policy.
October 2,2008	68	DRAFT
-------