             

      UNITED STATES ENVIRONMENTAL PROTECTION AGENCY

                                     WASHINGTON, D.C.  20460

			

                  OFFICE OF  

PREVENTION, PESTICIDES AND  

         TOXIC SUBSTANCES

November 1, 2006

MEMORANDUM

SUBJECT:	Aldicarb (List A Case 0140, Chemical ID No. 098301).  HED
Response to Comments Received During the Public Comment Period.  DP
Barcode No. D331538.

FROM:	Felecia Fort, Chemist

		Linda Taylor, Ph.D, Toxicologist

		Jeff Dawson, Chemist

Reregistration Branch 1

Health Effects Division (7509C)

THRU:	Michael Metzger, Branch Chief

Reregistration Branch 1

Health Effects Division (7509C)

TO:		Tracy Perry, Chemical Review Manager

Sherrie Kinard, Chemical Review Manager 

Special Review Branch

Special Review and Reregistration Division (7508W)

The responses presented are in reference to comments received following
a 60-day public comment period for aldicarb.  The comments discussed
herein are from the registrant, Bayer CropScience  (BSC) and the Natural
Resources Defense Council (NRDC) and pertain to the Environmental
Protection Agency's Revised Preliminary Human Health Assessment and
Disciplinary Chapters for the Reregistration Eligibility Decision (RED)
Document for aldicarb dated May 12, 2006.  The HED Human Health
Assessment and Dietary Exposure Assessment have been revised to reflect
comments and errors noted by the registrant.  No changes have been made
to the Product and Residue Chemistry Chapter, the Toxicology Chapter or
Occupational Chapter.  All comments pertaining to drinking water
exposures will be addressed by the Environmental Fate and Effects
Division (EFED).  

Response to Bayer CropScience Comments

Document #1: Bayer CropScience Response to the HED Revised Preliminary
Human Health Risk Assessment for the Reregistration Eligibility Decision
Document (RED) and Re-Evaluation of Dermal Absorption, Sensitivity, and
Reversibility Issues: Report of the Hazard Identification Assessment
Review Committee

 BCS lists 6 main issues; two of which are related to toxicology and are
discussed below.

A.  Toxicological Endpoint and Reference Dose: BCS believes that the RfD
for aldicarb should be derived from RBC cholinesterase data in the human
study. This is based on their conclusion that the doses in the human
study are closely spaced in the region of interest and thus provide the
most appropriate data for calculation of RBC BMD10/ BMDL10 estimates.
BCS contends that the doses in the rat studies produced saturated levels
of enzyme inhibition, and these are not appropriate for BMD calculation.

HED Response: Consideration of all available toxicity data was used to
determine the toxicity endpoints and reference doses appropriate for the
aldicarb risk assessment. Criteria utilized in determining the
appropriate toxicity endpoints and reference doses included the quality
and reliability of the studies and the presence of sound dose-response
data. There is a complete toxicology database of oral studies including
a human oral study. HED's previous risk assessment (RA) reported risks
using multiple endpoints, including those from the human study, to fully
characterize risks, but focused on results using the rat RBC
cholinesterase inhibition [ChEI] endpoint.  This decision reflected the
Agency's interpretation of the conclusions drawn by the HSRB prior to
issuance of the final report.  Based on the final report, which clearly
concluded that use of the human study endpoint was appropriate for human
health risk assessment, the current risk assessment continues to provide
results using all three endpoints considered, but focuses on the results
of the human study since these data best reflect human response to the
chemical. Because these human data are considered reliable, and the
study is considered scientifically valid, at this time the Agency
considers the human study to be the most suitable for risk assessment
purposes for this single-chemical risk assessment.

With respect to the BCS contention that the rat data are not appropriate
for BMD calculation, see detailed responses to comments surrounding BMD
analyses below (pages 10-14).

B. Dermal Absorption: BCS believes that the use of  a 100% dermal
absorption factor in the assessment of occupational risk is an
inappropriately conservative assumption. BCS concludes that sufficient
data exist to reduce this default value, and a roughly-defined dermal
absorption factor default as high as 50% can be adopted with complete
confidence.

HED Response: HED agrees that 100% is conservative, as stated in the RA;
however, the registrant has not provided any additional data to
support/allow the use of a lower dermal absorption factor. The agency
has in the past clearly articulated [HED Document No. 013265, 1999] the
reasons for the use of 100%. All of the available data were used by the
Agency in an attempt to provide a more realistic dermal absorption
factor. However, the data do not allow for any credible alternative at
this time.  Most importantly, BCS has not provided any documentation to
support their proposed use of 50%. Therefore, no alteration in the
dermal exposure assessment is appropriate at this time.

GENERAL COMMENTS from BCS BCS Line-By-Line Review of the Aldicarb RED
Document (only toxicology issues are addressed here). These are
identified as to the page of the BCS document where they appear. 

(a) On page 9 of 40. EPA’s Page: 7  Paragraph: 3 Lines: 21-22 of EPA
RA, states “… aldicarb induced ChEI has been shown to be reversible
in less than 24 hours.”

BCS contends that the reversibility time stated in the RA is misleading.
BCS states that Cholinesterase time to recovery is dependent on the
dose. BCS notes that toxicity studies are conducted at doses
significantly higher than experienced via the diet to elicit a response;
i.e., greater than 10% depression of ChE from baseline. At higher doses,
longer recovery times to less than 10% baseline are observed. Their
analysis of the data from an aldicarb human volunteer study demonstrated
that at 0.019 mg/kg, the recovery time was around one hour (Wyld et.
al., 1992). BCS indicated that this was the lowest dose for which a
recovery time could be predicted. Since maximum dietary exposure occurs
at levels more than 2 orders of magnitude below this level, BCS
concludes that at dietary exposure levels, ChE will return to baseline
levels extremely rapidly.

HED Response:  Aldicarb toxicity is characterized by maximal inhibition
of cholinesterase which occurs rapidly followed by recovery typically
occurring within hours.   A key consideration in risk assessment is
appropriate matching of the duration of exposure with the duration of
the toxic effect.  Typically, HED’s food and water exposure
assessments sum exposures over a 24 hour period.  This 24 hour total is
typically used in acute dietary risk assessment.  In the case of the
aldicarb, because of the rapid nature of aldicarb toxicity and recovery,
it may be appropriate to consider durations of exposure less than 24
hours.  Conceptually, a physiologically-based pharmacokinetic model
and/or biologically-based dose-response model would be available to
account for the dynamic nature of exposure, absorption, toxicity,
recovery, and elimination of aldicarb in animals and humans.  However,
such a model does not exist at this time.  In the interim, HED has
developed an analysis using information about external exposure, timing
of exposure within a day, and half-life of ChE inhibition from rats and
humans to estimate risk to aldicarb at durations less than 24 hours. 
Specifically, HED has evaluated individual eating and drinking occasions
and used the ChE half-life information to estimate the residual effects
from aldicarb from previous exposures within the day.  

Table 6 below provides information on the recovery of ChE inhibition in
rats and human subjects.  For both species, the recovery half-life for
RBC ChE inhibition is approximately two hours.  At high doses in rat,
the half-life is up to approximately 6 hours in females.  The estimates
of half-life at the lower doses are most relevant for risk assessment
and are thus the focus here.  As can be seen in the table, the recovery
half life of aldicarb-inhibited AChE in the human is estimated to be on
the order of 2 hours using RBC AChE activity   This 2 hour recovery
half-life is what is used in this refined dietary exposure assessment
which incorporates information on eating/drinking occasions.  There is
some uncertainty associated with the use of the two hour recovery
half-life.  As discussed in detail below, infants and children are the
focus of the current analysis.  Although there are dose-response ChE
data in juvenile animals exposed to aldicarb, there are no such data to
characterize ChE recovery in the young.  As such, the Agency has assumed
that the half-life to recovery in the young is similar to that seen in
adults.  The Agency is requiring such data in young animals to confirm
this assumption.

 

Table 6.  Recovery half-life information for ChE inhibition following
oral exposure to aldicarb in rats and human subjects

Chemical	Brain	RBC

	Recovery Half-Life Estimate (hrs)	Upper & Lower Confident Intervals
(hrs)	Recovery Half-Life Estimate in hrs

(dose; mg/kg)	Upper & Lower Confident Intervals (hrs)

Rat	1.52	1.16-1.99	F (< 0.1)  1.10 

(0.1,0.3)  2.91

(0.3,0.5) 3.39

(>0.5) 5.90

M (<0.1) 1.91

(0.1,0.3) 1.20

(0.3,0.5) 1.62

(>0.5) 1.50	F  0.50-2.40

1.96-4.33

2.35-4.90

3.52-9.91

M  1.31-2.79

0.87-1.64

1.19-2.21

0.80-2.82

Human	N/A	2.07	1.74-2.46



 (b) On page 9-10 of 40. EPA’s Page: 7 Paragraph: 4 Lines 30-36 on
increased sensitivity in PND 17 pups. BCS states that the only increased
sensitivity seen in the PND 17 animals was in the dose needed to achieve
“the MTD (more accurately should be called the lethal dose)” and
brain ChE depression. Administration of a bolus dose directly to pups at
dose levels that produce a high degree of ChE inhibition can
misrepresent the effects associated with lower levels that might occur
with relevant dietary exposures.

According to BCS, this difference in outcome arises because the
increased sensitivity of neonatal animals to AChE inhibition results not
from pharmacodynamic differences in target enzyme (AChE) sensitivity
(relative AChE inhibition for a given tissue concentration of the
proximal inhibitor), but from age-dependent pharmacokinetic differences
in the capacity of the metabolizing enzymes that determine the tissue
concentration of the proximal inhibitor. The generally lower capacity of
metabolic enzymes in neonatal animals may result in a greater internal
exposure for the same administered dose. Importantly, due to the nature
of capacity-limited metabolism, as described by the Michaelis-Menton
equation, the impact of these age-dependent metabolic differences can be
much greater at doses that exceed the relatively limited metabolic
capacity of the neonate than at doses that are well below saturation of
metabolism. BCS concludes that an FQPA safety factor is not required for
aldicarb.

HED Response: Although these are reasonable assumptions, sufficient
kinetics data to demonstrate the lack of a difference at the “much
lower levels of exposure that are relevant for an FQPA risk
assessment” Additionally, a 2X difference in response was observed in
both sexes at each dose level tested [see table below], including the
0.05 mg/kg dose level where a “high degree of inhibition” was not
demonstrated. At the lowest dose tested [0.05 mg/kg], the largest
difference between the age groups was observed; i.e., the difference in
males was 2.3 and difference in females was 2.4. Additionally, the ratio
of the BMDs for adults and pups indicates a 2X difference [brain BMD10s
ranged from 0.014 to 0.020 in juvenile animals and 0.024 to 0.031 in
adult animals].  

Moser Replication Study - Brain Cholinesterase Data



Age/Sex/Dose [mg/kg]	

Males	

Females



ADULT

0

0.05

0.1

0.2

0.3	

4.96±0.16

4.37±0.12 [12]

3.64±0.17 [27]

2.78±0.31 [44]

2.41±0.56 [51]	

4.80±0.10

4.30±0.12 [10]

3.78±0.18 [23]

2.49±0.27 [48]

2.19±0.19 [54]



PND 17

0

0.05

0.1

0.2

0.3	

3.95±0.34

2.83±0.11 [28] (2.3)^

2.38±0.27 [40] (1.5)

1.04±0.17 [74] (1.7)

0.64±0.07 [84] (1.6)	

3.80±0.19

2.89±0.23 [24] (2.4)

2.02±0.11 [47] (2.0)

1.17±0.18 [69] (1.4)

	
 [% inhibition]; ^ (ratio of  pup:adult %
inhibition)

It is to be noted that the peak effect time used in the Moser study was
determined for both age groups based on clinical signs. It is to be
noted that clinical signs were observed more frequently in adult animals
versus the young animal in other studies on aldicarb. Although the same
peak-effect time was determined for both groups, it is possible that the
adult animal reached a similar % ChEI as the pup but at a different time
post dose than was measured in this study. Data that are not available
to address the issue of sensitivity include an assessment of whether the
adult animals reach a similar % ChE inhibition as the pups at a
different time point (earlier or later) post exposure. In order to
address this aspect and others, a comparative cholinesterase assay would
be necessary for aldicarb. 

Although none of the guideline studies on aldicarb showed evidence of
sensitivity, it is to be noted that ChEI is not monitored in the rat and
rabbit developmental toxicity studies, and measurement of ChEI in the
reproduction study occurred in blood samples collected only at study
termination; ChE data between young and adult animal were not compared; 
time sample collected post exposure for ChE activity assessment is not
known. 

BCS believes that the developmental neurotoxicity study [DNT] is the
most appropriate study for assessing sensitivity, and sensitivity was
not demonstrated in the DNT on aldicarb. HED agrees that the DNT is an
appropriate study for assessing sensitivity. However, in the case of
aldicarb, the assessment of ChE activity was performed on samples
collected 2 hours after dosing of the dams. Effects on ChE following
aldicarb exposure are expected to peak sooner. Data from other studies
on aldicarb indicate that peak ChE inhibition occurs as early as 30-45
minutes post dose. Based on BCS own arguments on reversibility, this
2-hour time point for the assessment of ChEI is not appropriate in this
case. Because of the timing of sample collection in this DNT, it is
unclear whether accurate ChEI values were obtained or that peak
inhibition was measured. It also complicates interpretation of the
maternal ChEI values and the apparent lack of ChEI at the low- and
mid-dose levels. Additionally, since there was no indication of whether
the pups nursed prior to their ChE assessment, and there is no
information regarding the timeframe for aldicarb to reach the milk. The
apparent lack of ChEI in the pups cannot be considered definitive
evidence that the young animal is not more sensitive. For aldicarb, the
DNT does not provide a definitive assessment of sensitivity. The FQPA
safety factor of 2X is supported by reliable data and is retained. 

With respect to the statement that the route of administration in the
DNT was the most appropriate for assessing sensitivity in infants and
children, it is to be noted that dosing was to the adult (dam) via
gavage. The young animal was not dosed directly. Although some infants
are nursed, children and other infants are directly exposed to food.

(c) On page 11 of 40. EPA’s Page: 7 Paragraph: 5 Lines 38-42 regarding
EPA’s conclusion that there is a two-fold difference between animals
and humans with respect to toxic responses. BCS points out the similar
level of inhibition in both species at the common dose level of 0.05
mg/kg. Since the 2-fold difference is based on the BMDL10s calculated
for the two species [0.02 mg/kg and 0.013 mg/kg], it is BCS contention
that since enzyme inhibition was saturated at the majority of the rat
doses used in the BMDL10 analysis, the rat data are not reliable for
comparison with the human BMDL10 analysis.

HED Response: Based on the current use of the human data in the risk
assessment, this is a moot point for the single chemical assessment.
However, the Agency does not agree that the rat BMDL10 analysis is not
reliable [see detailed responses to comments surrounding BMD analyses
below (pages 10-14)].

(d) BCS states that a graphical illustration of the dose-response data
in human versus rat suggests an apparent continuation of the response
across both species from 0.01 to 0.1 mg/kg and further demonstrates the
similarity in the magnitude of the response in both species at the
common dose level of 0.05 mg/kg. 

HED Response: It is to be noted that the lowest dose tested in the rat
studies is 0.05 mg/kg. There are no data at the 0.01 mg/kg dose level
for the rat. 

(e) On page 13 of 40. BCS comments on EPA’s page 8, paragraph 2, lines
14-21, BCS states that “the implication in EPA’s revised preliminary
risk assessment document that the Human Studies Review Board (HSRB)
considered the human data appropriate only to inform the interspecies
factor is misleading.” BCS states that the HSRB report contains no
specific comment by the Review Board regarding the use of the human data
only to inform the interspecies factor.

HED Response:  HED’s charge to the HSRB: Please comment on the
scientific evidence that supports whether the aldicarb human study is
sufficiently robust for reducing the inter-species uncertainty factor.
The HSRB’s response to this charge was: 

 

(1) The HSRB concluded that the aldicarb human study appears to be a
scientifically valid study, suitable for use in both the aggregate risk
assessment and the cumulative risk assessment. 

(2) The Board stated that the dose-response data from the human study
appeared such that BMD and BMDL can be calculated. 

(3) The Board concluded that the results of the human study could be
used in the WOE analysis to determine a NOEL for RBC cholinesterase and
clinical signs in males and that the RBC cholinesterase demonstrated a
dose-dependent and time-dependent pattern of inhibition in both males
and females. 

While HSRB did not address directly the charge; i.e., whether the study
could inform the interspecies factor, the HSRB’s draft report
indicated that (1) the human study is sufficiently robust (appears to be
scientifically valid) and suitable for use in risk assessment; (2) it
can be used to calculate BMD/BMDL, which HED used to compare to the rat
BMD/BMDL; and (3) it can be used in the weight of analysis [WOE]. The
Agency's interpretation of the draft HSRB final report was that the
human study could be used to inform the inter-species factor. 

(f) On page 15 of 40, BCS states that none of the rat data have been
subjected to the rigorous reviews and extensive deliberations to which
the human data have been subjected. 

HED Response: This is not an accurate statement. HED has extensive
experience in the review of animal toxicology data. All of the rat data
on aldicarb have been through numerous assessments over the years by the
Agency review committees, including the HED HIARC, RfD, and TES
committees, CPRC/CARC, and an Agency non-OPP expert review committee.
Additionally, the acute neurotoxicity, subchronic neurotoxicity, and
developmental neurotoxicity studies are all guideline studies, which
were all performed using adequate numbers of animals per dose group for
statistical assessment. The non-guideline studies by Moser used the more
appropriate method for cholinesterase measurement for a carbamate
chemical, and an Agency non-OPP expert review committee reviewed these
studies.  Therefore, the Agency believes that both the human and animal
data have gone through similar thorough analysis.

Document #2: Bayer CropScience’s Assessment and Rebuttal of the
Reference Dose as Proposed by EPA in the Health Effects Division Revised
Preliminary Human Health Risk Assessment of Aldicarb Issued on May 12,
2006

It is to be noted that several of the issues/concerns in this document
are the same as in  Document #1 above and are not repeated here. 

(a) On page 6 Weight of the Evidence Establishment of the RfD: Regarding
the extensive toxicological database on aldicarb and the fact that most
of the data are in the rat, the registrant notes that the doses tested
in these studies typically produced ChE inhibition of 30% and greater
with often maximal or near maximal inhibition at several doses. In
contrast, the human study is noted to cover a dose-response range that
includes doses around the benchmark response of 10%, as well as dose
levels causing clear inhibition of ChE. Studies with one or more doses
near the level of the BMR are generally considered to provide a better
estimate of the BMD while studies with responses at or near the maximal
response level should not be considered  for BMD analysis due to
saturation. 

HED Response: With regard to the issue of appropriate BMD analysis, [see
detailed responses to comments surrounding BMD analyses below (pages
10-14)]. 

(b) On page 8 of 16. The authors state that the RBC ChE results indicate
that the 10% BMR is below 0.05 mg/kg in both rats and humans, and
inhibition of the enzyme begins to plateau above 0.1 mg/kg in both
species (emphasis added).

HED Response: With regard to the statement that in both rats and humans,
inhibition of enzyme begins to plateau above 0.1 mg/kg (also on page 16
of 86), it is to be noted that human data at the 0.1 mg/kg dose level
are not available from the human study reviewed by the HSRB. Also, the
statement here regarding a plateau in both species is in contrast to the
rebuttal statement on page 14 of 40, which indicates that the
“dose-response does not appear to produce saturation of cholinesterase
in humans compared to the rat”. These two statements are conflicting.

(c) On page 9 of 16, the authors state that, given the similarity of
ChEI response between rat and human at the common dose of 0.05 mg/kg,
one would expect that a BMD analysis of data in the same dose-response
range for each species should produce similar BMD10 and BMDL10 values
for both rat and human. EPA’s calculation of higher BMD estimates for
rats versus human is most likely an artifact due to the inclusion in the
analysis for the rat of multiple dose levels producing saturation of ChE
inhibition at the high doses. Combining the rat studies reinforces the
bias of each study’s weakness, i.e. multiple dose levels producing a
saturated dose-response, and distorts the BMD estimates.

HED Response: [see detailed responses to comments surrounding BMD
analyses below (pages 10-14)].

(d) Also on page 9, the authors state that since the BMD estimates
determined for the rat cannot be considered reliable due to the
distorted influence of a saturated dose-response, any quantitative
conclusions regarding interspecies differences derived from them are
equally unreliable.

HED Response: This is a moot point. However, see detailed responses to
comments surrounding BMD analyses below (pages 10-14).

(e) On page 10 of 16 (in discussing Figure 1), it is stated that the
dose-response does not appear to produce saturation of cholinesterase
inhibition in humans compared to the rat. 

HED Response: This statement is not supported, based on the lack of
human data at dose levels where an apparent saturation was observed in
the rat. As can be seen from the figure, at 0.05 mg/kg the response is
comparable (human and rat) and at the next higher dose for each species
(human 0.06-0.075 mg/kg vs rat 0.1 mg/kg), the response is also quite
similar. Lacking a higher dose in the human, one cannot speculate that
at 0.2, 0.4, 0.5, and 0.6 mg/kg dose levels the human would not have a
similar response as the rat. Having shown a similar response at two
similar dose levels, there does not appear to be a basis for concluding
there would be a difference in response between the two species at dose
levels both above and below these two dose levels.

(f) In the Sensitivity to Infants and Children section (page 12), it
states that the only increased sensitivity seen in the PND 17 animals
was in the dose needed to achieve “the MTD (more accurately should be
called the lethal dose)” and brain ChE depression. Administration of a
bolus dose directly to pups at dose levels that produce a high degree of
ChE inhibition can misrepresent the effects associated with lower levels
that might occur with relevant dietary exposures. Additionally, the
authors conclude that the overall weight-of-evidence supports that an
extra FQPA safety factor is not required for aldicarb.

HED Response: This comment is the same as in Document #1 and is
addressed above.

Detailed Responses to Comments Surrounding BMD Analyses

Responses to comments surrounding BMD analyses including issues such as
“saturation” of the ChE inhibition and Bayer CropScience’s
preference for the analyses provided by OPCum Risk over the more
complicated analysis combining datasets

There are two main themes in the comments from Bayer CropScience
surrounding the BMD analysis of the rat data:  saturation of ChE
inhibition distorts BMD estimates at the low end of the dose response
curve and evaluating one data at a time is preferred over combining data
from multiple studies.

Bayer CropScience makes the statement that “studies with responses at
or near the maximal response level should not be considered adequate for
BMD analysis due to saturation.”  The Agency’s BMDS web tutorial is
listed as the reference for this statement.  HED looked at this
reference and contacted ORD scientists involved in developing BMDS.  The
following quotes were extracted verbatim from the web tutorial and seem
to be the source of the Bayer CropScience statement.  

“Assess the Fit of the Models. Retain models that are not rejected
using a p-value of 0.1. Examine the residuals and plot the data and
models; check that the models adequately describe the data, especially
in the region of the BMR. (Sometimes it may be necessary to transform
the data in some way or to drop the highest exposure group(s) (e.g., if
the behavior at high exposures can be attributed to early mortality or
enzyme saturation effects) and repeat the modeling in order to get a
good fit.)

Are the data appropriate for a BMD analysis? 

Once the critical endpoints have been selected, data sets are examined
for the appropriateness of a BMD analysis. The following constraints on
data sets to use for BMD calculations should be applied: 

There must be at least a statistically or biologically significant
dose-related trend in the selected endpoint; and 

The data set should contain information relevant to dose-response for
modeling. determination of the amount of information about the
dose-response that is available need not be quantitative or technical.
For example: A data set in which all non-control doses have essentially
the same response level provides limited information about the
dose-response, since the complete range of response from background to
maximum must occur somewhere below the lowest dose: the BMD may be just
below the first dose, or orders of magnitude lower. When this situation
arises in quantal data, especially if the maximum response is less than
100%, it is tempting to use a model like the Weibull with no
restrictions on the power parameter, because such models reach a plateau
of less than 100%. This situation can result in seriously distorted
BMDs, because the model predictions jump rapidly from background levels
to the maximum level. In principle, other models could be found that
force the BMD to be anywhere between that extreme and the lowest
administered dose. Thus the BMD computed here depends solely on the
model selected, and goodness of fit provides no help in selecting among
the possibilities. (See the quantal data examples in the Technical
Guidance (EPA, 2000) Appendix for an example of this situation): “

For some of the quantal models in the BMDS (like the example provided in
the BMDS tutorial text), extreme high dose data can sometimes cause a
distortion of the BMD at the lower end of the dose response curve. 
Because of this, it may be appropriate to drop a high dose in a BMD
analysis for some datasets.  Dropping the high dose is allowed for in
the OPCum Risk program.  However, as Bayer CropScience also accurately
states, the Agency’s exponential dose response model is sufficiently
flexible to allow for a plateauing of ChE inhibition at a response level
lower than 100% thus significantly reducing this potential distortion.  

Moreover, in the aldicarb analysis, the Agency has used data from
multiple studies simultaneously which provides a more robust analysis of
the entire dose response curve compared to evaluating one dataset at a
time.  By modeling the combined datasets, the Agency has taken advantage
of the increased number of dose levels to get a better idea of the shape
of the dose-response curve.  This translates into the ability to fit
more complicated models which more accurately reflect the true
dose-response shape.  As Bayer CropScience shows in Figure 1 (pg, 10),
there seems to be a ‘shoulder’ at the very low end of the dose
response curve where the slope is fairly shallow for rat data.  The
OPCum Risk program can not account for this shallow slope at the low end
of the dose response curve.  In 2001 when the Agency was in the early
stages of developing its OP cumulative risk assessment, the Agency
developed the OP Cum Risk program.  This program has been reviewed by
the SAP and continues to be very helpful in risk assessments of ChE
inhibiting chemicals.  However, this program has some limitations. 
Specifically, based on comments from the SAP and the stakeholders, the
Agency’s revised dose-response assessment refined its analysis to
include the possibility of a low dose shoulder region.  The Agency has
shown that approximately ½ of the OPs exhibit this low dose shoulder
and approximately ½ do not.  Furthermore, the Agency learned that
combining data from multiple studies improved the ability to model the
complete shape of the dose response curve, including the low dose
shoulder region.  The Agency’s experience with the OPs informed the
analysis for the N-methyl carbamates.  Specifically, the sophisticated
analysis included in the N-methyl carbamate cumulative assessment
includes the possibility for this low dose shoulder.  

Based on the rat ChE data, ‘saturation’ or plateauing of ChE
inhibition seems to occur at or near 80% inhibition.  Here is a plot of
doses available in the four studies used in the rat BMD analysis.  Note
that two of the studies have only three dose levels (including control),
but, together, there are 8 dose levels ranging from near the BMD10 to
much higher levels of inhibition.  The estimated BMD and 95% confidence
limits are indicated with vertical lines:

  

Here is a plot of the predicted RBC AChE inhibition, with the
dose-levels in the different studies indicated by vertical dashed lines.


Because of the data from 8 different dose levels parameters for both the
low-dose shoulder and the high dose plateau (that is, lg and tz) could
be estimated.  Mathematically, in the analysis of the aldicarb RBC data,
the power parameter (actually, the log of it, lg) of the dose response
curve was estimated to be greater than 1, and is fairly significant
(nominally, P ~ 0.00033 for the test against the null hypothesis that
the power is 1). OPCumRisk is unable to estimate the power parameter,
effectively fixing that parameter to 1.0—and thus overpredicting ChE
inhibition at the low dose end to some degree.  

The Agency was able to estimate both the low-dose shoulder and the high
dose plateau (that is, lg and tz) because data from several studies that
span a range of doses were used.   Thus, by combining the information
from multiple studies, the Agency was able to fit the entire dose
response curve.  If the Agency were to follow Bayer’s recommendation
not to combine studies, and tried to fit models to individual datasets,
information from several dose levels would be ignored---because they
would be useless for estimating dose-response parameters.  A couple of
the studies would contribute at most one useful dose level for the
dose-response analysis if the analysis were conducted one at a time,
whereas, when analyzed together, a more complete description of the
dose-response curve is achieved.  The FIFRA SAP has supported the
Agency’s combining of data from multiple studies in dose-response
efforts with the ChE inhibiting chemicals.  For example, in the 2001
report, the panel states:

	

“There are several advantages of combining estimates from multiple
datasets compared to using estimates derived from single
datasets/points. Combining data increases the precision of estimates,
incorporates the variability among data sets into the overall estimate
of uncertainty (standard errors or confidence limits), and maximizes the
use of the available information.”

Another component of the Agency’s analysis is the ability to use all
the time course data in estimates of half-life to recovery---time course
analysis is not a component of the OPCumRisk program.  Lastly, the
Agency’s analysis accounts for repeated measures (ie, RBC ChE
inhibition from the same animal more than one time) and thus includes
within animal variability.  Because one dataset is analysed at a time in
OPCum Risk, repeated measures is not accounted for when using OPCum
Risk.

The Agency acknowledges that the statistical procedures used in the
aldicarb analysis (and for other members of the N-methyl carbamate
common mechanism group) is highly sophisticated.  The Agency has
presented its approach to the FIFRA SAP at two meetings (February and
August, 2005) and received positive feedback from the panel.  The
approaches are described in detail in Chapter B of the Preliminary
Cumulative Risk Assessment of the N-methyl carbamates and the associated
“B” appendices.  The “B” appendices contain numerous files,
including the R code used in the statistical analysis, the actual ChE
data used, and the output in pdf format.   On August 18, 2006, the
Agency confirmed the public availability and ease of access to these
files by extracting them from   HYPERLINK
"http://www.epa.gov/scipoly/sap/meetings/2005/index.htm#feb22" 
http://www.epa.gov/scipoly/sap/meetings/2005/index.htm#feb22  (“All
Meeting Documents”).  The Agency successfully downloaded the datasets
in excel format (II.B.1-1 & -2) and the pdf versions of the “B”
chapter and associated appendices, including the outputs of the aldicarb
rat BMD estimates (II_B_2_a & _b).  

Endpoint Selection 

Bayer’s comments seem to infer than they believe that data from plasma
or whole blood ChE were used in the Agency’s BMD analysis.  This is
incorrect.  The Agency did not include data from plasma or whole blood
in the BMD analysis.  As indicated on pg 4 of the EPA’s June 30, 2005
analysis:  

“We need to select out the records for red blood cells of animals
dosed via gavage. Drop records with missing values for cheact, and any
The one aggregated study, 43442305, used multiple chemicals: aldicarb,
aldicarb sulfoxide, and aldicarb sulfone. Drop everything but the active
ingredient. However, the control group in that study used NA to code the
control chemical, so we need to retain those records. Finally, make sure
we have a value for sd for records with n greater than 1.

> dta <- CleanUp(subset(wrkdata[["AggData"]], mrid %in%
MRIDSforRBCDR[["Agg"]] &

+ sctn %in% "RBC" & dsmtd == "GAVAGE" & !is.na(cheact) & (n ==

+ 1 | !is.na(sd)) & (is.na(ta) | ta %in% c("AI")), select =
KeepVars))”

Document #3: Responses to: Bayer Crop Science Response to the Revised
Anticipated Residues and Dietary Exposure Analyses for the HED Human
Health Risk Assessment 

Bayer CropScience noted four areas where they considered inappropriate
data had been used or data excluded as follows:

Application of Potato Processing Factors

Conversion of residues to aldicarb sulfone equivalents rather than
aldicarb equivalents

Sweet potato Residues Data Source

Incorporation of  Effects of Cholinesterase Reversibility

In response the comments made by Bayer CropScience, the dietary
assessment was revised to include updated sweet potato PDP data, 
missing cooking factors,  and to convert from aldicarb sulfone
equivalents to parent aldicarb. 

The sweet potato PDP data used were from the years 1999, 2003 and 2004. 
These data were used for all partially blended sweet potato food forms. 
In previous assessments, potato PDP data were used for all food forms. 
For all nonblended foodforms, potato data from the 1997 Special Survey
were used since HED use of single serving data for non-blended
commodities is more protective.  See Attachment 3 for RDF 56SWmcsp for
residue inputs.

Several cooking factors for boiled and fried potatoes were erroneously
not used in the previous assessment as not by the registrant.  This was
corrected in this revised assessment. 

HED also converted risks estimates from aldicarb sulfone equivalents to
aldicarb per se.  HED concurs with the registrant that since current
methods analyze parent aldicarb and metabolites sulfoxide and sulfone as
individual chemicals and no longer utilize the common moiety method
which converted parent and metabolites to sulfone equivalents, it is no
longer appropriate to convert all residues to aldicarb sulfone. 
Consequently, HED multiplied all risk estimates by 0.86 to convert the
sulfone equivalents to aldicarb per se. 

With respect to the issue of incorporation of the effects of
cholinesterase reversibility, the Agency acknowledges receipt of the
dietary assessment which utilized the CARES-Dietary Minute Module.  The
Agency is currently reviewing the submitted data and assessment and will
make a determination on its applicability to the Aldicarb Human Health
Assessment at a later date. See also response under Document #5 in this
memorandum.

Document #4: Bayer CropScience Response to the Update of Incident Data
Review of April 10, 1996.

Based on our review of the BCS submission, Bayer’s comments raise no
substantive issues and do not warrant a response from us at this time. 
Since the previous review of the PCC/TESS data and IDS data only covered
up through 1998 and 1999, respectively, CEB recommends updating the
Incident Review PCC/TESS data through 2005 when the 2004 and 2005 data
become available in the near future.   A review of the IDS data on
aldicarb should also occur at this time.  

Document #5: Aldicarb Acute Dietary Risk Assessment Including Food and
Drinking Water and Document #6  Human Health Risk Assessment for Use on
Currently Registered and Pending registrations as of May 2006.

HED acknowledges the receipt of the acute dietary risk assessment and
human health risk assessment conducted by the registrant using its
Dietary Minute Module.  The Agency is currently reviewing the submitted
data and assessment and will make a determination on its applicability
to the Aldicarb Human Health Assessment at a later date. 

HED has; however, further refined the acute aggregate risk from food and
groundwater by incorporating the time and amounts consumed for each
eating occasion from the USDA CSFII food diaries to estimate exposures
and risks on each eating occasion throughout the day.  This refined
assessment also incorporated the available toxicological data which
indicates that the estimated half-life for cholinesterase inhibition
resulting from aldicarb exposure is 2-hours or less.  Exposures and
risks using this approach were calculated using the DEEM model coupled
to a SAS® program which accounted for cholinesterase regeneration.  To
verify these DEEM-based eating occasion results, the Agency’s Office
of Research and Development’s Stochastic Human Exposure and Dose
Simulation (SHEDS) model was also used to conduct an eating occasion
analyses for aldicarb.  The SHEDS eating occasion results are similar to
the DEEM-based results, providing additional assurance regarding the
accuracy of these computations.  SHEDS was also used to conduct further
sensitivity analyses on the half-life parameter, as well as addressing
issues regarding both direct and indirect drinking water consumption. 
Detailed information on the methods used to derive the aggregate
exposures are presented in the document titled “Aldicarb: Acute
Dietary Exposure Assessment to Support the Reregistration Eligibility
Decision” [S. Nako and J. Xue, 11/01/06].   

Table 13 presents the respective DEEM-FCID® and SHEDS®’ estimated
risks at the per capita 99.9th percentile using a 2 hour half-life for
cholinesterase inhibition.  These eating occasion results are based on
several major assumptions: (i) 2 hour half-life, (ii) allocation of
direct drinking water consumption based on 6 equal and fixed occasions,
and (iii) no modifications to the amount of indirect drinking water
consumed as reported in the CSFII diaries for infants.  Direct water is
water that is consumed from the tap and indirect water is considered
water that is used in the preparation of food. 

It should be noted that incorporating eating occasion analysis and the 2
hr. recovery half life for aldicarb into the Food Only analysis  does
not significantly change the risk estimates when compared to baseline
levels (for which a total daily consumption basis – and not eating
occasion - was used)  From this, it is apparent that modifying the
analysis such that information on eating occasions and aldicarb half
life is incorporated results in only minor reductions in estimated risk:
generally  on the order of several percent, at most, for all age groups.
 However, risk estimates for which food and drinking water are jointly
considered and incorporated are reduced considerably (by a factor of 2
or more in some cases) compared to baseline and is not unexpected:
infants receive much of their exposures from indirect drinking water in
the form of water used to prepare infant formula.  

                                         

National Resources Defense Council (NRDC) Comments 

NRDC disagrees with the use of human study to inform interspecies
uncertainty factor, stating that the Human Studies Review Board (HSRB)
“warned that the human study was inadequately powered and thus the
possibility of a false negative, that is, the inability to detect an
adverse effect where it occurs, is “a real concern”.” Also, NRDC
contends that the HSRB concluded that the study failed to fully meet the
specific ethical standards prevalent at the time the study was
conducted. Therefore, NRDC concludes that “EPA’s consideration of
and reliance on the Aldicarb human study is arbitrary and capricious
since the study’s ethical violations are significant.”

HED Response: Although the HSRB indicated that the human study had
several limitations, the final report stated the following: (1) the
aldicarb human study appears to be a scientifically valid study,
suitable for use in the risk assessment; (2) the dose-response data from
the human study appeared such that BMD and BMDL can be calculated; (3)
the results of the human study could be used in the WOE analysis to
determine a NOEL for RBC cholinesterase; and (4) the RBC cholinesterase
data from the human study were reliable and demonstrated a
dose-dependent and time-dependent pattern of inhibition in both males
and females. 

NRDC questions why EPA dropped the 10X LOAEL to NOAEL uncertainty
factor, which HED presented to the HSRB as the endpoint for risk
assessment.

HED Response: As stated in the May, 2006 revised preliminary RA,
dose-response modeling is preferred over the use of NOAELs/LOAELs since
NOAELs and LOAELs do not necessarily reflect the relationship between
dose and response for a given chemical, but instead reflect dose
selection. Dose-response modeling is considered a higher tier assessment
providing Points of Departure (PODs) that are more accurate than the use
of NOAELs/LOAELs and uncertainty factors. In order to evaluate the
appropriate point of departure (PoD) for ChEI, the Agency considered the
benchmark dose estimates developed from both the rat and human data. 

The assessment of the data presented to the HSRB was based on the use of
NOAELs/LOAELs and default uncertainty factors; e.g., LOAEL to NOAEL. The
May, 2006 and the current assessment are based on dose-response
modeling. Therefore, the 10X LOAEL to NOAEL uncertainty factor is not
appropriate.

NRDC disagrees with the use of the oral data for inhalation exposure
assessment

HED Response: 

Aldicarb is a Toxicity Category I compound for all routes of exposure

ChEI is the toxicity of concern; the oral assessment is based on ChEI
and is being used pending submission of an acute inhalation study to
assess ChE activity and recovery.

An inhalation Toxicity Category of I results in label requirements of
protective equipment for workers. There are no residential uses

A subchronic inhalation study was a datagap. Although this study is no
longer required (an acute inhalation study is now required), it is
Agency policy not to apply an additional uncertainty factor for lack of
this route-specific study. 

There is a potential for inhalation exposure from aldicarb-treated
tobacco, but there are no residential uses or agricultural uses that
would result in residential exposure to the general population. Although
exposures can occur for occupational handlers loading or applying
aldicarb granulars, no postapplication exposure is expected because
aldicarb is soil-incorporated at planting.

4. NRDC disagrees with the conclusion that chronic exposure assessment
is not needed.

There is a complete database on aldicarb, and there is no indication
that toxicity increases on long-term exposure. In fact, the NOAEL for
effects other than ChEI  is greater than that for ChEI following various
exposure scenarios. Therefore, performing a chronic risk assessment
based on an endpoint other than ChEI would not be protective and would
not make sense. Additionally, the NOAEL following repeat exposure is
higher than that following acute exposure, which supports the decision
that chronic exposure can be considered a series of acute exposures and
a separate assessment is not needed.

5.  NRDC implies that the HSRB had “a real concern” (NRDC’s
emphasis) with respect to the statistical analysis and the fact that the
study was likely under-powered.

HED Response: HSRB discussed the statistical analysis of the human
study, which was considered a weakness of the study. However, they
concluded that, although “it is clear that the statistical power of
the study was low, the data do show a very clear and predictable dose-
and time-dependency in the RBC cholinesterase data.” Further, it was
concluded that the results of the human study could be used in the WOE
analysis to determine a NOEL for RBC cholinesterase. 

General Comments: Several respondents reiterate BCS contention that the
animal data have not been subjected to the rigorous reviews and
extensive deliberations to which the human data have been.

%

)

4

Q

o

˜

›

œ

¨

¹

Ò

ã

ä

4

Å

Æ

Ý

õ

␃愀$摧笄¡

耀

耀将

ô

␃愃̤

ༀ肄㜄$␸䠀$葞Ҁ摧㸡è

␃༃梄币梄愁̤摧

̤̀☊଀Ն愀̤摧

␃愃̤摧

摧

摧

̤̀☊଀цༀ㢄帄㢄愄̤摧

摧

␃愃̤摧

摧

e group for statistical assessment. The non-guideline rat studies by
Moser used the more appropriate method for cholinesterase measurement
for a carbamate chemical, and an Agency non-OPP expert review committee
reviewed these studies. Therefore, the contention that the animal data
review was of a lesser quality than that of the human study is an
inaccurate characterization.

Page   PAGE  1  of   NUMPAGES  18 

 

