§
<
30
\
^tD STa%
ro
z
o
^ c>v°
P R 0^
OFFICE OF INSPECTOR GENERAL
Catalyst for Improving the Environment
Evaluation Report
Opportunities to Improve
Data Quality and
Children's Health through the
Food Quality Protection Act
Report No. 2006-P-00009
January 10, 2006


-------
Report Contributors:
Jerri Dorsey
Alice Fong
Montira Pongsiri
Abbreviations
CFR	Code of Federal Regulations
EPA	Environmental Protection Agency
DHHS	Department of Health and Human Services
FFDCA	Federal Food, Drug, and Cosmetic Act
FIFRA	Federal Insecticide, Fungicide, and Rodenticide Act
FQPA	Food Quality Protection Act
OIG	Office of Inspector General
OMB	Office of Management and Budget
OPP	Office of Pesticide Programs
OPPTS	Office of Prevention, Pesticides, and Toxic Substances
ORD	Office of Research and Development
STAR	Science to Achieve Results
USDA	U.S. Department of Agriculture
Glossary: A glossary of the technical terms used in this report is in Appendix A.
Cover Photos: The Food Quality Protection Act emphasizes the need to protect children
from pesticides (EPA photos).

-------
<
33
\
^t0SrX
&
V PRO^4-0
o
2
Lll
o
U.S. Environmental Protection Agency
Office of Inspector General
At a Glance
2006-P-00009
January 10, 2006
Why We Did This Review
We sought to determine the
impact of the 1996 Food
Quality Protection Act
(FQPA) on the need of the
Environmental Protection
Agency (EPA) for scientific
data on how pesticides impact
children's health. We
evaluated whether EPA
enacted guidelines and
procedures, and addressed
new aggregate exposure and
cumulative risk assessment
efforts. We also sought
opportunities for
improvement.
Background
FQPA changed the way EPA
regulates pesticides, including
the introduction of aggregate
exposure and cumulative risk
assessments. FQPA required
the Office of Pesticide
Programs (OPP) to take into
account children's unique
patterns of exposure and
vulnerability regarding
pesticides. Additional data
needs were identified to
achieve the Act's mandates.
For further information,
contact our Office of
Congressional and Public
Liaison at (202) 566-2391.
To view the full report,
click on the following link:
www.epa.qov/oiq/reports/2006/
20060110-2006-P-00009.pdf
Catalyst for Improving the Environment
Opportunities to Improve Data Quality and Children's
Health through the Food Quality Protection Act
What We Found
To meet the requirements of FQPA, EPA instituted numerous data
requirements designed to provide infants and children with better protection
against the health risks of pesticides. FQPA established a single, health-
based standard that eliminated discrepancies, and emphasized infants and
children.
FQPA resulted in the revision of many regulations, guidelines, and
procedures. OPP made substantial changes to the aggregate risk assessment
process, which considers multiple routes and pathways of exposure for a
particular pesticide, to acquire more and better data on children's exposure.
OPP also took steps to collect data on the cumulative effects of pesticides
sharing a common mechanism of toxicity, which represent the combined
risks to children from a group of pesticides.
Significant challenges nonetheless remain. EPA's required testing does not
include sufficient evaluation of behavior, learning, or memory in developing
animals. There is no standard evaluation procedure for interpreting results
from developmental neurotoxicity tests (involving substances that damage a
developing nervous system, including the brain). OPP has requested data on
developmental neurotoxicity for certain pesticides, but to date no summaries
have been released or conclusions drawn. OPP is unable to collect sufficient
data on aggregate risk due to time and cost constraints and relies on other
agencies for data. Specific opportunities for improvement involve finalizing
Science Policy papers, assessing alternative testing strategies, using logic
models, and developing a multi-year strategic plan.
What We Recommend
We made recommendations to EPA for improving data collection. EPA
should develop a standard evaluation procedure, evaluate certain testing
methods, and take steps to reduce uncertainties. EPA can take various steps
to improve its aggregate exposure and cumulative risk assessments,
including updating databases and expanding partnerships with other Federal
organizations. EPA can also take steps to enhance accountability, act on
Science Policy papers, try alternative testing strategies, and develop an
overarching logic model and long-term strategic plan. The Agency
concurred with many of our recommendations but expressed concern with
certain issues raised.

-------
$ A \
\ Wi
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
OFFICE OF
INSPECTOR GENERAL
January 10, 2006
MEMORANDUM
SUBJECT:
Opportunities to Improve Data Quality and Children's Health
through the Food Quality Protection Act
Report No. 2006-P-00009
FROM:
Jeffrey K. Harris /s/
Director for Program Evaluation, Cross-Media Issues
TO:
Susan Hazen
Acting Assistant Administrator,
Office of Prevention, Pesticides, and Toxic Substances
This is the final report on the subject review conducted by the Office of Inspector General (OIG)
of the U.S. Environmental Protection Agency (EPA). This report contains findings that describe
the problems the OIG identified and corrective actions the OIG recommends. This report
represents the opinion of the OIG and the findings contained in this report do not necessarily
represent the final EPA position. Final determinations on matters in the report will be made by
EPA managers in accordance with established resolution procedures.
Action Required
In accordance with EPA Manual 2750, you are required to provide a written response to this
report within 90 days of the date of this report. You should include a corrective actions plan for
agreed upon actions, including milestone dates. We have no objections to the further release of
this report to the public. For your convenience, this report will be available at
http ://www. epa. gov/oi g/.
If you or your staff have any questions, please contact me at (202) 566-0831 or Jerri Dorsey,
Assignment Manager, at (919) 541-3601.
cc: George Gray, Assistant Administrator, Office of Research and Development
Jim Jones, Director, Office of Pesticide Programs

-------
Table of C
At a Glance
Chapters
1	Introduction		1
Purpose		1
Background		1
Scope and Methodology		3
Results of Review		3
2	FQPA Inspired Numerous EPA Data Requirements		5
National Research Council Report Emphasized Need for Better Data		5
FQPA Resulted in Numerous Key Changes		5
Various Registration Data Required		7
3	OPP Lacks Consistent Data on the Developing Nervous System
to Determine Potential Adverse Effects		8
Assessing Developmental Neurotoxicity Important		8
Developmental Neurotoxicity Testing Issues		9
Weaknesses in Toxicity Testing Guidelines		10
Proposals Made for Changes to Testing Requirements		11
Recommendations 		13
Agency Response and OIG Evaluation		13
4	OPP Made Substantial Changes to Address Aggregate Risk,
but Challenges Remain		15
Substantial Changes Made in Aggregate Risk Assessment 		15
Data on Children's Nondietary Pesticide Exposure Limited		16
EPA Relies on Public and Private Sources of Dietary Exposure Data,
and More Needed		18
Recommendations 		19
Agency Response and OIG Evaluation		20
5	OPP Moving to Assess Cumulative Risk, but Complexities
and Concerns Remain		22
Many Factors Impact on Assessing Cumulative Risk		22
New Science Needed to Measure Effects of Concurrent Exposures		23
Models, Computer Tools Can Enhance Cumulative Risk Assessments		24
Recommendations 		25
Agency Response and OIG Evaluation		25

-------
6 Opportunities Exist to Better Manage FQPA Implementation		26
OPP Science Policy Papers Not All Finalized		26
EPA Needs to Continue Pursuing Alternative Testing Efforts		27
Logic Models Could Better Guide Efforts 		28
Multi-year Strategic Plan Can Support Goals		29
Recommendations 		30
Agency Response and OIG Evaluation		30
Appendices
A	Glossary of Terms		31
B	Details on Scope and Methodology		33
C	Response from the Agency		34
D	OIG Comments on Agency's Response		57
E	Toxicity Testing Issues		68
F	Data and Tools for Estimating Dietary Exposure		70
G	Distribution	 73
ii

-------
Chapter 1
Introduction
Purpose
We performed this review to examine the impact of the Food Quality Protection
Act of 1996 (FQPA) on the Environmental Protection Agency's (EPA's) need for
scientific data and predictive tools, particularly in relation to children's health.
This report is the second in a series of three reports on FQPA's impact on EPA
regarding children's health. We specifically sought in this review to determine:
•	What data requirements were required by FQPA;
•	Whether testing guidelines, requirements, and evaluation procedures allow
EPA's Office of Pesticide Programs (OPP) to determine the potential adverse
effects of pesticide exposure on the developing nervous system;
•	What challenges OPP overcame and what opportunities exist for OPP to
acquire better pesticide exposure data to aggregate risks;
•	What challenges exist and what opportunities are available for OPP to
improve cumulative risk assessments; and
•	What opportunities exist to better manage pesticide health risk for children.
Background
The FQPA of 1996 established a single, health-based standard that eliminated
discrepancies from the past. The Act requires that standards for agricultural
pesticides be set at levels that protect the health of infants and children. The
FQPA altered the way OPP regulates pesticides. OPP must now ensure that the
pesticide residue limits in food (or tolerances) are at safe levels, and that there is a
reasonable certainty of no harmful developmental effects1 for children before a
pesticide can be registered.
The mission of OPP, within the Office of Prevention, Pesticides, and Toxic
Substances (OPPTS), is to protect human health and safeguard the environment
from unreasonable adverse effects resulting from the use of pesticides. OPP is to
ensure that pesticides are regulated fairly and efficiently while reducing pesticide
risks, especially for infants and children. The role of EPA's Office of Research
1 Development effects are adverse effects such as altered growth, structural abnormality, functional deficiency, or
death observed in a developing organism.
1

-------
and Development (ORD), as the principal research arm of EPA, is to provide the
critical science for environmental decision-making through its problem-driven
and core research projects. In support of OPP/OPPTS, ORD provides scientific
tools that can be used to characterize, assess, and manage risks in implementing
the FQPA requirements. FQPA emphasized the need for three types of
information for pesticide regulatory decision making:
•	Developmental toxicity data (the adverse effect pesticide exposure will have
during prenatal development and after birth);
•	Aggregate exposure risk data (all routes and pathways of exposure) for a
pesticide; and
•	Cumulative risk data for pesticides with common mechanisms of toxicity.
Prior to FQPA's enactment, OPP treated exposures to pesticides from different
pathways as independent events. The pathway represents the physical course a
pesticide takes from the source to the person (such as through food or residential
use pesticides). In reality, exposures to pesticides do not occur as single, isolated
events, but as a series of sequential or concurrent events. As a result of FQPA,
exposures from food, drinking water, and residential uses of a single pesticide are
combined when completing an aggregate risk assessment (see Figure 1.1).
Figure 1.12: Pathways, Routes, and Conditions in Aggregate Risk Assessment
EXAMPLES OF PATHWAYS, ROUTES, AND CONDITIONS TO CONSIDER IN
AN AGGREGATE EXPOSURE AND RISK ASSESSMENT
Dietary
Pathway
Drinking Water
Pathway
Pathway
Exposure and
Postapplication
Exposure and Risk froi
Exposure and
Exposure and
Incidental Ingestion of:
Exposure and Risk from
ii nting
Detergents |
Hand Soaps
Fogging
Swimming
2 Source: US EPA, Nov 28, 2001. General Principles for Performing Aggregate Exposure and Risk Assessments.
2

-------
OPP performs an aggregate exposure risk assessment for each chemical in the
common mechanism group before undertaking a cumulative risk assessment on
pesticides sharing a common mechanism of toxicity. The aggregate exposure risk
assessment process includes identifying significant contributors to risk and
estimating risk using probabilistic modeling (determining distribution of possible
values). In assessing cumulative risks, OPP evaluates the potential for people to
be exposed to more than one pesticide at a time from a group of pesticides with an
identified common mechanism of toxicity. The objective is to appropriately
match and combine estimates of pesticide exposures in a way that considers
factors associated with exposure (i.e., time, region, and age groups). The
potential for concurrent exposure to multiple chemicals by multiple pathways
(including food; drinking water; and exposure to air, soil, grass, and indoor
surfaces) would be included in the cumulative risk assessment.
Scope and Methodology
We performed our evaluation generally in accordance with Government Auditing
Standards, issued by the Comptroller General of the United States. Our review
focused on existing data and interviews. We did not examine internal controls.
Our field work occurred between July 2004 and July 2005. We generally covered
events from 1993, when the National Research Council released a report, entitled
Pesticides in the Diets of Infants and Children, which had many recommendations
that were incorporated into the FQPA of 1996.
This is the second in a series of three reports on FQPA's impact on EPA
regarding children's health. The first report, Changes Needed to Improve Public
Confidence in EPA 's Implementation of the Food Quality Protection Act
(2006-P-00003), was published October 19, 2005. A third report is planned to
address the measures and indicators for measuring progress in implementing the
FQPA.
Further details on our scope and methodology are in Appendix B.
Results of Review
FQPA resulted in the revision of many regulations, guidelines, and procedures
related to protecting infants and children from the health risks of pesticides. To
meet the requirements of FQPA, EPA instituted numerous data requirements that
should provide better protection. Additionally, EPA took steps to develop science
policies, develop methods and tools, and collect required data on aggregate
exposure and cumulative risk. Nonetheless, significant data gaps remain. Data
collected from developmental neurotoxicity tests need to have summaries released
and conclusions drawn. EPA needs to collect more data on aggregate exposure
risk and apply better methods to collect data on and assess cumulative risk.
Opportunities for improvement involve system accountability, finalizing Science
Policy papers, assessing alternative testing strategies, using logic models, and
3

-------
developing a multi-year strategic plan. We made various recommendations to
EPA for improving data collection.
The Agency concurred with many of our recommendations. However, the
Agency expressed concern that the report focused on issues that are "minor and
relatively insignificant within the overall scope of FQPA implementation,"
"characterized incorrectly," or "outside the control of the Agency." Also, the
Agency stated that "OPP had already identified and begun working on many of
the issues discussed" in this report. We summarized the Agency's response and
provided our comments on the response at the end of each chapter that contained
recommendations. The full text of EPA's response is in Appendix C. Appendix
D provides the full text of our comments on the Agency's response.
4

-------
Chapter 2
FQPA Inspired Numerous EPA Data Requirements
To implement the requirements of FQPA, EPA had to institute numerous data
requirements. FQPA established a single, health-based standard and requires that
allowable residue levels for food use pesticides be protective of the health of
infants and children. EPA data requirement changes involved emergency
suspension procedures, data collection activities, registration renewal, and
tolerance reevaluation. FQPA requires EPA to perform risk assessments
differently, in that it must now assess aggregate and cumulative risks of pesticides
instead of just the risks for one pesticide and one medium at a time. EPA must
consider all non-occupational sources of exposure, including drinking water, and
exposure to other pesticides. Such additional steps should provide improved data
and potentially result in better protection against pesticides for infants and
children.
National Research Council Report Emphasized Need for Better Data
Many of the provisions in the 1996 FQPA originated from recommendations
made in a 1993 National Research Council report, Pesticides in the Diets of
Infants and Children. This report brought attention to how better data on dietary
exposure to pesticide residues should be combined with improved information on
the potentially harmful effects of pesticides on infants and children. The report
emphasized the need for testing procedures and that "testing must be performed
during the developmental period in appropriate animal models, and the adverse
effects that may become evident must be monitored over a lifetime."
The National Research Council called for the development of new risk assessment
methods that would incorporate better data on children's exposure to pesticides
during fetal development, infancy, and childhood. Furthermore, it recommended
the use of exposure distributions, expansion of exposure assessment to consider
exposure to multiple chemicals with multiple routes of exposure, and the
development of pharmacokinetic models (for determining and quantifying the
time or absorption, distribution, biotransformation, and excretion of pollutants)
that could incorporate the unique physiological features of developing children.
FQPA Resulted in Numerous Key Changes
According to EPA, FQPA provides a more consistent pesticide regulatory scheme
by amending the Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA)
and the Federal Food, Drug, and Cosmetic Act (FFDCA). Primary authority for
pesticide regulation at EPA is through FIFRA and FFDCA:
5

-------
•	Under FIFRA, EPA registers pesticides for use in the United States and
prescribes labeling and other regulatory requirements.
•	Under FFDCA, EPA establishes tolerances for pesticide residues in food, both
domestic and imported. A tolerance is the maximum level of pesticide residue
allowed in or on human food and animal feed. These tolerances are enforced
by the Food and Drug Administration for most foods and by the
U.S. Department of Agriculture for meat, poultry, and some egg products.
The following outlines many of the requirements and amendments to FIFRA and
FFDCA resulting from FQPA, as interpreted by EPA:
Table 2.1: Key Changes and Additions for Pesticide Regulations Due to FQPA
FIFRA
Permits emergency suspension of a pesticide without simultaneously issuing a notice of intent to
cancel.
Prescribes data collection activities to ensure health of infants and children:
•	Collection of adequate data on food consumption patterns of infants and children.
•	Improved data collection on occurrence of pesticide residues in foods most likely consumed by
infants and children.
•	Evaluation of pesticide usage information and improved information gathering.	
Requires registration review and renewal, once every 15 years.
Establishes special provision for minor use pesticides, including public health pesticides.
Links tolerance reassessment to reregistration.
Establishes special provisions for antimicrobial pesticide registration.
Establishes mandate for continuing expedited consideration of application for pesticides meeting
one or more criteria for reduced risk pesticides.3
Establishes a Scientific Advisory Board to assist in the scientific peer reviews conducted by the
FIFRA Scientific Advisory Panel.
FFDCA
Delinks pesticides from Delaney clause and places all pesticide authority in FFDCA.
Establishes standard for establishing a tolerance based on whether tolerance is "safe," defined as
"a reasonable certainty that no harm will result from aggregate exposure to the pesticide chemical
residue, including all anticipated dietary exposures and all other exposures for which there is
reliable information."
Defines types of exposure to be aggregated for risk assessment purposes.
Requires consideration of cumulative effects of pesticides having common mechanism of toxicity.
Requires tolerance reassessment in three phases - review 33 percent within 3 years of FQPA
enactment; a second 33 percent within 6 years; and the remaining number within 10 years.
Specifies an additional 10-fold margin of safety for infants and children for threshold effects.4
Requires development of an estrogenic substances screening program.
Requires development and distribution of consumer information on pesticide risks and benefits.
3	This provision was overtaken by the 2004 Pesticide Registration Improvement Act, which establishes specific
timeframes for reduced risk pesticides.
4	Threshold effects are those effects considered to have exposure doses at some identifiable level which are likely to
be without appreciable risk of deleterious consequences.
6

-------
Various Registration Data Required
The 1993 National Research Council report brought attention to the uncertainty
about childhood exposure and toxicity at different stages of development. Under
FQPA, EPA must ensure that the pesticide residue limits in food tolerances are at
safe levels, and that there is a reasonable certainty of no harmful developmental
effects for children before a pesticide can be registered. Pesticide registration is
dependent upon the fulfillment of a series of data requirements. The number and
types of studies to be conducted vary with the intrinsic chemistry, anticipated
inherent toxicity, and proposed use pattern of the pesticide. Pesticides of
conventional chemistry proposed for use on agricultural commodities generally
require the greatest amount of information, whereas those for non-food use
generally require less.
Title 40 of the Code of Federal Regulations (CFR) Part 158 presents the
regulatory roadmap specifying the types and amounts of data needed by EPA to
decide whether to approve an application for a new or amended registration or
reregi strati on under FIFRA. The data requirements specified in Part 158 cover
areas such as product chemistry, toxicology for human health and domestic
animals, wildlife and aquatic toxicology, nontarget insects, environmental fate,
aerial drift evaluation, reentry protection, plant protection, product performance,
residue chemistry for food uses, and biochemical and microbial pesticides. The
type of data required is dependent on the product's proposed pattern of use, the
results of earlier studies, and other circumstances.
7

-------
Chapter 3
OPP Lacks Consistent Data on the
Developing Nervous System to Determine
Potential Adverse Effects
OPP's testing and evaluation procedures need improvement to better determine
the potential adverse effects of pesticide exposure on the developing nervous
system. EPA's current required toxicity testing does not include evaluation of
behavior, learning, or memory in developing animals until triggered by predefined
effect conditions in other required toxicity studies. Also, there is no standard
evaluation procedure for interpreting results of such tests. OPP had requested
data on developmental neurotoxicity (involving substances that damage a
developing nervous systems, including the brain) for certain pesticides in 1999,
but to date no summary has been released or conclusions drawn from the data.
Also, data requirements for pesticide registrants have not been comprehensively
revised since 1984; although EPA published proposed changes in March 2005,
EPA was awaiting public comments prior to amending the proposed rule or
promulgating a final rule. In the proposed changes, developmental neurotoxicity
data tests are proposed to only be "conditionally required."
Assessing Developmental Neurotoxicity Important
One of the conclusions in the National Research Council report was that the
toxicity testing strategies used by regulatory agencies were inadequate for
assessing toxicity to a number of organ systems, including neurodevelopmental
processes. The report indicated that pesticide exposures may disrupt the normal
development of a child's brain and nervous system, and recommended regulatory
agencies such as EPA revise published guidelines on testing as new information is
obtained.
Pesticide chemicals can easily enter the brain of fetuses and young children
because the blood-brain barrier is not fully developed. In the developing brain,
billions of cells must form, move to their positions, and establish precise
connections with other cells. If cells in an infant's brain are destroyed, or
connections between brain cells fail or send false signals to the developing
reproductive organs, nervous system or reproductive dysfunction may result that
can persist throughout life.
ORD's human health research has characterized the differential response of
younger animals to the neurotoxic effects of cholinesterase inhibiting pesticides
(cholinesterase is one of many important enzymes needed for the proper
functioning of the nervous systems of both humans and animals). ORD's
8

-------
research indicated that when younger animals are more sensitive to these
chemicals, they are also less efficient in detoxifying the pesticides. According to
ORD, this information has been used by OPP to limit the use of selected
pesticides, and helped form the basis for the data call-in process to collect
comparative sensitivity data for all registered organophosphate pesticides (a group
of pesticides that act to inhibit acetylcholinesterase enzymes, which result in
dysfunction in the nervous system). Through this data call-in process, OPP seeks
data from appropriate pesticide manufacturers. OPP issues a data call-in when
there is no existing, reliable information to characterize a pesticide's risk or
exposure, or otherwise complete a risk assessment. OPP implemented the data
call-in process for the cholinesterase-inhibiting organophosphates based on
known neurotoxicity concerns.
In an Overview5 that provided the basis for evaluating the Human Health
Research Program at the Agency, ORD indicated that EPA had resolved a variety
of groundbreaking policy and scientific issues in conducting the organophosphate
pesticides cumulative risk assessment. However, to protect the health of children,
ORD strongly recommended that OPP change its approach to require a
developmental neurotoxicity study for pesticide registration, and that in the
absence of this study, OPP should consider applying the traditional uncertainty
factor.
Developmental Neurotoxicity Testing Issues
While developmental neurotoxicity tests have helped to characterize risks to
young animals, particularly effects on learning/memory, auditory response, motor
activity, and neuropathology, external and internal stakeholders have raised
concerns regarding this test data. Industry representatives have said that
developmental neurotoxicity study results are difficult to interpret and expensive
to conduct. Public health and children's advocates expressed concerns about EPA
being slow in its review of the developmental neurotoxicity data call-in studies
and that the review results were not publicly available. OPP scientists reported
that call-in data are inconclusive because there is great variability among the
various sets of developmental neurotoxicity data and uncertainty in laboratory
conditions.
On August 6, 1999, EPA published in the Federal Register6 that it was requiring
registrants of neurotoxic pesticides to conduct acute, subchronic, and
developmental neurotoxicity studies and submit the results to EPA via the data
call-in process. As of March 11, 2005, a total of 50 developmental neurotoxicity
studies have been received for review by OPP. However, to date, no summary
has been released or conclusions drawn from the data.
5	From page 27 of "Section 007 Overview" in the ORD CD containing the briefing book and poster abstracts
provided to the Board of Scientific Counselors for Review of ORD's Human Health Research Program on
February 28, 2005, to March 2, 2005 in Research Triangle Park, North Carolina.
6	64 Federal Register 42945
9

-------
We found that EPA lacks a Standard Evaluation Procedure for developmental
neurotoxicity studies, even though the final Standard Evaluation Procedures for
developmental toxicity and a draft procedure for reproductive toxicity are now
available. OPP scientists reported that interpreting developmental motor activity
is relatively easy, but interpreting learning and memory tests from the
developmental neurotoxicity data call-in information is difficult, because such
tests are more qualitative in nature and not sensitive enough to determine if there
is a cause and effect. However, developmental neurotoxicity tests have helped to
characterize risks to young animals, particularly effects on learning/memory,
auditory response, and motor activity.
Weaknesses in Toxicity Testing Guidelines
Non-EPA scientists reported in literature7 published post-FQPA on the
insufficiency of current testing requirements for assuring children's safety from
most food-use pesticides. Examples of the weaknesses noted are:
•	EPA's core testing has included no requirement for specific testing of
developmental neurotoxicity in developing animals and immunotoxicity in
adult or developing animals.
•	EPA's core testing includes no adequate assessment of the effect of toxicity
on the function of developing animals (possibly apart from reproduction)
involving behavior, learning, or memory.
•	All but two core toxicity tests EPA required for food-use pesticides are
performed in adult animals, including the only test of metabolism.
•	EPA requires no data on pharmacokinetics (rate of absorption and distribution
of toxin in the body) or pharmacodynamics (sequence of events in the cell
leading to a toxic response) of the pesticide in developing animals, and its risk
assessments include no such information.
•	The exposure period recommended by EPA's developmental neurotoxicity
guidance may be too short to reflect the entire vulnerable period of brain
development in children, and statistical procedures to define the minimal
number of animals in a test group are lacking.
A 2002 EPA review by a technical panel8 also identified numerous gaps in testing
guidelines, and the panel suggested that the Agency develop alternative strategies
7	Claudio L. et al. 1999. Assessment of the US Environmental Protection Agency Methods for Identification of
Hazards to Developing Organisms, Part II: The Developmental Toxicity Testing Guideline. American Journal of
Industrial Medicine 35:554-563. Schettler, T et al. January 2001. In Harm's Way: Toxic Threats to Child
Development. Greater Boston Physicians for Social Responsibility/Clean Water Fund. Slotkin, TA. 2004.
Guidelines for Developmental Neurotoxicity and Their Impact on Organophosphate Pesticides: A Personal View
from an Academic Perspective. Neurotoxicology 25: 631-640. Wallinga, D. April 1998. Putting Children First:
Making Pesticide Levels in Food Safer for Infants & Children. Natural Resources Defense Council.
8	US EPA. December 2002. Review of the Reference Dose and Reference Concentration Processes. Risk
Assessment Forum, EPA/630/8-02/002F.
10

-------
and guidance to allow more targeted testing. The panel indicated that current
testing protocols do not provide information collected at different life stages - that
is, comparison of effects of exposure during infancy, adulthood, or old age. We
believe the case of Chlorpyrifos9 demonstrates this point and the intrinsic
incompleteness of scientific evidence in the existing regulatory test guidelines.
For example, as scientists probed deeper into the activity of this organophosphate
pesticide in the laboratory, they found previously unknown effects on the
development and function of the brain and nervous system in embryos, fetuses,
and young animals (including possible serotoninergic10 and dopaminergic11
effects).12
When assessing risk of developmental neurotoxicity, EPA may need to identify
more sensitive endpoints, or indicators, accompanied by the analytical methods to
test for them. Cholinesterase inhibition is currently the driving endpoint for
organophosphate pesticides, but there is concern in the scientific community that
this parameter alone may not be enough to assess the consequences of exposure
for some pesticides. This raises the issue of how EPA can begin to evaluate the
many different pesticides with potentially overlapping but different mechanisms
and outcomes. It has been suggested in literature that using an in vitro approach,
or lower organisms, might enable a high-throughput screening for developmental
neurotoxicants.13 Proposed model systems include neural cell cultures,
invertebrate (such as sea urchin), and non-mammalian systems (like the
zebrafish).
See Appendix E for further details on issues related to toxicity testing.
Proposals Made for Changes to Testing Requirements
The data requirements for pesticide registrants in 40 CFR Part 158 have not been
comprehensively revised since 1984. Other than some minor changes and
additions in the "Maxi-Regs" final rule14 published on May 4, 1988, relative to
9	See Chapter 4, section on Substantial Changes Made in Aggregate Risk Assessment.
10	Serotoninergic effect: The serotoninergic system is known to modulate mood, emotion, sleep, and appetite and
thus is implicated in the control of numerous behavioral and physiological functions.
11	Dopaminergic effect: Neurons or brain cells in the mid-brain produce dopamine which is a neurotransmitter that
controls movement and balance and is essential to the proper functioning of the central nervous system.
12	Slotkin, TA. 2006. Developmental Neurotoxicity of Organophosphates: A Case Study of Chlorpyrifos. In:
Toxicity of Organophosphate and Carbamate Pesticides. RC Gupta, Elsevier: (In press). Aldridge, JE et al. 2005.
Developmental Exposure to Terbutaline and Chlorpyrifos: Pharmacotherapy of Preterm Labor and an
Environmental Neurotoxicant Converge on Serotonergic Systems in Neonatoal Rat Brain Regions. Toxicology and
Applied Pharmacology 203: 132-144. Qiao D et al. 2003. Fetal Chlorpyrifos Exposure: Adverse Effects on Brain
Cell Development and Cholinergic Biomarkers Emerge Postnatally and Continue into Adolescence and Adulthood.
Environmental Health Perspective 111 :536-544.
13	Slotkin, TA. 2004. Guidelines for Developmental Neurotoxicity and Their Impact on Organophosphate
Pesticides: a Personal View from an Academic Perspective. NeuroToxicology 25: 631-640. Slotkin, TA. 2004.
Cholinergic Systems in Brain Development and Disruption by Neurotoxicants: Nicotine, Environmental Tobacco
Smoke, Organophosphates. Toxicology and Applied Pharmacology 198: 132-151.
14	53 Federal Register 15951
11

-------
data formatting and flagging of certain toxicology studies, the requirements have
remained unchanged.
On March 11, 2005, EPA published proposed changes to the data requirements
regulation.15 OPP established a docket for this action, which includes the
proposed rule revisions, background and supporting documents, and comments
filed by outside individuals. The Agency was awaiting public comments prior to
either amending the proposed rule or promulgating a final rule. Public comments
were due June 9, 2005, but the deadline was extended to September 7, 2005.
Although 40 CFR Part 158 has remained virtually unchanged, there have been
major changes in the testing guidelines16 and de facto data requirements imposed
by the Agency. Since 1984, EPA has issued additional test guidelines, first under
the old Pesticide Assessment Guidelines, then as the OPPTS Harmonized
Guidelines.17 However, Part 158 has never been revised to reflect these additions
and changes. The new proposed rule attempts to codify the changes and make
new additions, changes, and revisions.
EPA proposed adding new requirements for developmental neurotoxicity test data
to the toxicity testing battery as part of the chemical registration process. The
original documents submitted to Office of Management and Budget (OMB) for
review indicate that current studies do not include an in-depth assessment of the
nervous system, and that the proposed developmental neurotoxicity study would
evaluate "functional and behavioral deficiencies, as well as structural alterations
to the nervous system, that may result from pesticide exposure that occurs in utero
and/or during early postnatal life."
However, in the proposed rule for which OMB completed its review on February
28, 2005, developmental neurotoxicity studies were changed to "conditionally
required," meaning that they would only be required under certain conditions.
Notes in the revision of the proposed requirements indicated the change was a
response to questioning by OMB. In a letter dated March 14, 2005, OMB
expressed its concerns on the increasing amount of resources devoted to pesticide
registration and the amount of data required to support a new registration. OMB
asked EPA to provide a specific plan on how the Agency will improve its current
toxicity testing data requirement for chemical registration, including considering
the International Life Sciences Institute's approach to testing pesticides.
According to OMB, the plan should contain information on critical actions, target
timelines (including the timing of the Part 158 final rule for conventional
15	70 Federal Register 12276
16	According to EPA, guidelines only provide suggested protocols which a registrant can follow to meet data
requirements posed in Part 158. Guidelines are not enforceable as they are not codified.
17	These guidelines were issued in 1998. They include measures of gross morphology in the brain; tests for evidence
of deficits in neurobehavioral functions (for example, auditory startle, habituation, associative learning and memory,
and monitoring of motor activity); and measurement of surrogate biomarkers specific to certain classes of neurotoxic
chemicals, such as the use of plasma cholinesterase inhibition as a biomarker for organophosphate pesticide
developmental neurotoxicity.
12

-------
pesticides), and stakeholders' engagement in the process. EPA briefed OMB on
October 6, 2005, about its Integrative Toxicology Testing Strategy, but had not
responded in writing to OMB.
Recommendations
We recommend that the Acting Assistant Administrator for Prevention,
Pesticides, and Toxic Substances:
3-1 Develop a Standard Evaluation Procedure to assess results of
developmental neurotoxicity testing. Within this Procedure, incorporate a
discussion on the developmental neurotoxicity data call-in results and
address which indicator, or combination of indicators, is considered most
sensitive and meaningful for assessing developmental neurotoxicity from
exposures during critical windows of development.
3-2 Evaluate the utility of using alternative toxicity testing methods to
evaluate developmental neurotoxicity. For example, assess whether
relying on primary work in cell culture, invertebrate, or non-mammalian
species, followed by more targeted examinations of specific processes in
mammalian species, may benefit the assessment of developmental
neurotoxicity and improve testing efficiency.
3-3 Implement the recommendations made by National Research Council in
its report to reduce uncertainty in neurodevelopment effects of exposure
during critical windows of development by:
•	ensuring that developmental neurotoxicity tests are conducted on
developing animals in addition to young adult animals;
•	assuring that developmental neurotoxicity test information is collected
at different life stages and that there is comparison of effects of
exposure during infancy, adulthood, or old age; and
•	revising the developmental neurotoxicity testing guidelines to better
assess risks of chemical exposure during the critical period of rapid
human brain development.
Agency Response and OIG Evaluation
The Agency agreed with Recommendation 3-1 and said that a formal internal
Standard Evaluation Procedure is expected to be completed in 2006. The Agency
also explained that "OPP is currently involved in the final stages of an ILSI
[International Life Sciences Institute] project which is developing approaches to
evaluating/interpreting" different sensitive parameters to address issues of
sensitivity and meaningfulness of the developmental neurotoxicity data.
The Agency agreed with the principle of Recommendation 3-2 but disagreed with
13

-------
the recommendation as drafted, citing the lack of well accepted methods for
testing. While we acknowledge the complexity of the issue, we encourage the
Agency to review the research in the area. We have modified Recommendation
3-2 to offer the Agency flexibility in its evaluation of new strategies for detection
of developmental neurotoxicant actions of suspected pesticides.
The Agency offered no statement of agreement or disagreement for
Recommendation 3-3 but claimed that most recommendations listed have been
appropriately addressed. It stated that the "current developmental neurotoxicity
test guideline includes testing during the major phases of development (i.e.,
during early lactation and around the time of weaning) as well as in young
adults." The Agency acknowledged that "there is no current test guideline in
which exposure occurs continuously from conception through old age" but that
required studies assess neurotoxicity at a variety of lifestages spanning the full life
span. Also, the Agency acknowledged that its current developmental
neurotoxicity test guideline recommends exposures up to post-natal day 10, that
an extended exposure period up to post-natal day 21 would be incorporated into
the next guideline revision, and that during the organophosphate developmental
neurotoxicity data call-in in 1999 it recommended manufacturers to dose animals
up to post-natal day 21. It mentioned that it has not yet developed the
developmental immunotoxicity testing guideline, although it has proposed
requiring the adult immunotoxicity test data for pesticide registration.
The full text of the Agency's response is in Appendix C, and our detailed
comments on that response are in Appendix D.
14

-------
Chapter 4
OPP Made Substantial Changes to Address
Aggregate Risk, but Challenges Remain
OPP has made substantial changes to the aggregate risk assessment process to
acquire more and better data on children's exposure to pesticides, but challenges
remain. Aggregate exposure risk assessments, as required by FQPA, specify that
all routes and pathways of exposure for a given pesticide be considered when
assessing risk. Although OPP has taken various actions, data are needed to
understand pesticide levels based on crop cycles, and more effort is needed to
consider food consumption information. Sufficient data are not always collected
because of time and cost constraints, and because EPA is often dependent on
other agencies for data. Without adequate and timely information to perform
aggregate exposure risk assessments, there will be uncertainty surrounding
pesticide licensing decisions and how pesticides impact children's health.
Substantial Changes Made in Aggregate Risk Assessment
Since FQPA's passage, EPA has made progress in acquiring exposure data from
young children and developing tools to use such data. Besides updating its
pesticide testing guidelines to include animal study data on reproductive and
developmental effects, EPA developed methods and generated laboratory data
that had an impact on the risk assessment of the pesticides methyl parathion,
chlorpyrifos, and diazinon. As a result, EPA cancelled the use of methyl
parathion on all fruits and many vegetables, eliminated the manufacturing of
chlorpyrifos for nearly all residential uses, and eliminated all indoor and garden
uses of diazinon as well as uses on about 20 different food crops. Also, EPA:
•	Identified and prioritized aggregate exposure data needs for other tolerance
reassessments and determinations.
•	Conducted and funded research to develop biomarkers18 for pesticides and
collect data on children's residential pesticide exposures, physical activity
patterns, and food handling practices.
•	Collaborated with the Agricultural Research Service of the U.S. Department
of Agriculture (USD A) to collect additional food consumption data from
children in USDA's Continuing Survey of Food Intakes by Individual, and to
generate a Food Commodity Intake Database from the survey data.
•	Collaborated with USDA's Agricultural Marketing Service to obtain
additional pesticide residue data for children's foods.
18 Biomarkers are substances in blood, other body fluids, or tissues that can be used to indicate exposure to
chemicals or diseases.
15

-------
•	Refined its drinking water assessment approach, and initiated research to
collect available data on pesticides in drinking water and effects of drinking
water treatment on degradation products toxicity.
•	Developed predictive methods for pesticide dermal and inhalation
determinations and exposure dose estimations.
Data on Children's Nondietary Pesticide Exposure Limited
FQPA requires EPA to consider all nonoccupational exposures from a single
pesticide when conducting aggregate risk assessment. Figure 4.1 illustrates some
nondietary routes of exposure and related data that are needed to assess risks.
Several computer models have been developed and are being used for exposure
and risk assessments. Such probabilistic models enable risk assessors to answer
questions about the sources, pathways, and factors that contribute to aggregate
exposures and risks.
Children engage in behaviors and consumption that can increase their risk of
pesticide exposures compared to adults. They eat more food, drink more water,
and breathe more air than adults on a body-weight basis. Further, the risk to
children is increased because they are generally lower to the ground than adults,
and they often engage in hand-to-mouth behavior that further adds to oral
pesticide exposure. Literature19 indicates that data on children's exposures and
the factors that affect their exposures are limited and generally not adequate to
assess children's exposures to a wide array of chemicals in their homes and other
environments. Both OPP and ORD scientists agree that better data on infants'
and toddlers' physical activity patterns would improve aggregate risk
assessments. Elements still missing on nondietary exposure as identified by the
scientific community and ORD include:
•	An understanding of the most important pathways of exposure for young
children.
•	Approaches for evaluating exposure for critical pathways, such as dermal and
indirect ingestion exposure.
•	Protocols for generating the exposure data.
•	Approaches for determining the exposure factor and data on fate-and-transport
(how a pesticide is applied and then moves in the environment).
19 Daston, G et al. February 2004. A Framework for Assessing Risks to Children from Exposure to Environmental
Agents. Environmental Health Perspectives 112(2): 238-256. Gitterman, BA and CF Bearer. October 2001. A
Developmental Approach to Pediatric Environmental Health. Pediatric Clinics of North America 48(5): 1071-83.
Goldman, L et al. April 2004. Environmental Pediatrics and Its Impact on Government Health Policy. Pediatrics
113(4): 1146-1157. Landrigan, PJ. October 2001. Children's Environmental Health. Pediatric Clinics of North
America 48(5): 1319-1330. Moya, Jetal. April 2004. Children's Behavior and Physiology and How It Affects
Exposure to Environmental Contaminants. Pediatrics 113(4): 996-1006. Weiss, B et al. April 2004. Pesticides.
Pediatrics 113(4): 1030-1036.
16

-------
To meet these needs, priority research needs include:
•	Pesticide use patterns.
•	Spatial and temporal distribution of pesticides in residential dwellings
(movement of pesticide chemicals across space and time).
•	Dermal uptake (pesticides entering through the skin from touching surfaces).
•	Nondietary/indirect ingestion (swallowing substances through such nondietary
means as hand-to-mouth activity or swallowing swimming pool water).
•	Oral exposure assessments (how often children lick but not eat food, bite a
toy, eat dirt, etc.) which include children's food handling practices.
Figure 4.1
20
NONDIETARY EXPOSURE DATA REQUIREMENTS FOR ASSESSING RISKS
Pesticide Use Sources
Residential: Indoor and Outdoor
Agricultural
Commercial
Occupational
Other sources
Fate and Transport: Exposure Media
Outdoor
Air
Water
Soil
Other outdoor surfaces
Plants
Indoor
Air
Water
House dust
Plants
Indoor surfaces
Clothes
Activities
Mouthing
Activity
Activities
Nondietary
Ingestions
Inhalation
The Human Body
Dermal
Uptake
Respiratory
U ptake
G.I. Tract
Absorbed Dose
Target Tissue Dose
Blood
Hair
Bone
Nail
Biological Media
Fat
Urine
Organs
Feces
2" OIG staff developed this figure based on EPA's documents on exposure data requirements.
17

-------
According to OPP and ORD officials, longitudinal consumption and activity data
are important for the assurance that short-term measurements relate to long-term
exposures and the validation of various modeled assumptions in the probabilistic
exposure models used by OPP. ORD scientists designed and implemented small
pilot field studies to capture exposure data and work out methods for two larger
children's health and pesticide exposure longitudinal studies (across a string of
time rather than a "snapshot" at a given time), but these projects have encountered
delays. The National Children's Study will examine the effects of environmental
influences on the health and development of more than 100,000 children across
the United States, following them from before birth until age 21. The long term
support for the National Children's Study21 is still under debate.22 Meanwhile, the
Children's Health and Environmental Exposure Research Study (CHEERS),
designed to study how children are exposed to household-use pesticides, was
cancelled by EPA on April 8, 2005.
EPA Relies on Public and Private Sources of Dietary Exposure Data,
and More Needed
Collecting national food consumption data is costly and complex, and EPA relies
on other Federal agencies for such data. FQPA contains specific provisions for
cooperative activities between EPA and USD A. USD A provides data on food
consumption, food commodity, and pesticide residue through such databases as
the Food Commodity Intake Database. However, there are various data gaps in
the collection of food consumption data for infants and children. Some food
consumption monitoring activities and gaps are discussed below; more details are
in Appendix F.
Since 1999, USD A integrated its food intake survey with another large survey
known as the National Health and Nutrition Examination Survey, which is
conducted by the Department of Health and Human Services (DHHS). However,
OPP has not incorporated the new consumption data into its risk assessment work,
but it reported planning to do so in 2006.
In 2002, Gerber Products Company conducted its latest Feeding Infants and
Toddlers Study, which surveyed 3,022 children from 4 to 24 months of age. This
study is the most comprehensive, largest, and nationally representative study on
food consumption for this age group. Additionally, USDA's School Lunch and
Head Start programs may be other worthwhile avenues for collecting data on
young children.
21
The National Children's Study is led by a consortium of Federal agency partners: DHHS, National Institutes of
Health, Centers for Disease Control and Prevention, and EPA. The Web site for additional information on the study
is http://nationalchildrensstudv.gov.
22	Kehn, BM. 2005. Children's Health Study Closer to Launch: Lack of Funding Could Cause Delays. JAMA 294:
2154.
18

-------
FQPA requires that drinking water be considered as a pesticide exposure pathway
in human health risk assessment. Preliminary data suggest conventional water
treatment processes do not appear to remove most pesticides, and chemical water
softening and disinfection processes may cause chemical transformation of some
pesticides into toxic by-products. Because the regulated communities are
responsible for generating the necessary data for pesticide risk assessments,
standard testing protocols and strategies for evaluating water treatment effects on
pesticide removal and transformation need to be developed and provided to
pesticide manufacturers. ORD's research shows that water treatment processes
are highly variable among community water systems.
The main data sources for dietary residues data are from monitoring studies
conducted by the Pesticide Data Program at the USD A Agricultural Marketing
Service and field studies of residues on commodities in their raw state at the farm.
Also, the Food and Drug Administration within DHHS maintains both regulatory
and incidence/level monitoring on particular commodity/pesticide combinations
and carries out its market basket survey known as the Total Diet Study. USDA's
Pesticide Data Program has generated extensive pesticide residue data on over 50
foods out of hundreds of key foods eaten daily in the United States, and many of
the foods not tested may be important in the diets of infants and children.
EPA's National Human Milk Monitoring Program studies, done in the 1970s to
evaluate the extent of human milk contamination with organochlorine and other
pesticides, are the most recent nationwide studies on breast milk. Since mother's
milk is a staple food for many newborns, and lactation is one major route of
elimination of endogenous and exogenous substances, knowledge of
concentration of pesticides in breast milk would reduce uncertainty and yield
better aggregate risk assessments.
Recommendations
We recommend that the Acting Assistant Administrator for Prevention,
Pesticides, and Toxic Substances:
4-1 Update the dietary exposure databases used in probabilistic models for risk
assessments as soon as the food consumption data from the 2003-2004
National Health and Nutrition Examination Survey become available in
2006. EPA should also update the Food Commodity Intake Database with
the latest food consumption survey data, and if possible use data such as
the Gerber Products Company's Feeding Infants and Toddlers Study.
4-2 Continue collaborating with USD A and with assessing whether there are
additional foods consumed frequently by children that should be included
by USD A in its Pesticide Data Program testing based on consumption
results reported from the National Health and Nutrition Examination
Survey.
19

-------
4-3 Expand its partnerships with USD A and DHHS to further reduce
uncertainty in aggregate risk assessments and
•	develop the methodologies for collecting longitudinal food intake and
physical activity information from children, and utilize such
methodologies
¦	to validate modeled assumptions in the probabilistic exposure
models used by OPP,
¦	to improve the current tools for estimating longitudinal
exposures, and
¦	to better understand timing or patterns of exposure;
•	collect more current data on pesticide concentration in human breast
milk among lactating women;
•	evaluate dietary and nondietary exposures among children at schools
and day cares; and
•	continue to conduct research to characterize effects of dietary and
nondietary exposures of pesticides on children's cognitive functions
and performance.
Agency Response and OIG Evaluation
The Agency generally agreed with the OIG recommendations in Chapter 4 and
the overall message of updating its food consumption databases, developing
methodology for collecting longitudinal food intake and physical activity
information from children, surveying human breast milk, and collaborating with
its Federal partners to acquire more and better dietary and nondietary exposure
data. However, the Agency commented that implementation of Recommendation
4-1 should wait until later in 2006, when it expects the release of "the first full set
of 2-day data" from the 2003-2004 National Health and Nutrition Examining
Surveys. We modified Recommendation 4-1 to meet the Agency's need for
"sufficient data to provide a database comparable to USDA's Continuing Survey
of Food Intakes by Individuals (CSFII) currently used by OPP."
In addressing our recommendation on assessing whether additional foods should
be analyzed for pesticides by the USD A, the Agency responded that the
"recommendation does not adequately take into account that sampling budgets are
finite and that infrequently sampling a broader swath of foods with less frequency
may not yield improved or more accurate dietary risk assessments." Also, EPA
commented that "the methodology for collecting longitudinal data is extremely
difficult," that longitudinal data are "extremely expensive and difficult to obtain,"
and that "long-term, intra-individual eating patterns through extensive consumer
surveys does not appear sufficiently promising, at this time, to justify further
pursuit of methodology development."
20

-------
We recognize that resource limitations constrain each agency's research efforts in
measuring residual and dietary consumption patterns. However, in the area of
dietary pesticide exposure, EPA has a major responsibility because it sets the
tolerances and registers the pesticide chemical use on food.
The full text of the Agency's response is in Appendix C, and our detailed
comments on that response are in Appendix D.
21

-------
Chapter 5
OPP Moving to Assess Cumulative Risk,
but Complexities and Concerns Remain
EPA has initiated steps to perform cumulative risk assessments for pesticides, and
more action is needed. FQPA had directed EPA to include in its assessment of
pesticide safety the risks associated with the cumulative effects of chemicals.
Conducting cumulative pesticide exposure risk assessments is complex because
children are continuously exposed to mixtures of low-dose pesticide chemicals
through many sources and routes. Organophosphate pesticides were the first class
of pesticides that OPP evaluated, but OPP does not expect to publish the final
cumulative risk assessment until later in 2006. ORD and OPP scientists have
identified the need for better methods and more sensitive tests to estimate the
amount of chemicals in humans. Concerns about adverse effects of chronic low-
dose, concurrent exposures continue.
Many Factors Impact on Assessing Cumulative Risk
Both ORD and OPP recognize that questions related to assessing cumulative risk
and evaluating risks to children are complex. Such assessments require an
integrated approach to understand the linkages along a continuum, from source to
exposure, and dose to adverse outcomes or disease. In a recent ORD Human
Health Research review, ORD scientists indicated that improved tools (methods,
data, models, risk assessment guidance, and toxicity testing methods and
protocols) are needed to assess cumulative risks. These tools need to look at
children's exposure to different pesticides with the same mechanism of toxicity,
through the air, water, soil, and dust; from eating and drinking; and from touching
and hand-to-mouth activities. EPA is currently designing and collaborating in
research studies to fill gaps. To understand the linkages along this continuum,
research is needed on:
•	the effects of concurrent exposure to pesticides and other chemicals with like
mechanisms of action;
•	the nature of chemical (especially pesticide) interaction in producing a toxic
response;
•	the methods by which cumulative risk to chemicals (especially pesticides)
with common mechanisms of action may be assessed; and
•	when and how concurrent, low-dose effects experienced during stages of
development could result in an adverse effect over the course of a person's
lifetime.
22

-------
Figure 5.1 shows the complex scientific steps and the data necessary for human
health risk assessments. The challenge in such assessments begins with knowing
the chemical source, its concentration in the environment, the media carrying the
chemical, activities contributing to the actual human exposure, the dose that
results from exposure, and the biological and other impacts on a person.
Figure 5.1 Source-to-Outcome Continuum
SCIENTIFIC ELEMENTS OF HUMAN HEALTH RISK ASSESSMENT
Source / Stressor
Formation
Chemical
Physical
Microbial
Magnitude
Duration
Timing
Transport /
T ransformation
Dis persion
Kinetics
Thermodynamics
Distributions
meteorology
Environmental
C h aracteriz atio n
Early Biological
Effect
Altered Structure/
Function
Ai Edema
Arrhythmia
	— Enzymuria
Necrosis
etc.
Cancer
Asthma
Infertility
etc.
Exposure
Air
Water
Diet
Soil & Dust
Pathway
Route
Duration
Frequency
Magnitude
Activity Patterns
Molecular
Biochemical
Cellular
Organ
Organism
Absorbed
Target
Internal
Biologically Effective
Individ ual
Commu nity
Popu lation
Statistical Profile
Reference Population
Susceptible Individual
Susceptible
Subpopulations
Population Distributions
From ORD's Safe Food Multi-year Plan (FY2003-2008): Figure 1
New Science Needed to Measure Effects of Concurrent Exposures
Exposure to multiple pesticides in homes, schools, public areas, food, and
drinking water is a routine part of life for children. Concerns about adverse
effects of chronic low-dose, concurrent exposures on children continue to surface.
Factors that complicate the evaluation of effects of chronic low-dose, concurrent
pesticide exposures include:
•	difficulties in measuring actual exposure levels;
•	length of time between exposure and appearance of symptoms;
•	the diversity of the symptoms; and
•	the nature of exposures experienced in the environment over time.
23

-------
Research confirms the need to perform more sensitive tests to measure effects of
chemical exposure. Application of such sensitive techniques as biomonitoring
methods and neuroimaging technology may be needed to detect significant, subtle
sub-symptomatic, and below-pesticide poisoning health effects.23 Based on
findings by an EPA technical panel,24 OPP's current testing methods used for
assessments of reproductive toxicity and developmental toxicity should be
revisited to determine whether changes are needed regarding animal-model
selection and the end points being measured.
Models, Computer Tools Can Enhance Cumulative Risk Assessments
ORD and OPP scientists have identified the need for better methods of estimating
internal doses in target tissues to determine cumulative risk assessments.
According to OPP, since real-world data to characterize risk of human exposure
often cannot be collected, it relies on computer models and other computer
simulation tools to assess and predict cumulative risks from pesticide exposures.
ORD researchers have developed a physiologically-based pharmacokinetic model
called the Exposure-Related Dose-Estimating Model (ERDEM). It simulates the
human organism and its ability to absorb, metabolize, store, and eliminate
chemicals. Pharmacokinetic describes a process to determine and quantify the
time course of distribution, biotransformation, and excretion of pollutants. This
model has been used by EPA risk assessors to simulate the reaction of multiple
pesticides. Discussions are now underway on how this model can be interfaced
with ORD's probabilistic human exposure and dose simulation model, SHEDS
(Stochastic Human Exposure and Dose Simulation pesticides exposure model), to
provide enhanced dose estimates.
Computational toxicology can provide tools that OPP can use to assess
differences and similarities between children's and adults' responses to chronic,
concurrent low-dose pesticide exposures. According to ORD, computational
toxicology uses computing approaches to link chemical transformation and
metabolism, exposure indicators, dose metrics, toxicity pathways, systems
biology, and modeling programs. This approach can improve testing efficiency
and reduce uncertainties in such areas as genomics (how an individual's genes
interact with each other and the environment); literature suggests that genetics can
play an important role in how different people are affected by pesticides. ORD
has developed a new Computational Toxicology Program that uses computational
chemistry, genomics, bioinformatics, and systems biology to:
•	improve understanding of the linkages in the continuum between the source of
a chemical in the environment and adverse outcomes;
•	develop approaches for prioritizing chemicals for screening and testing; and
23	Such effects may include those resulting from chronic low-dose, concurrent exposures.
24	US EPA. Review of the Reference Dose and Reference Concentration Processes. Risk Assessment Forum,
EPA/630/8-02/002F, December 2002.
24

-------
• produce better methods and predictive models to assess risk.
OPP is awaiting products from ORD's Computational Toxicology Program to
improve the pesticide regulatory program by incorporating bioinformatics,
genomics, and toxicogenomics in its work. Computational toxicology may be the
new scientific tool that could bring interdisciplinary work (e.g., public and
environmental health, agricultural sciences) and data (e.g., human exposure
information, chemical structure, fate and transport information) on a broader
ecological and biological context for OPP. OPP's Health Effects Division plans
to build a team of multi-disciplinary experts with diverse background to meet
future work demands from computational toxicology.
Recommendations
We recommend that the Acting Assistant Administrator for Prevention,
Pesticides, and Toxic Substances:
5-1 Follow through with ORD to finalize the integration of probabilistic
modeling outputs with physiologically based pharmacokinetic modeling to
better address cumulative risk from concurrent exposure to pesticides and
other chemicals with like mechanisms of action.
5-2 Continue to execute plans and strategies on how computational toxicology
outputs from ORD's Computational Toxicology Program will integrate
into OPP's regulatory process; monitor, assess and document progress.
Agency Response and OIG Evaluation
The Agency agreed with Recommendation 5-2, but indicated that OPP's past and
continuing actions fully address Recommendation 5-1. The Agency also
commented that the "draft report does not recognize" additional on-going efforts
by EPA to link probabilistic exposure models with physiologically-based
pharmacokinetic models. While we agree with the spirit of outsourcing software
development for probabilistic models or funding external researchers to seed
research and develop the next generation of environmental scientists, we do take
the position that the Agency should fully utilize the expertise of ORD scientists to
develop exposure estimate models (including physiologically-based
pharmacokinetic modeling) for its core work. We maintain our position for
Recommendation 5-1 that OPP needs to coordinate its probabilistic efforts with
ORD. The full text of the Agency's response is in Appendix C, and our detailed
comments on that response are in Appendix D.
25

-------
Chapter 6
Opportunities Exist to Better Manage
FQPA Implementation
Since 1996, both OPP and ORD have generated a number of FQPA-related
scientific outputs; however, there are additional opportunities to better manage
health risk for children and reduce uncertainty in pesticide decisions.
Specifically, OPP can:
•	Take action on finalizing science policy papers.
•	Continue working with organizations to assess alternative testing strategies.
•	Use logic models to guide efforts.
•	Develop a multi-year strategic plan to support goals.
These efforts should result in improved data and better protection against
pesticide exposures for infants and children.
OPP Science Policy Papers Not All Finalized
The implementation of FQPA required OPP to revisit some of its existing policies
related to the determination and regulation of dietary risk, and raised a number of
new issues for which policies needed to be created. Since 1996, OPP has
developed and refined nine science policy areas25 identified as key to
implementing the FQPA. These activities were done in collaboration with other
Agency offices and programs, and with external stakeholders from industry,
environmental groups, and other interested entities. We found that a number of
the papers in the nine science policy areas are still in draft format. OPP officials
said they consider these draft papers as operating documents that they will update
as new information becomes available. Papers in draft form, as shown on EPA's
Web site as of August 2005, are shown in Figure 6.1.
25 The nine policy areas are: FQPA 10-Fold Safety Factor; Dietary Exposure and Risk Assessment; Threshold of
Regulation; Drinking Water Exposure; Residential Exposure; Aggregate Exposure and Risk Assessment;
Cumulative Risk Assessment for Pesticides with a Common Mechanism of Toxicity; Cholinesterase Inhibition End
Point; and Use and Usage Information.
26

-------
Figure 6.1: Science Policy Areas Identified as Key to Implementing FQPA but
Papers Still in Draft
Key Science
Policy Areas
Science Policy Papers Still in Draft
FQPA 10-Fold Safety
Factor
- Standard Operating Procedures (SOPs) for Use of the
FQPA Factor
Dietary Exposure and
Risk Assessment
-	Guidance for the Submission of Probabilistic Human
Health Exposure Assessments to the Office of Pesticide
Programs
-	Use of the Pesticide Data Program (PDP) in Acute
Dietary Assessments (not accessible via the Web site)
Drinking Water Exposure
-	Drinking Water Screening Level Assessment, Part A:
Guidance for Use of the Index Reservoir in Drinking
Water Exposure Assessments
-	Standard Operating Procedure for Incorporating
Screening-Level Estimates of Drinking Water Exposure
in Aggregate Risk Assessments
-	Water Treatment Effects on Pesticide Removal and
Transformation
Residential Exposure
-	Standard Operating Procedures (SOPs) for Residential
Exposure Assessment
-	Framework for Assessing Non-occupational / Non-
dietary (Residential) Exposure to Pesticides
Cumulative Risk
Assessment for
Pesticides with a
Common Mechanism
of Toxicity
- Application of the 10X safety Factor in Cumulative Risk
Assessment.
EPA Needs to Continue Pursuing Alternative Testing Efforts
EPA is working with external scientific organizations to assess alternatives to the
current pesticide data requirements: the International Life Sciences Institute and
the National Academy of Sciences.
EPA started discussions with the Health and Environmental Sciences Institute at
the International Life Sciences Institute on alternative testing strategies in 2001,
and created a cooperative agreement to work with the Agricultural Chemical
Safety Assessment Technical Committee under these groups. ORD and OPP
scientists are members on the three task forces that prepared a proposal for a new
testing strategy and three white papers for publication. Under the new strategy
being designed by the Institute, the following data could be used to inform a more
targeted testing approach in the design of studies or to support waiving specific
toxicology tests:
• Data on toxicity and dose-response;
27

-------
•	Mechanism or mode of action of the chemical;
•	Pharmacokinetic data;
•	Data on age-related sensitivity or susceptibility to chemical exposure; and
•	Information on potential or actual exposure to humans.
The National Academy of Sciences project is a two-part study funded by EPA to
advance current approaches to toxicity testing to meet regulatory data needs.
Part 1 of the study was designed to provide a report reviewing selected aspects of
several relevant reports provided by EPA and others on toxicity testing and
assessment. Part 2 is to be funded separately at EPA's option, and is to present a
long-range vision and strategic plan for advancing the practices of toxicity testing
and human health assessment for environmental contaminants. If funded, the
second report is anticipated for fall 2006.
Logic Models Could Better Guide Efforts
OPP has not developed a logic model as part of the strategic planning process.
Program logic models represent the logic underlying a program's design. They
can indicate how various components are expected to interact, the products or
services they produce, and how they generate the desired results. Significantly, a
logic model distinguishes between outputs (the specific tasks performed) and
outcomes (the actual results).
Figure 6.2 provides a logic model we produced that could assist EPA in gathering
the scientific data needed to implement FQPA. We have described in the logic
model our findings on EPA's response to the scientific data needed for
implementation of the FQPA. This logic model distinguishes between tasks
performed by OPP and ORD. Operations managers within OPP should focus on
the production of high quality outputs in the model we produced, but managers
who are concerned with overall FQPA implementation must look beyond outputs
to outcomes. In our model, outputs are valuable because they lead to benefits.
The long-term outcomes in the model are examples of how these ultimate criteria
can be used to gauge EPA's effectiveness in implementing FQPA.
28

-------
Figure 6.226
Logic Model on EPA's Implementation of the FQPA and Response to Data Needs
Customers
and
Resources
Legislators
Enacted
FQPA
Regulations
Interpreted
and
Enforced
Compliance
with New
and Existing
I
EPA Scientists,
Risk Assessors
and Risk
Managers
Identified and
Prioritized
Solutions to Fill
Data Gaps
NonEPA
Scientists From
Other Federal
Agencies,
Universities,
and
Nonprofit
Research
Organizations
Contributed
Suggestions
and Ideas for
Research
OPP Activities and Outputs
Policy Papers on
FQPA 10X Safety
Factor
Dietary Exposure &
Risk Assessment
Drinking Water
Exposure
Residual Exposure
Threshold of
Regulation
Aggregate Exposure
& Risk Assessment
Cumulative Risk
Assessment
Cholinesterase
Inhibition End Point
Use and Usage
Information
Collaboration with USDA to
Capture Consumption and
Pesticide Residual Data
Toxicity Testing Results and
Aggregate Risk Assessments
Chemical Reregistrations and
Tolerance Determinations
OPP Workforce Training on
Aggregate & Cumulative Risk
Assessments; Probabilistic
Models; PBPK Models
Annual Work Plans
Preparation for
Transformation to a
Multidisciplinary Workforce
Endrocrine Disruptor Screening
Program
ORD Activities and Outputs
Strategic Planning
ORD Multiyear Research
Plans
•	Exposure Research
•	Toxicity Research
•	Endocrine Disruptor
Screening Research
•	Children's Health
Research
Funding STAR Grants
National Children's Study and
CHEERS Research Planning
Children's Environmental
Health & Disease Prevention
Research Centers
Risk Assessment Workshops/
Presentations
Children's Exposure
Workshops/ Presentations
Probabilistic & PBPK Models
Development
Chemical Nominations for
Biomonitoring at CDC
Computational Toxicology
Research
Research Publications
Immediate
Outcomes
Intermediate Long-Term
Outcomes Outcomes
EPA Shifts
its Risk
Assessment
Paradigm
EPA Regulates
Pesticide in/on
Food Based on
Reassessed
Tolerance
Levels
EPA
Regulators
Strategically
Plan
Regulations for
Inexpensive,
Yet Sensitive
Toxicity Testing
Data
Requirements
Researchers
Develop
Scientific
Methods and
Produce Data
To Close Data
Gaps in the
J Source-
Exposure-Dose
- and - Effect(s)
Continuum
EPA Enforces
j New Toxicity i
Testing
Regulations i
EPA Regulates [
Pesticide i
Chemicals j
j Based on i
Knowledge of [
Pathology, i
Genomics, [
Toxiconetics, ki
Function,
Performance and i
Health Effects of [
Chemical i
Exposure '
~I	' '
| EPA and Other	]
Federal	i
Agencies Track	j
Body Burdens	i
and Health	^
Effects	i
Information	'
EPA Integrates
Multi-
disciplinary
Scientific Data
for Better Risk
Assessment
and
Management
Decisions
Public
Health of
Children
and the
Environ-
ment
Protected
Manufact-
urers
Produce
Safer
Pesticidal
Chemicals
End-users
Use
Suitable
Pesticides
Appro-
priately
Anticipated
Actualized
Multi-year Strategic Plan Can Support Goals
We found OPP has not developed a multi-year strategic plan that would enable it
to more efficiently achieve FQPA requirements and the ultimate goal of
preventing the exposure of children to pesticides. OPP has divisional annual
work plans that focus on annual work outputs, but no overarching, program-wide
26 OIG staff developed this figure based on data collected during this evaluation.
29

-------
strategic plan with immediate, intermediate, and long-term goals and expected
outcomes. According to OPP, FQPA's mandate of registration renewal every 15
years is expanding OPP's work plans to encompass long-term strategic planning.
Among the annual work plans that we reviewed, only the Fiscal Year 2004 plan
from the Health Effects Division described short-term, intermediate, and long-
term strategic projects. However, nearly all projects in this plan were labeled as
"high priority," with limited personnel resources designated. To more efficiently
plan its FQPA-required work and the resources needed to accomplish that work,
OPP should develop a multi-year strategic plan.
Recommendations
We recommend that the Acting Assistant Administrator for Prevention,
Pesticides, and Toxic Substances:
6-1 Either finalize all of the Science Policy issue papers, or change the word
"draft" to "operational" and schedule annual updates.
6-2 Sustain the development of an alternative testing strategy, ensuring that
risks are assessed across the entire life cycle of development.
6-3 Develop an overarching logic model and long-term strategic plan across
divisions to identify and link immediate work outputs to outcomes.
Agency Response and OIG Evaluation
The Agency agreed with all recommendations in this chapter. However, it
commented that the OIG report mischaracterized the status of some of its science
policy papers and that, specifically, Figure 6.1 fails to list three other draft policy
papers. We made modifications to Figure 6.1 based on comments provided by the
Agency. However, we have concerns over the "Science Policy Issues &
Guidance Documents" Web site being unclear, out-of-date, and misleading.
Additionally, we noted that there are four science policy papers posted on the
Web page for which OPP plans to issue Federal Register Notices announcing their
withdrawal. It is our opinion that managing the currency of science policy papers
and Web sites should be an Agency priority. We believe the Agency's Web site
is a tool that can be used to demonstrate how OPP applies sound science to reduce
uncertainty in its regulatory decisions. The full text of the Agency's response is
in Appendix C, and our detailed comments on that response are in Appendix D.
30

-------
Appendix A
Glossary of Terms
Aggregate Exposure: The combined exposure of an individual or defined population to a
specific agent or stressor via relevant routes, pathways, and source.
Aggregate Risk: The risk resulting from aggregate exposure to a single agent or stressor.
Cholinesterase: One of many important enzymes needed for the proper functioning of the
nervous systems of both humans and animals.
Common Mechanism of Toxicity: When two or more pesticide chemicals or other substances
cause a common toxic effect(s) by the same, or essentially the same, sequence of major
biochemical events (for example, a mode of action).
Computational Toxicology: The application of mathematical and computer models to better
understand the mechanisms through which a given chemical or exposure induces harm and
predicts adverse effects.
Concurrent Exposure: Potential human exposure by all relevant pathways, durations, and
routes that allow one chemical to add to the exposure of another chemical such that the total risk
is an estimate of the sum of the exposures to the individual chemicals.
Cumulative Risk: The combined risks from aggregate exposures to multiple agents or stressors.
Cumulative risk is the risk of a common toxic effect associated with concurrent exposure by all
relevant pathways and routes of exposure to a group of chemicals that share a common
mechanism of toxicity.
Data Call-ins: A process through which EPA seeks data from appropriate pesticide
manufacturers.
Deterministic Model: A model that contains no random elements. The model provides a point
estimate of exposure, assuming that a typical child eats an assumed mass of food per day with a
given concentration of a pesticide residue.
Developmental Neurotoxicity: Involving substances that damage a developing nervous system,
including the brain.
Developmental Toxicity: Adverse effects on the developing organism that may result from
exposure prior to conception (either parent), during prenatal development, or postnatally until the
time of sexual maturation. The major manifestations of developmental toxicity include death of
the developing organism, structural abnormality, altered growth, and functional deficiency.
Dose: The amount of a substance available for interactions with metabolic processes or
biologically significant receptors after crossing the outer boundary of an organism.
31

-------
Endpoint: An observable or measurable biological event or chemical concentration (e.g.,
metabolite concentration in a target tissue) used as an index of an effect of a chemical exposure.
Mechanism of Action: The complete sequence of biological events that must occur to produce a
toxic effect.
Pathway of Exposure: The physical course a pesticide takes from the source to the person
exposed (e.g., through food or drinking water consumption or residential pesticide uses).
Pharmacodynamics: The determination and quantification of the sequence of events at the
cellular and molecular levels leading to a toxic response to an environmental agent (also called
toxicodynamics).
Pharmacokinetics: The determination and quantification of the time course of absorption,
distribution, biotransformation, and excretion of pollutants (also called toxicokinetics).
Physiologically-Based Pharmacokinetic Model: A model that estimates the dose to a target
tissue or organ by taking into account the rate of absorption into the body, distribution between
target organs and tissues, metabolism, and excretion.
Probabilistic Model: A system whose output is a distribution of possible values; the model
considers the range of estimates and provides a probability distribution of exposures.
Risk: The probability of adverse effects resulting from exposure to an environmental agent or
mixture of agents.
Risk Assessment: The evaluation of scientific information on the hazardous properties of
environmental agents (hazard characterization), the dose-response relationship (dose-response
assessment), and the extent of human exposure to those agents (exposure assessment).
Route of Exposure: The way a chemical enters an organism after contact (e.g., ingestion,
inhalation, or dermal absorption).
Statistical Significance: The probability that a result is not likely to be due to chance alone.
By convention, a difference between two groups is usually considered statistically significant if
chance could explain it only 5 percent of the time or less.
Subchronic Exposure: Exposure to a substance spanning approximately 10 percent of the
lifetime of an organism.
Tolerance: The maximum level of pesticide residue allowed in or on human food and animal
feed.
Toxicology: The study of harmful interactions between chemical, physical, or biological agents
and biological systems.
32

-------
Appendix B
Details on Scope and Methodology
During our evaluation, we conducted a literature review on FQPA and the potential health
outcomes of prenatal and childhood exposures to pesticides. We reviewed documents pertinent
to risk assessment prepared by EPA and other Federal agencies. We examined dietary
assessment methods and examined food consumption databases, pesticide residue data sources,
and probabilistic models to better understand how such data integrate in risk assessments
performed by OPP. We reviewed the annual work plans prepared by OPP and the multi-year
research plans by ORD. We collected data to:
•	Report the types of scientific tools and data needed to meet the regulatory challenges posed
by FQPA;
•	Identify and evaluate OPP's existing toxicity testing strategy and tools;
•	Assess and report the types of dietary and nondietary exposure data OPP has and needs for
aggregating risks to establish tolerances that protect children;
•	Identify challenges and assess new science, technology, and research that could enhance
OPP's strategy to assess cumulative pesticide exposure risks; and
•	Identify and recommend strategies to assure successful implementation of the FQPA.
We reviewed EPA's infrastructure and Human Health Research Strategies to determine ORD's
role in supplying scientific data and tools for OPP's pesticide regulatory work. We attended
ORD's Human Health Research Program review meetings and public sessions on proposed
revisions to 40 CFR Part 158. Furthermore, we interviewed various administrators, science
policy directors, scientists, and risk assessors from various EPA offices, as well as experts from
other Federal agencies and outside organizations, to capture expert viewpoints, clarify our
interpretations, and confirm our findings (see below). We visited various offices in Washington,
DC, and Research Triangle Park, North Carolina.
Organizations for Experts Reviewed
EPA
•	Office of Pesticide Programs
•	Office of Prevention, Pesticides, and Toxic Substances
•	Office of Research and Development
•	Office of Children's Health Protection
Other Federal
• Department of Agriculture
Agencies
• Department of Health and Human Services

• Office of Management and Budget
Nongovernmental
Sources
•	Environmental health scientists, toxicologists, and epidemiologists
•	Pediatricians/physicians
•	Environmental law professors from universities
•	Children's Environmental Health and Disease Prevention Research Centers
•	Environmental health groups
33

-------
Appendix C
Response from the Agency
October 17, 2005
MEMORANDUM
SUBJECT: Draft Evaluation Report: Opportunities to Improve Data Quality and Children's
Health through the Food Quality Protection Act
FROM: Jim Jones /s/
Director, Office of Pesticide Programs
TO:	Jeffrey K. Harris
Director for Program Evaluation, Cross Media
Office of Inspector General
Thank you for the opportunity to comment on the draft report, dated September 1, 2005, by the
Office of Inspector General on EPA's Implementation of the Food Quality Protection Act
(FQPA). The Office of Pesticide Programs (OPP) appreciates the Office of Inspector General's
review of OPP's implementation of FQPA.
I. Introduction to OPP's Response to OIG's Report
The EPA's OIG draft report "Opportunities to Improve Data Quality and Children's Health
through the Food Quality Protection Act" focuses upon the following issues: (1) What data
requirements were required by FQPA; (2) Whether testing guidelines, requirements, and
evaluation procedures allow EPA's OPP to determine the potential adverse effects of pesticide
exposure on the developing nervous system; (3) What challenges did OPP overcome and what
opportunities exist for OPP to acquire better pesticide exposure data to aggregate risks; (4)
What challenges exist and what opportunities are available for OPP to improve cumulative risk
assessments; and (5) What opportunities exist to better manage pesticide health risks for
children. The OIG's specific recommendations and OPP's comments on the draft report and the
recommendations are below.
34

-------
Overall, the OIG draft report confirms OPP's many accomplishments to implement FQPA
mandates to strengthen the human health protections for infants and children:
1.	The draft report notes that OPP developed, issued and implemented a vast number of
key science policies and regulations, improving our ability to assess risks.
2.	The draft report recognizes that OPP has successfully incorporated into its pesticide
risk assessment procedures new cutting-edge methodologies and tools, including
aggregate exposure and cumulative risk assessment.
3.	The draft report corroborates OPP's risk management priorities (focusing on high risk
pesticides such as organophosphates) and acknowledges the improved public health
outcomes that have resulted from OPP's risk management decisions (such as the
cancellation of certain chlorpyrifos products due to risks to children).
4.	The draft report highlights forward-looking steps that both OPP and OIG believe are
important, exemplified by the ongoing National Academy of Sciences (NAS) and
International Life Sciences Institute (ILSI) work on improved approaches to human
health risk assessment, which incorporates more targeted testing and computational
toxicology to refine and reduce animal testing.
However, OPP believes that the draft report is uneven, and sometimes
misleading, in its evaluation of its progress in implementing the
provisions of FQPA. The draft report glosses over significant scientific
accomplishments of the past nine years (i.e., since 1996 when FQPA was enacted). It then tends
to focus on issues that are 1) minor and relatively insignificant within
the overall scope of FQPA implementation, 2) characterized incorrectly,
or 3) outside the control of the Agency.
Likewise, while the draft report recognizes potential opportunities for collaboration between
OPP and ORD and other organizations, it generally fails to acknowledge that such collaboration
has been on-going for years, and provides little indication of the
significant scientific gains that have been realized through these efforts.
The following comments are provided to clarify the specific issues, and
suggestions are made in order to add balance to the report and provide
emphasis on opportunities and challenges that remain for continued implementation of FQPA.
Of the thirteen (13) recommendations provided by the OIG in their draft report, OPP agrees with
eight of them because, in fact, we are already implementing most of them (i.e., six of the eight
are already implemented by OPP). OPP appreciates the OIG validating its current work
processes. OPP believes, however, that the final report would be more complete and balanced by
acknowledging that OPP had already identified and begun working on many of the issues
discussed in the draft. On the other hand, OPP disagrees with five of the recommendations
mainly because we feel, at this time, these recommendations have been made either prematurely
or with little practical or scientific basis.
See OIG Comment
in Appendix D,
Note 1
See OIG Comment
in Appendix D,
Note 2
See OIG Comment
in Appendix D,
Note 1
35

-------
II. Chapter 1 - Introduction to the Report
The first chapter of the report provides background on the scope and methodology of the report
and on FIFRA and FQPA generally, and also includes a brief summary of the results of the
OIG's review. This summary acknowledges that to meet the requirements of FQPA "EPA
instituted numerous data requirements that should provide better protection. Additionally, EPA
took steps to develop science policies, develop methods and tools, and collect required data on
aggregate and cumulative risk." However the summary concludes with "significant data gaps
remain" and summarizes the report's general recommendations
concerning collecting and using data. While we work actively to
improve the databases supporting our decisions, we generally disagree
with the draft report's unsupported, sweeping conclusion that
"significant data gaps remain."
See OIG Comment
in Appendix D,
Note 3
One specific technical correction we would like to draw your attention to is on page 3, lines 5-7
of the draft report. The definition/explanation of cumulative exposure is incorrect. It states
"Cumulative risk information for a given common toxic effect is calculated separately for each
exposure route and duration and then combined." It is not the toxic 	
effect, or adverse outcome, that defines the ability to group chemicals
for a cumulative assessment; rather, it is a common mechanism of
toxicity (per OPP's peer-reviewed guidance on this topic).
See OIG Comment
in Appendix D,
Note 4
III. Chapter 2 - FQPA Inspired Numerous EPA Data Requirements
Chapter 2 of the report highlights the FQPA inspired changes to data requirements. FQPA
established a single, health-based standard and requires that allowable residue levels for food use
pesticides be protective of the health of infants and children. Chapter 2 includes a table outlining
key requirements and amendments to FIFRA and FFDCA resulting from FQPA, as interpreted
by EPA.
We would like to draw your attention to the following technical corrections:
1. Page 5: Table 2.1
Table 2.1 describes selected FQPA statutory changes; however, it
misstates some of the changes. We have prepared a substitute table that
accurately reflects the FQPA changes, and have added citations to the
appropriate sections of the statutes.
See OIG Comment
in Appendix D,
Note 4
36

-------
Table 2.1 Key Changes and Additions for Pesticide Regulations Due to FQPA
FIFRA
¦	Permits emergency suspension of a pesticide without simultaneously issuing a notice of
intent to cancel. - FIFRA sec. 6(c)(3).
¦	Requires Registration Review, with a goal of once every 15 years - FIFRA sec. 3(g).
¦	Establishes special provisions for minor use pesticides, including public health pesticides
- FIFRA sec. 3 and sec. 4.
¦	Links tolerance reassessment to reregi strati on - FIFRA sec. 4(g).
¦	Establishes special provisions for antimicrobial pesticide registration - FIFRA sec. 3(h).
¦	Establishes mandate for continuing expedited consideration of application for pesticides
meeting one or more criteria for reduced risk pesticides - FIFRA sec. 3(c)(10).
However, this provision was overtaken by 2004 Pesticide Registration Improvement Act
(PRIA), which establishes specific timeframes for reduced risk pesticides.
¦	Establishes Science Review Board to assist in scientific peer-reviews conducted by the
FIFRA Scientific Advisory Panel - FIFRA sec. 25(d)(2).
FFDCA Sec. 408
¦	Delinks pesticides from the Delaney clause (FFDCA sec. 409) which prohibited
establishment of tolerances for carcinogens, by placing all pesticide authority in FFDCA
sec. 408.
¦	Establishes standard for establishing a tolerance based on whether tolerance is "safe",
defined as "a reasonable certainty that no harm will result from aggregate exposure to the
pesticide chemical residue, including all anticipated dietary exposures and all other
exposures for which there is reliable information."
¦	Defines types of exposures to be aggregated for risk assessment purposes.
¦	Requires consideration of cumulative effects of pesticides having common mechanism of
toxicity.
¦	Requires tolerance reassessment on phased schedule: 33% within 3 years of enactment;
a second 33% within 6 years; and the remaining number within 10 years of enactment.
¦	Creates a presumption in favor of applying an additional 10-fold margin of safety for
infants and children for threshold effects.
¦	Requires development of an estrogenic substances screening program.
¦	Requires development and distribution of consumer information on pesticide risks and
benefits.
FQPA sec. 301-305
Prescribes data collection activities to ensure the health of infants and children:
¦	Collection of adequate data on food consumption patterns of infants and children.
¦	Improved data collection on occurrence of pesticide residues in foods most likely to be
consumed by infants and children.
¦	Evaluation of pesticide usage information and improved information gathering.
37

-------
2. Page 6: First paragraph, lines 3-5
The statement that FQPA requires "that there is a reasonable certainty of no harmful
developmental effects for children" is incorrect. FQPA does not limit the assessment or
consideration of adverse effects to only those that are considered "developmental." The correct
safety standard is ".. .that there is a reasonable certainty that no harm
will result from aggregate exposure to the pesticide chemical residue,
including all anticipated dietary exposures and all other exposures for
which there is reliable information."
3. Page 6: Second paragraph, lines 5-6
The precise text of the CFRuses "domestic animals" (not "terrestrial mammals")
IV. Chapter 3 - OPP Lacks Consistent Data on the Developing Nervous System to
Determine Potential Adverse Effects
Chapter 3 provides an overview of the importance of developmental neurotoxicity data and the
concerns which have been raised by stakeholders concerning conducting developmental
neurotoxicity tests such as the expense of conducting such tests and the difficulty in interpreting
the results. Below are some specific comments we have on this chapter.
Recommendation 3-1: Develop a Standard Evaluation Procedure to assess results of
developmental neurotoxicity testing.
OPP agrees with this recommendation and has been actively addressing it over the past few years
and expects to have a final Standard Evaluation Procedure (SEP) in 2006.
Over the past few years, a standard Data Evaluation Review format was developed which
harmonized with Health Canada's Pesticide Management Regulatory Agency (PMRA) and
included detailed guidance on the summary/reporting, evaluation, and interpretation of all
endpoints typically reported in a Developmental Neurotoxicity (DNT) study.
In the summer of 2004, OPP directed resources to the assessment of 29 unreviewed DNT studies
by having the OPP's DNT Committee assemble a package of references and guidance that was
provided to all OPP toxicologists responsible for reviewing the studies.
A special two-day training session was held to ensure that all
toxicologists would be familiar with the documented evaluation
procedure.
Now that these reviews have been completed, the informal guidance package is being used as the
nucleus of a more formal, internal Standard Evaluation Procedure (SEP) that is expected to be
completed in 2006.
See OIG Comment
in Appendix D,
Note 5
See OIG Comment
in Appendix D,
Note 6
38

-------
Finally, there is an on-going International Life Sciences Institute (ILSI)ZRisk Science Institute
(RSI) project on DNT data interpretation (including OPP and ORD staff experts in DNT testing
and evaluation) that is addressing DNT methodology and performance, variability, positive
control data, and statistical data analysis; the results of these efforts will be published in the near
future.
Recommendation 3-2: Evaluate whether relying on primary work in cell culture,
invertebrate, or non-mammalian models, followed by more targeted examinations of
specific processes in mammalian models, may benefit the assessments of developmental
neurotoxicity and improve testing efficiency.
Since there was no discussion of this strategy in the draft report (it is
mentioned in passing at the top of page 10), the basis for this
recommendation is unclear. OPP actively supports the development of
test methods that do not use (or use fewer) animals (i.e., in vivo testing)
to assess the potential toxicity of a pesticide. To our knowledge, however, there are no well-
accepted methods of the types listed in the recommendation (i.e., in vitro, invertebrate, or non-
mammalian models), that have been shown to have clear linkages to adverse functional outcomes
for toxicity to developing nervous system in mammals. Consequently, while OPP agrees with
the principle behind the recommendation, OPP disagrees with this recommendation, as drafted.
See OIG Comment
in Appendix D,
Note 7
OPP is currently working with ORD/National Health and Environmental
Effects Research Laboratory (NHEERL)/Neurotoxicology Division on
several research projects relevant to this recommendation:
See OIG Comment
in Appendix D,
Note 7
¦	Development of a high throughput screening battery (primarily using cell culture
techniques) focused on detecting chemicals likely to be developmentally neurotoxic; and
¦	A proposal to evaluate the utility of considering molecular markers for some classes of
neurotoxic pesticide chemicals, to determine whether such markers could be more
sensitive/equivalently sensitive than the current DNT guideline for detecting adverse
effects on the developing nervous system.
Finally, EPA (both OPP and ORD) have been part of the steering committee for an upcoming
symposium to look at alternative ways to perform DNT-type studies. The TestSmart
Developmental Neurotoxicity (DNT) Test symposium, sponsored by the Johns Hopkins
University, the National Institute of Environmental Health Sciences
(NIEHS), and others, will be held March 13 - 15, 2006 to discuss the
DNT and alternative methods. More information is available at:
http://caat.jhsph.edu/dnt/.
See OIG Comment
in Appendix D,
Note 8
39

-------
Recommendation 3-3: Evaluate which indicator, or combination of indicators, is most
sensitive and meaningful for assessing developmental neurotoxicity consequences of
exposure during critical windows of development.
OPP agrees generally with this recommendation and we are addressing it a
Evaluation Procedure (SEP) for Developmental Neurotoxicity (DNT)
tests. DNT studies involve dosing pregnant animals (generally rats)
during gestation, allowing them to deliver their pups, and continuing
dosing the mothers during the lactation period (up to 21 days after birth).
In many cases, the pups are dosed directly beginning a few days after birth
(approximately 21 days after birth). This exposure period generally encom
windows of development of the rodent brain.
Based on a review of dozens of DNT studies, OPP believes that it is
unlikely that a single 'most sensitive' parameter would be found.
During the course of the experiment, pups are examined daily (clinical
observations) and are also tested at various time points for the following
indicators of developing nervous system functions: auditory startle habituation, functional
observations, motor activity, learning and memory, and brain pathology (brain weight,
neuropathology and brain morphometric measurements). Part of evaluating a DNT study is
understanding that each of these parameters is important and needs to be assessed in a "weight-
of-the-evidence" approach. Different chemicals with different modes of action are likely to
affect different functional systems. OPP is currently involved in the final stages of an ILSI
project which is developing approaches to evaluating/interpreting these different parameters
which will help address the issue of the sensitivity/meaningfulness of DNT data.
s part of the Standard
See OIG Comment
in Appendix D,
Note 9
up to weaning
ipasses the critical
See OIG Comment
in Appendix D,
Note 9
Recommendation 3-4: Implement the recommendations made by National Research
Council in its report to reduce uncertainty in neurodevelopment effects of exposure during
critical windows of development by:
a)	ensuring that developmental neurotoxicity tests are conducted on developing
animals in addition to young adult animals;
b)	assuring that developmental neurotoxicity test information is collected at different
life stages and that there is comparison of effects of exposure during infancy,
adulthood, or old age; and
c)	revising the developmental neurotoxicity testing guidelines to better assess risks of
chemical exposure during the critical period of rapid human brain development.
OPP takes the National Research Council's (NRC's) report seriously and believes that most of
the recommendations listed by the OIG already have been appropriately addressed, as
summarized below.
40

-------
a)
See OIG Comment
in Appendix D,
Note 10
With respect to bullet "a", the current Developmental
Neurotoxicity (DNT) test guideline includes testing during the
major phases of development (i.e., during early lactation and
around the time of weaning) as well as in young adults (around
post-natal day (PND) 60). Thus, testing according to the current guideline addresses this
recommendation.
b) With respect to bullet "b", while there is no current test guideline i
occurs continuously from conception through old age, required stu
neurotoxicity at a variety of lifestages spanning the full life span.
The DNT and subchronic neurotoxicity guidelines assess
neurotoxicity during development/early life and adulthood,
respectively. Neurotoxicity-related parameters are assessed
about half-way (approximately one year) into the current two-year chronic rat study
guideline. OPP is currently part of discussions to address a variety of endpoints in a
lifestage approach to toxicity testing underway through the International Life Sciences
Institute (ILSI)/Health Environmental Sciences Institute (HESI) project identified in our
response to Recommendation 6-2.
l which exposure
lies assess
See OIG Comment
in Appendix D,
Note 10
c)
With respect to bullet "c", the current (1998) DNT test guideline recommends exposures
up to post-natal day (PND) 10; however, since the organophosphate DNT data call-in in
1999, OPP has recommended investigators dose up to PND 21.
The latest OECD draft guideline also includes exposure through
PND 21. This longer exposure period will be incorporated into
the next guideline revision/OECD harmonization. The text of
the report (p. 9, 5th bullet) should be revised to reflect OPP's
practice.
See OIG Comment
in Appendix D,
Note 11
Other OPP Comments for Chapter 3
Page 7, section title: "Assessing Developmental Neurotoxicity Important"
The title should be revised to clearly articulate the topic of the following section. We suggest the
title simply be "Assessing Developmental Neurotoxicity."
Page 8, line 3 of Developmental Neurotoxicity Testing Issues section
In addition to the neurobehavioral tests cited, it is important to include "neuropathology," which
is assessed in the DNT study.
Page 8, line 8 of the same paragraph:
OPP suggests that the text be changed to read as follows, "Data received through the data call-in
were difficult to interpret because of: (1) significant variability seen in some data sets of
developmental neurotoxicity data; (2) differences in studies conducted across laboratories; and
(3) incomplete reporting of methods and results in some study reports."
41

-------
Page 9, first paragraph in the "Weaknesses in Toxicity Testing Guidelines" section.
OPP suggests deleting the first four words of the first sentence (i.e., "Independent scientists
reported in") and start the sentence with "Literature reports	The reason for this is because
the four articles referenced were not all written by "independent scientists." Two of the articles
were from peer-reviewed journals; the other two articles were from Non-Governmental
Organizations.
Page 9, bullets under "Weaknesses in Toxicity Testing Guidelines"
The first bullet in this section states that there is no requirement for immunotoxicity testing,
either in adult or developing animals. While this is true, it fails to address the fact that the adult
immunotoxicity testing guideline was first finalized in 1998, and that OPP proposed in March
2005 (in the revisions to 40 CFR Part 158) that this test be required for
pesticide registration. A developmental immunotoxicity (DIT) testing
guideline has not yet been developed; however, OPP and ORD scientists
have participated in three public workshops on this topic which resulted
in peer-reviewed published proceedings (2001 ILSI/HESI workshop, Holsapple, 2002; 2002
National Institute of Environmental Health Sciences (NIEHS)/National Institute for
Occupational Safety and Health (NIOSH) workshop, Luster et al, 2004; & 2003 ILSI/HESI
roundtable, Hosapple et al., 2005) and are in the process of drafting a DIT guideline and
background document.
Page 10, 2nd paragraph in the section "Proposals Made...": "The Agency is awaiting public
comments prior to either amending or promulgating the proposed rule."
OPP recommends that the sentence be revised to read: "The Agency will consider public
comments prior to promulgating a final rule." The wording about promulgation needs to be
corrected. Only final rules are "promulgated."
Page 10, 3rd paragraph in the section "Proposals Made..."
OPP suggests that this paragraph should include a specific statement that says the Guidelines are
not part of Part 158, nor are they proposed to be incorporated into Part 158. The Guidelines
provide protocols that can be used to satisfy the data requirements but they are not the only
protocols that can be used. This is critical because as written it sounds like the Guidelines are
regulatory in nature.
Page 10, last paragraph, 1st sentence: "However, in the proposed rule released on
February 28, 2005, developmental neurotoxicity studies were changed to "conditionally
required," meaning that they would only be required under certain conditions."
OPP believes that the OIG text is unclear. OPP believes that the sentence is referring to the end
of the Office of Management and Budget (OMB) review of the draft proposed rule. The
proposed rule was publicly released as a Notice for Proposed Rulemaking (NPRM) in the
Federal Register on March 11, 2005 (70 FR 12276). OPP suggests that the sentence be revised
See OIG Comment
in Appendix D,
Note 12
42

-------
as follows: "However, in the proposed rule that completed OMB review on February 28,
2005...."
p. 10, last paragraph, discussion on the Part 158 (conventional pesticides rulemaking)
proposed rule
EPA is proposing to conditionally require DNT studies for all neurotoxic pesticides and for
pesticides that meet other criteria indicating a potential for toxicity to the developing nervous
system, based upon a weight-of-evidence evaluation of the toxicological database. The first two
sentences give the misleading impression that OPP changed the substance and scope of the
proposed Developmental Neurotoxicity Test (DNT) requirement based on
In actuality, the draft proposed rule was clarified in response to OMB
questions, but neither the substance nor the scope of the proposed DNT
required was altered. We recommend deletion or revision of this
discussion.
In the draft proposed rule submitted to OMB, a DNT was "Required" and the notes denoted the
limited instances when the test would be necessary. The preamble of the draft proposed rule also
mentioned the weight-of-evidence approach, without an extensive discussion of the approach.
After discussions with OMB to clarify the criteria associated with the requirement to conduct a
DNT study, OPP realized that the rule should describe the DNT study as "conditionally
required" to reflect the limited conditions when the data requirement would be imposed.
Accordingly, OPP developed a more extensive discussion of its weight-of-evidence approach in
the preamble to the proposed rule; expanded one of the notes to include the weight-of-evidence
approach; and changed the "Required" to "Conditionally Required" to better reflect the
frequency that EPA would impose the requirement. Thus, the change was elicited by OMB's
request for clarification but the substance of the requirement was not changed.
Page 11: "EPA had not responded to OMB as of August 30, 2005."
Since response can include conversations, this sentence should be changed to read: "EPA had not
submitted a plan to OMB as of August 30, 2005."
V. Chapter 4 - OPP Made Substantial Changes to Overcome Aggregate Risk
Challenges, but Challenges Remain
Recommendation 4-1: Update its dietary exposure databases used in probabilistic models
for risk assessments. EPA should use the dietary consumption data compiled from the
DHHS's National Health and Nutrition Examination Survey, update the Food Commodity
Intake Data Base with the latest food consumption survey data, and if possible use data
such as the Gerber Products Company's Feeding Infants and Toddlers Study.
While OPP is indeed moving toward using dietary consumption data compiled from the DHHS'
National Health and Nutrition Examination Survey (NHANES) survey and updating the Food
Commodity Intake Database (FCID), we believe implementation of this recommendation should
questioning by OMB.
See OIG Comment
in Appendix D,
Note 13
43

-------
wait until 2006 when we expect NHANES to release sufficient data to provide a database
comparable to the USDA's Continuing Survey of Food Intakes by Individuals (CSFII) currently
used by OPP. Therefore, while OPP generally agrees with updating our food consumption
database to reflect the NHANES data, we believe the OIG
recommendation, as drafted, is premature. In fact, EPA has been
planning to update the food consumption database since the NHANES
effort began, and we intend to move beyond our current exploratory
analysis of the NHANES dietary consumption data after the first full set of 2-day consumption
data for 2003-2004 is released in 2006.
With respect to the DHHS' NHANES survey, we note that only the 1999-2000 and 2001-2002
data have been released so far and full survey integration between the NHANES and USDA
CSFII methods only began with the NHANES 2002 data. We further note that DHHS offers
several cautions about combining these data sets and it released publicly only part of the data due
to confidentiality concerns. In addition, different interview systems/ methodologies were used
between the 1999-2000 data, the 2001 data, and the 2002 data. For these reasons and others,
OPP has so far only conducted initial exploratory analyses on these data while anticipating the
release of the next set of data.
This next set of data, consisting of the 2003-2004 dietary consumption data, will be the first
complete, integrated 2-day dietary consumption data set released by the NHANES program.
This is expected to be made publicly available in 2006. OPP will begin examining this data
closely upon its release, particularly for kids' foods and for comparison with the 1998
Supplemental Children's Survey. We expect differences to be minimal for the fresh (raw)
commodities which we have generally found to contribute most to pesticide dietary burden.
Furthermore, as suggested by the OIG, we intend to update the FCID to incorporate new
(generally processed) foods that have come onto the market since the earlier USDA survey and
are being reported in new food consumption surveys. We are beginning to plan for updating this
database, but the majority of this work will not occur until after the work associated with
meeting the August 2006 FQPA tolerance reassessment deadline is completed.
Several years ago, OPP decided that practical considerations prevented the use of the Gerber
Feeding Infants and Toddlers Study (FITS), which is a survey of the eating habits and nutrient
intakes of more than 3,000 infants and young children, conducted during the five-month period
from March-July 2002. A study summary was published as a supplement to FITS in the January,
2004 issue of the Journal of the American Dietetic Association. Foods
are reported in this survey on an "as eaten" basis. OPP contacted the
Gerber Company several years ago to obtain this data; however, Gerber
provided only limited data quantity and format. Specifically, the data
was provided to us in a voluminous hardcopy (not electronic) form and the data were expressed
on a food form rather than a commodity form basis. Recipes (aka 100 gram files) were
proprietary and not provided for translating the foods (e.g., baby food applesauce) to its
component raw agricultural commodities (e.g., apples, sugar, etc.) on which our pesticide residue
data and exposure assessments are based.
See OIG Comment
in Appendix D,
Note 14
See OIG Comment
in Appendix D,
Note 15
44

-------
Given the nature of and issues associated with the FITS data, the fact that the FITS survey was
conducted only four years after the USDA-conducted 1998 Supplemental Children's Survey
associated with the CSFII, OPP's active involvement in the early stages of the replacement
NHANES dietary survey, and the reasonably comparable size of the CSFII survey for the
subpopulation of interest, OPP decided then and continues to believe that resources were more
appropriately invested in working jointly and cooperatively with other U.S. government agencies
toward a large scale, nationally-representative NHANES dietary survey.
Recommendation 4-2: Continue to collaborate with USDA to develop the methodology for
collecting longitudinal food intake and physical activity information from children.
EPA recognizes that data from a longitudinal dietary consumption survey could provide a better
basis for estimating longer-term dietary risk than the food consumption databases now available.
However, the methodology for collecting longitudinal data is extremely difficult. It requires that
individuals track and record their consumption of food and beverages
with respect to the identity of each food, the amount consumed, and
(generally, to be most useful) the time at which it is eaten. While this
might be possible and feasible over shorter periods of time (e.g., one or
two weeks), it becomes very difficult to collect reliable, accurate information over the longer
periods of time (months or years) which are of particular interest to OPP. Over extended periods
of time, difficulties understandably arise with respect to differential drop-out rates and
adequacy/accuracy of recorded eating occasions. Also, the very process of continually recording
eating patterns over an extended period of time is thought to potentially change those eating
patterns. Longitudinal food consumption data is also extremely expensive and difficult to obtain.
While there is general and widespread interest and support among a multitude of governmental
agencies and others with respect to general eating patterns and eating patterns over shorter terms,
there is less interest in collecting information on detailed eating patterns over extended periods of
time. Considering these issues, gathering extensive data on long-term, intra-individual eating
patterns through extensive consumer surveys does not appear sufficiently promising, at this time,
to justify further pursuit of methodology development.
As discussed in the response to Recommendation 4-4, we are pursuing alternative approaches to
improving our estimation of longer term dietary risk, including the statistical modeling
techniques and simulations which have developed under collaborative efforts with consulting
firms and science research organizations. Our efforts also include monitoring the work of the
National Cancer Institute (NCI) in this area.
Recommendation 4-3: Assess which additional foods frequently consumed by children
should be included by USDA in its Pesticide Data Program testing.
OPP believes that its current practice fully addresses this
recommendation and that no change is warranted. To the extent this
recommendation concludes that EPA and the USDA's Agriculture
Marketing Service (AMS) have failed to consider appropriately which
See OIG Comment
in Appendix D,
Note 16
See OIG Comment
in Appendix D,
Note 17
45

-------
food commodities should be included in the Pesticide Data Program (PDP) monitoring (and with
what frequency), we disagree. OPP also believes that this recommendation does not adequately
take into account that sampling budgets are finite and that infrequently sampling a broader swath
of foods with less frequency may not yield improved or more accurate dietary risk assessments.
OPP has an extensive and ongoing relationship with USD A with respect to determining which
foods are included in USDA's PDP program. As stated in PDP's annual summary reports:
"AMS works closely with the US EPA to select commodities and pesticides for
PDP testing. Commodities selected are those most often consumed by U.S.
consumers, with emphasis on foods consumed by infants and children."1
We disagree strongly with the OIG statement that USDA's PDP
program has generated extensive residue data on over, "50 foods out of
hundreds of key foods eaten in the U.S." but that, "many of the foods
not tested [by PDP] may be important in the diets of infants and
children." This statement incorrectly implies that EPA may be underestimating kids' exposure
because EPA is "missing" many important foods consumed by children. The sampled PDP
commodities for children 1-2 years old directly or indirectly represent approximately 90% of
children's diets. Moreover, EPA does include in its risk assessments a contribution from
consumption of foods not tested in the PDP. For those foods, EPA uses field trial residue data.
While these data tend to overestimate potential exposure, they rarely represent a significant part
of the total exposure.
Determining which foods to sample and not to sample should be (and is) based on careful
consideration of food consumption patterns, residue levels, and frequency of pesticide detections.
Professional judgment, grounded in extensive experience performing dietary exposure
assessment, should guide the choice of foods that collectively are likely to account for the
greatest amount of exposure. It would be unwise, given finite sampling resources, to design the
PDP survey program so that a greater number of commodities representing a high cumulative
(total) percentage of the diet are sampled. This is particularly true if this means important high
consumption children's commodities such as apples, oranges, grapes, or potatoes would be
sampled less frequently (i.e., at longer intervals) or with less intensity. In general, we attempt to
sample high consumption foods for 2-3 consecutive years at a time and at an interval which does
not exceed 4-5 years.
Recommendation 4-4: Expand its partnerships with USDA and DHHS to collect data on:
¦	longitudinal food consumption information;
¦	pesticide concentration in human breast milk among lactating women, through the
National Health and Nutrition Examining Survey and/or USDA's WIC Program
1 Pesticide Data Program: Annual Summary Calendar Year 2003 United States. Department of
Agriculture, Agricultural Marketing Service, Science and Technology Programs. February 2005.
http://www.ams.usda.gov/science/pdp/Summary2003.pdf
See OIG Comment
in Appendix D,
Note 18
46

-------
(Special Supplemental Nutrition Program for Women, Infants, and Children);
¦	dietary exposures among children at schools and day cares (perhaps through the
USDA's School Lunch and/or Head-Start programs); and
¦	effects of dietary and nondietary exposures of pesticide on children's cognitive
functions and performance.
With respect to the first bullet, while the OIG report presents no analysis or discussion of this
issue in the main body of its text, EPA agrees that understanding longitudinal dietary
consumption is important. OPP believes our actions have fully	See qig Comment
addressed this recommendation. We have communicated our interest in	in Appendix D,
longitudinal consumption information to DHHS. However, due to cost	Note 16
and various logistical considerations, they (i.e., the designers of the
NHANES study) are not able to provide dietary recall information for more than two (2) days.
We understand and agree with that position; we do not think it is realistic to extend the number
of days of consumer reporting.
OPP currently relies on the USDA's Continuing Survey of Food Intakes by Individuals (CSFII).
This survey is being replaced by DHHS's National Health and Nutritional Examination Survey
(NHANES) which is its successor survey. Both surveys are large, complex, multi-stage cross-
sectional surveys of 2-day consumption patterns and are designed to be representative of the U.S.
population. These surveys serve a variety of interests and customers of which EPA is only one
of many.
As noted in our response to Recommendation 4-2, we agree that understanding longitudinal
patterns in food consumption is important. To that end, here are some current activities that we
have participated in with USD A and DHHS:
¦ EPA has made attempts to statistically simulate this data in its dietary exposure
assessments in as realistic a manner as possible. In addition, a
variety of software developers (DEEM, CARES, and LifeLine) see OIG Comment
have incorporated longitudinal consumption patterns into their	in Appendix D,
software. This, however, is not based on actual longitudinal	Note 16
consumption data, but rather on statistical "matching" and other
criteria which attempt to simulate, on an individual-by-individual basis, consumption
patterns over the long term (e.g., seasonal or yearly). This methodology has been
presented to the SAP for two of the models (CARES and LifeLine). OPP is beginning to
incorporate this information into its risk assessments.
In particular, EPA's Office of Research and	See OIG Comment
Development (ORD) is attempting to gather	in
longitudinal data on eating and activity patterns
through its STAR (Science to Achieve Results) grant program and OPP has
been closely involved in these activities. OPP has participated extensively on
the ORD review panels which recommend funding priorities for these kinds of
studies. For example, OPP was actively involved in the review of proposals
47

-------
for ORD's STAR grant project entitled "Aggregate Exposure Assessment for
Pesticides: Longitudinal Case Studies." These grants were awarded several
years ago and amounted to several million dollars. These grants funded a
variety of innovative small-scale pilot studies collecting (many times
simultaneously for the same individual) information on both food
consumption and physical activity. Many novel methodologies for collecting
this information were proposed which were designed to minimize study
participant effort and tedium. These methodologies hold promise for the
future.
¦	We note that we have pursued alternate routes toward obtaining or incorporating this
information into our risk assessments. One approach is through the ORD STAR grant
program which recently award grants for specifically looking at
longitudinal patterns of exposure. Specifically, EPA's ORD
sought grant proposals that described studies for assessing
pesticide exposure that incorporate estimates of temporal and
inter-individual variability and attempt to effectively include and address many exposure
issues using longitudinal studies. These are naturally long-term studies and OPP will
remain involved and looks forward to receiving this data when it becomes available.
¦	ORD is exploring the possibility of working with other Federal partners to collect
longitudinal food consumption data. In September, 2005, NERL sponsored "EPA's
Workshop on Analysis of Children's Measurements Data" that included discussions of the
major sources of children's exposures to pesticides, and approaches to analyze existing
data. Information from this workshop may be used in future discussions with other
agencies.
¦	Additional work on the issue of longitudinal consumption is
being performed by the National Cancer Institute (NCI) and
others who are attempting to merge two types of data to develop
better estimates of long-term consumption. Specifically, they are
combining 24-hour dietary recall information (e.g., "What did you eat today?") with food
frequency information (e.g., "How many times in the last 90 days did you consume
peas?"). By merging these two kinds of data sets, it should be possible to develop better
estimates of long-term dietary consumption on an individual-by-individual basis.
EPA will remain active in its pursuit of data to support longitudinal consumption estimates.
However, as discussed in our response to Recommendation 4-2, we recognize that obtaining
actual, longitudinal consumption data from a large, representative survey is an extremely costly
undertaking for which there are inadequate funds and insufficient widespread or general interest.
EPA will work with ORD, various software vendors, and others to make the best and most cost-
effective use of available data and most appropriate use of appropriate simulation methods.
With regards to the second bullet in this recommendation (collecting data on pesticide
concentrations in human breast milk among lactating women), OPP acknowledges that for
certain chemicals - particularly stable, highly lipophilic chemicals (i.e., persistent
See OIG Comment
in Appendix D,
Note 19
See OIG Comment
in Appendix D,
Note 16
48

-------
bioaccumulative and toxic chemicals) like the organochlorine pesticides - with specific physical-
chemical characteristics, availability of exposure information from breast milk provides
additional characterization of the dietary exposure to nursing infants. Fortunately, the vast
majority of currently approved pesticide chemicals do not exhibit these physical chemical
characteristics, and are not expected to be found in significant amounts in human breast milk,
and EPA has breast milk data for most or all of the chemicals that have
such characteristics. In the event that OPP identifies a chemical for
which it thought breast milk might be a significant source of exposure,
but for which it lacked adequate data, OPP could address the situation
using the FQPA Children's Safety Factor provision that directs EPA to
retain an additional 10X margin of safety when it has uncertainty about the exposures
experienced by infants or children.
See OIG Comment
in Appendix D,
Note 20
OPP is not opposed to working with USDA and DHHS to collect data on pesticide residues in
human breast milk but overall, OPP does not believe this should be a high priority item for the
majority of pesticide chemicals as discussed below. In most cases it would be more valuable to
have monitoring data on infant's and children's foods which are directly
treated with pesticide, and therefore have a higher potential to lead to	s®e OIG Comment
infant's and children's exposure (e.g., apples/apple sauce/apple juice).	^ote^o*
(Also, because animal feeds frequently have significantly higher
residues than human foods, residues in cow's milk would be expected to be greater than those in
human breast milk. In these cases, risks associated with cow's milk consumption would likely
be greater than risks associated with breast milk consumption, and could be captured in infant's
risks assessments.)
Finally, OPP notes that collection and analysis of useful breast milk
biomonitoring is very complicated. Care must be taken in choosing the
population to be sampled, the pesticide chemicals to be monitored, and
in developing other aspects of the sampling protocol. For example,
since the composition of human milk changes within a feeding, over the course of a day, and
over the course of lactation, the exact timing and method of sampling can influence measured
levels. Quantitative use of these data in a risk assessment will also be complicated by these
factors.
See OIG Comment
in Appendix D,
Note 21
With regards to the third bullet in this recommendation (collecting data	s®® ®IG^°r^n^nt
on dietary exposures among children at schools and day cares), the	^ote"^*
current CSFII and NFLANES dietary survey collection methodology 	
considers and includes food consumption at schools and daycare centers.	Consumption amounts
and items are already included in the CSFII and NFLANES surveys.
With regards to the fourth bullet in this recommendation (effects of dietary and non-dietary
exposures of pesticide on children's cognitive functions and performance), OPP has worked, and
will continue to work, closely with USDA on a number of dietary exposure issues relevant to
children's food consumption (see our responses to Recommendations 4-2, 4-3, and the first three
bullets for 4-4).
49

-------
In addition, children's non-dietary exposures to pesticides have been a focus of an intense ORD
research effort (http://www.epa.gov/heasdweb/children/children.htm ). It is not clear to
us whether the USDA or DHHS has any research/activities in this area,
but we would be happy to collaborate if such opportunities were made
available.
Finally, the effects of pesticide exposure (dietary or non-dietary) on children's cognitive
functions and performance is an important area that EPA, along with NIH from DHHS, strongly
support and should be addressed in the development of the National Children's Study (cited as
footnote 8 on page 14 of the OIG draft report). This study is designed to follow the lives of
100,000 children from gestation through 21 years of age to examine the effects of physical,
chemical, biological, and psychosocial environmental influences on health and development.
Although the OIG draft report notes that the long-term support for this longitudinal study is
"under debate", six study centers were chosen and recently awarded contracts on September 29,
2005 to begin work on this important project (http://www.nationalchildrensstudv.gov/). In
addition, the EPA/NIEHS children's research centers at UC Berkeley and University of
Washington are conducting epidemiologic studies comparing aggregate pesticide exposures with
neurobehavioral and cognitive development in birth cohorts living in agricultural communities in
the Salinas (CA) and Yakima (WA) valleys.
Other Comments for Chapter 4
Page 12, paragraph 1, lines 3-5 (2nd sentence)
OPP suggests that the sentence be revised to read as follows: "Aggregate risk assessments, which
are required by FQPA, specify that all routes and pathways of exposure for a given pesticide be
considered when assessing risk."
Page 12, paragraph 1, lines 9-11 (last sentence)
OPP suggests that the sentence be revised to read as follows: "Without sufficient information to
perform highly refined and meaningful aggregate risk assessment, there will be uncertainty...."
Page 12, lines 13-14 (2nd paragraph, 2nd sentence)
OPP suggests that the sentence be revised to read as follows: "Besides updating its pesticide
testing guideline to improve and expand animal study data on reproductive and developmental
effects, EPA developed methods...."
Page 13, line 6 in paragraph starting with "Children engage...."
OPP suggests that any and all citation(s) supporting the statement "Literature indicates..."
should be included in the text or as a footnote.
Page 13, bullets on missing elements and research needs
See OIG Comment
in Appendix D,
Notes 14 and 16
50

-------
This text should acknowledge that all of these identified needs are the subject of on-going
collaborative research efforts by ORD.
Page 15, 3rd paragraph, line 3 (paragraph starting with "In 2002, Gerber Products....")
OPP suggests that the sentence be revised to read as follows: "This is the most comprehensive,
largest, and nationally representative study on food consumption for this age group,..
VI. Chapter 5 - OPP Moving to Assess Cumulative Risk but Complexities and
Concerns Remain
Recommendation 5-1: Coordinate efforts with ORD to finalize the integration of
probabilistic modeling outputs with physiologically based pharmacokinetic modeling to
better address cumulative risk from concurrent exposure to pesticides and other chemicals
with like mechanisms of action.
OPP believes that its past and continuing actions fully address this recommendation. Chapter 5
of the OIG draft recognizes that ORD has begun discussions of linking two of their models. One
of these models, ERDEM (Exposure Related Dose Estimated Model), is a physiologically-based
pharmacokinetic (PBPK) model. The other model, SHEDS (Stochastic Human Exposure and
Dose Simulation), is a probabilistic exposure model. OPP acknowledges that PBPK models are
powerful risk assessment tools that can incorporate the dynamic nature of environmental
exposure, internal dose, toxic effect, and recovery. OPP and ORD have
been collaborating for several years on many efforts to develop PBPK
models for a variety of chemicals (e.g., carbaryl, malathion, and
pyrethroids). Furthermore, OPP will continue to encourage efforts to
link probabilistic exposure models with PBPK models. The OIG draft report does not recognize
these additional on-going efforts by EPA to link probabilistic exposure models with PBPK
models. For example, the LifeLine Group, under contract with OPP, developed a white paper
entitled "Designing Exposure Models that Support PBPK/PBPD Models of Cumulative Risk"
which was reviewed by the FIFRA Scientific Advisory Panel (SAP) in December 2004. This
paper highlights issues regarding linkage between probabilistic exposure models and PBPK
models. The FIFRA SAP's report can be found at http://www.epa.gov/scipoly/sap/index.html.
See OIG Comment
in Appendix D,
Note 23
See OIG Comment
in Appendix D,
Note 19
In addition, with funding from an EPA STAR grant, the University of
Washington is conducting a concordance analysis of probabilistic
aggregate exposure assessment and biomarkers of exposure. The overall
objectives of this study are to examine the accuracy of current pesticide
exposure assessment models, and to demonstrate a novel method for the development and
evaluation of such models. Researchers have proposed to conduct second order probabilistic
assessments of aggregate pesticide exposures in three existing data sets, characterize the
variability of biological exposure measures, and evaluate the concordance of these two exposure
assessment approaches. This novel analytical approach will produce new methods for
determining the validity of exposure and risk estimates.
51

-------
With funding from an EPA STAR grant, researchers at Battelle are developing a physiologically
based pharmacokinetic/ pharmacodynamic (PBPK/PD) model to quantitate biomarkers of
exposure to organophosphate pesticides. The project entails development and validation of a
PBPK/PD model for chlorpyrifos to quantitate biomarkers of dosimetry and pharmacodynamic
(PD) response (i.e., acetylcholinesterase (AChE) inhibition) in young rats and children.
Recommendation 5-2: Develop specific plans on how computational toxicology outputs
from ORD's Computational Toxicology Program will integrate into OPP's regulatory
process, and implement such a transformation.
OPP agrees with this recommendation and believes its past and continuing activities fully
address this recommendation. A strategic plan for developing and implementing advances in
computational toxicology research in the context of pesticide and industrial chemical regulatory
frameworks was developed and endorsed by OPPTS and ORD Office and Laboratory/Center
Directors in 2004. This strategic plan for developing an integrative, or intelligent, risk
assessment paradigm complements related efforts by the National Academy of Sciences (NAS),
the European Union (EU), and the Organization for Economic Co-operation and Development
(OECD). The strategic view considers the development of ORD products with the associated
implementation of new OPPTS risk assessment methods as part of the current Agency goals,
programs and planning processes. The plan outlines includes activities and outcomes in the
short- to long-term. OPP has already begun collaborations with ORD's National Center for
Computational Toxicology (ORD-NCCT) to improve risk assessment methodologies. For
example, the dose-response modeling used in the Preliminary Cumulative Risk Assessment for
the N-methyl carbamate pesticides (August, 2005) was performed by the NCCT. This work was
reviewed by the FIFRA Scientific Advisory Panel (SAP) in August, 2005. OPP and ORD,
including NCCT, continue to work collaboratively to develop research plans to develop and
apply new technologies which will meet the needs of the regulatory program. OPP notes that the
outputs of the Computational Toxicology Program will likely apply not only to cumulative risk
assessment (the topic of Chapter 5) but also single chemical assessments.
Other Comments for Chapter 5
Page 19, paragraph 1, line 4 (paragraph beginning with "Research confirms the need...")
OPP requests clarification on what the following phrase means: ".. .below pesticide poisoning
health effects." Is this intended to refer to "low-dose effects"?
Page 20, paragraph 1, lines 7-10 (sentence beginning with "For EPA scientists to use...")
We do not believe that this is a sentence. OPP suggests revisions to the sentence: "For EPA
scientists to use the data and products from the Computational Toxicology Program,
interdisciplinary backgrounds in such areas as biostatistics, molecular biology, metabolism,
systems biology, computational chemistry, toxicology, and bioinformatics will be required."
52

-------
VII. Chapter 6 - Opportunities Exist to Better Manage FQPA Implementation
Recommendation 6-1: Either finalize all of the Science Policy issue papers, or change the
word "draft" to "operational" and schedule annual updates.
OPP generally agrees with this recommendation. However, OPP notes that the OIG report fails
to recognize the extent of OPP's progress in completing science policy papers, contains a
number of errors, and contains no explanation for its recommendation. Moreover, the report
does not analyze the impact of leaving the documents in "draft" form or the priority that EPA
should give to finalizing the documents.
The draft OIG report recommends that EPA finalize the "FQPA science policy papers" that were
only issued in "draft" form and never revised. Although the OIG indicates there are twelve (12)
papers, in fact there are only nine such papers. The OIG report purports to identify in Figure 6.1
twelve (12) science policy papers that "are still in draft format," based on "EPA's website as of
August 2005." It neglects to say OPP finalized 16 major science
policies. While Figure 6.1 correctly identifies six science policy papers
never issued in "revised" form, Figure 6.1 mischaracterizes the status of
others:
¦	It lists as "draft" two papers that were revised: "User's Guide to Available EPA
Information on Assessing Dietary (Food) Exposure to Pesticides" and "Science Policy 5:
Estimating the Drinking Water Component of a Dietary Exposure Assessment."
¦	It lists two papers twice, giving the impression there are four "draft" papers when there
are only two: "Standard Operating Procedures (SOPs) for Residential Exposure
Assessment" and "Framework for Assessing Non-occupational / Non-dietary
(Residential) Exposure to Pesticides."
¦	It lists two documents that are not FQPA Science Policies: "Draft Toxicology Data
Requirements for Assessing Risks of Pesticide Exposure to Children's Health" and
"Draft Exposure Data Requirements for Assessing Risks of Pesticide Exposure to
Children's Health." These two documents are actually appendices to a revised science
policy document titled "Determination of the Appropriate FQPA Safety Factor(s) in
Tolerance Assessment." OPP will update its website to make the status of these two
papers clearer.
The OIG report fails to list three papers that EPA has issued only in "draft" form: "Standard
Operating Procedures (SOPs) for Use of the FQPA Factor"; "Use of the Pesticide Data Program
(PDP) in Acute Dietary Assessments"; and "Water Treatment Effects on Pesticide Removal and
Transformation."
Changing Figure 6.1 to correct these errors would result in the listing of nine science policy
papers that OPP has issued in draft form but never finalized. See the Table below.
See OIG Comment
in Appendix D,
Note 24
53

-------
OPP generally agrees with the recommendation and intends to finalize some, but not all, of the
science policy papers that remain in "draft" form. See the Table below. OPP, however, does not
regard these actions as a high priority because:
¦ OPP is subject to statutory and court-ordered deadlines to
complete regulatory decision-making that will require significant
resources, leaving limited resources to address this activity; and
¦ External stakeholders have not asked OPP to finalize these science policy papers.
OPP Science Policy PaDers that have
not been issued in revised form
ProDosed OPP Action
1. "Standard Operating Procedures
(SOPs) for Residential Exposure
Assessment"
OPP is planning to finalize this paper.
2. "Application of the 10X Safety Factor
in Cumulative Risk Assessment"
OPP is planning to finalize this paper.
3. "Framework for Assessing Non-
occupational / Non-dietary (Residential)
Exposure to Pesticides"
OPP is planning to finalize this paper.
4. "Drinking Water Screening Level
Assessment"
OPP is planning to issue a Federal Register Notice
announcing that it has withdrawn this paper because
it refers to a method of estimating potential drinking
water exposure that OPP no longer uses.
5. "Standard Operating Procedure for
Incorporating Screening-Level Estimates
of Drinking Water Exposure in
Aggregate Risk Assessments"
OPP is planning to issue a Federal Register Notice
announcing that it has withdrawn this paper because
it refers to a method of estimating potential drinking
water exposure that OPP no longer uses.
6. "Guidance for the Submission of
Probabilistic Human Health Exposure
Assessments to the Office of Pesticide
Programs"
OPP is planning to issue a Federal Register Notice
announcing that it has withdrawn this paper because
the guidance is generic, has been superseded by a
number of more specific policy documents, and is
adequately covered in other finalized Agency
documents.
7. "Standard Operating Procedures
(SOPs) for Use of the FQPA Factor"
OPP is planning to finalize this paper.
8. "Use of the Pesticide Data Program
(PDP) in Acute Dietary Assessments"
OPP is planning to issue a Federal Register Notice
announcing that it has withdrawn this paper because
the paper refers to a method of estimating potential
exposure through food that OPP no longer uses.
OPP will review additional data to determine
whether it needs to issue a new science policy paper
to explain its general approach in this area.
9. "Water Treatment Effects on
Pesticide Removal and Transformation."
OPP is planning to finalize this paper.
See OIG Comment
in Appendix D,
Note 24
54

-------
Recommendation 6-2: Sustain the development of an alternative testing strategy, ensuring
that risks are assessed across the entire life cycle of development.
OPP agrees with this recommendation. OPP scientists, together with other EPA colleagues, have
engaged in a number of activities on different fronts to improve the toxicology testing paradigm
for environmental chemicals, including pesticides. First, EPA conducted a review and published
a Risk Assessment Forum report in 2000 on the current reference dose and reference
concentration (RfD/RfC) processes, in particular with respect to how well children and other
potentially susceptible subpopulations are protected. One of the objectives of this EPA activity
was to consider new scientific issues that have become more important and of greater concern in
risk assessment, and to raise issues that should be explored or developed further for application
in the RfD/RfC process.
Second, OPP scientists participated in an effort sponsored by the International Life Sciences
Institute (ILSI)/Health Environmental Sciences Institute (HESI) to design a better testing
paradigm for pesticide chemicals. This ILSI effort examined whether: 1) life-stages (i.e.,
infancy, pre-adolescence, adolescence, reproductive stage, post-reproductive stage, elderly) are
adequately assessed by the current battery of studies; 2) scientific evidence exists that certain
life-stages may be comparatively more susceptible to the effects of exogenous chemicals; 3)
altered susceptibility in a life-stage is general or specific. The goal of this analysis is to identify
a hierarchy of study types, endpoints and triggers that might be used in a decision tree to guide
appropriate testing to determine the safety of a pesticide. This new testing proposal will be
published by early 2006.
Lastly, EPA sponsored a National Academy of Sciences (NAS) study to review evolving
regulatory needs, current toxicity testing guidelines, emerging science and new tools (e.g., -
omics, transgenics, bioinformatics, computational toxicology, in vitro testing, alternatives to
animal testing) and develop a strategy that incorporates more complex information (e.g.,
toxicokinetics, mechanisms of action, systems biology) into improving human health risk
assessment. The NAS report is anticipated 2007. OPP will consider all of these activities, as
well as its own computational toxicology program, as it moves forward in developing a
hypothesis driven paradigm that uses resources more efficiently and improves the assessment of
human health.
Recommendation 6-3: Develop an overarching logic model and long-term strategic plan
across divisions to identify and link immediate work outputs to outcomes.
The recommendation that OPP use logic model (e.g., the logic model provided in the draft OIG
report) to guide efforts is a good idea and in reviewing this section, we interpret that the OIG is
trying to connect "opportunities to improve data quality" data with performance accountability.
The draft report appears to suggest is that it could be beneficial to develop strong performance
measures that would "pass muster" in an Office of Management and Budget (OMB) Program
Assessment Rating Tool (PART) review. Developing such a logic model could aid in the
development of a strategic plan.
55

-------
Other Comments for Chapter 6
Page 23, "Logic Models..." section, paragraph 1, last sentence
OPP suggests the following revision: "Significantly, a logic model distinguishes between
outputs (the specific tasks performed) and outcomes (the actual results)."
OPP's comments for the Appendices
Appendix C: Toxicity Testing Issues
Page 29, Table of developmental tests for effects on offspring
The table states that offspring are not evaluated in the developmental
guideline. This is incorrect; the test is focused on fetal development.
While it is true that there are no post-natal tests and no functional tests
in this guideline, the fetus evaluation is an evaluation of the offspring.
Page 30, first paragraph, line 4
Behavior and functional tests are sensitive	
Page 30, last paragraph
OPP believes that the discussion is too specific and thus out of place in a discussion of general
toxicity testing requirements.
Page 33
OPP recommends that the OIG consider the relevance of the ERDEM (Exposure Related Dose
Estimating Model).
See OIG Comment
in Appendix D,
Note 25
56

-------
Appendix D
OIG's Comments on Agency's Response
1.	We disagree with the Agency' s statements that our report is "uneven" and "sometimes
misleading." We conducted this review to examine the impact of the Food Quality
Protection Act of 1996 on the Agency's need for scientific data and predictive tools,
particularly in relation to children's health. As an independent office within EPA, we
presented information based on facts uncovered and evidence found. In helping the Agency
identify where it needs more and better scientific data and tools in its implementation of
FQPA, we have strived to serve as a catalyst for protecting children's health, improving the
environment, and increasing the Agency's accountability.
Additionally, we disagree with the Agency's statements about this report "glossing over
significant scientific accomplishments of the past nine years." In this report, we have
outlined the Agency's major scientific accomplishments since 1996. For example, we have
discussed how the Agency made substantial changes to the aggregate exposure risk
assessment process, eliminated usage of some pesticides like chlorpyrifos and diazinon, and
initiated steps to perform cumulative risk assessments for pesticides. We have described the
contributions of ORD in assuring that the Agency meets the scientific challenges posed by
FQPA. We have highlighted in our logic model a list of OPP's and ORD's scientific
activities and outputs. Finally, we have mentioned how OPP worked with ORD, external
scientific organizations, and its Federal partners to acquire more and better data on children's
exposure to pesticides.
2.	We disagree with the Agency's statement that this "report tends to focus on issues that
1)	are minor and relatively insignificant within the overall scope of FQPA implementation,
2)	characterized incorrectly, or 3) outside the control of the Agency." The findings we have
presented in this report are based on facts and evidence we have uncovered during our
review. We agree with the Agency that EPA depends on others for much of the data it uses
in risk assessments. However, it is our opinion that the Agency has the responsibility to
identify and make known the quantity and quality of data needed for its risk assessments.
In the next few paragraphs we have outlined why issues like acquiring developmental
neurotoxicity and additional exposure data are important. We have also discussed how these
issues are neither minor nor insignificant within the overall scope of FQPA implementation.
¦ Acquiring developmental neurotoxicity test data is neither a minor nor insignificant
issue: Prediction of neurotoxic effects is a key feature in the toxicological profile of
chemicals. It is our opinion that the Agency may be vulnerable to legal challenges to the
10X additional FQPA safety factor if it does not consistently require developmental
neurotoxicity testing of chemicals. Congress acknowledged that protecting the
developing nervous system from toxic insult is important27 when it unanimously passed
27 21 U.S.C. §346a (b)(2)(C)(i)(II)
57

-------
FQPA. We believe it was not a minor or insignificant issue when the Agency had to
eliminate the manufacturing of chlorpyrifos for nearly all residential usage. The driving
force behind this change in regulatory policy was the recognition that chlorpyrifos exerts
untoward effects on the developing nervous system.28 Scientists and public health
researchers29 have further confirmed that this pesticide interferes with brain development
and children's growth. The lessons learned from chlorpryifos are that the developing
brain is highly vulnerable and that a common pesticide can interfere with a child's brain
development.30
We believe EPA has the responsibility to guide the development of alternative
developmental neurotoxicity testing protocols which yield cost-effective, efficient data
for pesticide regulation and children's health decisions. Beyond having optimal, cost-
effective testing of developmental neurotoxicity for chemicals, we believe EPA can
demonstrate its commitment to protecting children by ensuring that such data are
collected across life stages beginning at critical windows of development (e.g., pre-natal).
¦ Acquiring more and better dietary and nondietary exposure data for use in its risk
assessments is an EPA responsibility. We recognize the Agency depends on its Federal
partners for national dietary exposure data. We also recognize that resource limitations
constrain the Agency's research efforts in measuring residual and dietary consumption
patterns. We acknowledge the problem of managing pesticide residue in food lies within
a much larger food safety monitoring arena in which EPA is a minor player. However, in
the area of dietary pesticide exposure data, we believe the Agency has a major
responsibility because it sets the tolerances and registers the pesticide chemical use on
food. Likewise, for nondietary pesticide exposure, because EPA registers the pesticide
chemicals use in homes, schools, public areas, and gardens, it has the responsibility to
acquire residential and nonoccupational pesticide exposure data. Thus, it is our opinion
that EPA should assess the scale of monitoring required to know and manage the mixture
of pesticides dispersed into our food and water supplies and our environment.
3. Our review focused on existing data and interviews. During our evaluation, we conducted a
literature review on FQPA and the potential health outcomes of prenatal and childhood
exposures to pesticides. We reviewed documents pertinent to risk assessment prepared by
28	Slotkin, TA. 1999. Developmental Cho lino toxicants: Nicotine and Chlorpyrifos. Environmental Health
Perspective 107 (suppl 1), 71-80. Slotkin, TA. 2004b. Cholinergic Systems in Brain Development and Disruption
by Neuro toxicants: Nicotine, Environmental Tobacco Smoke, Organophosphates. Toxicology and Applied
Pharmacology 198, 132-151.
29	Needham, LL. 2005. Assessing Exposure to Organophosphorus Pesticides by Biomonitoring in Epidemiologic
Studies of Birth Outcomes. Environmental Health Perspective 113:494-498. Berkowitz, GS et al. 2004. InUtero
Pesticide Exposure, Maternal Paraoxonase Activity, and Head Circumference. Environmental Health Perspective
112:388-391. Eskenazi, B et al. 2004 Association of in Utero Organophopshate Pesticide Exposure and Fetal
Growth and Length of Gestation in an Agricultural Population. Environmental Health Perspective 112:116-1124.
Whyatt, RM et al. 2004. Prenatal Insecticide Exposures and Birth Weight and Length among an Urban Minority
Cohort. Environmental Health Perspective 112:1125-1132.
30	Slotkin, TA. 2006. Developmental Neurotoxicity of Organophosphates: A Case Study of Chlorpyrifos. In:
Toxicity of Organophosphate and Carbamate Pesticides. RC Gupta, Elsevier: in press. Colborn, T. Online 7
September 2005. A Case for Revisiting the Safety of Pesticides: A Closer Look at Neurodevelopment.
Environmental Health Perspectives, available at http://dx.doi.org.
58

-------
EPA and other Federal agencies. We examined dietary assessment methods and examined
food consumption databases, pesticide residue data sources, and probabilistic models to
better understand how such data integrate in risk assessments performed by OPP. We
reviewed the annual work plans prepared by OPP and the multi-year research plans by ORD.
Furthermore, we interviewed current and past administrators, science policy directors,
scientists, and risk assessors from EPA offices, as well as experts from other Federal
agencies and outside organizations, to capture expert viewpoints, clarify our interpretations,
and confirm our findings.
Based on our field work, examples of the types of data gaps we have uncovered include:
¦	The Agency has not published a summary of its findings from the developmental
neurotoxicity data submitted after the 1999 Data Call-ins.
¦	The Agency needs alternative developmental neurotoxicity testing models that are
targeted, efficient, and cost-effective.
¦	The Agency requires no developmental immunotoxicity testing data on food-use
pesticides.
¦	The Agency has no review of existing pesticide residual data from the Food and Drug
Administration, U.S. Department of Agriculture, and State residue monitoring programs
in terms of their reliability in describing the exposure of fetuses, infants, and other
children to potentially toxic pesticides. Also, it has no public, user-friendly national
residue database derived from data collected by these governmental partners.
¦	The Agency lacks protocols for generating exposure data.
¦	Data on non-dietary routes of exposure to pesticides are limited; missing data includes
exposure through pesticide use in homes and schools, as well as pesticide levels in air,
soil, surface water, or rainwater.
¦	The Agency needs additional scientific tools and data to provide an understanding of the
most important pathway(s) of exposure for young children.
¦	The Agency needs data on fate-and-transport and approaches for determining and
verifying the exposure factor.
¦	The Agency requires no data on pharmacokinetics or pharmacodynamics of pesticides in
developing animals, and its risk assessments include no such information.
¦	The Agency has not required chemicals to undergo endocrine disruption screening and
has little data on a pesticide's potential to disrupt the endocrine (hormonal) system.
However, EPA has a planned Endocrine Disruptor Screening Program in a validation
phase.
¦	The Agency requires no testing and has little test data to assess the interactive effects of
multiple chemicals or of chronic low-dose multiple chemicals.
We maintain our position that more scientific tools and better data are still needed to help the
Agency meet its regulatory challenges posed by FQPA.
4. We recognize the legal and sensitive nature of using terminology specific to the Agency's
science policy language. We have considered the Agency suggestions for technical or
editorial changes and made minor editorial changes throughout the report to minimize
59

-------
confusion for the Agency and our stakeholders. In the case of the Agency's comment on the
usage of the term "toxic effect" versus "common mechanism of toxicity," we substituted the
existing sentence with two new sentences to provide clearer discussion on that subject matter.
Likewise, we edited Table 2.1 for the purpose of clarity.
5.	We acknowledge that the Agency's suggestion of quoting FQPA statutory language would be
beneficial. However, we reject the Agency's view that our statement is incorrect. FQPA
charges the Agency to employ a new standard for establishing pesticide tolerances and
articulates a strong policy of protecting infants and children from reproductive and
developmental hazards. In establishing tolerances, EPA must assess risks to infants and
children on the basis of available information concerning consumption patterns among
infants and children, special susceptibility of infants and children, and cumulative effects of
exposures to infants and children.31 More importantly, "in the case of threshold effects," the
Act specifies that the Agency must apply an additional "ten-fold margin of safety" to take
into account"potential pre- and postnatal toxicity and completeness of the data with respect
to exposure and toxicity to infants and children,"32 The Agency may use a different
additional margin of safety, but "only if, on the basis of reliable data, such margin will be
safe for infants and children."33
6.	We believe expediting the completion of any remaining work from the developmental
neurotoxicity studies since the 1999 Data Call-ins is a demonstration of accountability. Our
opinion is that the Agency owes stakeholders a written summary of the findings and
conclusions from the developmental neurotoxicity data submitted by the manufacturers after
the 1999 Data Call-ins.
7.	We did not elaborate on the details of the alternative developmental neurotoxicity testing
methods in the draft report. However, during our field work we discussed alternative
developmental neurotoxicity testing methods with OPP and ORD representatives. ORD
scientists have mentioned their efforts in developing a high throughput screening battery
focused on detecting chemicals likely to be developmentally neurotoxic. Also, we have
suggested to OPP staff present at our exit briefing to consider usage of cell culture,
invertebrate, or non-mammalian models for primary testing prior to more targeted
examinations of developmental neurotoxicity. We have referred them to the research at
Duke University and one specific manuscript34 which contained suggestions for alternative
models with high throughput (e.g., rat embryo cultures, neurotypic and gliotypic cells,
zebrafish embryos, and/or sea urchin embryos). In this final report, we have expanded our
discussion and provided references suggesting possible strategies. We encourage OPP to
review the references we have passed along and consider Recommendation 3-2.
31	21 U.S.C. §346a (b)(2)(C) (1994 & Supp. IV 1988).
32	21 U.S.C. §346a (b)(2)(C)(ii)(II) (emphasis added) and as reference on page A-l inEPA's Science Policy paper,
Determination of the Appropriate FQPA Safety Factor (S) in Tolerance Assessment, OPP, USEP A, Washington, DC,
February 28, 2002.
33	Id.
34	Slotkin, TA. 2004. Guidelines for Developmental Neurotoxicity and Their Impact on Organophosphate
Pesticides: a Personal View from an Academic Perspective. Neurotoxicology 25, 631-640.
60

-------
8.	We believe when the Agency engages its staff in learning events like the TestSmart
Developmental Neurotoxicity Test symposium sponsored by the Johns Hopkins University
and its partners, it demonstrates commitment to children's health and continuous regulatory
performance improvements.
9.	The Agency commented that it is addressing the evaluation of sensitive and meaningful
indicators as part of the Standard Evaluation Procedure for Developmental Neurotoxicity
tests. We have modified Recommendation 3-1 to include information on sensitive and
meaningful developmental neurotoxicity indicators and deleted the issue as a stand-alone
recommendation.
Our recommendation suggests the Agency evaluate "which indicator, or combination of
indicators, is most sensitive and meaningful for assessing developmental neurotoxicity
consequences of exposure during critical windows of development." Hence, we are not in
disagreement that a single, most sensitive parameter might be unlikely for assessing
developmental neurotoxicity consequences. Risk assessment of a developmental toxicant
requires careful consideration of the end point of toxicity, the dose-response relationship, and
the relevance of the animal model to humans. Improvements in analytical laboratory
equipment and testing procedures have made it easier to detect pesticides and their
metabolites (breakdown products) at very low concentrations in animal and almost all human
tissue. However, some of the endpoints used in the laboratory to detect functional
impairment of the brain and nervous system are measured at the biochemical, gene, cell, and
physiological levels, requiring high tech instrumentation to quantify. Emerging fields, such
as medical imaging, nanotechnology, and sensor technology are beginning to generate insight
on effects of pesticide exposure on brain cell damage.35 We encourage the Agency follow
the findings in these areas of research.
10.	We disagree with the Agency that it ensures developmental neurotoxicity tests are conducted
on developing animals in addition to young adult animals. First, the Agency only
"conditionally required" developmental neurotoxicity tests; the trigger for nervous system
toxicity testing hinges on results from other, less specific, toxicological testing that generally
does not involve the nervous system. Second, the current developmental neurotoxicity tests
do not assess toxicant-induced alterations in the developing nervous system of fetuses and
the embryos.
The current developmental neurotoxicity test guideline suggests neuropathology with
morphometry of several brain regions on postnatal day 11 and at termination of study,
observation of offspring for motor activity "o//postnatal days 13, 17, 21, and 60 (±2 days),"
auditory startle response habituation and pre-pulse inhibition, and "a test of associative
learning and memory" "conducted around the time of weaning and around day 60."
However, morphologic and histopathologic assessment of toxicant-induced alterations in the
developing nervous system for human health risk assessment requires an understanding of
corresponding timeframes for the critical events in nervous system development of the rat
and human. Additionally, such morphologic and histopathologic assessments require careful
35 Research Advisory Committee on Gulf War Veterans' Illnesses. September, 2004 Scientific Progress in
Understanding Gulf War Veterans' Illnesses: Report and Recommendations.
61

-------
qualitative and quantitative evaluations, including such basic methods as determination of
brain weight and dimensions as well as the more complex approaches of linear, areal, or
stereologic measurements of brain sections.
11.	We fear the loss of public confidence in EPA's commitment to protect infants and children
from developmental hazards when the Agency allows 6 years to elapse before it revises the
developmental neurotoxicity test guideline. However, we confirm that in 1999, when EPA
issued a Data Call-In for neurotoxicity testing of a list of organophosphate insecticides, the
exposure period was extended from gestation day 6 through post-natal day 10 to gestation
day 6 through post-natal day 21.
12.	We do not disagree that EPA has an adult immunotoxicity testing guideline and proposed in
March 2005 that immunotoxicity testing be required for pesticide registration. However, as
the Agency states, it currently has no requirement for immunotoxicity testing in adult or
developing animals. Furthermore, it has not developed a developmental immunotoxicity
testing guideline in the 9 years since FQPA was passed. Unlike adults, a child's immune
system is a protective mechanism still in development. Pesticides may interfere with the
maturation of immune cells during childhood and cause abnormal development of the
immune system. Abnormalities of the immune system could potentially lead to allergies,
asthma, and autoimmune disease or increased susceptibility to infections. We encourage the
Agency to expedite the completion of a developmental immunotoxicity testing guideline.
13.	In its own response the Agency stated, "OPP realized that the rule should describe the DNT
[developmental neurotoxicity test] study as 'conditionally required' to reflect the limited
conditions when the data requirement would be imposed." Our concern is with "the
limited conditions" under which such data would be required. While it is true that "neither
the substance nor the scope of the proposed DNT [developmental neurotoxicity test] required
was altered," the Agency changed the condition under which it requires developmental
neurotoxicity testing. As stated previously, by labeling the developmental neurotoxicity
testing as "conditionally required," the Agency is in fact saying it will only recommend this
kind of testing after certain conditions (triggers) have been met. Non-EPA scientists have
criticized the Agency for using triggers that are inadequate or not enforced. Also, a former
EPA neurotoxicologist had been cited36 to point out that the triggers for recommending a
developmental neurotoxicity study in some cases depend on information best obtained from
the developmental neurotoxicity study itself. It is our opinion that the Agency takes
seriously the lessons learned from chlorpyrifos and accepts the weight of scientific evidence
that points to how exposure to common pesticides can damage the developing brain.
14.	We disagree with the Agency that Recommendation 4-1 is premature. However, we have
modified the recommendation to ensure that the Agency updates its food consumption data in
2006 when USDA and DHHS release the 2003 and 2004 food intake survey data sets. We
believe that after the 1998 children's consumption survey activity, the Agency neglected to
communicate to its Federal partners OPP's continuous need for children-specific
consumption data. We agree that OPP's and ORD's expertise is not in dietary assessment
36 Schettler, T et al. January 2001. "In Harm's Way: Toxic Threats to Child Development" Greater Boston
Physicians for Social Responsibility! Clean Water Fund, page 111.
62

-------
methodology. We also recognize that Agency depends on its Federal partners for national
dietary exposure data and that the National Health and Nutrition Examination Survey used
different interview systems/methodologies between the 1999-2000/2001 data and the 2002
data. For these reasons, we organized a meeting during our field work for Agency
representatives to meet with dietary assessment methodology experts from USDA's
Agricultural Research Service. We invited the Agricultural Research Service National
Program Research Leader for Human Nutrition and her staff to this meeting. In their
presentations, these scientists discussed the Agricultural Research Service's role in
developing the dietary methodologies used in the National Health and Nutrition Examination
Survey and in handling the analysis of the survey's dietary data. Our hope was that the
Agency officials would seize this opportunity to dialogue about it needs to utilize the various
years of the survey when methodologies are different and to communicate its need to link
consumption, commodity, and pesticide residual data more efficiently. We encourage the
Agency to communicate with the dietary methodology experts at the Agricultural Research
Service about the feasibility of using the dietary consumption data from the National Health
and Nutrition Examination Survey after the 1999 USDA/DHHS integration.
15.	The Agency mentioned that "practical considerations prevented the use of the Gerber
Feeding Infants and Toddlers Study." It further commented that it decided "resources were
more appropriately invested in working jointly and cooperatively with other US government
agencies...." We do not disagree with this thinking; however, we would like to point out for
OPP that if it is serious about using the Gerber study's data, it needs to dialogue with Gerber
during the survey planning phase, not when the results have been compiled for Gerber's
needs. Also, the Agency's response alludes to Feeding Infants and Toddlers Study data from
older studies, perhaps before the release of Gerber's most recent study. Since Gerber
conducts the Feeding Infants and Toddlers Study at least once every 5 to 10 years, the
Agency might consider communicating its interest in acquiring data from the next study now.
16.	The Agency recognized the importance of correlating or validating model-predicted
exposures with "real world" measures from longitudinal studies, especially, since for
humans, food consumption and pesticide usage patterns do vary across weeks in a month and
seasons within a year. We believe the choice of data sets for food intake, pesticide residues,
chemical use, and toxicity as presented in probabilistic models can have dramatic effects on
exposure and risk estimates. Also, while OPP utilizes pesticide exposure models (like
Lifeline, CARES, DEEM-Calendex) to predict long-term dietary exposure risk, such risks are
estimates and could easily be above or below real world levels.
After reviewing the Agency comments on longitudinal consumption data, we consolidated
our recommendations on developing methodology for collecting longitudinal consumption
data with the recommendation to collect such data. In so doing, we also renumbered the
other recommendations accordingly. We agree with the Agency that it has neither the
expertise to develop the methodology for collecting longitudinal consumption data nor the
resources/funding to collect and analyze longitudinal food and activity exposure information.
Therefore, as stated in note #14, during our field work we organized a meeting for Agency
representatives to meet with dietary assessment methodology experts from USDA's
Agricultural Research Service. At the meeting, the Agricultural Research Service scientists
63

-------
introduced research activities from their human nutrition research centers (two of which
focus on children's health), highlighted the compact tools developed by the Agricultural
Research Service's scientists for food intake and activity monitoring, and discussed the
Agricultural Research Service's role in developing the dietary methodologies used in the
National Health and Nutrition Examination Survey (including commenting on the Food
Propensity Questionnaire) and in handling the analysis of the survey's dietary data. We
encourage the Agency to dialogue with the experts at Agricultural Research Service about
opportunities for collaborative research for more and better children's dietary and nondietary
exposure data.
We are aware that the National Cancer Institute's Food Propensity Questionnaire was pilot
tested in the National Health and Nutrition Examination Survey and was included in the
survey starting in 2003. This data collection instrument is the National Cancer Institute's
attempt to improve the method of assessing long-term average, or "usual dietary intake."
The Food Propensity Questionnaire was designed to build on the strengths of both the
24-hour dietary recalls and Food Frequency Questionnaires. It is similar to the National
Cancer Institute's Diet History Questionnaire but without portion size questions. Such an
instrument is meant to supplement the 24-hour dietary recall methodology currently used in
the National Health and Nutrition Examination Survey. We have learned the propensity
method assumes that usual intake is a function of the propensity to consume (the probability
that a person will eat a specific food or beverage on a given day over a designated time
period) and the average amount consumed on a day when the food is actually eaten. Initial
validation studies have shown the Food Propensity Questionnaire accurately measures
propensity as definedfor this method and that combining the 24-hour dietary recall and the
Food Frequency Questionnaire is a more efficient way to estimate commonly eaten foods in
the U.S. diet. However, this approach is still subject to error associated with self-reporting of
intakes. We encourage the Agency to dialogue with the experts at the Agricultural Research
Service about its need for more and better longitudinal dietary exposure data.
17. We disagree with the Agency that its current practice fully addresses Recommendation 4-2.
During our review, the Agency provided no documentation and we found no evidence that it
assessed Food and Drug Administration and USDA residual data in terms of their reliability
in describing the exposure of fetuses, infants, and children (through adolescence) to
pesticides consumed through domestic and/or imported foods, including ethnic foods. We
recognize that resource limitations constrain the Agency's research efforts in measuring
residual and dietary consumption patterns. Since OPP relies heavily on the use of models
and assumptions to determine risk for setting pesticide tolerances, the choice of data sets for
food intake, pesticide residues, chemical use, and toxicity as presented in models can have
dramatic effects on exposure and risk estimates. Also, given evidence from recent studies
demonstrating that a switch to organic diets significantly lowered children's dietary exposure
to organophosphorus pesticides,37 our recommendation is for the Agency to assess which
additional foods frequently consumed by children (including adolescents) should be included
37 Curl CL et al. 2003. Organophosphorus Pesticide Exposure of Urban and Suburban Pre-school Children with
Organic and Conventional Diets. Environmental Health Perspectives 111 :377-382. Lu C et al. Online September
1, 2005. Organic Diets Significantly Lower Children's Dietary Exposure to Organophosphorus Pesticides.
Environmental Health Perspectives available at http://dx.doi.org.
64

-------
in the Pesticide Data Program testing. We encourage the Agency to work closely with
USDA and the Food and Drug Administration in this area.
18.	The Agency stated that "the sampled PDP (Pesticide Data Program) commodities for
children 1-2 years old directly or indirectly represent approximately 90% of children's diets."
We require additional support to accept the claim that sampled commodity foods "for
children 1-2 years old" would represent "90% of children's diets" when, according to the
Agency (in its response to our prior report38), OPP evaluates pesticide risks for every
pesticide in food for children including those between ages 3-5, 6-12, and 13-19. We
recommend the Agency assess which additional foods (including ethnic foods) frequently
consumed by children (through adolescence) should be included in the Pesticide Data
Program testing. Furthermore, we believe it would be beneficial to assess the scale of
monitoring required to know and manage the mixture of pesticides dispersed into the food
and water supplies. We encourage the Agency to examine food intakes by children (through
adolescence) from the National Health and Nutrition Examination Survey to assess whether
additional foods should be analyzed.
We agree that determining which food to sample and not to sample requires careful
consideration of food consumption patterns, residue levels, and frequency of pesticide
detections. We also agree that the problem of managing pesticide residue in food lies within
a much larger food safety monitoring arena in which EPA is a minor player. We understand
the Agency is hampered in doing adequate aggregate exposure assessment both by the lack of
data on individual routes of pesticide exposure and by a lack of biological monitoring data in
infants and children. However, in the area of dietary pesticide exposure, we believe EPA has
a major responsibility because it sets the tolerances and registers the pesticide chemical use
on food. We encourage the Agency to solicit assistance from USDA (for example, the
Agricultural Research Service and the Agricultural Marketing Service) and DHHS (for
example, the Food and Drug Administration) when determining which additional foods to
sample.
19.	We agree that ORD has been a partner in funding and delivering research results pertinent to
OPP's mission and needs. We have been informed that OPP is involved in drafting relevant
Requests for Applications and participating in the internal relevancy review of grant
applications. Also, we know OPP has requested more longitudinal data from ORD. We are
aware of the extramural research program which the National Center for Environmental
Research manages. We recognize that National Center for Environmental Research grants
support both individual investigator research and multi-disciplinary research grants and
centers. More specifically, the Science to Achieve Results (STAR) program is an extramural
funding program within ORD and it has funded pertinent pesticide exposure and children's
health projects. We know that the STAR program communicates grantees' research results
to scientists in EPA and the public through Web sites, meetings, and publications. However,
we learned during our evaluation from several sources that OPP risk assessors seldom
included STAR grant project results into their risk assessments, do not routinely attend
STAR program presentations or National Center for Environmental Research seminars, and
38 EPA OIG. October 19, 2005. Changes Needed to Improve Public Confidence in EPA's Implementation of the
Food Quality Protection Act. Report No. 2006-P-00003.
65

-------
generally did not "participate extensively" in the ORD STAR program activities perhaps
until recently. We encourage OPP to fully utilize relevant findings from the ORD STAR
projects in its pesticide risk assessment work.
20.	We agree with the Agency that it should continue to track pesticide residue in cow's milk.
Nonetheless, the Agency's data on human breast milk have not kept pace with environmental
effects. As the Agency concluded, "more recent information is needed on breast milk
consumption and the incidence and duration of breastfeeding."39 As a short-term solution
we can accept the Agency's proposal to apply the FQPA Children's Safety Factor "in the
event that OPP identifies a chemical for which it thought breast milk might be a significant
source of exposure." However, we do not see this solution as adequate for risk assessments
in the long term. In our opinion, up-to-date breast milk data should be considered when
aggregating various sources of pesticide exposure among nursing infants.
21.	In reviewing the Agency's comments on breast milk biomonitoring, we encourage the
Agency to include in its research plans research to determine the rates of elimination kinetics
for various classes of chemicals from the mother's body during lactation. Further, the
Agency might consider partnering with the Centers for Disease Control and Prevention to
identify human biomarkers of exposure, susceptibility, and effects to predict potential health
risks associated with environmental chemicals to breast-fed and formula-fed infants and
mothers.
22.	We disagree with the Agency's view that it has adequate dietary exposure data through the
Continuing Survey of Food Intake by Individuals and the National Health and Nutrition
Examination Survey from schools and daycares. Throughout our evaluation, we were
provided no evidence to illustrate that the Agency analyzed the Continuing Survey of Food
Intake by Individuals dataset to confirm that it contains sufficient sample size to draw
statistical conclusions about children's dietary exposure risk from school lunch programs and
daycare centers. Also, we would need additional evidence to accept the Agency's view about
the 2003-2004 NHANES dietary data providing adequate dietary exposure data on children
from schools and daycares when the Agency has not updated its food consumption database
since the NHANES effort began.
We recommended the Agency expand its partnerships with USDA and DHHS and consider
collecting data on dietary exposure at schools (elementary or secondary) and day cares so
that the Agency increases its batch of children-specific data. A recent manuscript40 reported
several findings of concern about the risks of pesticide use in and around the nation's schools:
¦	pesticide poisoning incidence rates among children increased significantly from 1998 to
2002;
¦	drifting pesticides applied off site were responsible for 31 percent of reported poisonings;
and
39	US EPA. September 2002. Child-Specific Exposure Factors Handbook. National Center for Environmental
Assessment, Office of Research and Development, Washington, D.C. EPA-600-P-00-002B. Seepage 1-22.
40	Alarcon WA et al., July 27, 2005. Acute Illnesses Associated With Pesticide Exposure at Schools. Journal of the
American Medical Association, Vol. 294, No. 4:455-465.
66

-------
¦ insecticides and disinfectants were the pesticides most frequently at fault.
The authors noted that no Federal requirement limits pesticide exposures at childcare centers
and elementary or secondary schools, and that their pesticide poisoning results "should be
considered low estimates of the magnitude of the problem because many cases of pesticide
poisoning are likely not reported to surveillance systems and poison control centers." Since
simultaneous collections of nondietary exposure and biomonitoring data could be beneficial,
we have modified our recommendation to include "nondietary exposure" data.
23.	While we agree with the spirit of outsourcing software development for probabilistic models
or funding external researchers to seed research and develop the next generation of
environmental scientists, we do take the position that, prior to seeking external support, the
Agency should fully utilize the expertise of ORD scientists to develop exposure estimate
models (including Physiologically-Based Pharmacokinetic modeling) for its core work.
ORD is the Agency's principal research arm; its role is to provide critical science and
scientific products for environmental decision-making through its problem-driven and core
research projects. We maintain our position that OPP coordinate efforts with ORD to
finalize the integration of the Exposure-Related Dose-Estimating Model (ERDEM) with the
Stochastic Human Exposure and Dose Simulation pesticides exposure model (SHEDS).
24.	We appreciate the Agency's thoroughness in reviewing Figure 6.1. As a result, the figure
was modified to reflect corrected information from the Agency. However, we have concerns
over the "Science Policy Issues & Guidance Documents" Web site being unclear, out-of-
date, and misleading. We noted that there are four science policy papers still posted on the
"Science Policy Issues & Guidance Documents" Web page for which OPP plans to issue
Federal Register Notices announcing their withdrawals. Although OPP is subject to statutory
and court-ordered deadlines that require significant resources, we believe managing the
currency of the science policy papers and associated Web site should be an Agency priority.
We believe the Agency's Web site is a means to communicate EPA's FQPA implementation
efforts and OPP's use of sound science to reduce uncertainty in its regulatory decisions
25.	We included this table as an example of the types of toxicity testing data gaps cited by
non-EPA scientists since the enactment of FQPA.
67

-------
Appendix E
Toxicity Testing Issues
The following table contains examples of data gaps identified in the report Putting Children
First: Making Pesticide Levels in Food Safer for Infants and Children, issued in April 1998 by
the Natural Resources Defense Council. The complete table is on page 9 of that report.
Tests
Test includes
in utero
exposure
Test includes
post-natal
exposures
Tests for
effects on
offspring
Tests Required for Food-Use Pesticides
Acute Tests'.
Oral toxicity (rat)
No
No
No

Dermal toxicity
No
No
No

Inhalation toxicity (rat)
No
No
No

Primary eye irritation (rabbit)
No
No
No

Primary dermal irritation
No
No
No

Dermal sensitization
No
No
No
Mutagenicity.
Gene mutation1
N/A
N/A
N/A

Structural chromosomal abereration
No
No
Yes2
Subchronic.
90-day feeding (rodent & non-rodent)
No
No
No
Chronic:
Feeding study (rodent & non-rodent)3
No
No4
No
Cancer:
Carcinogenicity3
No5
No5
No
Metabolic:
General metabolism tests
No
No
No
Developmental: Developmental toxicity (rat and rabbit)
Yes
No
No
Reproductive:
Two-generation study (rat)
Yes
Yes
Yes
Significant Tests, Occasionally or Rarely Required
Acute delayed neurotoxicity (hen)6
No
No
No
Developmental neurotoxicity study (rat)
Yes
Yes
Yes
1	Most gene mutation tests are done on cell culture systems. Per 40 CFR Part 158, there may be other tests of genotoxicity
done, but this determination is only made on a chemical-by-chemical basis.
2	Second-generation offspring are tested only indirectly, by examining effects on sperm of exposed animals.
3	The chronic feeding and carcinogenicity studies may be combined into one test using the same animals.
4	Exposure is recommended to start immediately after weaning.
5	Pre- and perinatal exposure of test animals may be required under certain conditions, according to guidelines, but is atypical.
6	Only required for organophosphate or structurally related pesticides.
In a 2002 EPA review by an EPA technical panel, Review of the Reference Dose and Reference
Concentration Processes,41 the panel identified numerous data gaps in the testing guidelines.
This panel suggested that the Agency develop alternative strategies and guidance to allow more
targeted testing. The panel found study design and data collection gaps in life stages,
41 Reference Dose/Reference Concentration Technical Panel, Risk Assessment Forum, EPA/630/8-02/002F, December 2002.
68

-------
particularly in terms of the exposure periods in the current guideline testing protocols. The panel
indicated that there is minimal evaluation of aged animals, especially after exposures that include
early development. This panel pointed out that there is a lack of information on toxicokinetics
(the determination and quantification of the time course of absorption, distribution,
biotransformation, and excretion of chemicals), and that available data are generally limited to
studies conducted of young adult animals. The panel found no guideline protocol for
toxicokinetic evaluations during development or older age related to exposures and outcomes.
In the data requirements for pesticide registrants in 40 CFR Part 158, the Agency's required
testing includes no evaluation of behavior, learning, or memory in developing animals. Learning
and memory testing is part of the developmental neurotoxicity study, but requiring such a study
is contingent upon predefined conditions or triggers. Behavior and functional tests are sensitive
parameters that one would presume might be essential for assuring safety to infants and children
and fulfilling the challenges posed by the FQPA. Also, although there is a developmental
toxicity study and a two-generation reproductive study required in the testing, the Agency's own
assessment of these two studies is that they "do not include an in-depth assessment of the
development of the nervous system."42 Also, the Agency's Science Advisory Panel
acknowledged that these testing criteria "were not adequate for identifying every potential
developmental neurotoxicant, supporting the Agency's concern about the criteria's limitations."
The current EPA guideline for developmental neurotoxicity study states that the "dosing period
covers the period from day 6 of gestation through day 10 postnatally." According to literature,
critical period of rapid human brain development is from the third trimester through the second
year of life, which for mice or rats ranges during the first 21-28 days of life. The exposure
period recommended by the guide may be too short to reflect the entire vulnerable period of
brain development in children. Also, statistical procedures to define the minimal number of
animals in a test group needed to give sufficient power to detect meaningful differences are
lacking in the guide. Nonetheless, the current developmental neurotoxicity guideline is EPA's
most sensitive validated means of examining unique endpoints that are not examined in other
standard toxicity protocols. It enables the detection of effects in the offspring following pre-
and/or postnatal exposure.
Finally, cholinesterase inhibition is the driving endpoint for organophosphate pesticides.
Organophosphate pesticides were thought to affect brain development through their ability to
elicit cholinesterase inhibition and cholinergic hyperstimulation. Cholinesterase inhibition can
be detected in brain tissue, plasma, and red blood cells, but there is controversy over which form
is the most sensitive method of measure and whether whole-brain cholinesterase assays could
readily miss brain regional and sub-regional cholinesterase inhibition. Some researchers are
concluding that cholinesterase inhibition alone is not enough to assess the consequences of
exposure. Recent research shows that chlorpyrifos disrupted developmental neurotoxicity
synthesis in the brain and synaptic signaling and function. This means chlorpyrifos has direct
effects on cellular processes that are unique to brain development, and that these effects are
mechanistically unrelated to inhibition of cholinesterase.
42 70 Federal Register 12275 (March 11 2005), page 122295; Docket Control ID OPP-2004-0387.
69

-------
Appendix F
Data and Tools for Estimating Dietary Exposure
Collecting national food consumption data is costly and complex, and EPA relies on other
Federal agencies for such data. FQPA contains specific provisions for cooperative activities
between EPA and USD A. The figure43 below illustrates some of the dietary exposure
requirements for assessing risks, while the text that follows discusses various collection methods
EPA relies upon to estimate dietary pesticide exposure risks.
DIETARY EXPOSURE DATA REQUIREMENTS FOR ASSESSING RISKS
Dietary Exposure Risk
Residues in//
on Food
Handling
Transfers
Source of
P urification
systems of Watei
Regularity of
Cons umption
Regularity of
Pesticide Residual Monitoring Activities
USD A. AMS
DHHS. FDA
Pesticide
Data
Program
US DA, ARS
2 nonconsecutive days.
\ 24-hr recalls /
2 nonconsecutive days.
V 24-hr recalls /
Total Diet Study
US DA/EPA
Estimated Dietary Exposure Calculation
Deterministic: Point Estimates
Probabilistic: Probability models
Simple Distributions: Subpopulations,
Regions, Hot Spots
Pesticide/Commodity Combinations
	C ^">	
Confirmatory Methods
Biomarkers
Duplicate Diet Studies
Longitudinal Dietary Intake Records
AMS: Agricultural Marketing Services	FDA: Food and Drug Administration
ARS: Agricultural Research Services	NHANES: National Health and Nutrition
CSFII: Continuing Survey of Food Intakes by Individuals Examination Survey
DHHS: Department of Health and Human Services	USDA: U.S. Department of Agriculture
U.S. Food Consumption Data Collection Activities
"What We Eat in America from NHANES (National Health and Nutrition Examination
Survey)": The National Nutrition Monitoring and Related Research Act of 1990 provided the
impetus for a coordinated effort to collect and report nutrition and health status data, and
stimulate research to develop uniform methodologies, technologies, and procedures for national
nutrition monitoring. In 1998, the National Center for Health Statistics of the Centers for
43 OIG staff developed this figure based on data collected for this evaluation.
70

-------
Disease Control and Prevention at DHHS and the Agricultural Research Services at USDA
signed a memorandum of understanding to integrate the National Health and Nutrition
Examination Survey and the Continuing Survey of Food Intakes by Individuals dietary data
collection activities into an integrated survey. The Agricultural Research Services processes all
dietary recall data collected and releases it under the title "What We Eat in America from
NHANES"
Since 1999, the Continuing Survey of Food Intakes by Individuals was integrated with the
National Health and Nutrition Examination Survey, which is a multistage, stratified area sample
that is representative of the civilian noninstitutionalized population of the United States. Certain
groups were over-sampled to allow for more precise estimates. Over-sampled groups include
adolescents 12-19 years, persons 60-plus years, African Americans, Mexican Americans, low-
income persons, and pregnant women.
"Nationwide Food Consumption Survey" and the "Continuing Survey of Food Intakes by
Individuals": Prior to the passage of FQPA, food consumption surveys accepted by the EPA as
sources for estimating food intake by individuals were from the USDA Nationwide Food
Consumption Survey (1977-78) and the Continuing Survey of Food Intakes by Individuals
(1989-91). These surveys were designed to provide nationally representative, multistage,
stratified samples of U.S. adults, but infants, children, and certain demographic categories were
under-surveyed. In 1998, after FQPA enactment, USDA and EPA collaborated to collect data
specifically from children.
Food Commodity Intake Database: The Food Commodity Intake Database was generated by
the USDA Agricultural Research Service for EPA from the data collected in Continuing Survey
of Food Intakes by Individuals 1994-96, 1998.
U.S. Pesticide Residual Monitoring Programs
USDA Agriculture Marketing Service's Pesticide Data Program: This program concentrates
its efforts in providing better pesticide residue data on foods most consumed by children. The
program supplies continuing information on pesticide residues in fruits, vegetables, grains, dairy
products, and meats. Food samples are collected by USDA's Agricultural Marketing Service
immediately before commodities are shipped to grocery stores and supermarkets, and prepared
by the laboratory as they typically would be for consumption.
Total Diet Study: This study, sometimes called the Market Basket Study, is an ongoing DHHS
Food and Drug Administration program that determines levels of various contaminants and
nutrients in foods. Since its inception in 1961, the study has grown to include analyses of
radionuclides, residues of pesticides, industrial chemicals, toxic and nutritional elements, and
folate. The number of different foods sampled in the study has increased from 82 food items
when the study was initiated to about 280 foods in the current program. Three samples of four
regional market baskets are collected each year from three cities per region. Study diets are
derived from the national food consumption survey data and are generally compiled in
conjunction with updates of the study's food list. In response to the FQPA, additional infant and
71

-------
toddler foods were added to the study to provide more information on levels of pesticides and
lead in the diets of young children.
Computerized Risk Assessment Computation Models
Pesticide exposure assessments can use models that are either deterministic, probabilistic, or
both. Deterministic models provide a point estimate of exposure, assuming that a typical child
eats an assumed mass of food per day with a given concentration of a pesticide residue. A
probabilistic model considers the range of estimates and provides a probability distribution of
exposures. It would calculate the range of mass of food and the range of food types eaten by a
particular group of children of a certain age and gender. Whether EPA uses a deterministic or
probabilistic approach, it will always be dependent on other Federal organizations for the large-
scale human food consumption data needed in its risk assessment and the biomarker data from
national biomonitoring studies to validate its modeling predictions.
OPP supports and encourages the development of freely available models and software tools that
could be used to conduct such assessments. Besides the ORD SHEDS (Stochastic Human
Exposure and Dose Simulation) pesticides exposure model, there are three other models
developed for OPP risk assessment purposes. These probabilistic risk assessment models each
project pesticide exposure for the U.S. population. However, they differ in their basic design in a
number of ways. Details on the four models follow.
•	SHEDS (Stochastic Human Exposure and Dose Simulation)-PESTICIDES: This is a
physically-based stochastic model (one that involves a random variable) that quantifies
exposure and dose of humans to multi-media, multi-pathway pollutants, and, in
particular, aggregate exposures of children to pesticides. To date the model has focused
on simulating aggregate exposures of children to pesticides. The model is being
expanded to address cumulative exposures.
•	DEEM/CALENDEX (Dietary Exposure Evaluation Model/Calendar-Based Dietary
and Non-dietary Aggregate and Cumulative Exposure Software System): These are
software systems developed and licensed by the private sector. The first provides a
probabilistic assessment of dietary exposure/risk for the U.S. population or subsets. The
second is an exposure assessment model that estimates aggregate pesticide exposures
from multiple pathways as required by the FQPA.
•	LIFELINE Software for cumulative and aggregate exposure assessment Version 2.0:
This is an exposure model that produces exposure estimates using 1994-1998 Continuing
Survey of Food Intake by Individuals and the USDA Food Commodity Intake Database to
investigate pesticide residues in diet, tapwater, and residential environments.
•	CARES (Cumulative and Aggregate Risk Evaluation System): This software consists
of several modules, including a Population Generator; Dietary, Water, and Residential
Modules; Aggregate and Cumulative Assessment Modules; and a Contribution and
Sensitivity Analysis Module. The Population Generator is used outside of the system to
generate a reference population of 100,000 individuals selected from 5,000,000
individuals who completed the long form of the 1990 U.S. Census.
72

-------
Appendix G
Distribution
Office of the Administrator
Acting Assistant Administrator for Prevention, Pesticides, and Toxic Substances
Assistant Administrator for Research and Development
Director, Office of Pesticide Programs
Acting Director, Office of Children's Health Protection
Associate Director, Field and External Affairs, Office of Pesticide Programs
Agency Followup Official (the CFO)
Agency Followup Coordinator
Audit Coordinator, Office of Prevention, Pesticides, and Toxic Substances
Audit Liaison, Office of Research and Development
Audit Liaison, Office of Pesticide Programs
General Counsel
Associate Administrator for Congressional and Intergovernmental Relations
Associate Administrator for Public Affairs
Inspector General
73

-------