U.S. Environmental Protection Agency
Revised OP (Organophosphate)
Cumulative Risk Assessment'
June 10, 2002
Executive Summary
Glossary / Acronyms
Table of Contents
I. Revised OP Cumulative Risk Assessment
A. Introduction
B. Hazard / Relative Potency Factor
C. Cumulative Risk from Pesticides in Food
D. Residential OP Cumulative Risk
E. Water OP Cumulative Risk
F. Cumulative Assessment
G. FQPA Safety Factor
H. Risk Characterization
I. Future Work
J References
This document was only published electronically.
Accessed 1/14/04 from:
See additional volumes for: http://www.epa.gov/pesticides/cumulative
m. Appendices
B. Hazard / Relative Potency Factor
E. Water Exposure Assessment (sections 1-6)
Cover page created by EPA Region 9 Library staff, January 14, 2004.
-------
Executive Summary
A. Introduction
In 1996 Congress enacted the Food Quality Protection Act (FQPA), which among
other things, requires EPA to take into account when setting pesticide tolerances
(maximum residue legally allowed on a food) "available evidence concerning the
cumulative effects on infants and children of such residues and other substances that
have a common mechanism of toxicity." Also, FQPA mandates that by 2006, EPA must
review the safety of all existing tolerances that were in effect as of August 1996. The
law requires EPA to place the highest priority for tolerance reassessment on pesticides
that appear to pose the greatest risk, such as the organophosphorus (OP) pesticides.
To implement the cumulative provision of FQPA, EPA has been working to develop
methodologies for conducting cumulative risk assessments and then conduct its first
such risk assessment on the OP pesticides. This has been a challenging task given
that historically, the potential health risks associated with exposure to pesticides has
focused on single pathways of exposure (e.g., exposure from food, or water, or
pesticide use in and around the home) for individual chemicals and not on the potential
for individuals to be exposed to multiple (common mechanism) pesticides by all
pathways concurrently, as is required under FQPA.
This scientific assessment of OP pesticide food safety contains good news for
American consumers. After years of rigorous scientific work, it strongly supports our
confidence that the United States has one of the safest food supplies in the world.
Specifically, with this groundbreaking work, EPA has evaluated over 1,000 OP pesticide
tolerances and virtually all of them are now consistent with the highest levels of safety.
Please note that EPA is still evaluating the tolerances for a few of the OP pesticides.
This finding comes after years of scientific work, countless scientific and public
meetings, and an existing regulatory process to ensure these pesticide tolerances meet
the tough food safety standard in the Food Quality Protection Act. In the last several
years, EPA has taken a variety of regulatory actions on the OP pesticides, ranging from
lowering application rates to complete cancellations of specific uses. These actions
have substantially reduced the risks and have contributed to the high level of safety
found in the cumulative risk assessment.
Executive Summary Page 1 of 11
-------
On December 3, 2001 EPA issued for public comment its "Preliminary OP
Cumulative Risk Assessment." That assessment was a preliminary review of the
results of a new way of analyzing data regarding potential exposure to pesticides. The
focus of the assessment was on the methods used to assess the risk. In contrast this
revised risk assessment describes the potential risks of OP's by presenting a range of
estimates that reflect the variability inherent in an assessment of this scope. Table 1
provides a side-by-side comparison of the major changes between the December 2001
and current documents. The changes were made due to comments submitted during
the public comment period, suggestions from the FIFRA Scientific Advisory Panel
(SAP), and issues that EPA was aware of at the time the preliminary cumulative risk
assessment was issued but had not yet addressed. These major changes are
discussed under "Hazard Assessment" and "Exposure Assessment," below.
With the release of this document the Agency has met its deadline obligation under
a Consent Decree with the Natural Resources Defense Council to issue a revised risk
assessment of the OP pesticides by August 3, 2002. As existing analyses are revised
or new information is obtained, EPA will review this assessment and will make further
changes as appropriate.
Not all of the changes result in
quantitative impacts on the risk
assessment. For example, in
February 2002 the FIFRA SAP
suggested that the Agency conduct depends upon the jnformation fed jnto , upon its
Sensitivity Analysis is the study of how the variation in
the output of a model can be apportioned to different
sources of van'ation-it aims to ascertain how the model
structure, and upon the assumptions made to build it.
Overall, sensitivity analysis is used to increase the
confidence in the model and its predictions by providing
an understanding of how the model response variables
respond to changes in the inputs.
http://sensitivitv-analysis.irc.cec.eu.int/default.htm
more "sensitivity analyses" to assure
the quality and robustness of the
model being used (see text box).
While these analyses provide valuable
information on the reliability of the
models, they do not change the
quantifications of risk (e.g., MOEs).
On the other hand, other changes do
impact the risk quantification. During the public comment period food processing
factors were submitted; EPA has updated its food exposure estimates using this
information.
It has become evident that addressing issues such as the FQPA Safety Factor and
the threshold of concern are both dependent on the available data. The decisions
made regarding these two issues involve risk management considerations and will be
made on a case-by-case basis. EPA intends to use a systematic approach in making
these decisions to reflect such factors as the quality of the available data and the
characteristics of the modeling analysis.
Executive Summary Page 2 of 11
-------
Table ES-1. Major Differences Between the Preliminary OP Cumulative Risk
Assessment and the Revised OP Cumulative Risk Assessment
December 2001
June 2002
Toxfcfty
Relative Potency
Factors (RPF's)
FQPA Factor
Treatment of Animal Data
Used best available data
Not addressed
Used means and
standard deviations
Additional RPF's were calculated:
chlorethoxyphos, phostebupirim, profenofos and
omethoate (a metabolite of dimethoate)
FQPA Safety Factors were assigned based on
available information; 1X for three OP's and one
metabolite; 3X used for the others
To see how the results would be affected, single
animal data were used in a sensitivity analysis
Food:
Processing Factors
Consumption Data
Residue Data
Impact of High-End
Exposure
Used best available data
Used the CSFII data "as is"
Did not use any
over-tolerance residues
Not considered
Duration of Exposure One-day
Populations
Considered
Drinking Water:
Number of Regions*
Sensitivity Analysis
Residential:
Populations
Considered
Type of Distribution
Number of Regions'
Pet Uses
The standard populations
13
Some performed
The standard populations
Uniform
13
Not included
Revised based on data provided during the public
comment period
Conducted a sensitivity analysis to look for
'extreme1 outliers
Included over-tolerance residue values
Conducted an analysis to determine whether
specific high-end consumption and/or residue
values are significantly responsible for the
exposure estimates at the higher percentiles of
the exposure distribution
One - and seven-day rolling average. Also, a
sensitivity analysis was conducted using 14- and
21-day rolling averages
Conducted a sensitivity analysis by looking at
additional subpopulations
7; EPA found that a number of the Regions could
be combined due to similarities among
geography, climate, and soil type
Extensive analyses conducted, as suggested by
the SAP
Conducted a sensitivity analysis by looking at
additional subpopulations
log normal, as recommended by the SAP
7; EPA found that a number of the Regions could
be combined due to similarities among
geography, climate, and soil type
New data on tetrachlorvinphos
Risk Quantification
Summary results; MOE's
for single-day exposures at
various percentiles of
exposure
Identified pestide/crop combinations that have
significant roles in the estimates. Risk presented
as ranges of MOEs at various percentiles
reflecting one- and seven-day exposures, and 14
and 21-day rolling averages
1A Note on "Regions." Because the United States is so climatologically and geographically diverse, EPA has divided
the country into different risk assessment "Regions" so that this diversity could be factored in to the assessments.
Executive Summary Page 3 of 11
-------
B. Hazard Assessment
1. RPF's
The RPF's were revised and relative potency factors for four additional
chemicals have been calculated (chlorethoxyphos, phostebupirim, profenofos, and a
metabolite of dimethoate).
2. FQPA Safety Factor
In the December 2001 preliminary cumulative risk assessment, EPA discussed
and characterized the potential multiple sources of exposure to children but did not
address the FQPA Safety Factor. The decision regarding the Safety Factor is
determined based on the available data for the specific chemicals in this
assessment. The Revised OP Cumulative Risk Assessment provides an analysis
on the sensitivity and susceptibility of infants and children to cholinesterase (ChE)
inhibition (the common mechanism of toxicity) caused by OP pesticides.
In summary, based on available information, the FQPA Safety Factor is 1X for
three OP's and one metabolite (dimethoate; omethoate, a metabolite of dimethoate;
chlorpyrifos; and methamidophos) and 3X for the remaining OP's. A summary of
the rationale is provided below; please note that these Safety Factors are
appropriate for this risk assessment only.
Q In making an FQPA Safety Factor decision, EPA considers both the potential for
pre- and postnatal toxicity and the completeness of the toxicology and exposure
databases (USEPA, 2002a). Looking at the exposure side of the equation-there
is a high degree of confidence in the exposure data and methodologies
used-EPA believes that it is not necessary to retain the default 10X FQPA
Safety Factor based on the exposure database.
Q The toxicity endpoints for this assessment were developed in consideration of a
10X uncertainty factor to account for interspecies variability and a 10X
uncertainty factor to account for intraspecies variability. Because some OP
pesticides show age-dependent sensitivity and there are missing comparative
ChE inhibition data in young animals for many of the OP's, EPA chose an FQPA
Safety Factor of 3X for most of the OP pesticides. There were a few whose data
supported a 1X FQPA Safety Factor:
- Age-dependent susceptibility data are available for seven of the OP's. The
data for dimethoate, omethoate (a metabolite of dimethoate), chlorpyrifos,
and methamidophos support an FQPA Safety Factor of 1X.
On June 25 to 27, 2002 EPA is consulting with the FIFRA Scientific Advisory
Panel on this sensitivity and susceptibility analysis for children. For more
information see: httD://www.epa.qov/fedrqstr/EPA-MEETINGS/2002/Mav/Dav-31/.
Executive Summary Page 4 of 11
-------
For future cumulative risk assessments the FQPA Safety Factor may be
retained, reduced, or removed, based on the available data which are specific to the
chemicals examined.
C. Exposure Assessment
1. Regions
Because the United States is climatologically- and geographically-diverse, EPA
divided the country into Regions so that it could account for factors such as weather
and soil type (these affect the amounts and types of pesticides used). In the
December 2001 analysis 13 Regions were used; the current analysis has seven.
The reason for this reduction is that EPA realized that some of the Regions were not
truly distinct so they were combined. Provided in Figure ES-1 is a map of the United
States that shows the seven Regions.
As mentioned in the "Introduction" sensitivity analyses were conducted for a
number of variables. The exposure data used for these analyses were from Region
A. EPA chose Region A because it has the highest estimated exposure.
2. Food
The amount of pesticide to which an individual is exposed (i.e., exposure) is
determined by combining the amount of pesticide that is in or on the food (i.e.,
ES-1. Regions Used for Exposure Assessment
OP Pesticide Cumulative Assessment Regions
-------
residue levels) and the amount and type of foods that people eat (i.e., food
consumption). In the Revised OP Cumulative Risk Assessment EPA conducted a
number of sensitivity analyses on the data and models supporting the food risk
assessment.
Consumption Data. EPA uses USDA's Continuing Survey of Food Intake by
Individuals (CSFII) for its food consumption data. One of the criticisms that has
been raised regarding the food consumption data is that it may include individuals
who have "extreme" diets. EPA scientists, including a nutritionist, have conducted a
sensitivity analysis on the food consumption database; no outliers were identified.
Consumption data that appeared unusually high and were associated with high
exposures in the cumulative risk assessment were fully investigated.
Although they did not identify any outliers, it is important that appropriate
sensitivity analysis be conducted so that any outliers are evaluated. Please note
that several individual OPs are still undergoing individual assessments and for these
pesticides future analysis on food consumption will continue.
Residue Data. All of the residue data in this assessment came from USDA's
Pesticide Data Program (PDP) and FDA's Center for Food Safety and Applied
Nutrition (CFSAN) monitoring.data. In the Revised OP Cumulative Risk Assessment
EPA incorporated over-tolerance residue values from the PDP data.
Impact of High-End Exposure. The December 2001 document pointed out that:
The data inputs and assumptions need to be verified) and the
results at the tail end of the distribution at the higher percentiles of
exposure for children's age groups need to be evaluated to ensure
they reflect reasonable consumption patterns. Additionally, OPP is
in the process of conducting sensitivity analyses that will permit a
fuller characterization of the contributors or sources of potential
risks associated with the food pathway.
The Revised OP Cumulative Risk Assessment includes an analysis of the upper tail
of the exposure distribution to determine whether specific high-end consumption
and/or residue values are significantly or mainly responsible for the exposure
estimates at the higher percentiles of the exposure distribution. In addition, a range
of percentiles of exposure as well as the percentiles at which the MOEs approach
100 (100 because the toxicity endpoints for this assessment were developed in
consideration of a 10X uncertainty factor to account for interspecies variability and a
10X uncertainty factor to account for intraspecies variability) are presented in the
body of the risk assessment. This information provides the basis for bounding and
characterizing exposures.
Executive Summary Page 6 of 11
-------
Duration of Exposure. In the December 2001 risk assessment EPA used one-
day as the duration over which an individual would be exposed to a pesticide
residue in food. However, this analysis overestimates risk because the toxicity data
and consumption reflect different time frames. For the current analysis EPA added
a second exposure duration, that of the seven-day rolling average in an attempt to
better match the time frames for the toxicity data with the consumption data which
are not directly comparable. EPA also believes using these time intervals will bound
the risk (i.e., the potential risk is best represented by a range of values for different
exposure durations). In addition the Agency evaluated 14- and 21-day averages for
one Region (Region A). EPA conducted these additional analyses to determine
whether estimates of average daily exposure changed significantly over longer
durations.
The chart provided below (Table ES-2) provides a discussion of how the one-
and seven-day durations are affected by four key factors.
Table ES-2. How the One- and Seven-Day Durations Are Affected
Factor ;
The degree to which the
exposure and toxicity time
frames correspond to each
other.
The degree to which the
Agency has captured the
previous day's cholinesterase
inhibition.
Day-to-day variation in
individuals' diets.
Possible correlation among
residues on different days.
Interpretation of Model Outputs
Impact on Durations
The use of a steady state hazard endpoint-based on toxicity studies that are '
21 -days or longer-tends to overstate the risk for the one-day analysis. Use of
the steady state value is more appropriate with the 7-, 14-, and 21 -day
analyses.
For the one-day analysis, the consideration of only a single day's exposure
may underestimate risk, to the extent an individual's previous days' exposures
continue to cause ChE inhibition. For the same reason, multiday exposures
may also underestimate risk.
Day-to-day variability in an individual's diet does not affect the one-day
estimate. Limited data about such variability requires EPA to make
assumptions that tend to underestimate the potential exposures for the seven-
day analysis.
EPA's multiday analyses do not account for the possibility that a person may
be more likely to encounter high residues in food because some portion of
their consumption comes from the same source. This limitation means that
multiday analyses may underestimate food exposure somewhat. This
limitation does not affect the one-day analysis.
The one-day analysis assumes that an individual is exposed to OP residues
from the tail of the distribution every day. This assumptions overestimates
risk. The seven-day analysis incorporates day-to-day variability in exposure
and is more representative of anticipated exposures.
The Agency believes the timeframe considerations, as they relate to both hazard
and exposure, to be among the most important for the OP cumulative assessment.
This is not surprising since the essence of the cumulative assessment is to estimate
likely co-occurrence in exposure to multiple chemicals and the likely combined effect
of those exposures.
Executive Summary Page 7 of 11
-------
Populations Considered. Standard population subgroups that EPA considers in
dietary risk assessment include: children one- to two-years-old; children three- to
five; adults 20 to 49; and adults 50 and older. Upon SAP's recommendation, EPA
looked at other subpopulations such as infants less than one year and teenagers.
This was done to demonstrate that indeed children one- to two-years-old are the
most highly exposed, due to their high consumption-to-body weight ratio.
3. Drinking Water
EPA evaluated the contribution to overall exposure resulting from OP pesticide
residues in drinking water across different Regions and found that drinking water is
not a significant source of exposure. EPA looked at the impacts that periods of
high-volume runoff (e.g., during the spring and storm events) have on the level of
pesticide residue estimated in drinking water. It was found that there are higher
concentrations of pesticides in the drinking water during such periods. The analysis
shows that, even considering such events, drinking water is not a significant
contributor to overall risk.
4. Residential
Populations Considered. The population subgroups that EPA considers for
residential exposure are the same as those considered for the food exposure.
Similar to the food assessment, EPA conducted sensitivity analyses by looking at
additional subgroups such as infants. This was done to see how including more
population subgroups would change the risk estimates. The Agency is still working
to evaluate individual residential uses (as part of the cumulative assessment) where
additional risk mitigation will likely be necessary. In the next several weeks, EPA will
continue the scientific and regulatory work to evaluate and address these potential
risks.
Type of Distribution. EPA reassessed residential exposure using log-normal
distributions of the available data (instead of a uniform distribution), wherever
possible. This change was made because, according to the SAP, a log-normal
distribution better represents the data set. Some of the resulting residential
exposure estimates, and in turn risk, are lower than the December 2001 estimates.
Pet Uses. New data on exposure from the pet uses of tetrachlorvinphos have
been used to quantitatively include tetrachlorvinphos in the residential assessment.
Executive Summary Page 8 of 11
-------
D. Risk Characterization
The risk characterization summarizes and integrates all of the information from the
various components of the assessment. Risk characterization looks at the strengths
and weaknesses of the data used, including any potential biases in input parameters
and the direction of that bias, reliability and availability of the data, as well as the
characteristics of the exposure models, and attempts to bound that uncertainty. The
revised assessment discusses in great detail what data have been used; how the data
have been used; and the strengths and weaknesses of the resulting analysis.
The risk estimates presented in this Revised OP Cumulative Risk Assessment are
the culmination of several years of Agency analyses, outside input, and risk mitigation
efforts on the part of the regulated community. Beginning in the summer of 1998 EPA
started to seek public input on its individual OP risk assessments by issuing Federal
Register notices asking for comment. In addition, EPA actively sought the advice of the
regulated community, environmental groups, and others through two Federal advisory
committees, the Tolerance Reassessment Advisory Committee (TRAC) and the
EPA-USDA Committee to Advise on Reassessment and Transition (CARAT).
Throughout this period of public review and scrutiny, a good deal of risk reduction
has been achieved through the risk mitigation measures taken on the individual OP's.
In 1996 49 OP pesticides were registered for use in agriculture and residential settings.
Today, 14 of those pesticides have been canceled completely and for another 28,
considerable risk mitigation actions have been taken. For example:
Q Methyl Parathion. Methyl parathion had been one of the most widely used OP's. In
1999 the registrants voluntarily canceled many methyl parathion uses that contribute
most to the children's diet. These included: apples, peaches, pears, grapes,
nectarines, cherries, and plums, carrots, succulent peas, succulent beans, and
tomatoes.
Q Ethyl Parathion. Before 2000, ethyl parathion had been one of the most highly
restricted pesticides registered for use in the United States. A 2000 agreement
canceled all remaining uses of the OP pesticide ethyl parathion, which included use
on nine agricultural crops. Use of parathion on corn grown for seed was to stop
immediately, with the use on other crops to be phased out over the next few years.
Q Chlorpvrifos. Before the risk mitigation measures were taken, chlorpyrifos had been
one of the most widely-used pesticides in and around the home. It is also one of the
most widely used OP pesticides in agriculture. In 2000 the registrants agreed to
cancel nearly all indoor and outdoor residential uses, as well as use on several food
crops that contributed most to children's dietary exposure.
Executive Summary Page 9 of 11
-------
Q Diazinon. Diazinon is one of the most widely used agricultural insecticides and until
2000, one of the most widely used insecticides for household lawn and garden pest
control. In 2000 all indoor residential uses were terminated; outdoor residential uses
will be phased-out over the next several years. Additionally, many agricultural uses
of diazinon also are being canceled.
Without these measures, pesticide exposure through food and in and around the
home would have been more significant. December's preliminary analysis and now the
revised analysis reflect all these important risk mitigation measures.
1. Risk Quantification
This version of the cumulative risk assessment presents results showing a range
of estimated risks depending on the exposure period considered (one-day or
seven-day average) and the percentile of exposure. Ranges of estimated risk at
various percentiles of exposure are also presented for 14- and 21-day averages for
Region A. The selection of the range for the percentile of exposure must take into
account the data from the particular group of chemicals in the assessment. For
most portions of the ranges presented from the different exposure periods, the
estimated Margins of Exposure (MOEs) do not represent levels of potential concern.
After careful analysis, the Agency believes that the potential exposures are bounded
by the estimates for the one- and seven-day exposure durations, and generally the
margins of exposure in this assessment do not represent major concerns.
In considering the possible need for risk mitigation actions, EPA believes that it is
important to consider the range of risk assessment values, which in turn take into
account different exposure periods, for different age groups, living in different
Regions, with risks shown at different percentiles of estimated exposures. It is also
important to consider risk characterization, including the factors that may tend to
overestimate or underestimate risk, and the identification of major sources
contributing to potential exposure.
It appears that one of the major factors influencing the results at the highest
portion of the range derives from the fact that, for a few individual OP's, risk
assessments and mitigation actions have not been finalized. This is particularly true
for DDVP and dimethoate. The Agency expects to complete these risk
assessments and possible mitigation actions very soon.
Finally, it is important to remember that portions of this document are currently
under review by the FIFRA SAP. For instance, EPA intends to present preliminary
results of cumulative risk using two additional models-CARES and Lifeline™-to the
SAP during the June 2002 meeting. EPA will evaluate SAP's comments, as well as
other comments or data that it receives, and will modify this assessment, as
appropriate. In addition, as existing analyses are revised or new information is
obtained, EPA will review this assessment and will make further changes as
appropriate.
Executive Summary Page 10 of 11
-------
E. Conclusion
This scientific assessment of OP pesticide food safety contains good news for
American consumers. Regulatory actions taken over the last six years have
considerably reduced the risks posed to Americans from OP residues that may be
found in food, drinking water, and in and around the home. After years of rigorous
scientific work, the Revised OP Cumulative Risk Assessment strongly supports our
confidence that the United States has one of the safest food supplies in the world.
F. Road Map
The Revised OP Cumulative Risk Assessment is divided into three parts: (1) the
actual risk assessment which draws on the regional risk assessments and the
supporting toxicology analyses (I. Revised OP Cumulative Risk Assessment); (2) the
seven regional risk assessments (II. Revised Regional Assessments); and (3) the
detailed toxicology analyses such as the derivation of the RPF's and how the FQPA
Safety Factors were determined (III/ Appendices).
Executive Summary Page 11 of 11
-------
Glossary
BMD10 is a Benchmark Dose associated with a 10% response adjusted for background.
Benchmark Response (BMR) is a designated level or percent of response relative to
the control level of response used in calculating a BMD
Common Mechanism of Toxicity pertains to two or more pesticide chemicals or other
substances that cause a common toxic effect(s) by the same, or essentially the same,
sequence of major biochemical events (i.e., interpreted as mode of action).
Comparative effect level (CEL) is a dose by which potency of chemicals may be
compared; e.g. the dose causing a maximum of 15% cholinesterase inhibition.
Cumulative Assessment Group (CAG) is a subset of chemicals selected from a
common mechanism group for inclusion in a refined quantitative estimate of risk.
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.
Dose additivity is the Agency's assumption when evaluating the joint risk of chemicals
that are lexicologically similar and act at the same target site. In other words, it is
assumed that each chemical behaves as a concentration or dilution of every other
chemical in the CAG (or chemical mixture). The response of the combination is the
response expected from the equivalent dose of an index chemical. The equivalent dose
is the sum of the component doses, scaled by each chemical's toxic potency relative to
the index chemical.
Index chemical is a chemical used as the point of reference for standardizing the
common toxicity of the chemical members of the CAG.
Lowest-Observed-Adverse-Effect Level (LOAEL) is the lowest dose in a toxicity
study resulting in adverse health effects
No-Observed-Adverse-Effect Level (NOAEL) is the highest dose in a toxicity study
which does not result in adverse health effects
OPCumRisk is a computer program developed at ORD's NHEERL to determine relative
potency estimates and PoDs for the index chemical.
Pathway of Exposure is the physical course a pesticide takes from the source to the
organism exposed (e.g., through food or drinking water consumption or residential
pesticide uses).
Page 44
-------
Point of Departure (PoD) is a dose that can be considered to be in the range of
observed responses, without significant extrapolation. A PoD can be a data point or an
estimated point that is derived from observed dose-response data. A PoD is used to
mark the beginning of extrapolation to determine risk associated with lower
environmentally relevant human exposures.
Relative Potency Factor (RPF) is the ratio of the toxic potency of a given chemical to
that of an index chemical in the CAG. Relative potency factors are used to convert
exposures of all chemicals in the CAG into their exposure equivalents of the index
chemical.
Route of Exposure is the way a chemical enters an organism after contact (e.g.,
ingestion, inhalation, or dermal absorption).
Steady state inhibition is the time point at which continued dosing at the same level
results in no further increase in cholinesterase inhibition.
Page 45
-------
Acronyms
A Estimate of A (background cholinesterase activity)
AChE Acetylcholinesterase
B Estimate of B (horizontal-asympote from July 2001 analysis)
B/A Ratio of estimate of B/estimate of A
BMD10 A Benchmark Dose associated with a 10% response adjusted for
background
BMDL Lower 95% confidence limit on the BMD10
BMR Benchmark Response -a designated level or percent of response relative
to the control level of response used in calculating a BMD
CEL Comparative effect level - Dose level used to compare potencies
ChEl Cholinesterase inhibition
CL Confidence limit
CNS Central nervous system
D Displacement parameter in expanded model
DER Data evaluation record, a review of a toxicity study
F Female
FIFRA Federal Insecticide, Fungicide, and Rodenticide Act
FQPA Food Quality Protection Act
GOF Model goodness-of-fit
HED Health Effects Division
/dose Scaled internal dose
m Estimate of absolute potency for a single cholinesterase measurement in
.the July 2001 analysis
/A Log of background cholinesterase activity
/m Log slope-scale factor
M Male
MOE Margin of exposure
MRID # Study identification number
NA Not available
NERL National Exposure Research Laboratory
NHEERL National Health and Environmental Effects Laboratory
nlme Non-linear mixed effects model
NOAEL No-Observed-Adverse-Effect Level - the highest dose in a toxicity study
which does not result in adverse health effects
OP Organophosphorous pesticide
OPCumRisk Computer program developed at ORD's NHEERL to determine relative
potency estimates and PoDs for the index chemical.
OPP Office of Pesticide Programs
OPPTS Office of Prevention, Pesticides, and Toxic Substances
ORD Office of Research and Development
PB Limiting value of minimum cholinesterase activity (horizontal asymptote)
PBF Female specific value of PB
PBM Male specific value of PB
PBPK Physiologically Based Pharmacokinetics
POD Point of Departure
Page 46
-------
PNS Peripheral nervous system
RBC Red blood cells
RfD Reference Dose - A dose not expected to cause adverse health effects in
humans
RPF Relative Potency Factor
S Shape
SAP Scientific Advisory Panel
tB Transformed horizontal asymptote
Page 47
-------
Parti.
Revised OP Cumulative
Risk Assessment
-------
I I. Revised OP Cumulative Risk Assessment
i
I A. Introduction
I The Food Quality Protection Act of 1996 significantly amended the Federal
I Insecticide, Fungicide, and Rodenticide Act (FIFRA) and the Federal Food, Drug,
= and Cosmetic Act (FFDCA). One of the major changes is the requirement that EPA
| consider risk posed by pesticides acting by common mechanism of toxicity.1 For
| such groups of pesticides, EPA's Office of Pesticide Programs has treated
I cumulative risk, under FQPA, as the risk of a common toxic effect associated with
I concurrent exposure by all relevant pathways and routes.
jfWWW £
tl I were:
Since the enactment of FQPA, EPA's Office of Pesticides Programs (OPP) has
been working to develop new methodologies in a number of risk assessment areas.
The steps necessary to complete the Revised OP Cumulative Risk Assessment
Q Development of approaches for grouping chemicals by a common mechanism of
toxicity (USEPA, 1999a) and
Q Conducting aggregate (USEPA 1999c and 2001 d) and cumulative risk
assessments (USEPA 2000a and 2001 a)
At each major step in development OPP consulted with the FIFRA Science
Advisory Panel (SAP) to seek expert review, advice, and recommendations. We
held several external peer review meetings with the SAP and asked for comment on
our approaches to grouping chemicals based on common mechanism of toxicity,
grouping chemicals for the purpose of cumulative assessment, improved methods
for exposure assessment, approaches to aggregating food, drinking water and
residential exposure and proposed models for combining these exposures. We also
held several meetings with the FQPA Federal Advisory Committee Act (FACA)
Groups of stakeholders (public interest groups, state agricultural agencies, pesticide
industry representatives, growers, USDA and others) to present our methodologies
as they were developed, and to seek comments and recommendations. All of the
new science policies which are a foundation of this assessment were proposed for
public comment. The work to develop the methodology was completed with the
publication of the Revised Guidance Document for Cumulative Risk Assessment
USEPA 2001 a).- The document was proposed for public comment on June 30, 2000
(65 FJR 127:40644-40650). The SAP and public comments were reconsidered and
the Guidance was revised in December, 2001. All of these documents played roles
in preparation at the Preliminary OP Cumulative Risk Assessment which was issued
'For details see The Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA), 7U.S.C. §§ 136 et
sea., and Federal Food, Drug, and Cosmetic Act (FFDCA) 21 U.S.C. § 346a.
LA Page 1
-------
I in December 2001 and presented to the public in a technical briefing in January
I 2002. The assessment was reviewed by the SAP in February 2002. Public
I comment was also solicited for ninety days. The comment period closed in March
I 2002.
:-yysN
OPP proceeded with the methodology and risk assessment development in a
step-by-step process. The approach to the risk assessment was evaluated using a
case study of three organophosphorus pesticides. That assessment was reviewed
by the SAP (September and December 1999 see FIFRA SAP 2000a,b), who
•f-v,, | recommended that OPP proceed with a more comprehensive case study. OPP
developed the hazard, dose-response and exposure assessment for 24 OP
pesticides and brought it to the SAP for comment in September and December of
2000 (FIFRA SAP 2001 a,b). Based on the comments, the hazard and dose-
response assessment was revised and again reviewed by the SAP in September of
2001 (FIFRA SAP 2001 c). The SAP was very supportive of the approach, calling it
both 'skillful' and 'creative.' The recommendations made by the September 2001
SAP were addressed in the Preliminary Risk Assessment. In February 2002, OPP
brought the Preliminary OP Cumulative Risk Assessment to the SAP. Once again
the panel was generally supportive of the preliminary document while suggesting
some revisions. These revisions have been incorporated in this current document,
the Revised OP Cumulative Risk Assessment.
j£ | Cumulative Risk Assessment is a complex analysis and OPP needs to
C/) 1 emphasize that the results are not a collection of numbers or bright lines.
^7 i Quantitative methods have been used throughout the analysis but the results need
wL, | t0 be interpreted with a full understanding of the assumptions made and the
Cj) | uncertainties introduced by making these assumptions. As the regulatory managers
> I and the stakeholders look for guidance in reading the document, it is especially
*4p I important to consider the Risk Characterization Chapter of the Risk Assessment.
•.w.%^1 —
I.A Page 2
-------
I. Revised OP Cumulative Risk Assessment
B. Hazard/RPF
1. Introduction
Since the passage of the FQPA, the Office of Pesticide Programs (OPP)
has presented proposed guidance, tools and methodologies for conducting
cumulative risk assessments to the FIFRA Scientific Advisory Panel (SAP).
Specifically, the hazard and dose-response sections have been presented to the
FIFRA SAP four times between 1999 and 2002 including the February 5-8, 2002
meeting on the methods used in the Preliminary Cumulative Risk Assessment
(PCRA) of the Organophosphorus Pesticides (FIFRA SAP, 2000a, 2001 a,
2001 b, 2002a). Following the previous SAP reviews, constructive comments and
recommendations have been incorporated into revisions and refinements of the
hazard and dose-response assessment for the Organophosphorus pesticides
(OPs). Key recommendations from SAP reports have included the utilization of
the exponential model for fitting the cholinesterase data, the derivation of relative
potencies from several relatively consistent studies rather than a single study,
and further exploration of low dose modeling issues. In collaboration with EPA's
Office of Research and Development (ORD) National Health and Environmental
Effects Research Laboratory (NHEERL), OPP released a Preliminary Dose-
Response Assessment for OPs on July 31, 2001 (USEPA 2001 b) followed by a
revised dose-response assessment on December 3, 2001. At the February 5-6,
2002 meeting, the SAP was very supportive of the approach used in the PCRA
for OPs. The panel commended the Agency for its progress in modeling of
dose-response relationships of OP exposure to cholinesterase inhibition. The
panel indicated that remaining issues concerning cumulative hazard and dose-
response assessment reflect the evolving nature of the field and do not need to
be specifically addressed in the cumulative risk assessment of OPs.
Revised relative potency factors (RPFs) for 33 OPs were released to the
public on April 17, 2002
(http://www.epa.gov/pesticides/cumulative/pra-op/rpf final.htm). EPA has
calculated RPFs for four OPs not included in the hazard and dose-response
assessment of the PCRA: chlorethoxyphos, omethoate, phostebupirim, and
profenofos. After issuing its PCRA, the Agency identified computer programming
errors in its statistical modeling procedure. EPA discussed them at the February
5-8, 2002 meeting of the FIFRA SAP. These errors impacted the curve-fitting
procedure for some OPs. In addition, EPA received additional toxicology data
for disulfoton, fenamiphos, phosalone, tetrachlorvinphos, and tribufos, which
were used in the revision of the RPFs for these chemicals.
I.B Page 1
-------
2. Methods
a. Overview
Before the cumulative risk of exposure to OPs can be quantified, the
relative toxic potency of each OP must be determined. The determination of
relative toxic potency should be calculated using a uniform basis of
comparison, by using, to the extent possible, a common response derived
from the comparable measurement methodology, species, and sex for all the
exposure routes of interest (USEPA 2001 a, 2002).
b. Endpoints and Toxicology Studies
i. Selection of Endpoints
As part of the hazard analysis, all relevant responses were
evaluated to identify the most appropriate endpoint pertaining to the
common mechanism of toxicity and to determine which endpoint(s)
provide(s) a uniform and common basis for determining the relative
potency of the cumulative assessment group. OPs exert their
neurotoxicity by binding to and phosphorylating of the enzyme
acetylcholinesterase in both the central (brain) and peripheral nervous
systems (Mileson et a/., 1998). There are laboratory animal data on
OPs for cholinesterase activity in plasma, red blood cell (RBC) and
brain, as well as behavioral or functional neurological effects in
submitted guideline studies. Measures of acetylcholinesterase
inhibition in the peripheral nervous system (PNS) are very limited for
the OP pesticides. As a matter of science policy, blood cholinesterase
data (plasma and RBC) are considered appropriate surrogate
measures of potential effects on PNS acetylcholinesterase activity and
of potential effects on the central nervous system (CNS) when brain
cholinesterase data are lacking (USEPA, 2000c). Behavioral changes
in animal studies usually occur at higher doses compared to doses
needed to inhibit cholinesterase. Also, behavioral measures are
limited in terms of the scope of effects assessed and the
measurements employed. Plasma. RBC. and brain cholinesterase
inhibition were considered potential endpoints for extrapolating
risk to humans in the OP cumulative risk assessment.
ii. Selection of Routes and Duration of Exposure for Potency
Determination
Humans may be exposed to the OPs through diet, in and around
residences, schools, commercial buildings, etc. Therefore, the
potency of OPs needs to be determined for the oral, dermal, and
inhalation routes of exposure. Cholinesterase inhibition can result for
single or short-term exposures. The Revised Cumulative Risk
I.B Page 2
-------
Assessment for OPs (RCRA) has evaluated both single-day and
multiple sequential days (i.e., 7-day rolling average) exposures for
integrating multiple sources of OPs.
Various toxicokinetic and toxicodynamic factors influence an
individual OP's time to peak effect of inhibition, persistence of action
following acute exposure, and the duration of exposure required to
reach steady state inhibition. OPP has elected to estimate relative
potencies and points of departure (POD) using measurements where
cholinesterase inhibition in the laboratory animal is not changing with
time. OPP defines this point where continued dosing at the same level
results in no further increase in enzyme inhibition as steady state. The
use of cholinesterase data for single-dose or short duration studies to
model the comparative potency is problematic because the extent of
inhibition is rapidly changing immediately following dosing. Measures
x*>v ••
f^; I of cholinesterase taken during this time will be highly variable and
| iii. Toxicity Database
.- .V.WJV S
> I
fl) I As stated previously, relative potency should be based whenever
possible on data from the same species and sex to provide a uniform
measure of relative potency among the cumulative assessment group
(USEPA, 2002). Under FIFRA, toxicology studies in various species
(e.g., dog, mouse, rat and rabbit) are submitted to OPP. For the OP's,
I.B Page 3
-------
I toxicology studies in the rat provided the most extensive
I cholinesterase activity data for all routes, compartments, and both
i sexes. Thus, the focus of this analysis was on cholinesterase
I activity data derived from male and (non-pregnant) female rats.
EPA used rabbit studies for five chemicals with
residential/nonoccupational exposure potential because dermal toxicity
data in rats were not available. The cholinesterase data considered in
this analysis were extracted from the study types listed in Table I.B-1.
Studies used in this analysis were identified by their source MRID
number. Studies submitted to OPP are reviewed for their quality of
cholinesterase measurements and consistency of their experimental
protocol with the OPPTS Guidelines
(http://www.epa.gov/opptsfrs/home/guidelin.htm).
When assessing cholinesterase activity, it is important to carefully
consider methodological issues that may affect the accuracy and
variability of the data. There are many methods available for
measuring cholinesterase activity. These methods include colorimetric,
electrometric, titrimetric, radiometric, fluorimetric, gasometric, and
immunochemical assays. The colorimetric method, based on the
Ellman reaction, is the most commonly used method for measuring
brain, plasma and erythrocyte cholinesterase activity (Ellman et al.,
1961; USEPA 1992; ASCP, 1994). For this preliminary assessment, if
the Data Evaluation Record (DER) for a particular study indicated that
the study was acceptable, it was assumed that the methodology was
also acceptable.
A comprehensive list of all the studies utilized in the present
analysis is given in Appendix III.B.2. The cholinesterase data are
available to the public at http://www.epa.gov/pesticides/cumulative/.
\ if =
V.-X- =
:
•X'l^'X^;; Z
WX s
ffi S
VV :
I.B Page 4
-------
Table I.B-1. Test Guideline Studies Evaluated for Cholinesterase Activity.
Test Guideline Studies Evaluated for ChoJinesterase Activity
Study Type
Guideline Type
Oral
90-day oral toxicity study in rat
Chronic oral toxicity in rat
Carcinogenicity in rat
Combined chronic toxicity/carcinogenicity in rat
Subchronic neurotoxicity study in rat
Range finding oral toxicity study in rat
Other —Special studies
OPPTS 870.31 00
OPP 82-1
OPPTS 870.41 00
OPP 83-1
OPPTS 870.4200
OPP 83-2
OPPTS 870.4300
OPP 83-5
OPPTS 870.6200
OPP 82-7
Not applicable
Not applicable
Dermal
21/28-Day dermal toxicity in rat or rabbit
90-Day dermal toxicity in rat
OPPTS 870.3200
OPP 82-2
OPPTS 870.3200
OPP 82-2
Inhalation
90-Day inhalation toxicity in rat
21/28-Day inhalation toxicity in rat
Inhalation carcinogenicity in rat
OPPTS 870.3465
OPP 82-4
OPPTS 870.3465
OPP 82-4
OPPTS 870.3320
OPP 83-5
T
$"">
.^•\
*«**
c. Collection of Cholinesterase Activity Data
i. Oral Route
Oral relative potency values were needed for all OP pesticides
included in the RCRA because of potential dietary exposures from food
and drinking water and hand to mouth exposures associated with
residential/nonoccupational uses. Numerous oral studies with
comparable methodologies were available and suitable for quantitative
dose-response analysis. An electronic spreadsheet is needed to perform
quantitative dose-response analysis. Study type, duration of exposure,
I.B Page 5
-------
number of animals per dose group, sex, compartment, and the measured
effect for each dose group (mean cholinesterase activity, activity units,
and standard deviation) were compiled into an electronic spreadsheet. In
feeding studies/average compound intake (mg/kg/day) over the entire
study was used. At least one oral toxicity study of the appropriate
duration was available for all the OPs. Time of measurement was
expressed as number of days on study where: number of days = number
weeks x 7 and number of days = number months x 30.
ii. Dermal and Inhalation Route
Relative potency factors were needed for 10 OPs with residential
exposure. Unlike the database of oral toxicity studies, the database of
dermal and inhalation studies with cholinesterase measurements is
limited. However, using the CEL approach is adequate for the RCRA.
Comparative effect levels (CELs) have been used to compare the relative
potency of the OPs. CELs are dose levels from a given study with a
defined range of effects. The CEL was defined as the dose causing a
maximum of 15% brain cholinesterase inhibition. Quantitative dose-
response analysis for estimating a common benchmark response is the
preferred method for determining relative potency.
d. Selection of Relative Potency Factors for the Female Brain
Cholinesterase Data Set
A key component of cumulative hazard assessment is to select an
endpoint pertinent to the common mechanism of toxicity that can be used
to quantify cumulative risk. In the July 2001 dose-response assessment,
OPP prepared a dose-response analysis for 25 OPs in which a large body
of toxicity data on a common mechanism endpoint for these OPs - the
ability to inhibit cholinesterase in plasma, RBC, and brain - was analyzed.
To determine which compartment would provide a strong basis for
determination of relative potency, OPP reviewed data in each
compartment. In the July 2001 analysis, RPFs based on the male RBC
database were proposed. It was stated in that document that the RBC
RPFs proved to be a reliable and sensitive endpoint considered protective
of both the peripheral and central nervous systems for the majority of the
chemicals. The major advantage of the RBC database was its large size
compared to the whole brain ChE database; this large database allowed
the examination of time course information and observation of a steady
state response.
After considering the comments from the September 2001 SAP
meeting in addition to the comments from the public and stakeholder
groups, OPP has:decided to use female brain ChE data for quantifying
cumulative risk for OPs. OPP has decided to estimate cumulative risk
based on RPFs and PODs from the female brain ChE database for
I.B Page 6
-------
several reasons. Principally, compared to relative potency estimates
based on RBC, estimates of relative potency based on brain ChE have
tighter confidence intervals and therefore will confer less uncertainty on
cumulative risk estimates. Also, these data represent a direct measure of
the common mechanism of toxicity as opposed to using surrogate
measures. The toxic potencies and PODs for brain cholinesterase
inhibition for these OPs are generally similar to the RBC data for the oral,
inhalation, and dermal exposures (USEPA, 2001 b). The SAP
recommended the Agency address the issue of repeated measures in its
revised analysis. This issue concerning repeated cholinesterase activity
measures only pertains to the plasma and RBC ChE data because blood
can be collected several times from a single animal, whereas brain ChE
can only be collected once. Finally, in the present analysis, although male
and female rats were equally sensitive for 30 OPs, female rats were more
sensitive to three OPs. Therefore, OPP has chosen to based its RPFs on
female brain measurements.
In the RCRA, potency estimates have been recalculated only from the
brain ChE database. The plasma and RBC databases were thoroughly
examined in the July 2001 analysis (USEPA, 2001 b). Re-analysis of the
plasma and RBC databases using the revised methodology is unlikely to
significantly change potency estimates from these compartments
(USEPA, 2001 c).
e. Determination of Chemical Potency: Oral Route
In their review of the September, 2000 pilot analysis, the SAP
suggested that EPA consider Michaelis-Menton kinetics or the exponential
model to fit cholinesterase data from OPs (FIFRA SAP, 2001 a).
Preliminary simulations using a subset of studies (one study per 24
chemicals) were performed using both the rectangular hyperbola (i.e.,
Michaelis-Mention kinetics) and the exponential function. The exponential
model was selected over the rectangular hyperbola based on statistical
criteria such as goodness of fit (USEPA, 2001 b). Based on the results
presented to the SAP in September, 2001, the panel indicated that no
alternative to the exponential model would be more appropriate at the
present time (FIFRA SAP 2001 b).
I.B Page?
-------
i. Exponential Equations Used To Determine Potency
The simplified and general exponential equation used for modeling the
effect of the OPs on cholinesterase activity is:
Equation I.B-1a
y = A
PB+(l-PB)Q
-mxDose
where y is cholinesterase activity,
Dose is the dose level of the OP, in mg/kg/day,
A is the background (similar to control) ChE activity,
m is the slope-scale factor,
and PB is the horizontal asymptote (i.e., limiting value of
minimum cholinesterase activity), expressed as a
fraction of the background activity.
Both y (cholinesterase activity) and dose were extracted from the oral
toxicity studies. Pe expresses the horizontal-asymptote as a fraction of
background cholinesterase activity. PB does not have any units. As
described in detail in the technical appendix (III.B.1), Equation I.B-1a was
reparameterized to Equation I.B-1b, where benchmark dose is an explicit
parameter, to simplify the statistical calculations.
I.B Page8
-------
Equation I.B-lb
y = A
BMD
where A is the level of cholinesterase activity in the
absence of exposure to organophosphate (i.e., control),
PB is the fraction of cholinesterase activity remaining at a
very high dose of organophosphate,
BMR is the level of inhibition at which to estimate the
benchmark dose (in this study, BMR is always 0.10),
BMD is the benchmark dose, and
Dose is the dose of organophosphate pesticide,
generally in units of mg/kg/day.
I.B Page 9
-------
Figure I.B-1. Plot of basic equation.
o
o
OH
Csl
O
O
<
LU
.c
O
<
0
Background ChE
!>lope scale factor
= A[pB+(i-PB)e
mxDose
Fraction of Activity
at High Dose
500 1000
Dose
1500
I.B Page 10
-------
The exponential function in Equation I.B-1a/b decreases in a linear
fashion in the low dose region (Figure I.B-1). Considerable discussion at
the August 2001 Technical Briefing and the SAP meeting of September 5-
6, 2001 centered around the potential for a flat region in the low dose
portion of the dose-response curve. This potential low-dose flat region
was explored and a revised equation was developed. This revised
equation is a modified version of the exponential function in Equation I.B-
1 b which includes two additional variables, S (shape) and D
(displacement). S and D together describe a low-dose flat region of the
dose-response curve (Figure I.B-2). The second equation results from
combining Equation I.B-1b with an equation which describes the
relationship between administered dose and calculated internal dose
(Equation I.B-2). The value /dose replaces Dose in Equation I.B-1b. The
SAP called this revised equation "elegantly simple" and pointed out that
the equation improved fit for many OPs with little response at low dose
levels. For ease of discussion, Equation I.B-1b will be called the 'basic'
model (low dose linear) and Equation I.B-2 will be called the "expanded"
model (low dose flat).
Equation I.B-2
idose = g ( Dose; S, D ) = 0.5 T ( Dose - S - D ) + ^(Dose -S-D)2 + 4xDosexS 1
where idose is the scaled internal dose,
Dose is the administered dose level (mg/kg/day),
S controls the low-dose shape of the curve, and
D controls the ultimate horizontal displacement of the
curve relative to the identity line (i.e., the line with idose =
Dose).
I.B Page 11
-------
As shown in Figure I.B-2, for the basic model, the low dose region
decreases in a linear fashion. For the expanded model, the low-dose end
of the dose-response curve has a more shallow slope (more flat). As S
grows small, or D grows large, the estimated benchmark dose increases
in magnitude. As S grows large, or D approaches 0, the relationship
between /dose and Dose approaches the line /dose = Dose. In other
words, as S increases or D decreases, the shape of the expanded
equation approaches the shape of the basic equation. The technical
discussion of the expanded model and its derivation are described in
more detail in Appendix IN.B.1.
.V.VfV
I.B Page 12
-------
I Figure I.B-2 shows the relationship between the basic and expanded models and also
| how the shape and displacement variables impact the dose-response curve.
^
o i
\-iWJ*
X-^v« |
iw-v-:
2000 -
1500 -
1
<
LJJ
o
1000 -
500 -
0 -
- 10
- 8
- 6
- 4
- 2
- 0
0
Dose
o>
CO
I
to
o
CO
| Figure I.B-2. Basic and expanded equations. The black solid curve is the basic
| equation of Equation I.B-2 with A = 2000, PB = 0.15, and m = 1. The colored solid
| curves show the results of the expanded equation with 3 different values of S and
| D=2. The dotted curves shows the relationship between /dose (blue, purple, and
| red) and Dose (black).
I.B Page 13
-------
ii. Joint Analysis of OP Cholinesterase Data
In the joint dose-response analysis, the Cholinesterase data for various
time points for a specific chemical are modeled together all at once. For
example, there are five measurements of female rat brain Cholinesterase
following exposure to methamidophos. All five datasets were analyzed
together to determine the benchmark dose (although studies are plotted
separately.in Appendix III.B.2). This approach allows information about
the shape of the dose-response curve to be "shared" among individual
studies and results in benchmark dose estimates which are representative
of a given OP. To perform the joint analysis of all the datasets for each
chemical, several aspects of the data need to be accommodated . First,
measurements of Cholinesterase activities can have different units (mainly
U/G, U/L, and ApH), which need to be accommodated in the same
analysis. Model parameters may also differ between males and females.
Finally, it is likely that model parameters vary randomly among studies
and within a study. When more than measurement of brain
Cholinesterase was available, the approach to nonlinear mixed effects
(nlme) modeling developed in Lindstrom and Bates (1990) was used to fit
the Cholinesterase data to Equations I.B-1b and I.B-2. Only one
measurement of brain Cholinesterase was available for four OPs; for
these OPs generalized least squares (gnls) was used to fit the
Cholinesterase data. Profile likelihood plots were used to estimate the
horizontal asymptotes, shape, and displacement parameters. All
estimated parameters, including the shape and displacement parameters,
were estimated separately for each OP and vary among OPs. The
technical statistical methodology used to fit the Cholinesterase data to the
exponential model are not discussed here. The statistical methodology
are discussed in detail in Appendix III.B.1.
Thirty-two OPs were fit to both the basic and expanded models. In
cases where the expanded model resulted in a significantly improved fit (P
< 0.05), the expanded model was used to estimate potency. The basic
model was used to estimate the potency of the remaining OPs.
Omethoate was modeled using only the basic model. At the time of public
release for the revised RPFs only one measurement of brain
Cholinesterase in female rats with the appropriate duration of exposure
was available for omethoate. In this dataset, all treatment groups
exhibited reduced brain ChE activity compared to the control. Three other
OPs have one dataset for female rat brain Cholinesterase inhibition was
available. For only one of these, dichlorvos, reduced Cholinesterase
activity was observed at all treatment groups. The expanded model did
not improve the fit for dichlorvos; the basic model was used to estimate
the potency. In addition, the potency of dimethoate, the parent chemical,
of omethoate was estimated using the basic model. At this time, it is
reasonable to assume that the expanded model would not improve the fit
for omethoate.
I.B Page 14
-------
**»»
iU I
iii. Use of BMD10 for Relative Potency Determination
Potency determinations of the OPs for the oral route exposure are
based on the benchmark dose where cholinesterase activity is reduced
10% compared to background activity (BMD10). The BMD10 was selected
as the effect level for potency determination because this level is
generally at or near the limit of sensitivity for discerning a statistically
significant decrease in cholinesterase activity across the blood and brain
compartments and is a response level close to the background
cholinesterase.
At the February 5-8, 2002 meeting of the FIFRA SAP some members
of the panel in addition to some public commenters discussed the
Agency's selection of the BMD10 as the benchmark response level. In
response to this discussion, the Agency analyzed the detection limits of
the studies assessing female brain cholinesterase levels used in the
RCRA of the OPs. This analysis has shown that generally these studies
can reliably detect around 10% cholinesterase inhibition, that such levels
were generally achieved in the studies, and that therefore, the Agency's
use of the BMD10 as the benchmark response is appropriate. This
analysis is described in detail in Appendix III.B.3
iv. Software Used in Oral Potency Determination
The programming code in R-language used to develop the relative
potency factors and the PODs for the index chemical in the current
analysis has been included in Appendix III.B.4.
In the July 2001 dose-response analysis, a computer program,
OPCumRisk, was used to determine relative potency estimates and PODs
I for the index chemical. OPCumRisk was developed at ORD's NHEERL
I specifically for use in the July 2001 OP dose-response assessment and is
| available at http://www.epa.gov/scipolv/sap/index.htm. OPCumRisk is
I written in R (lhaka and Gentleman, 1996), a freely distributable
| implementation of the S programming language available for download on
I the internet at http://www.R-project.org. Minor revisions recommended by
I the SAP have been incorporated into the OPCumRisk program (See
I Appendix III.B.3). The statistical methodology used in the present
| document has not been incorporated into the OPCumRisk program.
I.B Page 15
-------
f. Determination of Chemical Potency: Dermal Route
Chemical potency was determined using CELs for the dermal route of
exposure. These CELs are experimentaldose levels which elicit a similar
toxicological response to the selected endpoint.
O Cholinesterase activity data were collected from dermal toxicity studies for
nine chemicals with residential/nonoccupational exposure and the index
chemical (methamidophos). Five OPs were tested by the dermal route in
rats. Only rabbit studies were available for the other five OPs. Thus, it was
not possible to compare Cholinesterase activity data from dermal studies in
only one species. Of the chemicals with potential dermal exposure, only .
three chemicals (acephate, disulfoton, and naled) had more than one dermal
toxicology study which could be used for assessing relative potency. One
chemical, dichlorvos, had no dermal exposure study. The requirement for a
dermal toxicity study with dichlorvos was waived because the volatility of the
chemical renders it technically difficult to conduct such a study.
Relative potencies of the chemicals with residential/non-occupational uses
were determined by using CELs derived from data on inhibition of
Cholinesterase activity in female rat brain. The CEL was defined as the
lowest dose where a maximum 15% brain Cholinesterase inhibition
(compared to control) occurred.
g. Determination of Chemical Potency: Inhalation Route
Chemical potency was determined using CELs for brain Cholinesterase
activity for the inhalation route of exposure. Cholinesterase activity data were
collected from inhalation toxicity studies for seven chemicals with
residential/nonoccupational exposure and the index chemical
(methamidophos). Two inhalation exposure studies were available for
acephate whereas only one suitable study was available for the other OPs.
Although all of the inhalation studies were performed with the same species
(rat), four different strains of rats were used. Furthermore, the exposure
conditions varied among the chemicals tested. There were four whole-body
exposure studies, one head-nose, and three nose only exposure studies. No
inhalation toxicity study was available for three chemicals, bensulide,
fenthion, and tetrachlorvinphos.
Relative potency was calculated from CELs for brain Cholinesterase
activity determined from inhalation toxicity studies. The CEL was defined as
the lowest dose where a maximal response [brain Cholinesterase inhibition] of
15% (compared to control) occurred.
I.B Page 16
-------
h. Selection of the Index Chemical (Methamidophos)
The cumulative risk assessment guidance document (USEPA, 2002)
states that the index chemical should be selected based on the availability of
high quality dose-response data for the common mechanism endpoint and
that it acts toxicologically similar to other members of the common
mechanism group. High quality dose-response data allows the calculation of
points of departure (POD) for oral, dermal, and inhalation exposures with
confidence. A POD is a point estimate on the index chemical's dose-
response curve that is used to extrapolate risk to the exposure levels
anticipated in the human population. Thus, any error or uncertainty in an
index chemical's POD value will be carried forward in the cumulative risk
estimates. For the cholinesterase inhibiting OP pesticides, the ideal index
chemical should exhibit high quality dose-response data in plasma, RBC, and
brain for both sexes of a single species for all exposure routes of interest.
In the July 2001 dose-response assessment, methamidophos was
selected as the index chemical for the OPs. The selection criteria and the
potential candidates for the index chemical were discussed in detail in the
July, 2001 document (USEPA 2001 b). Methamidophos remains the index
chemical for the RCRA OPs because this chemical has a high quality
database for the common mechanism endpoint of inhibition of
acetylcholinesterase for the oral, dermal, and inhalation routes of
exposure.
i. Points of Departure (POD)
The oral, dermal, and inhalation PODs for the index chemical are based
on the benchmark dose where cholinesterase activity is reduced 10%
compared to background activity (BMD10). The BMD10 was selected as the
effect level for the POD because this level is generally at or near the limit of
sensitivity for discerning a statistically significant decrease in cholinesterase
activity across the blood and brain compartments and is a response level
close to the background cholinesterase.
I.B Page 17
-------
j. Calculation of Relative Potency Factors (RPF)
Oral RPFs were calculated from oral BMD10s for female brain
cholinesterase activity by the Equation I.B-3.
"--.vv.
Oral RPF Chemical x = BMD10 lndex Chemical / BMD10 Chemjcal x
Equation I.B-3
where BMD10 Chemicaix is the BMD10 for Chemical X
and BMD10 |ndexChemicalisthe BMD10of the index chemical.
CELs for brain cholinesterase activity measured in dermal studies were
determined in order to calculate RPFs. Dermal RPFs were calculated using
Equation I.B-4.
Dermal RPF Chemica| x = CEL |ndex Chemicai / CEL Chemicai x
Equation I.B-4
.v.vX*
CELs for brain cholinesterase activity measured in inhalation studies were
determined in order to calculated RPFs. Inhalation RPFs were calculated
using Equation I.B-5.
Inhalation RPF Chemical x = CEL lndexChemicai/ Chemlcalx
Equation I.B-5
I.B Page 18
-------
3. Results
a. Dose-Response Modeling: Oral Route of Exposure
The joint analysis using the exponential model served as good method for
determining potency and provided confident estimates of the benchmark
dose. The exponential model fits the cholinesterase data well. Plots of dose-
response data, residuals, and profile likelihoods for all 33 OPs are given in
Appendix III.B.2. BMD10s and RPFs for the OPs are listed below.
i. Basic vs. Expanded Models
A joint analysis using the basic (low dose linear) and/or the expanded
(low dose flat) equations of brain cholinesterase data for OPs was
performed. The potency of 17 pesticides listed in Table I.B-2 were
determined with the expanded model. The expanded model fit was
significantly improved; i.e., the P-value of the likelihood test for the
expanded model was <0.05 for all 17 chemicals. The potency of the
remaining 16 were determined with the basic model.
Table I.B-3 shows the dose-response model parameters for the
horizontal asymptote (Pe), shape (S), and displacement (D) parameters
for each OP. These parameters vary among OPs.
I.B Page 19
-------
Table I.B-2. Listing of OPs which were modeled with basic and expanded models
Listing of OPs which were modeled with basic and expanded models.
Chemical
Acephate
Azinphos-methyl
Bensulide
Chlorethoxyfos
Chlorpyrifos
Chlorpyriphos-methyl
Diazinon
Dichlorvos
Dicrotophos
Dimethoate
Disulfoton
Ethoprop
Fenamiphos
Fenthion
Fosthiazate
Malathion
Methamidophos
Methidathion
Methyl-parathion
Mevinphos
Naled
Omethoate
Oxydemeton-methyl
Ph orate
Phosalone
Phosmet
Phostebupirim
Pirimiphos-methyl
Profenofos
Terbufos
Tetrachlorvinphos
Tribufos
Trichlorfon
Expanded vs, Basic
Basic
Expanded
Expanded
Expanded
Expanded
Basic
Expanded
Basic
Basic
Basic
Expanded
Basic
Basic
Basic
Expanded
Expanded
Basic
Basic
Expanded
Expanded
Basic
Basic
Basic
Expanded
Expanded
Expanded
Expanded
Basic
Basic
Expanded
Basic
Expanded
Expanded
P value for the Improvement
Ir* Model Fit for
Expanded vs> Basic
' 0.999
3.04E-21
0.0002
7.05E-24
1.88E-13
0.96
8.05E-21
0.77
0.998
0.81
2.06E-10
0.78
0.46
0.998
2.73E-09
9.29E-13
0.17
0.86
1 .03E-07
0.0001
0.62
NA
0.9996
4.23E-28
0.01
5.20E-05
0.001
0.99997
0.9999
1.14E-32
0.39
8.79E-13
8.90E-06
I.B Page 20
-------
I Table I.B-3. Exponential model parameters for female and male brain
I cholinesterase data
Exponential model parameters for femate and male brain cholinesterase data
Chemicals
Acephate
Azinphosmethyl
Bensulide
Chlorethoxyfos
Chlorpyrifos
Chlorpyriphos-methyl
Diazinon
Dichlorvos
Dicrotophos
Dimethoate
Disulfoton
Ethoprop
Fenamiphos
Fenthion
Fosthiazate
Malathion
Methamidophos
Methidathion
Methylparathion
Mevinphos
Naled
Omethoate
Oxydemeton-methyl
Phorate
Phosalone
Phosmet
Phostebupirim
Pirimiphos-methyl
Profenofos
Terbufos
Tetrachlorvinphos
Tribufos
Trichlorfon
Displacement* (D)
--
0.597
22.066
0.603
0.764
-
18.725
-
-
-
0.043
--
-
--
11.560
1415.734
-
--
0.351
0.057
-
--
-
0.235
4.502
1.379
0.097
-
--
0.211
-
1.775
28.437
Shape b
-------
ii. Benchmark Dose Calculations
The BMD10s for brain cholinesterase measured in male and female
rats using the joint analysis procedures are listed in Table I.B-4 and
shown graphically in Figures I.B-3 and I.B-4. Among the OPs, BMD10s
range widely over approximately five orders of magnitude.
Ratios of the male to female BMD10s are plotted in Figure I.B-5. For
30 of 33 OPs the ratio is approximately one indicating that male and
female rats exhibit similar sensitivity to the OPs for brain cholinesterase
activity. For these three OPs (terbufos, tetrachlorvinphos, and trichlorfon)
the females rats were ~2- to 7-fold more sensitive compared to male rats.
I.B Page 22
-------
I Table I.B-4. Oral BMDins and BMDLs (mg/kg/day) estimated for brain ChE activity
Oral BMD1Bs And BMDLs (mg/kg/day) estimated for brain ChE activity
«
Chemical
Acephate
Azinphos-methyl
Bensulide
Chlorethoxyfos
Chlorpyrifos
Chlorpyriphos-methyl
Diazinon
Dichlorvos
Dicrotophos
Dimethoate
Disulfoton
Ethoprop
Fenamiphos
Fenthion
Fosthiazate
Malathion
Methamidophos
Methidathion
Methyl-parathion
Mevinphos
Naled
Omethoate
Oxydemeton-methyl
Phorate
Phosalone
Phosmet
Phostebupirim
Pirimiphos-methyl
Profenofos
Terbufos
Tetrachlorvinphos
Tributes
Trichlorfon
Female
BMD10
0.99
0.86
31.91
0.65
1.48
16.20
6.24
2.35
0.04
0.25
0.07
1.37
1.96
0.24
1.28
313.91
0.08
0.25
0.67
0.11
1.00
0,09
0.09
0.21
6.93
3.56
0.37
2.25
20.58
0.10
' 60.69
4.27
31.74
BMDL
0.53
0.79
30.44
0.61
1.26
4.77
2.89
1.61
0.04
0.22
0.06
0.70
0.69
0.21
0.32
221.12
0.07
0.17
0.50
0.10
0.82
0.07
0.09
0.20
6.27
2.03
0.24
1.61
17.64
0.08
20.97
3.31
28.62
Male
BMD,,
0.77
1.14
40.88
0.69
1.50
14.26
9.62
1.71
0.04
0.35
0.10
1.35
1.73
0.18
1.48
212.02
0.07
0.24
0.70
0.15
1.00
0.14
0.07
0.29
7.88
4.15
0.40
1.58
24.98
0.18
369.27
4.52
58.49
BMDL
0.41
0.98
37.11
0.62
1.27
4.21
5.39
0.08
0.03
0.31
0.09
0.69
0.63
0.15
0.38
119.31
0.06
0.16
0.51
0.13
0.82
0.12
0.07
0.26
7.05
2.25
0.26
0.93
21.86
0.17
102.31
3.47
45.39
wift^y
I.B Page 23
-------
I Figure I.B-3. BMD10s (mg/kg/day) for female brain ChE activity for 33 OPs
BMD10':s for
Female Brain ChEl Data
1000
***> =
f<.o- r
.(A
.*-
£
o>
(0 O
b C
cf 3
5 c
0. g
o
g
-------
1 Figure I.B-4. BMD10s (mg/kg/day) for male brain ChE activity for 33 OPs
BMD10'sfor
Male Brain ChEl Data
"
i
flj
O
b C
'
1000
100
10
0.1
0.01
;;;;«
^^^^^^^^EE.
<&
;:E;::::::::;;;::::E:;;;;;;;;;:::::::::::::;;;:;::;:::::::"::::::::r
,S|,
.'.':.'.'.*.vkv.
"T"O"
i
I
i
::::::::---7-:::--^v:-v.v
•
.::::::Jf-:::::iii::1::: ::::::::::::::::::::::::
^fc±EEEEEE
|;:::::=
Chemical Name
BMD,D's calculated from basic model
Index chemical
BMDIO's calculated from expanded model
I.B Page 25
-------
Figure I.B-5. Comparison of BMD10s (mg/kg/day) for female and male brain ChE
activity for 33 OPs
Comparison of
Female and Male BMD^s
to*
10
o
a
s
CO
_
(0
I 1
•&
tn
—
re
0.1
• ' • •
Chemical
I.B Page 26
-------
b. CELs Determined for Dermal Endpoints for OPs with
Residential/Nonoccupational Exposure
'^•vv.
Table I.B-5 lists CELs and the next higher dose levels for brain ChE
inhibition from dermal exposure studies of OPs with residential/occupational
exposure plus the index chemical, along with the level of ChE inhibition
(compared to control values).
Table I.B-5. CELs for brain and RBC cholinesterase activity from dermal
exposure studies (% cholinesterase inhibition compared to control value)
Chemical
Acephate
Bensulide
Dichlorvos
Disulfoton
Fenamiphos
Fenthion
Malathion
Methamidophos
Naled
Tetrachlorvinphos
Trichlorfon
Sp&«ies
rat
rat
Male Brain CEL
rngfltg/day
300
9%
500"
0-9%
Dilate Brain
Next Higher J?ose
ing/kg/day
>300*
9%
>500*a
0-9%
Female Brain
cei
mg/kg^day
300
14%
500*
2-10%
Female Brain
Mext Higher Dose
mg/kg/day
>300*
14%
>500*'
2-10%
Dermal exposure study waived due to volatility of compound.
rabbit
rabbit
rabbit
rabbit
rat
rat
rat
rabbit
1.6
7%
10*
0%
100
13%
300a
2%
0.75
0%
10
0%
1000
0%
1000
0%
3
55%
>10*
0%
150
65%
1000"
65%
11.2
41%
20
60%
>1000*
0%
>1000 *
0%
1.6
8%
0.5
0%
50
13%
50'
0%
0.75
5%
10
0%
1000
. 0%
100
4%
3
27%
2.5
18%
100
24%
300"
19%
11.2
38%
20
60%
>1000*
0%
300
18%
Highest dose tested.
I.B Page 27
-------
I c. CELs Determined for Inhalation Endpoints for OPs with
| Residential/Nonoccupational Exposure
[ Table I.B-6 lists CELs for brain cholinesterase inhibition determined for
| inhalation toxicity studies for OPs with residential/nonoccupational exposure
| plus the index chemical, along with the level of cholinesterase inhibition
\ (compared to control values).
[ Table I.B-6. CELs for brain and RBC cholinesterase activity from inhalation
| toxicity studies (% cholinesterase inhibition compared to control value)
Chemical
Acephate
Bensulide
Dichlorvos
Disulfoton
Fenamiphos
Fenthion
Malathion
Methamidophos
Naled
Tetrachlorvinphos
Trichlorfon
WettiQd
nose only
Male
SEU
{mg/kg/dayj
1.419
14%
Male
Next trfg Her
dose
{wg&gAteyfr
1.419*
14%
Female
OEU
mg/kg/day
1.492
13%
Female
Next w&hw
dose
0.928*
0%
0.458
11%
0.047
5%
0.984
0%
0.458
11%
0.410 '
28%
>0.984'
0%
No inhalation toxicity study was available for fenthion
whole body
head/
nose
whole body
115
3%
0.292
8%
0.354
0%
514
17%
1.432
29%
1.594
38%
121
8%
0.310
11%
0.378
4%
540
41%
1.520
25%
1.702
46%
No inhalation toxicity study was available for tetrachlorvinphos.
whole body
9.388
0%
27.44
21%
3.574
0%
9.96
27%
I *Highest dose tested.
I.B Page 28
-------
d. Points of Departure for the Index Chemical (Methamidophos)
Table I.B-7 lists the PODs and no-observed-adversse-effect-levels
(NOAELs) for the oral, dermal, and inhalation routes for methamidophos.
The PODs for all three routes were calculated with dose-response modeling
using the basic model of Equation I.B-1. OPP has used these endpoints in
the RCRA.
Brain cholinesterase was only measured once (at study termination) in the
methamidophos 21-day dermal and 90-day inhalation studies. Therefore
only one data set was available for calculation of the PODs for these routes.
Within route of exposure, the BMD10s for brain cholinesterase shown in
Table I.B-6 were similar for males and females. The values of the BMDLs
were close to the BMD10s. This observation increases the confidence not
only in the selection of methamidophos as the index chemical but also the
utilization of the central estimate of the female data (BMD10) for cumulative
risk extrapolation rather than its lower limit (BMDL). It is notable that the
BMD10 and BMDL values were similar to but slightly larger than NOAELs
established for the oral (chronic NOAEL used for RfD derivation), dermal, and
inhalation routes.
i Table I.B-7. Points of departure for index chemical (methamidophos) by route of
1 exposure for brain cholinesterase activity measured in female and male rats
Route of
Administration
Oral8
Dermalb
Inhalation0
Sex
F
M
F
M
F
M
BMD19
{mg/kg/clay}
0.08 d
0.07
2.12 d
1.88
0.39 d
0.30
BMDL
{mg/kg/day}
0.07
0.06
1.77
1.41
0.21
0.20
NOAELs
{rog/kg/day)
0.03*
0.75
0.31
0.29
: a
:
MRID nos. 41867201, 43197901, 00148452
| "MRID no. 44525301
I CMRID no. 41402401
I dPODs for RCRA of OPs.
E
*NOAEL used for chronic RfD derivation in the single chemical assessment.
I.B Page 29
-------
•i
e. Relative Potency Factors (RPFs)
Table I.B-8 provides the RPFs for the oral, dermal, and inhalation routes
of exposure based on brain cholinesterase in female rats which were used in
the RCRA for OPs. Figure l.B-6 shows the oral RPFs with 95% confidence
limits.
These values were calculated with Equations I.B-3,1.B-4, and I.B-5 for
oral, dermal, and inhalation routes, respectively, and using methamidophos
as the index chemical. BMD10s for all of the chemicals are listed in Table I.B-
4. Dermal and inhalation CELs are given in Tables I.B-5 and l.B-6. Although
a model-derived oral RPF was determined for fosthiazate, this is a new OP
that is not yet registered. Fosthiazate has no appropriate monitoring data to
support characterization of exposure from food, and therefore, was not
included in the quantification of cumulative risk.
I.B Page 30
-------
I Table I.B-8. Relative potency factors for the oral, dermal, and inhalation routes of
I exposure
Relative Potency Factors for Female Brain Cholinesierase Activity
Chemicals
Acephate
Azinphos-methyl
Bensulide
Chlorethoxyfos
Chlorpyrifos
Chlorpyrifos-methyl
Diazinon
Dichlorvos
Dicrotophos
Dimethoate
Disulfoton
Ethoprop
Fenamiphos
Fenthion
Fosthiazate
Malathion
Methamidophos
Methidathion
Methyl-parathion
Mevinphos
Naled
Omethoate
Oxydemeton-methyl
Phorate
Phosalone
Phosmet
Phostebupirim
Pirimiphos-methyl
Profenofos
Terbufos
Tetrachlorvinphos
Tribufos
Trichlorfon
Oral
0.08
0.10
0.003
0.13
0.06
0.005
0.01
0.03
1.91
0.32
1.26
0.06
0.04
0.33
0.07
0.0003
1.00
0.32
0.12
0.76
0.08
0.93
0.86
0.39
0.01
0.02
0.22
. 0.04
0.004
0.85
0.001
0.02
0.003
Dermal
0.0025
0.0015
0.47
1.5
0.015
0.015
1.00
0.075
0.00075
0.0075
Inhalation
0.208
0.677
6.596
0.315
0.003
1.00
0.82
0.087
*>•
*
I.B Page 31
-------
j Figure I.B-6. Relative potency factors for female brain ChE activity for 33 OPs
•qt-K&XA
Relative Potency Factors
for Female Brain ChE Activity
100
10
•
Si o
O C
II
in
0.1-
0.01
0.001 •
0.0001
f
Chemical Name
Female Brain RPFs (June, 02)
Female Brain RPFs (June 02)
Low dose modification
Index Chemical
I.B Page 32
-------
4. Discussion
a. Determination of Relative Potency
With the passage of the FQPA in 1996, EPA was faced with numerous
challenges such as the reassessment of 66% of all tolerances by 2002 and
notably the development of methodology for doing cumulative risk
assessment. As part of the methodology development, EPA has participated
in the public process with technical briefings and reviews by outside experts
who make up the SAP. The SAP has offered constructive and thoughtful
guidance in the development of the hazard and dose-response component of
cumulative risk assessment. With each review, EPA has taken the
recommendations into consideration and has made appropriate revisions or
refinements. The analysis performed for the OPs represents an innovative
and novel approach to hazard and dose-response assessment, and by taking
advantage of the large database of oral toxicity studies in adult rats available
to OPP, offer a comprehensive review of the common mechanism endpoint
(i.e., cholinesterase inhibition) from available toxicity studies in adult animals.
By incorporating dose-response information from multiple studies into one
estimate of potency for the oral route, potency estimates are representative
of the overall toxicity of each pesticide.
Adult cholinesterase data for many OPs has been extensively analyzed
for plasma, RBC, and brain ChE response (USEPA 2001 b, 2001 c). OPP has
generated an extensive database of ChE data that is available to the public.
This large database has allowed OPP to investigate sex differences among
rats, study-to-study variability for over 75 studies, time course data ranging
from. 21 days to > 2 years of exposure, and steady state response. The joint
analysis allowed the exploration of low dose issues using a sophisticated
model. The joint analysis using the exponential model resulted in high
;3i i confidence RPFs and PODs that are representative of the OPs.
The data for the inhalation and dermal routes were less extensive
compared to the oral route. Potency estimates using CELs from the dermal
and inhalation studies are not as robust as those calculated for the oral route
but are adequate for use in the cumulative assessment. It is also notable that
the relative order of estimated potencies for all three routes of exposure are
consistent with current knowledge about their toxicology.
The selection of methamidophos as the index chemical was supported by
the SAP. Methamidophos had the highest quality database for the common
mechanism endpoint in three routes of exposure and three biological
compartments. The PODs calculated with methamidophos have narrow
confidence limits which reduces overall uncertainty in the cumulative risk
assessments. In this assessment, administered dose was used to estimate
RPFs and PODs. At this time there are inadequate pharmacokinetic data for
I.B Page 33
-------
these OPs to incorporate information about dose at the target site or species
to species extrapolation.
b. Dose Additivity
The cumulative risk assessment for the OPs is based on the assumption
. of dose additivity. Dose additivity is the Agency's assumption when
^ I evaluating the joint risk of chemicals that are lexicologically similar and act at
1 the same target site (USEPA 2001 a). The SAP (FIFRASAP, 2001 a)
I indicated that substantial reliance would have to be placed on what is known
I about the mechanism of toxicity because it is very difficult to prove dose
I additivity at human exposure levels. They further pointed out that studies
I available on individual chemicals were usually not designed to address the
| issue of dose additivity.
| The mathematical definition of dose addition requires a constant
I proportionality among the effectiveness of the chemicals (USEPA 2001 a;
| Hertzberg et al.,1999). Thus, an important objective in the dose response
I assessment is to evaluate whether dose-response relationships are
i consistent with the assumption of dose additivity. There is some uncertainty
| surrounding the assumption. Two different versions of the exponential model
| have been used in this assessment. Approximately half of the pesticides
I were fit using a model with a flat low dose region while the remaining OPs
I were fit using a model which is linear in the low dose region. In addition, the
| OPs did not exhibit a common horizontal asymptotes (PB); rather the PBs vary
| among chemicals. Both of these factors indicate that the dose-response
| curves are not parallel.
I Dose additivity assumes that the common mechanism chemicals behave
| in a similar fashion (i.e., same pharmacokinetics and pharmacodynamics). In
| reality, these common mechanism chemicals may not behave ideally (i.e., the
I exact same pharmacokinetics and pharmacodynamics). Biotransformation of
I OPs is extremely complex and involves several metabolic systems in different
| organs (e.g., reactions involving cytochrome P450 isoenzymes, hydrolysis by
I esterases, and transferase reactions; see Nigg and Knaak, 2000). The
| differential activation and/or deactivation of OP pesticides has not been well
[ documented in the literature, nor have the human metabolic pathways
I (Mileson et al., 1998). At this time, these pesticides can not be separated
I into subgroups based on pharmacokinetic or pharmacodynamic
I characteristics. Thus, current information on OP metabolism does not
| provide a sufficient basis to depart from dose additivity at low levels of
I exposure anticipated to be encountered environmentally.
I The application of dose additivity requires the assumption of no
[ interactions other than additive among the chemicals at low doses. There are
a limited number of investigations of the toxicity of combinations of
organophosphorus substances, not necessarily pesticides, that are known to
I.B Page 34
-------
inhibit cholinesterase enzymes (For example see Dubois, 1961 and 1969;
Frawleyetal., 1957 and 1963; Calabrese, 1991; Cohen, 1984; Eto, 1974; Su
et al., 1971; Casida et al., 1963; Keplingerand Deichman, 1967; Rosenberg
and Coon, 1958; El-Sebee, et al., 1978; Seume and O'Brien, 1960; Singh,
1986; Mahajna et al., 1997; Serat and Bailey, 1974; Richardson, et al., 2001;
Karanth et al., in press; Abu-Qare , et al., 2001a; Abu-Qare et al., 2001b).
Most of the studies reviewed were high dose studies that investigated the
acute lethality (LD50) of combinations, mostly binary, and not the cumulative
effects of low exposure levels from multiple OPs. A number of these studies
were conducted using intraperitoneal (i.p.) administration which confounds
interpretations of effects that may be expected by the oral, dermal, or
inhalation routes.
Overall, the studies reported in the literature do not provide a basis for
concluding that interactions between OPs will result in significant departure
from dose addition at low doses. Nevertheless, this literature provides data
showing that different types of interactions can occur between OPs and that
the magnitude of the interaction appears to depend on the specific
combination of OPs investigated, the dose-levels administered, and also the
sequence of exposure (Singh, 1986; Pope and Padilla, 1990). In particular,
the data available are not sufficient to establish the nature of interactive
effects on cholinesterase activity that may be expected among OPs at low
exposure levels.
The OPs all act on the same target site- namely, the inhibition of
acetylcholinesterase by phosphorylation in nerve tissue, which elicits a variety
of cholinergic effects. Dose addition is regarded as a reasonable and
appropriate approach for estimating the cumulative risk associated with joint
exposure to the OP common mechanism group. At this time, there is not
sufficient basis to depart from dose additivity.
Although a biological or pharmacokinetic modeling approach would be
preferred to determine the cumulative risk for these OPs, the input
parameters for such an approach are not available. Thus, the
pharmacokinetic (PK) characteristics of the OPs could not be incorporated in
the dose-response assessment which would allow for a more refined
estimate of the combined risk to humans. Therefore, OPP has applied
simple dose addition and used an empirical curve fitting model (i.e., the
exponential model) to determine RPFs and PODs.
c. Future Directions in Cumulative Dose-Response Assessment:
Physiologically Based Pharmacokinetic (PBPK) Modeling
Physiologically based pharmacokinetic (PBPK) models, which describe
the time course disposition of chemicals and their metabolites, are well suited
to help assess cumulative risk. PBPK models are excellent tools to quantify
the cumulative toxicity that can result from multiple exposures (multiple
I.B Page 35
-------
exposures and multiple pathways) and from exposure to multiple chemicals
with a common mechanism or mode of action. These models typically are
systems of first order differential equations describing the mass balances and
disposition of the chemicals and their metabolites in the body. While these
models are excellent tools, numerous input parameters are necessary for
each chemical. Organ specific thermodynamic parameters (such as tissue to
blood equilibrium partition coefficients) are required for each pesticide
entering the body and for each of its metabolites. Additionally, values for all
of the metabolic rates governing all the biotransformation steps for each
pesticide would be necessary. The complex processes for the common
mechanism effect would be necessary. Using the OPs as an example,
compound specific inputs such as binding constants and values for the rates
of enzyme degradation, aging, and resynthesis would be needed.
ORD's National Exposure Research Laboratory (NERL) has formulated
such a model that has been used to simultaneously model the disposition of
three OPs and their metabolites (Blancato, et al., in review). Another PBPK
model has been developed to describe the complex pharmacodynamics of
acetylcholinesterase inhibition following OP exposure, based almost entirely
on in vitro information (Gearhart, et al., 1994). Timchalk et al. (2002)
developed a PBPK model for chlorpyrifos and and its major metabolites.
At present, these types of data/information on the majority of the OPs are
not available to EPA. PBPK modeling techniques offer good promise despite
the current limitations regarding the necessary input information. Continued
development and testing of the models is necessary and should be pursued.
Pharmacokinetic studies (in vivo and in vitro experiments to determine key
values for PK parameters and the time course disposition of the compounds
in the body) need to be done with many compounds to determine the key
parameters of use in PBPK modeling. It is anticipated that data and methods
will continue to improve and evolve as more experience is gained in this area.
I.B Page 36
-------
I I. Revised OP Cumulative Risk Assessment
| C. Cumulative Risk From Pesticides in Foods
I 1. Introduction to Food
| The cumulative dietary risk due to the use of Organophosphorus (OP)
| Chemicals on food crops was assessed using residue monitoring data collected
I by the United States Department of Agriculture's Pesticide Data Program
I (USDA-PDP) supplemented with information from the Food and Drug
I Administration Center for Food Safety and Applied Nutrition (FDA/CFSAN)
| monitoring data. The BMD10 for brain cholinesterase inhibition in female rats
| was chosen as the Toxicological Point of Departure (POD) for this assessment.
i Methamidophos served as the index chemical. The residue values for the other
I OP chemicals were converted to methamidophos equivalents by a Relative
| Potency Factor (RPF) approach. Residue data were collected on approximately
| 44 food commodities monitored by POP between the years of 1994 and 2000.
i Food processing factors were applied to specific chemical/commodity pairs to
I extend these data for use on cooked and processed food/food forms in the
analysis. The POP residue data were further extended to other commodities
identified as reasonable for translation of pesticide residue data per OPP/HED
SOP 99.3 (USEPA, 1999b); see Appendix III.C.4. Other food commodities, not
included in the PDP database, were incorporated using FDA monitoring data.
The residue estimates incorporated in the assessment represent approximately
97 percent of the per capita food consumption for children aged 1 to 2 years (the
most highly exposed age group) in the Continuing Survey of Food Intakes by
Individuals for the years 1994-1998.
The residue data were compiled as distributions of cumulative residues of
methamidophos equivalents and, after application of processing factors and
FQPA factors, were summed on a sample-by-sample basis. These residue
distributions were combined with a distribution of daily food consumption values
via a probabilistic procedure to produce a distribution of potential exposures for
multiple subpopulations in the CSFII 1994-1998 (Infants less than 1, Children 1-
2, Children 3-5, Children 6-12, youth 13-19, Adults 20-49, and Adults 50+ years
old). The most highly exposed age group was confirmed to be Children 1-2
years old.
I.C Page 1
-------
2. Sources of Residue Data
a. USDA-PDP
The POP program has been collecting pesticide residue data since 1991,
primarily for purposes of estimating dietary exposure. The program is
designed to focus on foods highly consumed by children and to reflect foods
typically available throughout the year. Foods are washed and inedible
portions are removed before analysis. This database is the primary source
for residue data used in the current assessment, and data collected between
1994 and 2000 were included. A complete description of the POP program
and all data through 2000 are available on the Internet at
http://www.ams.usda.gov/science/pdp. A summary of the POP residue data
on OP chemicals is shown in Appendix III.C.2. Appendix III.C.1 lists all of the
food forms for which estimated residues were based on POP data.
b. Market Basket Study of OP Residues in Apple Sauce
The Apple Processors Association sponsored a market basket study of
OP pesticide residues in apple sauce samples collected in 1999. These data
are incorporated in the current assessment for residue estimates on apple
sauce and baby food apple sauce. The residue data on these samples are
summarized in Appendix III.C.2.
c. FDA/CFSAN Surveillance Monitoring Data
The FDA Surveillance Monitoring Program is designed primarily for
enforcement of EPA pesticide tolerances on imported foods and domestic
foods shipped in interstate commerce. Domestic samples are collected as
close as possible to the point of production in the distribution system. Import
samples are collected at the point of entry into U.S. commerce. The
emphasis in sample collection is on the agricultural commodity, which is
analyzed as the unwashed, whole (unpeeled), raw commodity. Processed
foods are also included in the program. A description of the program and
residue data for recent years can be found on the Internet at
http://vm.cfsan.fda.gov/~lrd/pestadd.html. Because the emphasis of this
program is not on dietary exposure, it is being used in the current
assessment mostly as a semi-quantitative check on the potential for residues
and as support for data from other sources. The program has extensive data
available on eggs and fish and these data support the judgement that the OP
residues are negligible on these foods as consumed. Appendix III.C.1
indicates the food forms for which exposure estimates were supported by this
program.
I.C Page 2
-------
d. FDA/CFSAN Total Diet Study (TDS)
The TDS has provided data on dietary intake of food contaminants for
about 40 years. A program description and residue data can be found at the
same Internet site listed above for FDA Surveillance Monitoring Data. Foods
are purchased in grocery stores, generally 3 or 4 times a year, prepared and
cooked for consumption, and analyzed by highly sensitive multiresidue
methods. Between 1991 and 1999 there have been 26 market baskets
collected and approximately 260 foods analyzed for, among other things, OP
pesticide contamination. A disadvantage of these data is that only one
sample is analyzed of each food in each market basket. For this reason
these data have been used primarily as semi-quantitative support for
judgements on residues in foods. An exception is made in this assessment
for the estimate for residues in meats other than poultry. Multiple forms and
tissues of beef, pork, lamb, and meat byproduct cold cuts have been
analyzed in all of the market baskets with only limited residues of OP
pesticides on a few of the meats at low levels. In an effort to include residue
estimates for all highly consumed foods, a conservative estimate for meat
commodities was based on the TDS Data. A maximum residue level was
used for each meat based on the TDS. The meat commodities included on
this basis are identified in Appendix III.C.1 and the residue data are
summarized in Appendix III.C.3.
3. OP Pesticides Included in Cumulative Assessment
All of the OP analytes detected in the POP program are included in the
current assessment. See Appendix III.C.2 for a complete summary of the
analyses for OP pesticides and metabolites on each food commodity in the
database. There have been significant numbers of analyses for 67 OP active
ingredients, degradates, or metabolites between 1994 and 2000. A total of 39 of
these OP analytes have been detected in at least one of the foods analyzed.
After exclusion of data on pesticides that have been canceled or do not have
food uses, and combining data for metabolites and degradates, there are
positive analytical data being used for 20 OP pesticides. These are the
following:
acephate
diazinon
disulfoton
methidathion
oxydemeton-methyl
phosalone
terbuphos
azinphos methyl
dichlorvos
ethoprop
methamidophos
methyl-parathion
phosmet
tribufos
chlorpyrifos
dimethoate
malathion
mevinphos
phorate
pirimiphos-methyl
I.C Page3
-------
Naled has not been separately analyzed generally and residues from this use
would be reflected in the dichlorvos analyses. Bensulide is not included in the
PDF data; however, negligible residues would be expected in foods based on
field trial data submitted for registration purposes. Cadusafos is not represented
in the POP data but the only registered use that could potentially result in food
residues is as a nematacide soil application on bananas that are imported into
the United States. Field trial data submitted for registration/tolerances purposes
indicate that residues will not occur in the edible portion of the banana.
Chlorethoxyfos is not included in POP data but its only food use is soil
application to corn crops at a low rate; therefore, significant residues in edible
portions and processed foods from corn would not be expected. Dicrotophos,
not included in POP data, has one food use on cotton. Cottonseed oil is the only
food commodity of cotton and it is not included in the current assessment, but
the impact of the chemical on dietary (food) exposure is expected to be low due
to the extent of refining and blending of the oil. Tebupirimphos (phostebupirim)
has one food use on corn, mainly to control root worm. Significant contribution
to cumulative food exposure is not expected. Profenofos is used on cotton,
which is not included in the current assessment for the reasons stated above.
Trichlorfon has no food uses except for an overseas use as pour-on treatment of
beef cattle. Tetrachlorvinphos is used only on livestock or livestock premises.
Potential residues from the two latter livestock uses are anticipated to be
covered by the conservative cumulative residue estimate for meat commodities.
4. Foods Included in the Food Risk Assessment
The universe of foods included in the cumulative dietary exposure
assessment is defined by the USDA CSFII for the years 1994-1996 with
supplementary data on children obtained in 1998. The survey data, CSFII 1994-
1998, is integrated into DEEM-FCID™. Appendix III.C.1 lists all of the foods in
CSFII 1994-1998 in decreasing order of their relative per capita consumption by
children 1-2 years old and children 3-5 years old. Each food is assigned a
percent of relative consumption which was estimated in the following manner:
the per capita consumption of each food was summed for all children in the
. | survey in the two age groups. These consumptions were totaled for all foods in
| the survey and the individual sums for each food were expressed as a percent of
f | the total. This measure of relative consumption is used as a partial indication of
the potential significance of a given food in the diet of children.
According to the above described measure of relative consumption, the
available POP data were used either directly or with processing factors to
estimate cumulative residues in foods accounting for about 88% of the per capita
consumption of children 1-2 years old. POP data were used for the top 10
ranked foods and for 24 out of the top 30 foods. Apple sauce, which was
supported by special study data, account for about 1 % of the consumption by
children 1-2 years old.
I.C Page 4
-------
I Residues in other foods were estimated using translated POP data according
| to HED SOP 99.3, (USEPA, 1999b) as summarized in Appendix III.C.4.
| Translations included only residues for chemicals registered on the food being
I simulated. These foods account for about 1% of the per capita consumption of
i children 1-2 years old.
s
I Surveillance monitoring data from FDA include extensive analysis of eggs
T_ I and fish with the implication that OP residues would not be expected to occur in
| significant amount on these two categories of foods. The TDS data from FDA
| indicate a similar situation for livestock meats. In this case a conservative
I estimate of residues was incorporated into the assessment, i.e., meats were
I assumed to always be contaminated with OP residues equal to the maximum
I values found in the TDS market baskets (see Appendix III.C.3 for a summary of
I TDS data used). These foods being supported by FDA data, i.e., eggs, fish, and
i meat, account for about 5% of the per capita consumption of children 1-2 years
I old.
POP has analyzed high fructose corn syrup and found no OP residues but
has not analyzed any other sugar or syrup sources. The FDA TDS has analyzed
refined sugar and maple sugar and found no OP residues in 26 market baskets.
A knowledge of the highly refined nature of sugars and syrups supported by the
limited residue data mentioned above is the basis for assuming that negligible
residues of OP pesticides would be expected to occur in sugars and syrups.
Therefore, residues were assumed to be zero for these foods derived from
sugarcane, sugar beet, and maple. These foods account for about 2% of the per
capita consumption of children 1-2 years old.
| The food forms not included in the current assessment account for almost 3%
I of the per capita consumption of children 1-2, distributed among many food
forms. Table I.C-1 summarizes the relative consumption of foods in the
assessment for children 1-5 years old. The information is provided in detailed
form in Appendix III.C.1.
I.C Page 5
-------
Table I.C-1. The Proportion of the Diet of Children (1-5 years old) Covered in the
Cumulative Food Assessment
Source of Residue Estimate
PDF (RACs & processed)
Apple Sauce Study
Translation of PDF
FDA Monitoring and TDS
Assumed Negligible
Not Included in Current Assessment
Percent of Per Capita Consumption
Children 1-2
88.4
0.9
1.1
4.9
2.0
2.7
Children 3-5
85.0
0.7
1.3
6.3
3.1
3.6
5. Method of Estimation of Cumulative Dietary Risk
Dietary exposure was estimated using the Dietary Exposure Evaluation
Model (DEEM-FCID™) software. A joint distributional analysis was conducted
by combining representative data on concentrations of OP pesticides on foods
with distributions of anticipated consumption of these foods by different
segments of the U.S. population. The primary advantage of a joint distribution
analysis is that the results are in the form of a simultaneous analysis (i.e., a
distribution) of exposures that demonstrate both best-case and worst-case
scenarios of exposure. The inputs were distributions or point estimates for
residues, distributions for consumption, and a hazard endpoint. The output was
a series of distributions of one-day dietary exposures and distributions of
associated risks, i.e., margin of exposures (MOEs). The different components of
the input data are discussed further in the remainder of this section.
a. Manipulation of Residue Data for Exposure Assessment
Commonly, the following two equations are used for estimating exposure
and risk from a single chemical:
1) Exposure = Residue X Consumption
2) Risk = Hazard X Exposure
In the case of cumulative exposure assessment, the residue term in the
first equation is changed to Index Equivalent Residue (Residue^), and the
hazard end point in the second equation is based on the index chemical.
The calculated cumulative residue is a simple arithmetic addition of
residues of different chemicals that have different toxicities (potency) and
therefore simple addition of their residues is not appropriate. For that reason,
I.C Page 6
-------
s
the amount of residue of each chemical is adjusted by multiplying by a RPF
to get the equivalent residue of an index chemical. This new calculated
residue is termed ResidueIE and the exposure value resulting from combining
Residue,E and consumption is termed Index Equivalent Exposure
(ExposureIE). The new central equation for exposure will then become:
Exposure,E = Residue,E X Consumption
•IWAWW
and in the risk equation (second equation) the toxic end point of the index
chemical is used. The following discussion explains in more detail how this
was accomplished for this cumulative risk assessment.
b. Generation of Cumulative Equivalent Residue (Residue,E)
To determine a given one-day cumulative oral exposure to multiple OP
chemicals, first an Residu.e,E for each residue value is calculated. On a given
PDF sample, each residue value is multiplied by any applicable processing
factor (PF) for that chemical on food sample of interest and the RPF for the
same chemical to express it as a Residue,E for that chemical; this is step 1.
Stepl: Residue,E (per chemical n) = Residue X PFn X RPFn
The cumulative Residue,E for all chemicals detected on one POP sample
will then be the sum of all the Residue^ for all the chemicals on that sample;
this is step 2.
Step 2: Cumulative ResidueIE = ^ Residue,E (per POP sample)
For example, given 100 samples of apples, each analyzed for 22 OPs,
there will be generated 22 ResiduelE values for each sample. In step 2, each
set of 22 Residue,E for a sample is summed to generate a cumulative
Residue,E per one sample; hence 100 cumulative Residue,E points for 100
samples of apples are generated.
By summing on a sample-by-sample basis, the potential for capturing any
co-occurrence on the same commodity is enhanced. Another very important
advantage of this approach is that, using appropriate record keeping (see
next section), the complete history of each cumulative residue value in the
exposure assessment can be potentially traced back to its origins. All of the
sample collection and analytical information associated with a given POP
sample and all arithmetic adjustments incorporated in producing a Residue,E
can be traced in the process of sensitivity analysis or critical food commodity
contribution analysis.
I.C Page 7
-------
c. OPCRA Food Residue Database
The data manipulations necessary to prepare the PDF residue data for
input into the risk equation are in principle very simple; however, the task of
performing these calculations for multiple chemicals and food commodities is
problematic. The residue data used in this assessment consist of
approximately 1.5 million records of analytical data and sample information.
The processing factors account for several thousand additional records of
information. For this reason, and in anticipation of the need to make multiple
uses of the data, to keep track of them, and work backward from the
cumulative assessment results to determine contributors, all the data
manipulations were conducted using relational database techniques. The
OPCRA food residue database currently being used for this purpose consists
of, among other things, four major data tables:
t—• 5
f^ I 1 - Residue data table(s); about 1.5 million records containing essentially
C/| | all of PDP sample and analyses data for OP pesticides as well as
0} | other residue data compiled from FDA and the Apple Sauce Market
CP | Basket Survey.
I 2. Processing factor data table; containing all relevant processing factors
| for specific food form/chemical combinations. Appendix III.C.5 is
| extracted from these data.
E
| 3. RPF Table; containing the RPF for all chemicals of interest.
I 4. Translation Table; providing bridging links between POP commodity
I codes, such as AJ (apple juice), and all corresponding DEEM™ food
I forms, such as Apple, juice cooked:canned;cook meth N/S. This table
I allows the assignments of translation of data between POP
;~;;| I commodities also, such as cantaloupe data to watermelon food forms.
C 1 Appendix III.C.6 summarizes the links used in this assessment.
These four tables are linked through common fields, including pesticide
codes and.commodity codes. Calculation queries are coded into the
database so that all the pertinent residue records can be extracted, each
calculation outlined above can be performed, and the results can be sorted
and stored in various formats for further analysis.
A cumulative residue calculation query performs the two-step process
described earlier, extracting the various parameters needed from the four
tables described above. The calculation is performed on all of the food
samples that are of interest and the results are compiled in text files
containing the cumulative distributions for each food commodity of interest.
I.C Page 8
-------
Each text file contains a header with sample information (number of
values, number of detects, number of zeros, average of residues) and all of
the cumulative residue values for a single food form, sorted in descending
order.
Residue distribution inputs to DEEM™ are converted to single average
values for those foods that are highly blended before consumption.
By maintaining all of the calculation parameters in separate tables in the
database, it is possible to repeat the above process with new inputs by simply
replacing or adding data to the appropriate table. For example a specific
I chemical can be omitted from the entire process by assigning it a value of
| zero in the RPF table. Specific chemical/commodity combinations can be
| selectively omitted by entering a zero value for that pair in the processing
| factor table. These methods have been used extensively in the current
i assessment to adjust the inputs to reflect currently supported uses of OPs on
I food crops and to test the relative contributions of chemicals and
I commodities to the results of an assessment.
I d. Generation of Exposure Values
The cumulative Residue,E values (text files described in the previous
section) are treated as distributions of representative residues and linked to
all appropriate food forms; cumulative residue values are then randomly
picked and combined with a consumption record to generate a single
exposure value which is termed Exposure,E. This process (semi-Monte Carlo
in nature and conducted by DEEM™ software) is repeated many times per
each consumption record to generate a distribution of exposure values. This
process has been described in public documents and proceedings of the
FIFRA Science Advisory Panel
(http://www.epa.gov/oscpmont/sap/2000/#february). For the food forms that
are highly blended before consumption, the residue input consisted of the
average of all the cumulative residues, i.e., a single average residue value
was entered into the DEEM™ calculation.
e. Food Consumption Data
For this assessment, food consumption is being modeled on the USDA
CSFII, 1994-1998. The consumption survey is included as an integral
component of the DEEM-FCID™ software. The CSFII 1994-1998 contains
survey data on 20,607 participants interviewed over two discontinuous days.
It contains a supplemental children's survey conducted in 1998 in which an
additional 5,459 children, birth through 9 years old, were added to the survey.
This is the first dietary exposure assessment in which OPP has used this
survey.
I.C Page 9
-------
DEEM-FCID™ also has integrated new USDA/EPA recipes for conversion
of foods reported eaten in the survey to food commodities on which residue
data are available. These recipes, which are available to the public, replace
proprietary recipes used in previous versions of DEEM™.
Separate assessments were conducted on the various segments of the
population as represented in the CSFII 1994-1998. Among others, the
current assessment included the following age groups:
Q Infants less than 1 year old
Q Children 1-2 years old
Q Children 3-5 years old
Q Children 6-12 years old
Q Youth 13-19 years old
Q Adults 20-49 years old
Q Adults 50+ years old
^xxwc-x _
The most highly exposed population group in this cumulative assessment
is children 1-2 years old; subsequent analyses of the results reported in this
document will emphasize results for this age group.
f. Hazard Data used in the Cumulative Food Assessment
Section II describes the hazard portion of this risk assessment in detail.
Methamidophos was chosen as the index chemical for this assessment and
relative potencies of the OP chemicals were based on female rat brainv
cholinesterase inhibition. The point of departure (BMD10) was 0.08 mg/kg
body weight/day. The application of FQPA Safety Factors for this OP
cumulative assessment is made for each individual chemical in the
assessment. This is accomplished by incorporating these factors into the
relative potency factors for each chemical in the assessment.
6. Results ;
The revised cumulative food exposure assessment for OP pesticides on food
commodities was conducted for seven age groups, infants of less than one year,
children 2-3 years old, children 3-5 years old, children 6-12'years old, youth 13-
19 years old, adults 20-49 years old, and adults 50+years old.
I.C Page 10
-------
Appendix III.C.7 contains a complete listing of the food forms in the DEEM-
FCID™ software that were included in this assessment. This table also includes
summary information on the residue distributions that were prepared from the
OPCRA food residue database as input for each food form. Although most of
the data inputs in this table are defined as residue distributions (rdf files), for
highly blended commodities, a single average residue was estimated. The
actual DEEM™ input file and necessary rdf files will be made available on CD
ROM and on the internet for any interested party.
The most highly exposed age group in this assessment is Children 1-2 and
the subsequent results reported in this chapter will focus on this group. Figure
I.C-1a is a cumulative plot of the exposure distribution for this age group.
Figure l.C-1b expands the portion of the cumulative distribution between the 99th
and the 99.99th percentile. There are 4192 person-days (approximately half that
many individuals with two reported days of food consumption) represented in the
consumption records for this age group.
I.CPage11
-------
Figure I.C-1a. Cumulative Distribution of Food Exposure for Children 1-2 yrs
su i
o:oos
o.pu :
On*
I.C Page 12
-------
Figure I.C-1b. Cumulative Distribution of Food Exposure for Children 1-2 yrs - 99th percentile to 99.99th percentile
of Exposure
;100
•0.0008
0.0016 0,0024
One Day Bsposure Omg/kg)
O.OW
I.C Page 13
-------
•'•"' *
ixyvx?
w- i
^
vs^ =
a. Analysis of Significant Presence in The Upper Portion of the
Distribution
The DEEM software has a provision for analyzing the foods and food
forms that are contributing to the upper portions of an exposure distribution,
up to a maximum interval of 5 percentile. This provision was used in the
current assessment, in combination with the chemical/commodity specific
information maintained in the database described above, to assess both .
foods and chemicals present in the tail of the distribution. The data
summarized here were obtained by examination of the exposure distribution
interval from the 99.8th percentile to the 100th percentile. Table I.C.2 lists all
of the food forms appearing at or above the 99.8th percentile from a Monte
Carlo assessment of the exposure of children 1-2 years old.
I.CPage14
-------
Table I.C-2. Partial Summary of Foods and Food Forms Occurring in the Top 0.2
Percentile of Exposure to an Exposed Sub-population in OP Cumulative Risk
Assessment*
[Monte Carlo Iterations =1000. Number of actual records in this interval = 8247.
N=number of appearances in all records (including duplicates).]
POOS'
Grape
Pear
Apple, fruit with peel
Apple, juice
Tomato
Grape, raisin
Bean, snap, succulent
Pepper, bell
3ean, snap, succulent
Potato, tuber, w/o peel
Spinach
Bean, snap, succulent
Squash, summer
Bean, lima, succulent
Celery
Cucumber
Bean, lima, succulent
Spinach
Cucumber
Potato, tuber, w/peel
Pepper, bell
Bean, snap, succulent
Bean, snap, succulent
Potato, tuber, w/o peel
Pepper, bell
Potato, tuber, w/o peel
Tomato
Potato, tuber, w/o peel
Apple, juice - babyfood
Apple, juice
Strawberry
Bean, lima, succulent
Bean, snap, succulent
Cherry, juice
Bean, snap, succulent- babyfood
Tomato
Spinach
Peach
Grape, juice
Tomatillo
Apple, juice
Tomato, juice
Food; Form
Uncooked; Fresh or N/S; Cook Meth N/S
Uncooked; Fresh or N/S; Cook Meth N/S
Uncooked; Fresh or N/S; Cook Meth N/S
Uncooked; Fresh or N/S; Cook Meth N/S
Uncooked; Fresh or N/S; Cook Meth N/S
Uncooked; Dried; Cook Meth N/S
Cooked; Frozen; Boiled
Uncooked; Fresh or N/S; Cook Meth N/S
Cooked; Canned; Boiled
Cooked; Fresh or N/S; Boiled
Cooked; Frozen; Boiled
Cooked; Fresh or N/S; Cook Meth N/S
Cooked; Fresh or N/S; Boiled
Cooked; Frozen; Boiled
Uncooked; Fresh or N/S; Cook Meth N/S
Uncooked; Fresh or N/S; Cook Meth N/S
Cooked; Canned; Boiled
Cooked; Fresh or N/S; Baked
Cooked; Canned; Cook Meth N/S
Cooked; Fresh or N/S; Fried
Cooked; Canned; Cook Meth N/S
Cooked; Fresh or N/S; Boiled
Cooked; Fresh or N/S; Boiled/baked
Cooked; Frozen; Fried
Cooked; Fresh or N/S; Cook Meth N/S
Cooked; Fresh or N/S; Fried
Cooked; Fresh or N/S; Boiled/baked
Cooked; Fresh or N/S; Baked
Cooked; Canned; Cook Meth N/S
Uncooked; Frozen; Cook Meth N/S
Uncooked; Frozen; Cook Meth N/S
Cooked; Fresh or N/S; Boiled
Cooked; Fresh or N/S; Fried
Uncooked; Fresh or N/S; Cook Meth N/S
Cooked; Canned; Cook Meth N/S
Cooked; Fresh or N/S; Boiled
Uncooked; Fresh or N/S; Cook Meth N/S
Uncooked; Fresh or N/S; Cook Meth N/S
Uncooked; Fresh or N/S; Cook Meth N/S
Uncooked; Fresh or N/S; Cook Meth N/S
Cooked; Canned; Cook Meth N/S
Cooked; Canned; Cook Meth N/S
N
2600
1549
2177
1510
584
376
397
337
383
155
39
58
33
76
136
42
64
22
28
32
32
110
31
36
46
22
32
23
37
22
15
36
13
11
19
94
7
42
8
7
94
c
\J
Fraction of
Total
Exposure
0.33
0.16
0.13
0.10
0.05
0.04
0.03
0.03
0.02
0.02
0.01
0.01
0.01
0.01
0.01
0.01
<0.01
<0.01
<0.01
<0.01
O.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
O.01
O.01
<0.01
<0.01
<0.01
<0.01
<0.01
-------
Food
Potato, tuber, w/peel
Pepper, bell
Cucumber
Bean, snap, succulent
Apple, peeled fruit- babyfood
Lettuce, head
Potato, tuber, w/peel
Broccoli
Celery
Potato, tuber, w/o peel
Squash, summer
Squash, summer
Tomato
Spinach
Grape
Tomato
Apple, sauce
Orange
Grape
Tomato, puree
Orange, juice
Strawberry
Celery
Potato, chips
Pepper, bell
Orange, juice
Celery
Celery
Pear
Broccoli
Grape, raisin
Potato, tuber, w/o peel
Pepper, non-bell
Celery
Foo
-------
I To evaluate the presence of chemicals in the tail of the distribution, all of
| the food forms in the above table were linked with the corresponding residue
I distributions that had been generated for the cumulative assessment. The
I individual chemical contributors to these distributions were extracted from the
I OPCRA food residue database used to generate the distributions. Thus the
| relative percent contributions of food forms derived from DEEM were
| combined with the relative percent contributions of chemicals to each food
I form's residue distribution to give an estimate of the relative contribution of
| each chemical to the interval being examined. These data were further
| reduced by combining all food forms to the crop level, for example, fresh
| grapes, raisins, and grape juice were all combined under the crop name
I grapes, and so on. All metabolites, degradates, and isomers were combined
I for each active ingredient in that assessment. The linkage of the DEEM
1 output and the OPCRA food residue database information on chemical/food
| form specific contributions are summarized in Appendix III.C.8.
1 The most significant chemical presence in the exposure interval between
| the 99.8th percentile and 100th percentile is dimethoate/omethoate with a
| relative contribution of approximately 48% of the exposure, followed by
| azinphos methyl at 27%, acephate and its metabolite methamidophos at
I 11%, methamidophos, as active ingredient, at 5%, phosmet at 2.5%, phorate
I at 2.2%, and mevinphos at 1.8%. The most significant food crops are:
I apples, grapes, green peppers, pears, potatoes, spinach, succulent beans,
| and tomatoes.
l
[ 7. Discussion
| a. Changes Since the Preliminary Assessment
| A preliminary cumulative assessment was published on December 3,
| 2001. Since that time a number of changes have been made in the input
I data, some of them as a result of public comments. All of these changes are
I captured in the data summaries included in the Appendices to this document.
I A summary of the major changes or categories of changes is provided in this
| section.
| Q Processing factors were updated. Several factors were added or
I changed based on public comments on the preliminary assessment.
| Appendix III.C.5 provides a summary of the processing factors currently
| being used. It should be noted that the absence of a processing factor in
| Appendix III.C.5 or a factor of zero indicates that the specific food
| form/chemical pair does not contribute to any residue distribution
I estimates. In some cases the absence of a factor is simply due to the fact
| that there are no detectable residues of that chemical in the database but
| in other cases it is due to the fact that a specific use is being excluded
| from the assessment because it is not being supported. Several
I commodities are not entered in the table at all because the residue
I I.CPage17
-------
analyses conducted on these foods were uniformly below detectable
levels. Therefore, one must not use this table as a means of determining
the uses included in the assessment. The appropriate starting point for
this determination is Appendix III.C.7, which lists every food form included
in the assessment. A factor of zero in the processing factor table in some
cases is due to a correction of a former entry and in some cases, such as
for chlorpyrifos on apple and grape commodities, is a means of adjusting
the assessment to account for mitigation negotiations. In the example of
chorpyrifos, the use patterns are being altered to allow only pre-bloom
applications, which are expected to yield essentially no detectable
residues. The PDP data set contains many detects due to foliar
applications; therefore, the processing factor was used as a use flag that
was lowered to zero for this assessment. The residue data are still in the
database and can be reactivated by raising the use flag.
Q Apple sauce residue data have been added to the data set based on
MRID 45432001: Data are from The National Food Laboratory and were
sponsored by Apple Processors Association. The data have been
incorporated in the PDP Data set in the OPCRA food residue database.
Q Relative Potency Factors have been revised. Among the revisions,
omethoate now has a factor different from that of dimethoate.
Q FQPA Factors of 3 have been applied (via RPF adjustments) to all
chemicals in the assessment except methamidophos,
dimethoate/omethoate and chlorpyrifos, which have their FQPA factors
reduced to 1.
Q The bridging and translation of residue data from source to CSFII food
forms have been updated. Several adjustments and corrections were
made in these assignments. Among these was the changing of all tomato
processed food forms to be derived from canned tomato residue data
instead of fresh tomato residue data. All of the translations for the current
vv:J assessment can be seen in Appendix III.C.6.
Cl Some inappropriate residue data were removed from the assessment.
Lettuce residue data from 1994 were removed completely from the
assessment because they contained residues from use patterns of
methamidophos and mevinphos that are not current. Lettuce residues are
now based solely on data from 1999 and 2000.
Fenamiphos and chlorpyrifos-methyl have been removed from the
assessment based on planned phase-outs. Appendix III.C.9 should be
consulted for the complete summary of OP pesticide uses or import
tolerances that are currently being supported in the reregistration process.
It is important to note that this appendix provides the scope of use
patterns that are being considered as having potential for producing OP
I.C Page 18
-------
I residues on foods. This list is based on reregistration actions up through
I March of 2002. If an OP use is not listed in this appendix then it is not
I considered in the current assessment. >
[ Q Tolerance exceeding residues were added back to the residue data as a
[ result of discussion with SAP after release of the preliminary assessment.
[ These violative residues are not a significant contributor to the
I assessment.
<«x»»»x £
^v I b. Major Assumptions in the Revised OPCRA for Foods
E
z
I The following discussion of input assumptions is provided as a revision to
| the same discussion that was included in the preliminary assessment. The
I assumptions are revised to reflect their current status and thus some of the
I points covered in the previous section may be restated here.
| The processes for exposure and risk assessment in the Office of
| Pesticide Programs (OPP) have been undergoing a rapid evolution. A
1 number of choices and assumptions made in the conduct of the current
| assessment may differ from previous single-chemical assessments. The
I following discussion is intended to provide some background on the impact of
| choices that are unique to this assessment.
E
| i. Some POP Residue Data Were Excluded
i
| The assessment includes only chemical/crop combinations currently
I being supported for registration in the United States or with import
\ tolerances (see Appendix II.C.9). Therefore, residues representing
I canceled and phased-out uses are excluded. That is, residues in the
I OPCRA food residue database that do not represent supported section 3
| registrations, SLN uses, or supported import tolerances, are excluded
I ' .from the assessment. In a change from the preliminary assessment of
I 12/3/01, we are not excluding violative residues, i.e., tolerance exceeding
residues, from the assessment. The criteria listed in this paragraph are
intended to ensure that the cumulative assessment simulates the residue
pattern that will result from ongoing mitigation actions in the reregistration
of OP pesticides. Although this may appear to underestimate the food
exposure as reflected by available residue data, it should be kept in mind
that these data reflect past patterns of residue occurrence. The inclusion
of violative residues in the assessment has no significant effect on the
overall results of the assessment. Violative residues are rare in residue
monitoring data.
I.CPage 19
-------
ii. Composite Samples Were Used to Estimate Residues in Single-
Servings as Consumed
Only the residue data from composite samples were utilized in this
assessment. A single composite sample may contain several individual
serving of some foods. For purposes of the present assessment, it is
assumed that residues reported on composite homogenates adequately
^,v,,, reflect the residues in any given single-serving contained in that
homogenate. Therefore, no attempt was made to "decomposite" residue
values to simulate residues that might be present in the single-servings
contained in the POP composite sample. POP has conducted single-unit
sampling for apples, pears, and peaches since 1998. A comparison of
the residue levels on these single-servings to the residues on comparable
composite samples indicate that use of composite samples will not result
in a significant under- or overestimation of residues.
iii. PDP Samples Were Assumed to Reflect Residues in Foods
Prepared for Consumption
The PDP generally collects foods at wholesale distribution centers and
stores them frozen until analysis. Foods are washed and inedible portions
are removed before analysis but these foods are not further cooked or
processed. Processing factors (see Appendix III.C.5) were applied to the
residue data in this assessment. These factors were taken from the most
recent single-chemical dietary exposure assessments for the OPs.
Information on these factors is somewhat limited; therefore, some storage
or process related dissipation of residues may not be accounted for. In
response to the preliminary assessment we have had several public
comments with suggestions for improvements in the processing factors.
These suggestions have been incorporated in the current assessment as
appropriate. The processing factors in Appendix III.C.5 reflect these
changes. The processing factors still probably result in some
....... overestimation of residues in processed foods for which factors are not
available, but the impact on this assessment of this possibility appears to
be minimal according to the results reported here. Most of the food forms
that appear to be significantly present in the upper ends of the exposure
distribution are either uncooked food forms or are supported by residue
data on food forms that have been processed in a similar manner.
iv. Residue Data Were Assumed to Reflect Co-occurrence of OPs in
Single-day Diets
One reason for conducting the assessment of PDP residue data on a
sample-by-sample basis is to maintain the connections in multi-analyte
occurrences on these samples. In other words, it is assumed that the
PDP sampling and analysis protocols capture the co-occurrence of OPs.
Appendix III.C.10 demonstrates the extent of this measured co-
I.C Page 20
-------
I occurrence in the POP program between 1994 and 1999. It can be seen
| in this table that a majority of POP samples were reported as containing
I no detectable residues at all. For those that contained detectable
I residues, single residues were most prevalent but many multiresidue
I samples were found. The maximum number of OP analytes reported on
i a POP sample is 5 (this occurred on only 5 samples during the period
| 1994-1999).
WVS.SSSVL Z
^wxw I In addition to considering co-occurrence of different OPs on one food,
| the potential exists for co-occurrence from residues of one or more OPs
1 on different foods consumed in one-day. This assessment is using
I residue data collected over a seven year period, 1994 through 2000. This
i is necessary in order to maximize the number of food commodities in the
| assessment but this raises issues of lack of co-occurrence. Co-
| occurrence in the food is important from the standpoint of all the food
| consumed in the same time period. One may question if it is appropriate
I to model exposure based on bananas grown in 1994 and apples grown in
| 1998. On the other hand, the consistency in appearance of residues in
| the monitoring data over time suggest that the uncertainty in this choice is
I probably not more significant than those in other aspects of the model.
I A related choice in selection of residue data was to include all
| available data for a given food. This has resulted in data sets that span
I time periods of less than one year to as much as four years of data. At
I the time of the preliminary assessment we were exploring the impact of
I using reduced data sets for foods with a suggested maximum of 2 years
| of data for any given commodity. A decision was made to continue using
I the complete data set in the absence of specific information that a given
| subset of the data were inappropriate to consider for currently supported
| uses or import tolerances. Therefore, the complete data set is still being
I • used (1994-2000 POP) with the exception of data from 1994 on lettuce.
| These samples contained residues of methamidophos and mevinphos
I that could have resulted from applications that were not representative of
| current use patterns. The use of the complete data set from 1994 through
| 2000 increases the probability that variations in climate and pest
I pressures may have been captured in the residue distributions.
I.C Page 21
-------
W.V.W.A
v. It Was Assumed That AII.OPs of Concern on an Analyzed Food
Sample Were Accounted for in the Residue Analysis
All residue analyses are subject to the limitations of the sensitivity of
the analytical methods. Many of the samples analyzed are reported as
being below the analytical method reliable limit of detection (LOD). It has
been usual practice in Agency assessments on individual pesticides to
assume that residues in non-detectable samples are present at 1/2 LOD of
the analytical method in samples that were harvested from treated fields.
Thus, for purposes of estimating residues in samples reported as
-------
I vi. PDP Residue Data Were Translated in Some Cases to Foods for
I Which No Residue Data Were Available
5
In chemical-specific dietary exposure assessments the Agency
routinely translates residue data from one food commodity to related ones
if the pesticide use patterns are similar on these commodities (USEPA,
1999b). For example, data on cantaloupes is often used as surrogate
^ = data for watermelons and other melons. For a cumulative assessment, in
^x-xw | which a grower has a choice of several chemicals from the cumulative
assessment group, these translations of data become more difficult to
make. In the current assessment, translations of the residue data were
t I made using the translation scheme in HED SOP 99.3 (USEPA, 1999b) in
^ I order to ensure representation of the maximum number of commodities
^ I possible. The allowable translation are summarized in Appendix III.C.4.
•^ | In making these translations the only residues included were those that
C I could occur on the simulated food from current registrations of OP
CD | pesticides. The uncertainty in this scheme is not expected to have a
Cp | major impact on the assessment because the foods being translated
Cl) | comprise a relatively small portion of the per capita consumption by
^ I • children (See Appendix III.C.1 for confirmation of this fact). An analysis of
foods in the higher percentiles of exposure in this assessment has
confirmed that translated foods do not significantly impact that portion of
the distribution.
i
| vii.The Food Exposure Portion of This Cumulative Assessment is
| Considered to be Constant Throughout the Year and Across Regions
I It is currently assumed that the food distribution and storage systems
| in the United States result in essentially a national distribution of the major
| foods in our diet that is constant throughout the year. For some of the
I seasonal changes in availability of certain foods, PDP has designed its
| sampling program to concentrate on these time frames so that the residue
| data should reflect the foods as available to the consumer. This applies
| to imports also. For the water portion of dietary exposure it is recognized
that the potential for residues is not constant nationwide. The national
food estimate is combined with regional water assessments to provide a
series of regional dietary assessments.
viii. Some Residue Data are Under Consideration But Not Included
in This Assessment
A task force of pesticide producers has provided the Agency with an
OP pesticide market basket survey. The results of this market basket
survey, conducted in 1998, were submitted to the Agency in 2001. In this
survey 13 foods were analyzed for 29 OP analytes. Samples were taken
from grocery stores and single-serving size homogenates analyzed by
methods with very low limits of detection. The foods collected, all of which
I.C Page 23
-------
are also covered by PDF, were apples, broccoli, cherries, cucumbers,
green beans, grapes, peaches, sweet corn, lettuce, oranges, potatoes,
strawberries, and tomatoes. Preliminary examination of the data indicate
that cumulative exposure assessment using market basket survey data
are in general agreement with a similar assessment using POP data. The
impact of these data on the OP cumulative risk assessment are not
included in this assessment.
.
•Kvwv
w.v>"
I.C Page 24
-------
I I. Revised OP Cumulative Risk Assessment
I D. Residential OP Cumulative Risk
E
I 1. Introduction
f The Office of Pesticide Programs (OPP) has used a calendar based model
= (Calendex™) to address the temporal aspects of the residential use of pesticides
I in 7 geographic regions throughout the United States. These regions, based on
[ major crop growing areas and their influence on surface and ground water, also
I present an opportunity to consider the unique climate patterns, pest patterns and
I potential socioeconomic patterns that influence residential pesticide use and
I expected exposure.
s
Calendex™ allows the OPP to delineate the critical timing aspects of
seasonal uses of Organophosphate (OP) insecticides that result in exposure to
pesticides. Calendex also enables OPP to identify potential co-occurrences from
multiple sources. This includes the exposure from home lawn and garden
treatments, pesticides used on golf courses and applications made by
governmental entities for the control of public health pests such as wide area
mosquito sprays.
1 In nearly all cases, the residential exposure scenarios were developed using
CD | proprietary residue and exposure data. Exposure factors such as breathing
=y? | rates and durations of time spent indoors or outdoors were taken from various
JU I references including US EPA/ORD/NERL Consolidated Human Activity
(D | Database (CHAD), and the Agency's Exposure Factors Handbook (USEPA,
>> f 1997a). In this assessment, the full range of exposure values - expressed as
I uniform, log-normal or cumulative distributions -- are used, where appropriate,
I rather than relying solely on measures of central tendency. While the dietary
I and drinking water assessment address only the oral exposure route, the
I residential assessment considers the dermal and inhalation exposure routes as
| well as the oral route based on the mouthing behavior of young children.
| EPA registered labels, while useful for establishing site/pest relationships and
1 recommendations for applications, generally cannot provide the temporal
| aspects of regional pesticide use. Thus, OPP has relied on other sources of
I pesticide use information, including the National Home and Garden Pesticide
| Use Survey (NHGPUS) data and information available in State Cooperative
| Extension Service publications. These data resources were comprehensively
I used to identify information such as frequency of applications, the type of
I application equipment used, and the type clothing worn while making those
| applications. State Cooperative Extension Service recommendations were used
| to establish regional windows of pesticide applications based on the observed
| appearance of insects such as white grubs on lawns. For example, the timing for
| the treatment of white grubs occurs during mid-July in southern Texas (Region F-
I.D Page 1
-------
s
;v;v;;.;
Lower Midwest) and mid-August in areas such as New York (Region D - North
East/ North Central).
2. Scope of Regional Assessments
The residential and drinking water assessments were developed for 7 distinct
geographic Agricultural Production Regions (Figure I.D-1). .EPA included ten OP
I pesticides with residential uses and potential for significant exposures in its
I assessment. Not included in the cumulative assessment were certain OP uses
that result in low exposure and uses for which risk mitigation actions have been
taken.
Two OP pesticides are currently registered for use on pets, tetrachlorvinphos
(TCVP) (shampoo/dip and flea collar) and Dichlorvos (DDVP) (flea collars). OPP
had insufficient data on DDVP or TCVP exposure for flea collar uses to include
them in a calendar-based probabilistic assessment. However, OPP did assess
TCVP through the shampoo/dip and powder use and these results are
incorporated in this cumulative assessment.
Other OP uses were not included because they resulted in low exposures or
because their single chemical REDs showed low risk. These low exposure uses
include ant baits, paint additives and post application residential exposure from
sod farm application of pesticides. Ant baits are contained inside enclosed
packages. The treatment of individual fire ant mounds has very low applicator
exposure and the reentry or significant play on fire ant mounds is unlikely. Low
exposure is expected also because the treatments often take more than one day
to produce results.
In case of paint additives, the diazinon additives in outdoor paints result in
low potential for exposure because of the complexity of the paint/pesticide matrix
as well as the dilution of airborne concentrations in the outdoor environment. For
sod farm uses, post application exposure is mitigated by rapid dissipation of
residues, residue removal during harvesting (cutting, rolling or stacking), and
transportation. Installation of the sod requires considerable site preparation
which is followed by watering in, further lowering potential for significant
exposure in a post application scenario. OPP believes that children are unlikely
to enter the lawn area immediately following the sod installation.
Finally, for wide-area public health treatments, the more significant uses such
as fenthion, malathion and naled were included. Chlorpyrifos use for mosquito
control was not included because very low exposures were estimated in the
single chemical, screening level assessment.
I.D Page 2
-------
Figure I.D-1. OP Pesticide Cumulative Assessment Regions
OP Pesticide Cumulative Assessment Regions
1S>tQ. KO ED). 2002
I.D Page 3
-------
3. Residential Scenarios
The Residential Scenarios addressed in this document represent critical OP
uses that have the potential for significant exposure or risk when considered in a
cumulative assessment. These are:
Q Golf course and lawn care applications,
Q Home gardens,
Q Wide area Public Health sprays,
Q Pet Treatments (includes aerosol, liquid, and powder uses.), and
Q Impregnated pest strips (limited to closets and cupboards.).
Table I.D-1. Summary of Changes Between December 3, 2001 Preliminary
Cumulative Risk Assessment and Revised Cumulative Risk Assessment
Uses Included in the Preliminary Assessment
That Have Been Removed from the Final Assessment
Use Scenario
DDVP Crack and Crevice
DDVP Pest Strips
(currently under mitigation)
Malathion Lawn Spray
Malathion Golf
Malathion Vegetable Garden Dust ,
Trichlorfon Lawn Care Spray - Applicator
Scenario
Rationale for Change
Registrant is presently generating data
This assessment was limited to use of a small
pest strip in closets and cupboards only. The
uses in attics, basements, and garages were not
considered in this assessment.
Registrant is no longer supporting this use.
Registrant is no longer supporting this use.
Registrant is no longer supporting this use.
This use has been limited to professional
application by lawn care operators only. Only
post-application scenarios will be considered.
Uses included in the Final Assessment That Were not included in the Preliminary Assessment
Use Scenario
TCVP Aerosol, Powder, Pump Spray
Rationale for Change
Pet collars were not included in this assessment
and are believed to pose less risk than the
aerosol, powder, and pump spray uses
considered in this assessment.
I.D Page 4
-------
H
a. Golf Course and Lawn Treatments
Golf Course
Five OPs are registered for use on golf course fairways, greens and tees
and/or residential lawns. Of the five pesticides, four may be applied on golf
courses (Malathion golf course use is no longer supported by the registrant
and was thus not included in the assessment). These pesticides are
acephate, bensulide, fenamiphos, and trichlorfon. Acephate is used for
surface feeding insects, like the chinch bug, which invade primarily warm
season grasses such as St. Augustine grass. Bensulide is used for
germinating weeds such as crabgrass on fairways, greens, and tees.
Fenamiphos is a nematicide and is also watered. Trichlorfon is used for sub
surface or thatch dwelling insects such as white grubs.
Lawn Treatments
On lawns, two pesticides may be applied by homeowners or by
professional LCO. These pesticides are bensulide and trichlorfon. Bensulide
is an herbicide used to control germinating weeds, and trichlorfon is labeled
for insects such as white grubs, which damage turf when present in
significant numbers. Both of these pesticides need to be watered in for
effective control. Malathion is also registered for use on lawns applied as
surface sprays to control nuisance pests such as fleas; however, this use is
no longer supported by the registrant and was not included in the
assessment.
b. Home Gardens
The home garden scenarios include ornamental and edible food gardens
(including home fruit orchards). Due to the wide variety of plant/pest
relationships that can exist in any given region, it was assumed that
applications could be made throughout the growing season for a given area.
Acephate and disulfoton are insecticides that have systemic properties and
appear to be more widely recommended in the cooperative extension
publications. However, malathion continues to be recommended for aphids
by most cooperative extension services. In addition to use on ornamental
gardens, malathion is also registered for use on home vegetable gardens and
orchards.
c. Public Health Uses
Residential exposure from aerial and ground based applications for the
control of public health pests made by regional or state personnel was
addressed in this assessment. Malathion, fenthion and naled are applied to
control mosquitoes. Fenthion is also applied to control black flies.
d. Indoor Uses
DDVP is the sole OP pesticide with indoor registrations. DDVP is used as
both a crack and crevice spray and as a pesticide impregnated pest strip for
the control of flying insects. Since OPP is currently in negotiations with the
registrant regarding the use of DDVP crack and crevice applications, this
scenario was not evaluated. The DDVP pest strip scenario, however, was
I.D PageS
-------
evaluated but was limited to use of smaller strips to control insects in closets
and cupboards in which strip replacement occurred every 4 months.
e. Pet Uses
TCVP was evaluated in this assessment as an aerosol, pump, or powder
flea and tick treatment for pets. TCVP is also available in impregnated from
in pet collars. This assessment considered only TCVP pet treatment using
the aerosol, pump, or powder form (and not the impregnated collar form), as
these uses are believed to result in equal or higher exposures than the pet
collar use. This is based on the assumption that shampooing a dog will result
in greater exposure than merely securing a collar around a dog's neck. Post
application exposure to the collar is also expected to be lower due to a
smaller area being treated .(area around the neck rather than the whole
body).
4. Exposure Routes Considered
k- The routes of exposure considered in this cumulative assessment varied
depending on certain application and post-application exposure activities which
were determined to be age group-specific. The results of exposure are
described in detail below:
Post-Application Oral Route of Exposure: Oral ingestion via hand-to-mouth
activity of children was the only oral route of exposure considered in the
residential portion of this assessment. Specifically, oral hand-to-mouth ingestion
was considered only for the age groups Children 1-2 and 3-5 for their activities
on treated lawns. OPP acknowledges that there is very limited data on exposure
to very young children; in general, however, children ages six and older no
longer exhibit mouthing behavior to the degree seen in younger children such as
placing hands and /or objects into the mouth. In addition, while OPP recognizes
that non-dietary pathways other than through hand-to-mouth activities do exist
such as ingestion of soil and mouthing of grass, these latter two pathways are
not considered because they had little impact on the exposure assessment when
they were addressed in the individual chemical OP risk assessments.
Post-Application Dermal Route of Exposure: The dermal route of exposure
was considered for both children and adults; however, the calculation for children
adjusted by the appropriate surface area to body weight ratio. Children are
considered in a separate group from adults because of the potential for
additional exposures that result from a higher skin surface area to body weight
ratio. In general children six and older have a surface area to body weight ratios
that are similar to adults.
Post-Application Inhalation Route of Exposure: The inhalation route of
exposure was considered for adults and children.
5. Data Sources
Three basic types of data were considered in this assessment: pesticide use
data, residue concentration and dissipation/decay data, and residue contact and
exposure factor data. Together, this information can be used to predict the
potential for co-occurrence of exposure events in aggregate and cumulative
I.D Page 6
-------
assessments. These data are described in more detail below and in Table I.D-2
(by application scenario).
I Pesticide Use Data: Pesticide use information is critical to establishing
| windows of potential exposure when using a calendar-based exposure model.
| This information is needed to predict what pesticide will be used, the amount of
| pesticide which will be used, when the application will be made, how many times
I the pesticide will be applied (and for how long), and whether the applicator will
I be a professional or not. Other data such as frequency of applications, types of
T~" | application equipment used, and types of clothing worn while making the
| applications are also used in developing exposure scenarios.
| Several references were used to determine the application timing for lawn
I care pesticides and to estimate the number of pesticide users. To determine the
I percent of households that employ professional lawn care operators (LCO), the
I Agency used the 1996-1997 National Gardening Survey (Butterfield, 1997)
| conducted by the Gallup polling organization. For specific chemicals, regional
I percent of lawns treated were taken from the National Home and Garden
! Pesticide Use Survey (IMHGPUS) (USEPA, 1992). Two other data sources, Kline
I Professional Markets for pesticides and Kline Consumer Markets for pesticides,
1 were also used to check/confirm the NHGPUS estimates/data.
1 Residue Concentration Data: Residue concentration data and associated
1 pesticide decay/dissipation parameters are used to define the sources and
I magnitude of exposure resulting from human contact.
I Exposure Factor (Contact) Data: Exposure factors such as the amount of
| time spent in an area, whether the exposure is occurring indoors or outdoors,
1 and whether the residue source is a golf course or a lawn (and if the latter, its
I size) are critical for estimating exposures to a given substance.
I For example, an important variable for estimating home-owner applicator
1 exposure is the size of the lawn. OPP considered the average and median lawn
I sizes reported in a journal article by Vinlove and Torla (1995). The means and
| medians were ~13,000 ft2. However, the authors noted problems, interpreting
| the data since it is based primarily on low income houses and consists of
I adjustments of the lot size by the house's foundation (footprint) only. The data
I do not consider other structures such as decks or other green space such as
| gardens, which can reportedly reduce the lot size by up to 50%. Similar lawn
I sizes were noted in an extensive survey conducted by the Outdoor Residential
I Exposure Task Force (ORETF) with similar problems encountered with respect
I to confounding variables such as decks and other green spaces. For this
I assessment, OPP used a uniform distribution for lawn size bounded by 500 ft2
I and 15000 ft2.
-
I Another important variable for addressing post-application exposure from
| home lawn treatment is the duration of time spent on lawns. In this OP CRA,
I cumulative distributions of durations on lawns of up to two hours were used to
| address adult exposure on lawns. These data are presented in Table 15-64 in
| EPA's Exposure Factors Handbook; however, OPP notes that the percentiles
I above the 95th have the same values (121 minutes). A similar cumulative
I distribution was given for children ages one to four. In order to be protective of
I . children and to address the uncertainty of the upper percentiles of the exposure
I.D Page 7
-------
>• e
V?:.
..v
factor data, OPP selected a cumulative distribution from the Exposure Factor
Handbook's Table 15-80 with a bound of 3.5 hours for children.
This distribution represents the amount of time spent outdoors. This allows
for the time that children spend outdoors not only at home but also in parks and
nearschools.
6. Lawn Care Exposure Data
a. Lawn Applicator Dermal and Inhalation Exposure Data
Residential applicator exposure was assessed for the applicator scenarios
used in this assessment (i.e., commercial/professional applicator exposures
were not included in the assessment). Both dermal and inhalation exposures
were considered. Briefly, dermal exposures were calculated as the product
of the Unit Exposure (mg/lb ai handled), application rate (Ibs ai/ft2), and area
treated (ft2). Unit exposure and area treated were inserted in the calculation
as a log normal distribution and uniform distribution respectively, and
application rate as a point estimate. Inhalation exposures to applicators were
entered in the assessment as a uniform distribution bounded by high and low
measured values.
Data concerning Unit Exposures (UE) (through both the dermal and
inhalation routes) were generated by the ORETF. Specifically, this data
consisted of exposure data from 30 volunteers using a push-type rotary
spreader to apply 50 Ibs of dacthal product to treat 10,000 ft2 of turfgrass.
Exposure data from these studies were used to generate normalized values
expressed as milligrams exposure per pound of active ingredient of a
pesticide handled (referred to as UE). Volunteers participating in these
exposures studies were adult non-professionals who use pesticides on their
own gardens and lawns. Many of the volunteers selected as subjects in
these studies were members of garden clubs. All volunteers made their
applications without specific instruction from the study investigators. Unit
exposures from these studies were available for various clothing scenarios
that consider individuals wearing short pants and short sleeved shirts, to long
pants and long sleeved shirts. For this assessment, OPP assumed that all
applications were performed using short pants.1 Based on the Unit Exposure
values generated in this study, UE's used in this assessment for the dermal
and inhalation exposures were as follows: (i) for dermal exposure, a
lognormal distribution with arithmetic mean of 0.69 mg/lb ai handled and
arithmetic standard deviation of 0.36 mg/lb ai handled, truncated at the
estimated 99th percentile of 1.93 mg/lb ai handled and (ii) for inhalation
exposure, a uniform distribution bounded by the low and high measures
values of 0.00019- and 0.0096 mg/lb ai handled, respectively.
1 A survey conducted by Doane and Gallup (Johnson et al., 1999) on behalf of the ORETF identified
70% of those who treat their lawns wear short sleeved shirts while applying granular formulations.
Likewise, 32% reported wearing short pants while applying granular formulations. Sensitivity analysis
performed by OPP demonstrated that significant differences in Unit Exposures existed only between
long pants and short pants, and that no significant differences existed between any of the other clothing
scenarios (i.e., short- vs. long-sleeves did not impact estimated exposures). However, these significant
differences in unit exposure between long pants and shorts had negligible effect on total MOEs, and
thus, for this assessment, OPP assumed that all applications were performed using short pants.
*
I.D Page 8
-------
Table I.D-1. Pesticides and Use Scenarios Considered in the Residential/Non-Occupational Regional
Assessments
f><&m$e
Acephate
Bensulide
DDVP
Disulfoton
Fenamiphos
Fenthion
Malathion
Naled
TCVP
Trichlorfon
<3nff{JdMr$£
Used in Regions
A, E. F, and G
Used in Regions
A, C, D, E, F, and G
None
None
Used in Regions
A, C, E. F, and G
None
None
None
None
Used in Regions
C, D, E, F, and G
UWftO&f*
None
Used in Region F
None
None
None
None
None
None
None
All Regions
HOWfc G3Jttef&
Edible Foods: None
Ornamentals: All Regions .
Edible Foods: None
Ornamentals: None
Edible Foods: None
Ornamentals: None
Edible Foods: None
Ornamentals: All Regions
Edible Foods: None
Ornamentals: None
Edible Foods: None
Ornamentals: None
Edible Foods: All Regions
Ornamentals: All Regions
Edible Foods: None
Ornamentals: None
Edible Foods: None
Ornamentals: None
Edible Foods: None
Ornamentals: None
puwtmm*
None
None
None
None
None
Used in Regions
A, and G
Used in Regions
A, D, E, F, and G
Used in Regions
A and D
None
None
P$st$«1p 1
None
None
All Regions
None
None
None
None
None
None
None
pauses
None
None
None
None
None
None
None
None
All Regions
None
I.D Page 9
-------
The application rate used in the assessment was taken as a point
estimate equal to the maximum (label) application rate. For lawn size, OPP
selected a uniform distribution of lot sizes ranging from 500 to 15,000 ft2.
This range considers smaller lawns for residences such as town houses.
Information in a survey conducted by the ORETF also indicates that many
pesticide users make spot treatments of insecticides. The upper bound of
15,000 ft2 (-1/3 acre) appears reasonable given the type of application
equipment assumed to be used by residential applicators (rotary granule
spreaders). Information on frequency and timing of applications for
pesticides were obtained from Representative Cooperative Extension Service
publications and are described in each of the region specific section in Part II
of this assessment.
I.D Page 10
-------
Table I.D-2. Scenario-Specific Residential Exposure Data Inputs
Pa«8*x**f
v#6$
AS$trtftptfOP* , '
trjjmiFojwat
&a*4$68r
-------
Parameter
Value
A**w*«0i*
top* form*
p*t*$m,r<*
Ornamental Granular incorporated Treatment-Disulfoton
Unit
Exposure
-inhalation
-dermal
Application rate
Frequency of application
0.00001 mg/lb ai
mean of 0.18 and a
SD of 0.29 mg/lb ai
(trunc. at 99th%tile)
label
1-3
Based on 1/2LOQ
Distribution assumes the applicator is wearing
short pants and short sleeved shirt.
1 app.=63%, 2 app.=32%, 3
app.=5% at six-week intervals
point value
Lognormal distribution
point value
(Merricks, 2001)
(Merricks, 2001)
ORETF with the National
Garden Survey
0DVP-Pesi Strips i
Concentration in Air
0.005-0.11 mg/m3
over 1 20 days
samples taken at 1, 7, 14, 28, 56, and 91 day
intervals and adjusted proportionately to account
for smaller strips than measured in Collins and
DeVries, 1973.
uniform distribution
reflecting the range
each sample day
Collins and DeVries,
1973
.I.D Page 12
-------
;v.w$t
b. Post-application Dermal and Non-Dietary Exposure Data
i. Dermal Exposure-Residue Contact Data
The fate of pesticides applied to turf, and subsequent human contact,
is a key variable for assessing post-application dermal exposure and can
be an important exposure pathway to consider as part of a cumulative
assessment. This exposure pathway was evaluated here in the OP
Cumulative Risk Assessment by using data from a number of available
studies (described in more detail below). Briefly, post-application dermal
exposure (mg pesticide) is calculated by multiplying the transfer
coefficient (cm /hr) derived from literature and other studies by the time
spent on the lawn (hr) and the residue concentration on the lawn
(mg/cm2). For this assessment, the transfer coefficient and the time spent
on lawn were represented by a distribution of values while the residue
concentration on the lawn was represented by a time series of
concentration values (which accounted for residue degradation over time).
The transfer coefficients used in this equation were developed by dividing
the hourly dermal exposure (ug/hr) obtained from a set of activities by the
measurement commonly referred to as turf transferable residues (TTR)
(ug/cm2). Since none of the dermal exposure studies used to estimate
hourly exposure in the above chemical specific residue studies permitted
direct calculation of the TTR, the transfer coefficients for this assessment
were instead for this assessment developed by assuming a transfer
efficiency of 0.5% for granular formulations and 1% for spray formulation.
This was done for two reasons:
Q to make use of available dermal exposure measurements in the above
studies which are not influenced by TTR method, and
Q to make use of the available residue dissipation data for which there
are no corresponding dermal exposure transfer coefficients.
The values of 0.5% and 1% are within the range of efficiency for the
existing chemical specific TTR data. To account for the additional
uncertainty of assuming a certain'transfer efficiency to develop the
transfer coefficients, TTR data having transfer efficiencies lower than
0.5% (granular) or 1% (spray) were adjusted upwards to make up the
difference in efficiency. If the transfer efficiency of the TTR data was
higher than 0.5% for granular formulations or 1% for spray formulations,
they were not adjusted.
For a more detailed discussion of the relationship of transfer
coefficients and TTRs please refer to the "Overview of Issues Related to
the Standard Operating Procedures for Residential Exposure
Assessment" presented to the FIFRA Scientific Advisory Panel on
September 21, 1999.
Using the above-indicated calculation methodology, several exposure
studies were used to assess post application dermal exposure to
individuals reentering treated lawns. Separate studies are available, and
used, for kids and adults. These studies are described in additional detail
below:
I.D Page 13
-------
I Children's Exposure: Two studies were used to assess exposure to
| kids under granular and spray application scenarios. One study
I (Black, 1993) investigated dermal exposure values of young children who
I are exposed to a non-toxic substance used to represent a spray
| application scenario. In this study, children performed unscripted
I activities on turfgrass treated with a non-toxic substance used as a
i whitening agent in fabrics. The subjects of the study were 14 children
( aged four to nine years old. The children performing the unstructured
I activities were provided toys and were observed in the treated area for a
I period of one half hour. Activities recorded included the following
i classifications:
Q Upright (standing, walking, jumping and running)
Q Sitting (straight-up, cross legged, kneeling, crouching and crawling)
Q Lying (prone or supine)
Dermal exposure was measured by fluorescent measurement
technology described in Fenske et al., (1986). Measurements on various
body parts were expressed as /wg/body part (e.g., hand, face, etc.) and as
concentration (^g/cm2). These concentrations were normalized to
represent the surface area of children three to four years of age for use
with a standardized body weight of 15 kg. Standard surface area values
were taken from the Agency's Exposure Factors Handbook.
In a second study (Vaccaro, 1996) in which a granular formulation was
used, seven adults performed structured activities intended to mimic a
child's activities. These activities included:
Q Picnicking
Q Sunbathing
Q Weeding
Q Playing frisbee
Q Playing touch football.
The subjects performed these activities for a period of four hours
beginning after the turf had dried. Turf had been treated earlier with a
granular form of chlorpyrifos and exposure was estimated in the study by
monitoring the amount of a chlorpyrifos metabolite - 3,4,5, 6-TCP -
excreted over the following period of 6 days. This method directly
measures internal dose and was used to back-calculate a generic "to the
skin" transfer coefficient by using chemical specific dermal absorption
data for chlorpyrifos (Nolan et al., 1993) These data were further adjusted
to account for differences in surface area of adults vs. children.
The transfer coefficients (cm2/hr) for children estimated from these two
studies are summarized below in Table I.D-3:
I.D Page 14
-------
I Table I.D-3. Transfer Coefficients for Dermal Exposure to Lawn and Public Health
[ Uses for Children 1-6 Years of Age
Vacarro - Granular
{scripted)
714
1042
1042
1485
1736
2758
4785
Black - Spray
{unscripted}
2844
3594
3776
4051
4103
4357
4902
6812
8395
8746
9119
9885
10713
16008
A lognormal distribution was used to fit these transfer coefficients
values and an arithmetic mean and standard deviation for each
distribution was calculated2. Specifically (for children's exposures) the OP
cumulative assessment used a distribution for the transfer coefficient
defined as a lognormal distribution with an arithmetic mean of 7265 cm2/hr
and a standard deviation of 4621 cm2/hr for the spray application. For
the granular application, the distribution used to define the transfer
coefficient was a lognormal distribution with an arithmetic mean of 2225
cm2/hr with a standard deviation of 2162 cm2/hr. In each case, the
lognormal distribution was truncated at the calculated 99th percentile of the
distribution (i.e., 23,769 cm2/hrforthe spray application and 10,623 cm2/hr
for the granular application) in order to avoid a distribution which
contained values that were well-beyond those that are deemed
reasonable.
Adult Exposures: The Vaccaro study data discussed above were also
used to assess exposure to adults under granular and spray application
scenarios. These data are presented below in Table I.D-4:
= 2 See Appendix 3 of "Guidance for Submission of Probabilistic Human Health Exposure Assessments to
I the Office of Pesticide Programs [draft dated 11/4/98] available at
| http://www.epa.goV/docs/fedrgstr/EPA-PEST/1998/November/Day-05/6021 .htm for more information.
I.D Page 15
-------
I Table I.D-4. Transfer Coefficients for Dermal Exposure to Lawn and Public Health
Uses for Adults 18 Years of Age and Older
Vacarro -Spray (scripted)
3348
6770
7217
8779
9895
11243
13169
13243
Vacarro - Granular (scripted)
1229
2813
2813
4010
4688
7446
12920
A lognormal distribution was used to fit these transfer coefficients
values and an arithmetic mean and standard deviation for each
distribution was calculated (see footnote 2 in this chapter). Specifically
(for these adult exposures), the OP cumulative assessment used a
distribution of values for the transfer coefficient characterized by a
lognormal distribution with an arithmetic mean of 9,784 cm2/hr and a
standard deviation of 5,515 cm2/hr for the spray application. For the
granular application, the distribution used for the transfer coefficient was
characterized with a lognormal distribution with an arithmetic mean of
6,370 cm2/hr with a standard deviation of 7,789 cm2/hr. In each case, the
lognprmal distribution was truncated at the calculated 99th percentile of the
distribution (i.e., 28,907 cm2/hr for the spray application and 37,250 cm2/hr
for the granular application).
ii. Non-Dietary Exposure Data Hand-to-Mouth Behavior
The assessment also incorporated exposure from hand-to-mouth
activity by children on lawns. Briefly, exposure through this pathway is
calculated as the product of the following factors: hand-to-mouth contact
frequency (hr1), surface area of inserted hand parts (cm2), saliva
extraction efficiency (unitless), wet hand adjustment factor (unitless), and
hours spent on lawn (cumulative distribution)3.
= 3 The cumulative distribution used for hours spent on lawn by children was obtained from the Exposure
1 Factors Handbook and represents a cumulative distribution for "do-ers" only, i.e., a cumulative
| distribution for only those children that reported spending at least SOME time on the lawn (i.e., it does
not consider that some children on any given day DO NOT spend time on the lawn). Thus, the
cumulative distribution assumes that some time is spent on the lawn by each child. To the extent that
this overestimates time spent on the lawn, this overestimates exposure by this pathway. On the other
hand, this cumulative distribution for time spent on the lawn is not stratified by season. To the extent
that children spend time on the lawn during the seasons when applications occur, this may
underestimate exposure: On balance, however, OPP believes that the distribution used is a reasonable,
yet conservative estimate of time spent on the lawn during the relevant portions of the year.
I.D Page 16
-------
Surrogate data to evaluate non-dietary ingestion through hand-to-
mouth behavior in young chijdren consist, in part, of observations reported
in Reed et al., 1999 concerning the frequency of hand-to-mou.th activity.
This study addressed the mouthing behavior and other observations of
children situated indoors, ages three to six at day care (n=20) and children
ages two to five at home (n=10). The children were video taped and the
frequency of hand-to-mouth events were enumerated after the taping. The
hourly frequencies of the hand-to-mouth events reported were a mean of
9.5 events per hour, a 90th percentile of 20 events per hour and a
maximum of 26 events per hour. These data were used to construct a
uniform distribution to represent the frequency of hand to mouth activity
bounded by a low value of 0 events/hr and a high value of 20 events/hour.
The observations reported by Reed, and discussed above, are based
on children in real world settings. However, they provide little information
regarding the characterization of the hand-to-mouth event, residue
transfer efficiency, or extraction efficiency of the residues on the hands by
saliva during the mouthing event. For these values, additional
assumptions and studies to address the transfer efficiency of turf residues
by wet hands are needed. Variables addressing this exposure pathway
are discussed in the following below:
I Q Based on previous conversations with the SAP, each hand-to-mouth
1 event has been estimated to equal one to three fingers or 6.7-20 cm2
I per event. To account for the fact that a child may touch nothing
I between successive events, and the fact that the event may not result
I in insertion of fingers at all (Kissel et al., 1998), a uniform distribution
I of 0 to 20 cm2 per event was assigned.
i
I Q Hands wet from saliva are reportedly more efficient at residue transfer
I than dry hands. A uniform distribution of transfer efficiency multipliers
I of 1.5 to three times was selected to address the increased efficiency
| of wet hands. Wet hands had higher transfer efficiencies than dry
I hands and other TTR methods addressed in a study performed by
I Clothier et al., 1999. The TTR methods used in the study had similar
1 efficiencies as the chemical specific lawn residue data (TTR data)
I used in this assessment.
E
| Q To address the removal of residues from the hands by saliva during
| the mouthing event several studies were considered. The removal
I efficiency of residues on hands by saliva and other substances (e.g.,
I ethanol) suggests a range of removal efficiencies from 10% to 50%
I (Geno et al.,1995; Fenske and Lu 1994; Wester and Maibach 1989;
| Kissel etal., 1998). Thus a uniform distribution of 10% to 50% was
| used in this assessment.
:
I Q The time spent on the lawn was estimated as a cumulative distribution
1 ranging from 0.25 hours to 3.5 hours. This data was obtained from the
1 Exposure Factors Handbook and represents children aged 1 to 4
I years old. To be protective of children and to address the uncertainty ,
I of the upper percentiles of the exposure factor data, OPP selected a
I cumulative distribution from Exposure Factors Handbook Table 15-80
| with a bound of 3.5 hours for children. This distribution represents the
I.DPage17
-------
amount of time spent outdoors. This allows for the time that children
spend outdoors not only at home but also in parks and near schools.
The percent contribution to total exposure via non-dietary ingestion
continues to be difficult to quantify. This includes the variables discussed
above as well as issues regarding the utility of using children's hand-to-
mouth frequencies based on indoor activities for outdoor exposure
scenarios. There are also differences in mouthing behavior based on
active and quiet play with increased mouthing likely to be during activities
of quiet play. Limited data evaluated by Groot et al.,1998 suggests' there
can be longer durations of mouthing activities for children aged six to 12
months (exceeding 160 minutes per day) than children 18 to 36 months
(up to 30 minutes per day). However, children in this age group are not
likely to be engaged in the higher post application lawn activities which
^ OPP is currently modeling. Additional data for very young children (under
the age of two) are needed in addition to delineating the frequency
differences between hand-to-mouth events for children engaged in active
and quiet play. The Agency recognizes this is an evolving field of study
and that additional research is also needed to evaluate the distribution of
behaviors across different age ranges with a view towards the influence of
factors such as socioeconomic status.
7. Home Garden Applicator and Post Application Exposure Data
The US EPA National Home and Garden Pesticide Use Survey (1992), as
well as various proprietary data sources were used to estimate dermal and
inhalation exposure of individuals applying OPs to ornamental gardens, fruit and
vegetable gardens, and home orchards. In addition, post-application dermal
exposures were assessed for individuals harvesting or performing post
application maintenance activities in home fruit and vegetable gardens and
orchards. Both applicator and post-application scenarios are described in
additional detail below.
Applicator Exposures: As described for dermal lawn applicator exposure,
dermal exposures to applicators in home garden scenarios were similarly
calculated as the product of the Home Garden Unit Exposure (mg/lb ai handled),
application rate (Ibs ai/ft2), and area treated (ft2). Both Unit Exposure and area
treated were inserted in the calculation as a distribution, while application rate
was entered as a uniform distribution .
For spray applications, Unit Exposure was estimated from a surrogate study
with volunteers applying carbaryl to shrubs and trees using a small tank sprayer.
This data was used in developing unit exposures for application of acephate and
malathion to ornamentals. These data are presented below in Table I.D-5:
I.D Page 18
-------
I Table I.D-5. Applicator Unit Exposures for Using a Hand Pump Sprayer /ug/lb ai
i handled for Ornamental Uses of Acephate and Malathion (also for home
| Vegetable/fruit Gardens (malathion only)
Replicate
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
32
34
36
38
40
Short-Sleeved Shirt, Short Pants
25348.3
51515.8
125828.3
26598.1
354396.6
55550.5
118695.9
173841.6
45160.0
39757.8
46075.7
14886.1
35911.5
81656.0
76548.2
74890.0
46498.0
36582.8
25014.7
63485.8
Inhalation
2.3
2.7
2.0
2.1
3.2
2.9
2.3
9.3
5.7
9.2
2.1
2.3
2.3
2.0
14.2
6.6
2.1
13.4
2.0
10.8
V.VW
For dermal exposures, distributions for Unit Exposure through acephate and
malathion ornamental uses (log normal with an arithmetic mean of 78 mg/lb ai
handled, a SD of 76 mg/lb ai handled, truncated at 99th percentile value of 372
mg/lb ai handled, for application rate (uniform distribution specific to pesticide
being assessed and detailed in Part II of this document), and for area treated
(uniform distribution with a minimum value of 500 ft2, and a maximum value of
2000 ft2) were used. This latter value is based on US census data indicating a
median house area of 2,225 ft2. For this assessment, it was assumed this area
was for one floor having a perimeter of -200 linear feet. The ornamental beds
were assumed to be 2.5 to 10 feet wide.
For granular disulfoton applicator exposures through the dermal route,
chemical specific data measuring exposure of individuals using a shaker can of
disulfoton granules to the soil around roses followed by soil incorporation are
I.D Page 19
-------
available and were used in the OP CRA. Distributions for dermal unit exposure
for applications to shrubs were developed from the following data in Table I.D-6:
Table I.D-6. Dermal Unit Exposures (//g/lb ai handled) for Applicator Using
Disulfoton on Shrubs and Flower Beds
Replicate
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Shrub
134
304
187
150
35.3
172
45.1
16.3
94.6
360
41.1
245
13.9
69.9
161
Note: all inhalation replicates were non-detect/loq - LOQ = 1 .5 /^g.
Specifically, a lognormal distribution with an arithmetic mean of 0.18 mg/lb ai
handled, a SD of 0.29 mg/lb ai handled, and a truncation point at 1.31 mg/lb ai
handled (99th percentile) was used for dermal Unit Exposure and a point
estimate was used for application rate. This point estimate for application rate
was specific to pesticide of interest and is detailed in Part II of this document. A
uniform distribution was used for area treated bounded by a minimum value of
10 ft2, and a maximum value of 2000 ft2. As described above, this value is based
on US census data indicating a median house area of 2,225 ft2. For this
assessment, it was assumed this area was for one floor having a perimeter of
~200 linear feet. The ornamental beds were assumed to be 2.5 to 10 feet wide.
Based on ORETF-submitted data, applicator inhalation unit exposures were
represented by a uniform distribution for acephate and malathion ornamental
uses, and as a point estimate for disulfoton ornamental use. Specifically, Unit
Exposures for acephate and malathion ornamental uses were represented by a
uniform distribution with a lower bound value of 0.002 mg/lb ai handled and an
upper bound value of 0.0142 mg/lb ai handled (which represent the minimum ,
and maximum measured values as per Table I.D-5 above); application rate was
represented by a uniform distribution specific to the pesticide of interest and
detailed in Part II of this document; and area treated was considered as a
I.D Page 20
-------
uniform distribution with a minimum value of 500 ft2, and a maximum value of
2000 ft2 based on US census data indicating a median house area of 2,225 ft2.
For this assessment, it was assumed this area was for one floor having a
perimeter of ~200 linear feet.
For applicator inhalation exposures for disulfoton, point estimates were used
for unit exposure (0.00001 mg/lb ai handled) and application rate. The point
estimate for inhalation Unit Exposure represents !4 the j-OQ, since all measured
_ . inhalation unit exposures were less than the analytical limit of quantitation. The
% ™ point estimate for application rate is specific to disulfoton and detailed in Part II
of this document in the regional assessments. A uniform distribution with
minimum value of 500 ft2, and a maximum value of 2000 ft2 was used to
represent area treated. This value is based on US census data indicating a
median house area of 2,225 ft2. For this assessment, it was assumed this area
was for one floor having a perimeter of -200 linear feet.
Post-Application Exposures: Post-application exposure while harvesting or
performing post application maintenance activities in home fruit and vegetable
gardens and orchards was assessed using a wide range of transfer coefficients
to account for the diversity of gardens and types of activities. Specifically, post
application exposure was estimated as the product of a transfer coefficient
(cm2/hr), time spent in the activity (hrs), dislodgeable residue concentration
(mg/cm2), and the dermal absorption factor (unitless).
For the above calculation, the transfer coefficient was characterized as a
uniform distribution ranging from 100 to 5000 cm2/hr to account for and reflect a
wide range of tasks for gardeners. The time spent harvesting or performing
post-application maintenance activities was represented by a uniform
distribution ranging from 0.0833 hr/day to 1 hr/day. These estimates of time
spent in the garden performing post application activities (as well as the
frequency of applications) were based on survey data performed by the Outdoor
Residential Exposure Task Force (ORETF). Dislodgeable residue
concentrations (expressed in mg/cm2) were expressed as a time series of values
collected from chemical-specific dislodgeable residue data obtained from
studies performed in California (for Western regions) and Pennsylvania (for
Eastern regions) and detailed in the region specific sections in Part II of this
document. Timing and frequency aspects (on both a regional and chemical-
specific) of post-application gardening activities were based on information
available in representative state cooperative extension service publications, and
regional use data was based on information available in the National Home and
Garden Pesticide Use Survey and Kline Professional Markets Reports (1997-
1998).
W-*
8. Golf Courses Post Application Exposure Data
The potential dermal exposure of individuals playing golf on treated golf
courses was estimated using chemical-specific turf residue data and transfer
coefficients derived from surrogate dermal exposure data. Specifically, post-
application exposure to residues from golf courses (in mg) was calculated as the
product of transfer coefficient (cmz/hr), the time spent golfing (hr), and the turf
residue value (mg/cm2). The percent of the population playing golf and the
percent of golf courses that are treated with any specified OP was also
considered and incorporated into the assessment.
I.D Page 21
-------
The surrogate data used to derive transfer coefficients were based on two
measurements of four individuals playing golf on two golf courses treated with
chlorothalonil (Bailee, 1990), and the exposure of golfers (four volunteers) to
flurprimidol (Moran et al., 1987). The data are presented below in Table I.D-7:
Table I.D-7. Golfing Transfer Coefficients (^g/cm2) for Post Application Dermal
Exposure:
Chemical
Chlorthalonil
Flurprimidol
Transfer Coefficient
391
329
561
547
592
533
385
508
756
522
, 264
278
^>>vv>
For both studies, an assumed transfer efficiency of 1 % was used to calculate
the transfer coefficients, since the studies were conducted using spray-able
formulations. Based on these two studies, a lognormal distribution with an
arithmetic mean of 483 cm2/hr and an arithmetic standard deviation of 185
cm2/hr was used to represent the transfer coefficient. This distribution was
truncated at the calculated 99th percentile value of 1066 cm2/hr. The exposure
duration for individuals playing golf was assumed to be a uniform distribution
bounded at the low end by two hours and at the upper end at four hours. The
four-hour value was obtained from a 1992 survey conducted by the Center for
Golf Course Management.
To establish the percent of individuals playing golf, the above-mentioned
1992 study was used. It was reported here that an average of 12.2% of the
population plays golf. To determine the likelihood of playing golf on a treated
golf .course, percent of golf courses treated data provided by Doane's GolfTrak
(1998-1999) was used. These data indicated anywhere from 5 to 85% of golf
courses are treated with any specified OP depending upon the identity of the OP
and the region of use. Additional details concerning the chemical- and region-
specific use patterns used in the estimation of exposures through this pathway
are present in Part II of this document.
. I.D Page 22
-------
^.vw
9. Public Health Post Application Exposure Data
Assessment of post-application exposure to public health sprays was
conducted in a manner similar to the method used to assess post-application
exposure to lawn chemicals. That is, exposures to residues on lawns were
estimated using the same dermal transfer coefficients, hand to mouth variables,
and duration of time spent on the lawn. What differs between the public health
spray scenario and post application lawn exposure scenario is the source
strength of the residues deposited on the lawn from the public health sprays.
The amount of residues that contact and may be present on the lawn can be
predicted from the application rate for the various public health sprays and the
application specifics, such as equipment type and spray nozzle settings. The
percent of the application rate that is deposited on lawns following ground
applications of public health sprays is based on a study by Tieze, et al. (1995)
which measured the percent of the mosquito spray that is deposited on lawns
following ground applications. These deposition values ranged from 3.8 to -5%.
For aerial applications, the percent of the application rate that is deposited on
lawns were calculated using the spray drift model AgDrift which were reported
(an discussed) in the individual risk assessments for malathion, naled and
fenthion. These values ranged from approximately 15 to 30%. To address the
uncertainty regarding the percent of use by ground equipment and or aerial
equipment, a uniform distribution for deposition bounded by 3.8% and 30% was
used. Inhalation exposure to public heath mosquitocide use was not addressed
since there are no refined models to address this scenario. It is also expected
that near-infinite dilution based on the outdoor location mitigates this exposure.
Timing aspects and estimates of percent of use are based on conversations
with representatives of Florida Mosquito Abatement Districts (Whichterman)
Florida A&M (Dukes) and Health Canada (Dr. Burke) for Black Fly. For other
regions having public health spray uses, a spray schedule of once every two
weeks was assumed for the summer season. Additional region-specific details
regarding the application and timing of treatments and chemical-specific details
regarding degradation are presented in Part II of this assessment.
10. Indoor Uses Inhalation Exposure Data
The only OP pesticide registered for indoor use is DDVP. This was assessed
as a resin impregnated pest strip limited to unoccupied areas such as closets
and cupboards. Exposures through crack and crevice treatments with DDVP
were evaluated in the Preliminary Cumulative Risk Assessment, but not evaluate
here in the Revised OP Cumulative Risk Assessment (Revised OP Cumulative)
since OPP is currently in negotiations with the registrant regarding crack and
crevice use.
Furthermore, estimated exposures through the DDVP pest strips were
modified in this revjsed CRA to account for additional mitigation actions being
taken and/or negotiated by the Agency. Specifically, use is expected to be
limited to unoccupied areas such as closets and cupboards with the
corresponding size of the pest strip reduced to account for the smaller spatial
I.D Page 23
-------
volume being treated4. Exposure while handling the impregnated pest strips is
expected to be minimal.
Thus, only post-application inhalation exposure was estimated for adults and
children, with applicator exposure considered negligible and not evaluated.
Briefly, post application inhalation exposures (expressed in mg) were
calculated in the OP CRA using the following equation:
E = Cair x BMR x H x VQ X MET_TIME
E = Exposure through the inhalation pathway (mg)
Cair = residue concentration in air (mg/m3),
BMR= Basal Metabolic rate (MJ/hr) which is specific to a CSFII individual's
age, sex, and weight
H = 0.05 m3/MJ, a constant representing the volume of oxygen consumed
(at standard temperature and pressure) in the production of 1 MJ of
expended energy.
VQ= 27 (unitless), a conversion factor reflecting the ratio between the amount
of air breathed to the amount of oxygen obtained
MET_TIME (hr) which represents a distribution reflecting the sum (over a
day) of the product of an unitless activity-specific metabolic factor and the
amount of time spent in that activity (summed over all activities in a day).
The residue concentration in air (C.^) is represented by a time series of
calculated concentrations in homes using reduced -size DDVP pest strips in
closets and/or cupboards. Specifically, use of a smaller pest strip was assumed
to produce a proportionately smaller air concentration, thus, the air
concentrations in this revised CRA were estimated by multiplying the measured
concentration values found under a "whole-house" scenario following use of an
80 gram (as per Collins and DeVries, 1973) by either 0.26 or 0.066 to represent
use of Pest Strips of 21 g or 5.25 g size.5
The BMR term in the above exposure equation is calculated internally by the
Calendex software and represents a point estimate specific to and calcu|ated for
each individual in the CSFII based on his self-reported age, sex, and weight.
Both H and VQ in the above equation are constants as described above. The
| 4 Mitigation actions that are currently being negotiated do permit uses in additional unoccupied areas
I such as attics, basements, and garages, but for purposes of this cumulative assessment exposures
I through these uses were not assessed.
E 5
(21gor5.25g)/80g
I.D Page 24
-------
MET TIME variable is represented by an age group-specific empirical
distribution and accounts for the fact that an individual's breathing rate and the
specific activities an individual engages in on any given day are NOT
independent. That is, this factor (or term) accounts for the interrelationship that
exists between the activities that an individual engages in and the breathing rate
with which that activity is connected.6 The genesis and derivation of this
MET_TIME variable is discussed in additional detail below.
The MET TIME term: As indicated earlier, OPP in the OP Preliminary Risk
Assessment assumed independence between a person's daily activities, the
place in which these activities are conducted, and the amount of time spent in
these activities. OPP has refined these assumptions in this Revised CRA by
using information on each of the activities that an individual engaged in on that
day; as well as the location and duration spent in each micro environment
(activity-location combination). Thus, this revised CRA appropriately considers
the implicit relationship between a specific activity and its and duration.
Specifically, OPP obtained information on time-activity data from the US EPA
ORD Consolidated Human Activity Database (CHAD)
(http://www.epa.gOv/chadnet1/V This is a recently constructed meta-database
of time use survey data in which time and activity by each individual survey
participant is recorded in a chronological diary format. The database, in total,
consists of 22,968 person diary days from 10 different time use surveys; there
are 875,339 records in total, with each record containing detailed information for
each micro environment (activity-location). Since MET values vary by activity, it
is possible to calculate breathing rates for each distinct micro environment
(reading the newspaper, preparing meals, eating, cleaning the house, etc.) which
are weighted by the amount of time spent in that micro-environment. Therefore,
an individual who reported spending 24 hours indoors in bed (illness) will have
lower indoor inhalation exposure than if that individual had spent 24 hours
indoors engaged in various physical activities (8 hours sleeping, 2 hours
preparing meals, 2 hours exercising, etc.). In this manner, the calculated total
indoor inhalation would be consistent with the information available in the time
use diaries.
OPP generated a set of random MET values for each of the 875,339activities
reported by respondents in the CHAD database which were consistent with the
CHAD-defined distributional form of the activity category. These distribution
functions were developed based on a compilation and review of the published
literature. For example, the MET value for 'Sleep or nap' (CHAD activity code
14500) follows a lognormal distribution, with mean 0.9, standard deviation 0.1;
minimum 0.8 and maximum 1.1. The MET values for 'Prepare and clean-up
§ 6 In the preliminary OP CRA, an average daily indoor 'breathing rate' factor (MET) was assumed for
| each individual. This MET factor was assumed to be uniformly distributed and bounded by 1 and 2 (i.e.,
I MET ~U(1,2)). The time spent indoors (representing the duration of exposure) was drawn independently
| using the empirical distribution published in the Exposure Factors Handbook. That time spent indoors
i and average breathing factors are related was not explicitly considered. As discussed in the main body
| of this document, OPP has refined this calculation using the time-use surveys available from
I CHADS/NHAPES in a more comprehensive manner.
E
I I.D Page 25
-------
food' (CHAD activity code 11110) are exponentially distributed, with mean 2.8,
standard deviation 0.9, minimum of 1.9 and maximum of 4.7
OPP then multiplied each generated MET value by the corresponding
duration during which that activity was undertaken to maintain any correlation
between time spent indoors and corresponding activities. For each individual,
this MET x Time variable was summed over all records in that individual's daily
diary, in which the activity occurred indoors. This value is used in the equation
above to calculate that individual's daily indoor exposure. For each age cohort
(Age <1, 1-2, etc), a frequency distribution of this MET x Time variable was
calculated. The table I.D-8 below presents these distributions for each of the
age cohorts. There was no information on respondent age for 224 of the 22,968
person-days. Included in this distribution were individuals (n=74) who did not
report spending any time indoors (perhaps camping, or on vacation).
Specifically, the table below represents for each of six age groups (children 0-1
years old, children 1-3 years old, children 4-5 years old, etc.) the cumulative
distribution of the MET_TIME variable (e.g., 95% of children 1-3 years old have
METJTIME values of 56 or less, 98% of children in this age group have
METJTIME values of 65 or less, etc.). It was this cumulative distribution that
was used for MET_TIME variable in the above equation.
Table I.D-8. Distribution of MET Time Values, By Age Grou
Cum Pet
N
25%
50%
75%
90%
95%
98%
99%
100%
0-1
563
26
33
41
50
54
59
65
70
1»3
2,171
25
32
41
51
56
65
69
115
4-e
2,088
19
25
32
41
47
53
59
101
Mfc
3,930
16
20
25
32
37
44
49
84
P
13*17
1,192
15
18
23
28
33
39
42
120
18+
12,800
15
21
29
40
49
58
67
130
Use information for the number of households using DDVP pest strips
indoors was taken from the National Home and Garden Pesticide Use Survey,
1991. The use of pest strips was assumed to occur year round with these
7 The amount of information available for specific activities varied across the different activities. These
studies also varied with regards to the methodology and instruments used to measure MET
corresponding to the different activities.
I.D Page 26
-------
-ww
replaced once every 16 weeks. Based in part on information provided in the
National Home an Garden Survey, two percent of the homes were assumed to
use DDVP pest strips. Further, it was assumed that in those homes that used
pet strips, one 5.25 g strip was placed in a cupboard and one 21 g strip was
placed in a closet.
11. Pet Uses
The Cumulative Risk assessment also considered exposures through flea
and tick treatments. There are several products containing TCVP which are
available in aerosol, pump spray, and powder form. TCVP is also available in
impregnated form in pet collars. Exposure assessments were performed for
both applicators and non-applicators (i.e., post-application exposures). For
applicators, both dermal and inhalation routes were considered. For post-
application exposures, only the dermal and oral (hand-to-mouth) routes were
considered. Each of these routes is discussed in additional detail below.
Applicator Exposure Dermal and Inhalation: The data for the applicator
assessment scenarios are based on studies submitted to the Agency which
involved application of a flea and tick products to dogs. In this OP CRA,
applicator exposure was calculated as the product of Unit Exposure (in mg/mg ai
handled), application rate (mg ai handled/lb of animal), animal weight (in Ibs of
animal), and number of pets treated. Each of these terms was represented in
the calculation as a distribution. Unit Exposure (in mg/mg ai handled) was
represented by a cumulative distributions for powder and aerosol/pump spray
formulations. This empirical cumulative distribution is presented in Table I.D-9
for powder an aerosol/pump spray applications
I.D Page 27
-------
Table I.D-9. TCVP Applicator Unit Exposure (mg/mg ai handled) for Pets
Dog
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Pet
6.7%
13.3%
20.0%
26.7%
33.3%
40.0%
46.7%
53.3%
60.0%
66.7%
73.3%
80.0%
86.7%
93.3%
100.0%
Powder
Dermal
(mg/mg ai handled)
0.0016667
0.0017328
0.0021848
0.0022796
0.0023325
0.0023699
0.0024669
0.0028417
0.0030423
0.0034921
0.0040102
0.0040917
0.0050375
0.0052139
0.0149053
Inhalation
(mg/mg ai handled)
0.0000004
0.0000005
0.0000007
0.0000013
0.0000025
0.0000028
0.0000028
0.0000049
0.0000052
0.0000068
0.0000081
0.0000110
0.0000114
0.0000220
0.0000238
Aerosol & Pwrop Spray
Dermal
(mg/mg ai handled)
0.0028700
0.0043400
0.0050300
0.0053300
0.0054400
0.0056000
0.0058800
0.0061700
0.0077300
0.0093800
0.0098400
0.0099600
0.0102200
0.0143200
0.0270900
Inhalation
(mg/mg ai handled)
0.0000001
0.0000080
0.0000090
0.0000110
0.0000150
0.0000150
0.0000170
0.0000230
0.0000230
0.0000280
0.0000280
0.0000280
0.0000460
0.0000470
0.0000550
AWASW
Application rate in this equation was represented by a uniform distribution
depending upon the formulation, as follows:
Q for powder, the application rate was represented by a uniform distribution
bounded by 21 and 25 mg ai handled/lb of animal;
Q for aerosol, the application rate was represented by a uniform distribution
bounded by 11 and 15 mg ai handled/lb of animal; and
Q for pump spray, the application rate was represented by a uniform distribution
bounded by 9 and 10 mg ai handled/lb of animal.
I.D Page 28
-------
I Animal weight and number of pets were each represented by a empirical
I cumulative distribution. Animal weights were drawn from an empirical
I distribution represented in Table I.D-10 and which ranged from 3 to 148 Ibs/pet.
I
f Table I.D-10. Pet Applicator Exposure Variable Dog Weights
Cum, POT of Dogs
1%
10%
20%
30%
40% .
50%
60%
70%
80%
90%
95%
99%
100%
Dog Weight (fos)
3
11
16
20
23
30
43
55
70
80
89
108
148
Source: Boone, Tyler, Chambers: 1999 So I
446584-01; and MRID 446584-01.
Poster session; CarbaryTStudy MRlD
ww.1-*
I.D Page 29
-------
Pet owners were assumed to treat between one and four pets of identical
size as per the information presented in Table I.D-1 1 .
I Table I.D-11. Applicator Exposure Variable Number of Dogs Treated
Cum. PCT of Dog Owner-Apps
50%
75%
90%
100%
Number of Dogs
1
2
3
4
Applicator exposures through the inhalation route were calculated in a
similar manner to the applicator dermal exposures described above, except that
Unit Exposure (in mg/mg ai handled) were specific to inhalation exposures.
These empirical unit exposures through inhalation are also presented in Table
I.D.9, above. All other terms relating to application rates, animal weights, and
number of pets treated remained as described earlier.
Frequency, timing, and probability of TCVP applications to pets were also
considered in the OP Cumulative Risk Assessment. Based on the US EPA
Home and Garden Pesticide Use Survey, less than one percent of the homes
reported using TCVP products for flea and tick treatment on pets. In addition,
this survey reported that between 13 and 19 percent of all households reported
using a pet collar for control of fleas and tick. This estimate, however, includes
all pet insecticide collars, not just those that contain TCVP. These use
estimates are consistent with proprietary marketing data published by Kline,
Incorporated. Since recent estimates for use of TCVP pet collars are not
available, the percent of households applying TCVP flea and tick treatments
was set in the OP CRA at 15 percent, and was assumed to be equally split (at
5% each) between each of the three (powder, aerosol, and pump spray) TCVP
formulations to account for use of both the flea and tick treatments and the
TCVP impregnated collars. This is believed to be protective, since high-end
exposures from flea and tick treatments are expected to be higher than high-end
exposures from pet collars. The household applicators were assumed to reapply
TCVP flea and tick treatments every 8 weeks, with use occurring all year in the
Southern regions (A,C, E, F, and G), and between April through mid-August
(three applications) in the Northern regions (B and D).
Post-Application Exposure - Dermal and Oral (Hand-to-Mouth): Dermal and
oral (hand-to-mouth) post-application exposures from TCVP flea and tick
treatments were also considered in the OP CRA. Dermal exposures scenarios
were considered to be applicable to both adults and children while non-dietary
oral exposure scenarios (oral hand-to-mouth) were assumed to apply only to
infants and children <6 years old.
Dermal Post-Application exposure: Dermal Post-Application exposure (to
adults an children) was calculated as the product of Residue concentration
(mg/cm2), the Transfer Coefficient (in cm2/hr), and the Time spent (in hrs/day).
Residue concentration values on the application day were estimated from the
I.D Page 30
-------
Day 0 residue measurement from a study conducted by Hartz for TCVP
Reregistration purposes. Residues measured on Day 0 (4 hours after treatment)
ranged from 0.224 mg to 0.413 ug/cm2 for the powder formulation, 1.1 to 1.9
ug.cm2 for the aerosol formulation, and 1.2 to 3.5 ug/cm2 for the pump spray
formulation. This information is presented below in Table I.D.12:
[ Table I.D-12. Post-Application Residues on Day 0 (day=0.167), (Empirical
I Distribution)
Obs
Applicator A
Applicator B
Applicator C
Applicator D
Applicator E
Powder (ugjfcm2)
0.413
0.224
0.395
0.299
0.230
ASroSol (ug/crn*)
1.603
1.947
1.750
1.559
1.061
Spray {ug/cm*)
2.433
1.348
1.416
3.595
1.267
.WAV
Memo S. Hanley to D. Fuller, March 22, 2002, Re-Issue: HED^TReview of
Determination of the Dislogeability of Tetrachlorvinphos (TCVP) from the Fur of Dogs
Following the Application of an Insecticide Powder; Pump Spray or Aerosol; MRID
454855-01. C Code 083701; DP Barcode D277543, Submission S597121. Tables 5b,
6b, 7b; 8 and 9 (half life).
These residues were assumed in the OP CRA to persist for a period of 32
days with a half life of 3 days (both as estimated from the submitted study).
Thus, residue value inputs for dermal post-application exposures were assumed
to be a time-series of concentrations values represented initially by a measured
Day 0 value which is dissipated over a 32 day period in a manner consistent with
a half-life of 3 days.
The Transfer Coefficient used in this assessment of dermal post-application
exposures to adults and children was derived from a carbaryl groomer exposure
study in which sixteen different veterinary technicians treated/handled eight dogs
each, over a two to five hour time period. These transfer coefficients are
presented in Table I.D-13 for adults and children and were derived assuming an
average transfer efficiency of 2.97% (calculated as the average transfer
efficiency of powder (0.62%), aerosol (3.3%) and pump spray (5%)) and an
allometric scaling factor to estimate transfer coefficients specific to children.
I.D Page 31
-------
Table I.D-13. Post-Application Transfer Coefficients for Dermal Exposure to Pet
Fur Residues (Empirical Distribution)1
Groomer
WSF
exposure
8796
6199
1408
2914
5667
2527
2,348
2961
1135
14872
1026
13490
4275
4461
1511
111
Duration;
hrs
2.88
2.58
3.07
2.48
3.08
3.18
2.93
2.72
4.03
3.88
3.17
4.05
4.92
3.45
3.03
3.00
MS/hr
3054
2403
459
1175
1840
795
801
1089
282
3833
324
3331
869
1293
499
259
ai
deposited
^gfcrrt2
37.5
31.0
18.6
36.4
32
19
15.9
. 7.75
14.8
28.8
16.6
56.98
25
, 42.25
8.87
48.6
Dislodged:
2.97%
efficiency
assumed *
ugfema
1.114
0.921
0.552
1.081
0.950
0.564
0.472
0.230
0.440
0.855
0.493
1.692
0.743
1.255
0.263
1.443
Average
Transfer
Coefficient
(adults)
Cms/hr
2742
2610
831
1087
1936
1409
1696
4731
642
4481
657
1968
1170
1030
1894
179
1817
Transfer
Coefficient
(children)
Cm'/hr*
1016
967
308
403
717
522
628
1752
238
1660
243
729
433
382
702
66
673
Source Carbaryl Groomer Exposure Study (activity - wash/dip/groom). Each vet tech
treated/handled 8 dogs: held small dogs w/arms and torso; some dogs climbed on
person's shoulders while grooming etc.
^Average transfer efficiency 2.97% =(powder (0.62%) + aerosol (3.3%) +pump spray
The transfer coefficients derived from this study were adjusted by an allometric scaling
factor based on the relative size of children to adults to derive an appropriate transfer
coefficient for children Adult:Child surface area ratio - 2.7:1 (avg. Adult 3169: avg child
1174)
Finally, the time spent in this activity was assumed to follow a triangular
distribution with minimum value of 0.0333 hours, a most likely value of 0.108
hours, and a maximum value of 1.025 hours (as per Freeman et al, JEAEE,
2001,11:501-509).
Oral (Hand-to-Mouth) Post-Application Exposure: Post-application exposure
through the oral (hand-to-mouth) route was also assessed (for children only) in
I.D Page 32
-------
the OP CRA. Specifically, exposures through the hand-to-mouth route were
calculated as the product of the Residue value (in mg/cm2), the surface finger
area (cm2), the frequency of events (hr-1) and the time spent (hr). The residue
value was obtained from TCVP residue studies. Surface finger area (per event)
was assumed to follow a uniform distribution bounded by 0 and 20 cm. The
frequency of events was assumed to follow a triangular distribution with a
minimum value of 0.4 hr-1, a most likely value of 9 hr-1, and a maximum value of
26 hr-1. The time spent with the pet was assumed to follow the same
distribution described above for dermal post-application exposures.
As under the Dermal and Inhalation Exposure Scenarios discussed above,
applications were assumed to re-occur every 8 weeks, with use occurring all
year in the Southern regions (A,C, E, F, and G), and between April through mid-
August (three applications) in the Northern regions (B and D).
12. In Summary
In summary, this assessment relied upon the best available data from all
sources that could be identified. Sources included chemical specific and task
force generated data, as well as data from the scientific literature. When
available, regional distinct residue dissipation data were used for the lawn and
garden uses. Additional Region-specific information is presented in Part II of this
document.
I.D Page 33
-------
I I. Revised OP Cumulative Risk Assessment
E. Water OP Cumulative Risk
1. Introduction: Incorporating Water Exposure Into the OP Cumulative
Assessment
FQPA, passed in 1996, imposed an expansion of the risk assessments for
food use pesticides by requiring that the Agency perform cumulative risk
assessments, i.e., that the Agency assess the risks from different pesticides
having a common mechanism of action and focusing on the likelihood that a
person will be concurrently exposed to multiple pesticides from multiple sources
(food, drinking water, and residential uses). Ideally, data to support the water
side of this exposure calculation would provide information on multiple
pesticides, and their transformation products, collected from drinking water
sources, both surface and ground water, throughout the U.S. at a sufficient
frequency to reflect daily and seasonal patterns of pesticide occurrence in water.
However, due to the great diversity of geographic-, climatic-, and time-dependent
factors that affect pesticide contamination in water, this approach is not possible.
For the organophosphorous (OP) pesticides cumulative assessment, the Office
of Pesticide Programs (OPP) must rely on both available monitoring data and
modeling to develop sufficient data for use in the exposure assessment.
Because drinking water is local, the national exposure assessment for
drinking water must address localized areas of the country where exposure to
OPs may occur due to drinking water contamination. The methods described in
this chapter account for the fact that pesticide concentrations found in drinking
water are not random, but are in large part determined by the amount, method,
timing and location of pesticide application, the physical characteristics of the
watersheds in which the community water systems (CWS) are located, and other
environmental factors (such as rainfall) which cause the pesticide to move from
the location where it was applied.
OPP is using a probabilistic, calendar-based approach to appropriately match
and subsequently combine estimates of pesticide residues in food with estimates
of pesticide residues in drinking water to determine reasonable approximations
of the amount of OP pesticides ingested in the diet on a daily basis. This
approach looks at each individual day of the year and allows appropriate
temporal matching of exposures through food and drinking water on a daily
basis. Each single day assessment serves as a "building block" for the
construction of multiple consecutive day average exposures. This method
accounts for the temporal aspect of exposure to OPs due to expected seasonal
pulses and seasonal use-patterns.
I.E Page 1
-------
\XvW°
To realistically estimate exposures, the assessment must take into account
which OPs can and do occur together in time and place to account for co-
occurrence. Only those exposures which are likely to occur in the same location,
in this case a watershed, are combined. Those exposures that are likely to occur
on different days and in different locations will be separated. Although multiple
OP pesticides may be registered for use on the same site, they may not
necessarily be used at the same time.
Risk is a function of both hazard and exposure, and estimation of the
exposure portion for drinking water requires data on concentrations of the
pesticides in the drinking water and consumption of drinking water for different
demographic populations on a daily basis. Drinking water is locally derived and
concentrations of pesticides in source water fluctuate over time and location for a
variety of reasons. Pesticide residues in water fluctuate daily, seasonally, and
yearly as a result of the timing of the pesticide application, the vulnerability of the
watershed to pesticide runoff, spray drift and leaching, and changes in the
weather. Changes in concentrations also result from the method of application,
the location and characteristics of the sites where a pesticide is used, the
climate, and the type and degree of pest pressure. Given the data needs and
the number of variables that can affect the outcome of the predictive model, it is
apparent that the development of daily distributions of concentrations of co-
occurring OPs in drinking water for various regions of the US is far-reaching in
scope and complexity.
The goal of the drinking water exposure assessment is to provide estimates
of distributions of residues (concentrations in drinking water) for use in
probabilistic exposure assessment that account for
Q daily and seasonal variations in residues over time due to time of
application(s) and runoff/leaching events
Q year-to-year variations due to weather patterns
Q variability in residues from place to place, resulting from the source and
nature of drinking water and from the regional / local factors (soil, geology,
hydrology, climate, crops, pest pressures, usage) that affect the vulnerability
of those sources
Q the potential for co-occurrence of more than one OP in location and time only
when this is likely to happen
I.E Page 2
-------
The section that follows discusses briefly what we know about OP occurrence
in drinking water sources from available monitoring data and how OP residues in
drinking water may be affected by conventional drinking water treatment
processes. Based on the needs of the probabilistic cumulative exposure
assessment and the information gained from this assessment of monitoring data,
OPP designed a drinking water assessment that provides multiple years of daily
residue concentrations from drinking water sources in twelve regions across the
country. These methods, and a characterization of the results of this
assessment, follow the monitoring assessment.
s
:
i 2. What We Know About OP Occurrence in Drinking Water
I The drinking water exposure assessment for the OPs would ideally be
| performed using direct drinking water data, or at least using extensive surface-
| and ground-water monitoring data as a surrogate. With few exceptions, the
I quantity, quality and relevance of available monitoring data analyzed in each of
| the individual OP risk assessments were considered inadequate to support a
| drinking-water exposure assessment. For many of the OPs, limited or no
1 monitoring data are available. For some OPs, no detections were reported from
I a limited monitoring set, but it is unclear whether these non-detects signify a lack
I of transport, or insufficient or non-targeted sampling.
| The first part of this section briefly summarizes available surface-water and
I ground-water monitoring studies that included multiple OP pesticides. Additional
I monitoring data that focused only on a single OP pesticide are summarized in
| the individual chemical risk assessments (available through the Office of
| Pesticide Programs web site at http://www.epa.gov/pesticides/op/status.htm ).
I This is followed by a review of published literature and registrant-submitted
I studies on the effects of water treatment on OP residues in drinking water. The
I section concludes with an evaluation of the extent to which the monitoring data
| fulfill the needs of the cumulative water exposure assessment.
I a. Summary of Monitoring Information
| Evidence from the available monitoring studies confirms that OP
| pesticides do occur in drinking water sources. The frequency of detections is
| generally low, except for chlorpyrifos, diazinon, and malathion, and the
I magnitude generally ranges from sub-parts per billion to a few parts per
1 billion. Significantly greater frequencies of detection occur in the limited
I number of targeted monitoring studies.
I.E Page 3
-------
I These OP pesticides can occur together in the same water source at the
I same time. Chlorpyrifos, diazinon, and malathion are most likely to occur
| together. However, other OP pesticides may also occur with one or more of
| these three in local areas. The USGS NAWQA study detected multiple OP
| pesticides in the same water samples at the same time in almost all of its
| study units. In some instances, up to 7 of the 11 OP pesticides included in
I the monitoring study were detected together (see Appendix III.E.1).
In general, surface water sources are more likely to be vulnerable to OP
contamination than are ground water sources. OP pesticides are found in
streams draining through predominantly urban/residential as well as
agricultural watersheds. Chlorpyrifos, diazinon, and malathion are frequently
detected in urban streams. While the residential uses of Chlorpyrifos and
diazinon are being cancelled, residential uses for malathion remain.
Although monitoring for OP pesticides in treated drinking water is very
limited, the weight of evidence from available studies is that chlorination may
transform the OPs to oxons, sulfoxides, and sulfones, which are of
toxicological concern. A few studies indicate that the oxon transformation
product will be stable in chlorinated water for at least 24 to 48 hours after
treatment.
b. Surface Water Monitoring
Available monitoring has shown that OP insecticides contaminate surface-
water resources from both agricultural and urban use. Maximum contaminant
levels (MCLs) under the Safe Drinking Water Act have not been developed
for the OP pesticides, and OPs will be included on the Unregulated
Contaminant Monitoring List for the first time in 2002. As a result, States and
public water supplies (PWS) have not often included OPs in surface-water
monitoring. Therefore, with the exception of results from the pilot USGS-EPA
Reservoir Monitoring Study, few studies include analyses of OP insecticides
in raw and finished drinking water.
Available surface-water monitoring for OPs represents a range of surface-
water bodies, from agricultural drainage ditches to outflow samples from the
largest rivers in the nation. Monitoring data from bodies such as small
streams may not represent direct drinking water sources, but can give an
indication of possible surface-water concentrations in high OP-use areas.
Sampling from streams that are used for drinking-water supply gives an
indication of possible concentrations in drinking water. Without direct data at
a drinking water intake downstream, however, a risk assessor cannot assume
potential exposure at concentrations above or below that detected.
I.E Page 4
-------
i. Sources of Surface-Water Data
Although the number of "ambient" surface-water monitoring studies
which have included OP pesticides as analytes is extensive, extensive
monitoring data is not available for all OPs. The largest available source
of surface-water monitoring for OPs, the USGS NAWQA Program,
includes only nine active OPs: chlorpyrifos, diazinon, malathion, phorate,
methyl parathion, disulfoton, terbufos, azinphos-methyl and ethoprop. Two
other OPs - fonofos and parathion - included in the study have been
voluntarily cancelled.
The NAWQA program includes monitoring data for 76 pesticides and
covers "more than 50 major river basins and aquifers covering nearly all
50 states" (Figure I.E-1) (http://water.usgs.gov/nawqa/nawqa home.html
X. Results of the individual NAWQA study units are highlighted in the
appropriate regional assessments and in more detail in Appendix III.E.1.
Figure I.E-1. Location of USGS NAWQA study units (Source: USGS).
The USGS National Stream Quality Assessment Network
(NASQAN) program monitors water quality in the Rio Grande, Mississippi,
Columbia, and Colorado Rivers, four of the nation's largest rivers. This
study monitors for the same OPs included in NAWQA. NASQAN was
designed to measure the mass flux of constituents such as pesticides and
nutrients in these rivers, and so the 41 sampling stations are located at
the mouths of these rivers, at the confluence of tributaries entering the
I.E Page 5
-------
rivers, and at the intake and outflow of reservoirs along their path. Any
detection of OPs in these studies is significant because detection in such
large water bodies indicates that a large mass of the pesticide has run off
to surface water. The relatively small number of stations and relatively
infrequent sampling make it more difficult to connect detections in this
study to specific OP uses.
O | .
,^I- | State surface-water monitoring programs are most likely to include
analytes required by the Safe Drinking Water Act, but may include OPs if
consistent with budget priorities and local needs. When available, State
monitoring programs are important additions to NAWQA data for a full
understanding of possible OP exposure in drinking water. State programs
•*~* are described in detail in Appendix III.E.2.
The USEPA Office of Pesticide Programs (OPP), USEPA Office of
Ground Water and Drinking Water (OGWDW), and USGS National Water
Quality Assessment (NAWQA/USGS) initiated a reservoir monitoring
project to assess pesticide concentrations in untreated and finished
drinking water derived from surface water reservoirs. Twelve drinking
water reservoirs were selected from a list of candidate drinking water
reservoirs which were potentially vulnerable to pesticide contamination.
Vulnerable reservoirs are considered to be located in small watersheds
with high pesticide use areas and high runoff potential. A summary of the
results of this study occurs later in this section and in more detail in
Appendix III.E.3.
ii. Completeness of the Surface-Water Monitoring Data Set
Monitoring data is most extensive for chlorpyrifos, diazinon and
malathion, the three OP pesticides most frequently detected in agricultural
and urban surface waters. States that did include more OPs generally did
so as part of a wider screen, using a multi-analyte method, rather than
specifically monitoring for the OPs in specific areas of OP use.
Many of the OP parent compounds not included in broad surface-
water surveys are short-lived, and degrade by aerobic soil metabolism,
photolysis or hydrolysis to longer-lived transformation products. Some of
these short-lived compounds transform into degradates of toxicological
concern that are more persistent and mobile than the parent compounds.
The transformation of disulfoton to its sulfoxide and sulfone degradates is
an example. Unfortunately, the toxic transformation products are, by and
large, not included in monitoring studies.
I.E Page 6
-------
Detection of pesticides in surface water is most likely when the
sampling corresponds at least roughly to the timing and location of
pesticide use. Several monitoring studies illustrate this:
Q A series of studies by the California Department of Pesticide
Regulation (.CDPR) and the USGS investigated OP contamination
from winter use as a dormant spray to tree fruits and tree nuts. The
frequency and concentrations of OP detections in these studies were
both relatively high. Among OPs detected in these studies were
methidathion and dimethoate, which are rarely included in other
monitoring programs.
Q Diazinon and chlorpyrifos in urban streams represents the OP
contamination most frequently detected in NAWQA surface water,
followed by detection of malathion in urban streams. Since urban uses
of these pesticides can occur year-round, and every NAWQA study
monitored streams in watersheds dominated by urban or mixed land
use, these studies were targeted to the timing and location of these
uses.
Q A study in the USGS San Joaquin River Basin (SJRB) further
confirmed the importance of timing of sampling. Sampling three times
per week in this study was more likely to detect higher concentrations
than once per week. Sampling once per week was more appropriate
for determining the median concentration.
iii. Effects of Study Design
In general, the surface-water studies which included OP pesticides as
analytes were not specifically designed to correspond with times and
locations of agricultural OP use. For instance, the same suite of nine OPs
was included in NAWQA sampling programs nationwide. Azinphos-
methyl was detected in surface water in the NAWQA Lower Susquehanna
River Basin study unit, an area where azinphos-methyl is used in
orchards. NAWQA also included azinphos-methyl as an analyte in three
study units that it identified as part of the "Corn Belt." Surface-water
sampling in the Lower Illinois River Basin study was specifically targeted
to "two watersheds with greater than 90 percent row-crop agriculture and
the basin inflow and outflow sites." Azinphos methyl is not used on corn,
and it was not detected in any surface-water samples from these three
study units. The USGS notes this effect of design in its analysis of OP
occurrence in surface water and ground water from 1992 to 1997,
reporting that azinphos methyl and ethoprop were not widely distributed in
NAWQA and NASQAN studies, but that they "were detected in 43 and 69
percent, respectively, of samples from a few small agricultural watersheds
in western irrigated valleys."
I.E Page 7
-------
WAVAV
«* =
s
The design of the available programs determines their utility for the
cumulative drinking water exposure assessment. While the NAWQA
program samples in almost all states, a good number of the studies were
designed to answer locally important questions for each river basin, and
are not uniformly designed. The USGS Pesticide National Synthesis
Project elaborates on why the studies are not specifically designed to
produce a statistically representative analysis of national water-quality
conditions ( http://wwwdwatcm.wr.usgs.gov/ccpt/pns data/data.html).
In comparison to NAWQA, NASQAN includes relatively few sites and
samples each year, and is designed to allow an assessment of mass flux
from some of the largest rivers. State studies were even more limited,
and were most likely to include diazinon and chlorpyrifos in monitoring
programs, if OPs were included at all. States that did include more OPs
generally did so as part of a wider screen, using a multi-analyte method,
rather than specifically monitoring for the OPs in specific areas of OP use.
iv. USGS-EPA Reservoir Monitoring Project
The USGS-EPA Reservoir Monitoring Study (Blomquist et al., 2001;
available through the USGS web site at
http://md.water.usgs.gov/nawqa/OFRj31-456.pdf) was designed to
evaluate potential concentrations of a variety of pesticides and
transformation products in untreated and treated drinking water derived
from reservoirs. This study included twelve reservoirs covering a range of
pesticide use areas across twelve states (Figure I.E-2). The study
focused sampling during the period of the year with highest pesticide
runoff vulnerability and variability in the post pesticide application season.
Each, reservoir was sampled quarterly for one year, as well as biweekly for
a 4 month post-application period. Two sites were sampled at weekly
intervals for 6 months post-application-season to improve the estimate of
peak concentrations for short-lived compounds. Additional data collected
for each site provided information on watershed properties, water
treatment information, and reservoir characteristics.
I.E Page 8
-------
I Figure I.E-2. Location of reservoirs in the USGS-EPA Reservoir Monitoring
I Project
While both untreated (raw) and treated (finished) water samples were
1 taken at each sample time, the sampling scheme does not account for the
travel time of the pesticide and its transformation products through the
water treatment plant. Therefore, the occurrence and magnitude of
pesticides in raw and finished waters cannot be directly correlated. This
will likely exaggerate variability in removal efficiencies and limit direct
linkage of degradation and formation patterns of pesticides during water
treatment.
I The pilot reservoir monitoring study provides two years of sampling,
with 602 to 626 samples for each of 31 active OP parent and
transformation products included in this cumulative assessment. This
program included some rarely-monitored OPs, such as tribufos,
phostebupirim, profenofos and dicrotofos, and rarely analyzed
transformation products.
i
Thirteen of these 31 compounds were detected in either raw or
finished drinking water samples, in spite of extreme drought conditions in
| 6 of the 12 watersheds in 1999 (see Appendix 11 I.E.3 for details).
Diazinon, the most frequently detected OP, was found in 35% of 323 raw
water samples but in none of the 227 finished water samples. Although
the lack of truly paired raw and finished water samples precludes
definitive conclusions, these results suggest that diazinon may be
removed by the treatment process. However, the likely transformation
product, diazoxon, was not included as an analyte in the pilot program.
Results for other OPs suggest that parent OP compounds are
1 transformed during water treatment. For instance, malathion was
detected in raw water samples (2%) while malaoxon was detected in
finished water samples (5%). Chlorpyrifos was detected in 5% of raw
water samples; neither chlorpyrifos nor its oxygen analog were detected in
I
I.E Page 9
-------
I finished water. Azinphos-methyl and its oxon were both found in raw and
I finished water but the difference between the number of detections for
| each is insufficient to draw conclusions on treatment effects, especially
| since azinphos methyl and its oxon were only found in the same reservoir
| once (Missouri in 2000). While the actual transformation process is
i difficult to assess because raw and finished water samples were not
I temporally paired, the conversion of some OPs to oxon transformation
,y.,,., | products is consistent with published data and recent studies submitted by
I OP registrants.
| A small number of detections of other transformation products are
1 consistent with expectations based on the environmental fate properties of
[ the parent chemicals. Fenamiphos and disulfoton were not detected in
I this limited sampling program, but both the longer-lived sulfoxide and
| sulfone transformation products were detected in one or two samples
I each. While their detection in raw water is an indication that drinking
| water contamination is possible, detections were few enough that the lack
I of detections in finished water is not a clear indication of removal by
| treatment.
[ Diazinon was detected in 10 of 12 reservoirs, and chlorpyrifos was
I detected in 6, which likely reflects their widespread use. No other OP was
I detected in more than three reservoirs in this limited sampling. Azinphos-
[ methyl had the highest concentration detected of all parent products
i (0.114 ug/l in South Carolina raw water). Azinphos-methyl was found in
I 46% and 32% of samples taken in South Carolina in 1999 and 2000.
I Azinphos-methyl oxon was detected at a maximum concentration of 0.263
1 ug/l in Oklahoma, and was detected in 20% of samples in New York and
| Missouri in 2000. Malaoxon had the highest concentration detected of all
1 analytes with maximum detections in Louisiana of 0.556 ug/l in 2000, and
f 0.204 ug/l in 1999.
I Phostebupirim, which is very rarely included in any monitoring studies,
1 was detected in 10% and 8% of 1999 raw water samples in Missouri and
I Pennsylvania, respectively. The concentrations were low (0.003 to 0.007
| ug/l), but serve as a reminder that OPs may be transported to surface
| water bodies, even if few monitoring data are available to confirm this.
I Although the reservoir monitoring study was not specifically targeted to
| high OP-use areas, it included more OPs than any previous study.
I Therefore, it is useful for considering the possibility of exposure to multiple
| OPs. Of 314 intake samples considered, 137 (44%) had one or more
I detectable OPs. Of the 137 with detectable OPs, 16 (12%) included more
I than one detected OP. Of 67 outfall samples considered, 17 (25%) had
I one or more detectable OPs, two of those samples (12%) having more
| than one OP detected. Of 218 finished samples considered, 24 (11 %)
I.E Page 10
-------
had one detectable OPs, and none of the finished samples considered
here had more than one OP detected.
The pilot reservoir monitoring program confirmed the utility of sampling
for a wide range of OPs and transformation products in drinking water,
using low levels of detection. Continued and expanded monitoring should
improve understanding of potential drinking water exposure, and of the
effects of degradation in the field and from drinking water treatment.
c. Ground Water Monitoring
Due to the chemical properties of most of the OP insecticides, drinking-
water exposure through contamination of surface-water resources is
generally more likely than through contamination of ground water. However,
even in regions where surface water is the predominant source of drinking
water for most of the population, a significant portion of homes derives
drinking water from relatively shallow domestic wells. In some areas of the
country, where soils are especially permeable and depth to unconfined
ground water particularly shallow, domestic wells represent some of the
drinking-water sources most vulnerable to pesticide contamination.
Most OPs were described as unlikely to leach to ground water in the
individual risk assessments completed over the last few years. This is due
mainly to the relatively short aerobic soil-metabolism half-life of many OPs.
However, there are some important exceptions. Several OPs are described
as having the potential to contaminate ground water, but lack, the data to
sufficiently evaluate the magnitude of this risk.
Fenamiphos and its degradates, fenamiphos sulfoxide and fenamiphos
sulfone, are the best examples of this problem. These chemicals have been
^3 detected at high levels in ground-water studies conducted in Florida, and to a
lesser extent in California. Concentrations of fenamiphos and its
transformation products detected in the Central Ridge area of Florida ranged
as high as 246 ppb (204 ppb fenamiphos sulfoxide) in a retrospective ground-
water study.
However, recent ground-water monitoring which includes fenamiphos is
scarce. The USGS undertook a fenamiphos ground-water study at seven
golf courses in Florida, and reported maximum detections of < 1.0 ug/l each
for fenamiphos and its transformation products. The State of Florida reports
that its database includes only two wells with detections of fenamiphos
sulfoxide in its ground-water database. California collected samples from 40
drinking water wells in fenamiphos use areas during the early and mid 1990s,
but did not detect fenamiphos (another round of sampling is currently
underway). Hawaii, Michigan arid North Carolina report that fenamiphos was
not detected in a total of fewer than 100 drinking water or monitoring wells,
and fenamiphos is not included among analytes in the NAWQA program.
I.E Page 11
Xsv-XwJ
-------
Therefore, while fenamiphos has been detected in vulnerable to very
vulnerable soils in Florida and California, sufficient data is not available which
could allow a more detailed monitoring assessment for other areas of the
country.
i. Sources of Recent Ground-Water Monitoring Data
,,.,.„ = The Agency contacted pesticide lead agencies and other agencies
I in all 50 States to inquire whether OPs were included in surface-water or
[ ground-water monitoring (either ambient or drinking water) programs over
I the last decade. OPP requested recent data since 1) earlier data are
more likely to be included in the aggregate assessments of individual
OPs, 2) recent data are more likely to reflect current use rates and use
areas, and 3) such data are more likely to be in electronic format,
accessible either over the Internet or as an e-mail attachment.
Government scientists in nearly all States offered to describe or provide
summaries of current monitoring programs, or directed the Agency to data
which are available online.
As a result of the relative non-persistence of most OPs in soil and the
limits on funding for monitoring in State and Federal programs, few OPs
are included in ground-water monitoring programs conducted over the last
decade. Chlorpyrifos, diazinon and malathion are the OPs most
commonly included as analytes in State ground-water monitoring
programs. In some States, multiple OPs are included as part of a general
screen under EPA methods 507 or 525.5. In such cases, the levels of
detection are often higher than in more chemical-specific analyses.,
The voluntary cancellation of non-agricultural uses of chlorpyrifos and
diazinon affects the ground water assessment for these chemicals. While
many of the agricultural uses remain, the Agency believes that most of the
ground water monitoring detections of these chemicals are associated
with uses that have been cancelled. The termiticide use of chlorpyrifos,
which is currently being phased out, represents the use that has led to the
highest known concentrations of any OP in ground water. The
concentrations of chlorpyrifos measured in wells affected by the
termiticide use ranged as high as 2090 /ug/l, significantly higher than
concentrations found in agricultural areas, which generally are below 1
The USGS NAWQA program is the other major source of ground-
water data for the OPs. While the NAWQA program has provided a very
valuable ground-water data set, it has several important limitations with
respect to the cumulative OP drinking water assessment:
Q Only nine OPs included in this cumulative assessment are included.
I.E Page 12
-------
<&
Many NAWQA ground-water studies included only a single sample of
each well in the network. Even if wells were located in OP use areas,
the monitoring was not timed to correspond specifically to account for
pest pressure and OP application for that particular year.
A number of land-use studies in the program were focused on urban
areas. The phase-out of homeowner uses of chlorpyrifos and diazinon
renders such data less useful for our assessment.
Finally, the design of the ground-water studies differs between each
study unit, reflecting the local aspect of ground-water quality that was
being investigated in each monitoring program. For instance, monitoring
in the Eastern Iowa Basins study unit included 65 domestic wells in order
to assess the water-quality of the most heavily used aquifers in the study
unit. By contrast, one of the ground-water studies in the Ozark Plateaus
study unit was designed to evaluate water-quality in domestic wells in
cattle and poultry-producing regions. One of the studies in the Southern
0} Florida study unit included wells less than 15 feet deep and located in the
drip line of citrus trees, where the depth to the water table was 2 to 4 feet
below the land surface. In addition, a study in the Central Arizona Basins
study unit included domestic, public supply, and other wells that draw
older water (at least pre-1953) from a confined aquifer, which to this point
is considered to have had very little hydraulic connection with potentially
contaminated shallower ground-water above the confining layer. The
differing design among the different ground-water monitoring
studies limits the applicability of statistical methods to the combined
NAWQA ground-water dataset for a national OP drinking-water
assessment.
Some OPs are not included in any ground-water monitoring supplied
to the Agency, such as phostebupirim, chlorethoxyfos and tribufos. Other
OPs have only very limited monitoring data from the 1980s in which a
small number of ground-water detections are reported. One example is
methamidophos, which was detected in four wells near a Maine potato
field in 1986 at concentrations up to 10 ug/l. Such data may not well
represent current use or use rates, but may also have underpredicted
possible ground-water contamination due to higher analytical detection
limits. Older studies which revealed ground-water contamination indicate
that exposure to rarely analyzed OPs is possible. However, the lack of
extensive, recent ground-water data for some compounds makes it very
difficult to quantify the potential risk nationwide.
With few exceptions, ground-water monitoring programs which include
OPs are surveys which are not targeted specifically to assess the effects
of OP use on ground-water quality. Examples of exceptions include
chlorpyrifos termiticide use studies and fenamiphos studies near Florida
golf courses. The results of survey studies give some indication of the
I.E Page 13
-------
possible exposure to populations as a whole. However, since survey
studies usually include sampling of wells in areas where OPs are not
used, they are less useful for quantifying potential drinking-water exposure
in OP use areas.
Few ground-water studies include OP transformation products as
analytes. The fenamiphos prospective ground-water studies and the
USGS golf-course study mentioned above are rare exceptions. Lack of
monitoring for transformation products might be important for other OPs
which form sulfoxide and sulfone degradates, such as disulfoton, phorate
and terbufos. If these OPs follow the same pattern as fenamiphos, the
sulfoxide moieties of these chemicals may be a greater concern for
ground-water contamination than the parent compound.
d. Effects of Drinking Water Treatment on OP Pesticides
The weight of evidence from open literature, a registrant-sponsored study,
| an ORD/EPA laboratory study, and the USGS-EPA reservoir monitoring
| program show parent OP pesticide residues in water are likely to be
I transformed during drinking water treatment. The most probable pathway is
[ transformation by oxidation through chlorination and not physical removal.
I . Oxidative transformation products of toxicological concern, such as sulfones,
I sulfoxides, and oxons, have been detected in finished water samples from
1 water-treatment plants. Although not all oxons were tracked, the USGS-EPA
| reservoir study suggests that malathion may have been converted to
I malaoxon as a result of treatment.
Studies have shown oxons to be relatively stable in chlorinated drinking
water for at least 48 hours. Although the detection frequencies of oxidative
degradation products were low in the reservoir monitoring data, they were
more frequently detected in finished water than in raw water. These data
suggest oxidative degradation products such as oxons, sulfones, and
sulfoxides have a high likelihood of occurrence in finished drinking water.
Appendix III.E.4 provides additional detail on removal and transformation
of organophosphorus pesticides and certain degradation products through
water treatment. The review extends the OPP water treatment literature
review (http://www.epa.gov/scipoly/2000/September/sept-00-sap-dw-
0907.pdf). Documents in this report include open literature, registrant-
sponsored water treatment data, and the USGS-OPP pilot reservoir
monitoring data.
Available information indicates that two common water-treatment methods
lead to transformation of some OPs:
I.EPage14
-------
Q Treatment of water by chlorine and chlorine compounds for
disinfection can result in transformation of parent OP compounds. The
P=S bond of OPs can be oxidized to a P=O bond leading to the formation
of oxon transformation products. According to Magara et al (1994),
several OPs are transformed to their corresponding oxons in this manner.
For instance, diazinon is oxidized to diazoxon, which is relatively stable in
chlorinated water for at least 48 hours. In a laboratory study at EPA-
ORD's AWBERC facility in Cincinnati, Ohio, about 90% of chlorpyrifos-
methyl was removed by chlorine treatment. The removal was most
probably due to oxidation of the insecticide to oxons and other products.
G In areas where water softening treatments add lime and soda ash to
reduce calcium and magnesium levels in water, the pH can increase to
I about 10-11. This high pH can lead to base-catalyzed hydrolysis of the
| OPs which are susceptible to hydrolysis under alkaline conditions. In the
I ORD treatment study, more than 99% of malathion was removed during
| softening treatment. The effects of softening may not be so dramatic for
I all OPs; although phorate has a 3-day hydrolysis half-life at pH 9, lower
I removal (20%) of phorate was observed.
| A complete review of a registrant-sponsored jar test study on the potential
| effects of chlorination on six OP pesticides and four oxons (Tierney, et al.,
I 2001) was hindered by incomplete information on the experimental
I procedures (particularly, water quality data, the impact of sodium thiosulfate
7 | on water chemistry, storage stability, and clarification regarding pesticide
concentrations above the limit of detection and below the limit of
quantification). Despite the lack of information on experimental methods, the
data indicate organophosphorus pesticides (acephate, azinphos-methyl,
chloropyrifos, diazinon, malathion, and methamidaphos) are transformed in
chlorinated drinking water. Chemical oxidation of the organophosphorus
compounds led to the formation of oxons for azinphos-methyl, chloropyrifos,
diazinon, and malathion. Chloramines were formed during the experiment,
and because chloramines have lower oxidizing potential than hypochlorus
acid, the extent of degradation and formation of oxidative degradation
products (oxons) may be different under conditions of higher free chlorine
concentrations.
e. Suitability in Meeting Cumulative Assessment Needs
While the available monitoring studies provide a profile of OP occurrence
in water, critical limitations preclude basing the cumulative water exposure
assessment solely on monitoring. In particular, the monitoring studies were
not designed to characterize daily concentration profiles and are not robust
enough to provide daily distributions. Nor have the studies been conducted
over a long period of time (typically less than three years) necessary to
characterize year-to-year fluctuations due to weather patterns. While the
NAWQA study units coincide with a number of high OP-use areas, not all of
I.E Page 15
-------
the major OP use areas have monitoring data. Lack of monitoring for some
compounds make it difficult to completely assess co-occurrence. Finally,
monitoring provides a snapshot in time and does not reflect recent mitigation
actions, such as lower application rates and fewer applications or cancellation
of certain uses or chemicals, initiated for individual chemicals during the risk
management phase.
Despite these,limitations, water monitoring was used in the cumulative
assessment to help identify vulnerable surface water sources, characterize
OP residues in ground-water sources, compare relative impacts of OP use on
water resources in different locations across the country, and provide a
baseline comparison for estimated OP concentrations used in the
probabilistic exposure assessment. Appendices III.E.1 and III.E.2 compare
estimated OP concentrations with available local monitoring. Significant
trends between estimated concentrations and monitoring are highlighted in
the regional assessments in Part II.
With the publication of data from the nationwide set of NAWQA study
units, more surface-water data for the OPs is available than ever before.
However, the cumulative OP drinking-water exposure assessment requires.
the estimation of simultaneous daily drinking-water exposures to multiple
pesticides, which is something that has never been attempted before.
Although the available data is extensive, the cumulative drinking-water
exposure assessment cannot be solely based on monitoring.
Therefore, the daily drinking water exposure estimates have been
generated using the simulation models PRZM and EXAMS. A description of
the use of these models for the cumulative OP drinking water exposure
assessment follows. The use of models allows estimation of possible
concentrations of OPs not included in monitoring programs, or in areas for
which monitoring for locally important OPs was not available. As described in
the Risk Characterization section, peak values from the modeling are not
always as high as some seen in small streams in the NAWQA program.
However, the models allow the Agency to estimate a cumulative exposure
assessment for all OPs used in representative scenarios for each region,
even if they do not consistently match all the highest detections for each
individual chemical.
3. Drinking Water Assessment Methods
The goal of the cumulative assessment is to aggregate exposure from the
organophosphorous (OP) pesticides over multiple routes of exposure (food,
drinking water, residential) in a manner that is consistent in time (i.e., those
exposure routes that are likely to occur on the same day are combined; those
that are not likely to occur on the same day are not combined) and in location
(i.e., only those exposures that may potentially occur in the same location are
considered together). The Agency needs reasonable approximations of daily
I.E Page 16
-------
I distributions of OP residues (concentrations) in drinking water to combine with
1 food and residential exposures using a probabilistic, calendar-based approach.
| This cumulative risk assessment represents the first attempt to quantify
1 possible drinking water exposure to multiple chemicals at the same time.
| Available surface-water monitoring is not sufficient to allow estimation of
I potential daily drinking water exposure to the OPs included in this assessment.
I No currently-available model is specifically designed to simulate the
| simultaneous application and transport of multiple pesticides in a watershed.
f Therefore, the Agency looked to available tools to provide these daily exposure
I estimates for consideration with food and residential exposures.
i Because drinking water is local, the national exposure assessment for
[ drinking water must address localized areas of the country where exposure to
| one or more OPs may occur due to drinking water contamination. The
I consideration of OP use in specific regions of the country will facilitate the
I assessment of potential co-occurrence of different OPs in drinking water, leading
I to a cumulative assessment of OPs in drinking water on a regional basis.
I The sections that follow describe the steps OPP has taken to generate
I . regional drinking water exposure assessments as a part of the cumulative OP
| assessment.
| a. Chemicals and Uses Included in the Cumulative Assessment
I
| Table I.E.1 lists the parent OP, transformation product(s) of toxicological
d) i concern, and approach for considering the contributions of the transformation
!> | products to the cumulative water exposure. Detailed chemical-specific
inputs, based on environmental fate studies submitted by the OP registrants,
are documented in Appendix III.E.5. These inputs are based on the individual
chemical assessments that were published in the REDs.
WkVk-
I.E Page 17
-------
Table I.E-1. OP Pesticides and Toxic Transformation Products Included in the
Cumulative Water Exposure Assessment
P&sttekte
toephate
Azinphos Methyl
Bensulide
Chlorethoxyfos
Chlorpyrifos
Diazinon
Dichlorvos (DDVP)
Dicrotophos
Dimethoate
Disulfoton
Ethoprop
Vlalathion
vlethamidophos
Vlethidathion
Vlethyl Parathion
Naled
ODM
3horate
Phosmet
3hostebupirim (also known as
febupirimphos)
Profenofos
Terbufos
Tribufos
Trartsformatkm Products of
Toxtcological Concern
Vlethamidophos
Oxon
Oxon
Oxon
Oxon
Diazoxon, Hydroxy-diazinon
slone
vlonocrotophos
Oxon
Sulfone, Sulfoxide
SME, OME, M1
\/lalaoxon
vlone
^Jone
Methyl Paraoxon
Dichloivos (DDVP)
Sulfone
Sulfone, Sulfoxide
3hosmet Oxon
Oxon
^one
Sulfone, Sulfoxide
vJone
Approach for Including
Transformation Product
Conversion from parent to product;
max rate based on fate studies
:ormed by treatment
Formed by treatment
Formed by treatment
Formed by treatment
rormed by treatment
ia
Not in field studies
Formed by treatment
Combined residues
Not modeled; negligible residues;
)arent relatively stable
Formed by treatment
la
la
:ormed by treatment
Conversion from parent to product;
nax rate based on fate studies
^Jot modeled; negligible residues
Combined residues
Formed by treatment
rormed by treatment
la
Combined residues
la
i. Parent Chemicals and Uses
The drinking water exposure assessment includes those OP pesticides
with registered outdoor uses that may potentially impact surface- or
ground-water sources of drinking water (Table I.E.1). Those pesticides or
pesticide uses that are being cancelled and/or phased out as a result of
agreements between the Agency and the specific OP registrants, and
I.E Page 18
-------
those OPs with uses that are unlikely to reach drinking water were not
included in the water exposure assessment. Those agreements in place
as of May 1, 2002, were considered in this assessment. Revisions since
the preliminary assessment in December 2001 include exclusion of
fenamiphos and azinphos methyl use on cotton, both of which are being
voluntarily cancelled.
I ii. Transformation Products
E
E
Those OP transformation products identified as being of toxicological
I concern (Table I.E.1) were included in the cumulative drinking-water risk
I assessment when environmental fate studies indicate that these products
I may be formed in the environment or may form as a result of water
| treatment. Some OP risk assessments did not consider the transformation
| products quantitatively because no environmental fate data was available,
I while others assumed that the characteristics of the transformation
I products were equivalent to that of the parent, or combined limited data
| with conservative assumptions for a screening assessment.
I Sulfoxide and Sulfone Products: The sulfoxide/sulfone products of
| disulfoton, phorate, and terbufos are often more persistent and mobile
than the parent compounds. Full environmental fate profiles are not
available for the sulfoxide/sulfone transformation products, requiring some
assumptions to be made about their physicochemical properties. The
parent OP and two transformation products were modeled as "total toxic
residues". Formation and decline curves from aerobic soil-metabolism
studies allowed the assessment team to fit a single modeling half-life for
the combined residues. However, this required the assumption that all
three chemicals were equally toxic, and that the sulfone and sulfoxide had
the same soil-water partitioning coefficient as parent.
Oxon Products: Table I.E.1 identifies ten OP pesticides which form
oxon transformation products. While the oxons are generally not found at
significant levels in the environment, available studies suggest they are
formed by water treatment - in particular, through chlorination of the
parent OP, as noted earlier. Based on the available studies, OPP
assumed that oxons were not formed in the environment and, for the most
part, would not be found in signficant levels in untreated drinking water
sources. This assumption was supported by the results of the USGS-EPA
reservoir monitoring study, in which oxons were detected in the treated
water samples but not in samples taken at the drinking water intake.
| Transformation To Another Active OP: Acephate transforms to
methamidophos and naled transforms to diclorvos (DDVP). For these
pesticides, OPP assumed a conversion from one OP to the other based
on the maximum percent transformation from available environmental fate
studies. Thus, OPP assumed that 25% of applied acephate transformed
I.E Page 19
-------
f r
V,
>>
x-:«M«* g
-.v.v.y —
into methamidophos and 20% of applied naled transformed into DDVP.
The transformed OP as modeled separately, with an application rate that
reflected the appropriate percent conversion of the parent OP (with
adjustment for differences in molecular weights). The timing of the
simulated "application" was off-set by one half-life period. In the case of
acephate, this amounted to two days (e.g., the timing of the formation of
25% methamidophos was simulated as occurring 2 days after acephate
was applied). Because the half-life for naled was less than 1 day, the 20%
DDVP load was assumed to form on the same day as application.
iii. Accounting for Water Treatment By-Products
Limited scientific evidence (section I.E.d) suggests that many parent
OP pesticides may be transformed during drinking water treatment,
primarily by oxidation through disinfection. The oxidative transformation
products of toxicological concern - sulfones, sulfoxides, and oxons - have
been detected in treated water. Limited data suggest that these treatment
by-products may be stable for sufficient periods of time (for least 24 to 96
hours) to move through the distribution system.
The information is not sufficient to make quantitative adjustments to
the cumulative exposure estimates. OPP estimated maximum potential
impacts to determine whether additional information is needed by
assuming that all OP parents that form oxons, sulfoxides, or sulfones (see
Table I.E.1) are completely transformed into those products as a result of
oxidation. Where the transformation is less than complete, and where
non-toxic products are also formed, the such an assumption will
overestimate drinking water exposure. For a preliminary evaluation, OPP
did not assume removal of any of the other OP parent pesticides. OP
assumed that the sulfoxide and sulfone products are equal in toxicity to
I the parent and that the oxon products are ten times more toxic than the
parent. A comparison of the RPFs for dimethoate (0.32) and omethoate
(0.96), the oxon of dimethoate, suggests that this assumption would be
protective. The impacts are addressed in the risk characterization (I.G).
b. Regional Approach for the Cumulative Water Exposure Assessment
The Agency used a regional approach as a first step in addressing the
impacts of regional and localized variability in site, environmental, and
management practices that effect pesticide concentrations in water. The
USDA Farm Resource Region map (Heimlich, 2000) provided a framework
for focusing the cumulative assessment (see Appendix III.E. 10). By providing
general groupings according to similarities in key environmental factors
affecting runoff and leaching, such as precipitation, irrigation practices, and
soil types, these farm resource regions provide a framework for identifying
one or more locations which represent an area of the greatest concern for
I.E Page 20
-------
-
.VMV$
drinking water exposure in each region. In this way, the Agency chose a set
of locations to represent drinking water sources throughout the US.
Within the regions, drinking water exposure will vary locally due to OP
use, agricultural practices, nature and vulnerability of drinking water sources,
and weather patterns. Thus, the water exposure assessment focused on one
or more specific geographic areas within each region in a manner that would
be realistically protective of all sites within the region. OPP selected locations
where OPs in drinking water sources are likely to be of concern. If OP levels
in water from these vulnerable sites are not major contributors to the total
regional cumulative OP exposure, then the Agency can reasonably conclude
that drinking water exposures will not be a concern in other, less vulnerable,
portions of the region. If drinking water exposure from one or more of these
vulnerable sites is a significant contributor to the total cumulative exposure,
then additional assessments are necessary to characterize the extent of the
potential exposure.
Based on results of the preliminary cumulative risk assessment, OPP has
condensed the twelve farm regions into seven regions (Figure I.E-3). Table
I.E-2 compares the combined regions with the original regions.
| I.E Page 21
-------
Table I.E-2. New and Old Regions and Representative Vulnerable Sites Used in
the Cumulative Water Exposure Assessment
New Region
A - Florida
B - Northwest
C - Arid/Semiarid West
D - Northeast/
Northcentral
E - Humid Southeast
F - Lower Midwest
G - Midsouth
Old Region '
Fruitful Rim, SE (12)
Fruitful Rim, NW (10)
Fruitful Rim, SW (7)
Basin & Range (8)
Northern Great Plains (3)
Heartland (1 )
Northern Crescent (2)
Southern Seaboard (6), east
Eastern Uplands (5), east section
Prairie Gateway (4)
Fruitful Rim, TX (11)
Mississippi Portal (9)
west sections of E. Uplands, S.
Seaboard
Representative Vulnerable Site
West Palm Beach Co (FL) *
Willamette Valley (OR) *
Central Valley (CA) counties of
(a) Merced, San Joaquin, Stanislaus
it
(b) Fresno, Tulare, King, Kern
none (Red R. Valley surrogate)
Red River Valley (ND/MN) *
Central IL
Southcentral PA
Coastal Plain, northern NC *
Western NC
Central TX Hills *
Central TX Hills (surrogate)
Northeast LA, west-central MS *
none
ca
•;•:•:•:•>> 2
V m,
.•:•:•.•;•.•. £
S
: Scenario used to represent new region in revised OP cumulative risk assessment.
B(10;
I Figure I.E-3. Regions used in OP Cumulative Risk Assessment, based on USDA
| Farm Resource Regions
I.E Page 22
-------
Cs
5
c. Selection of Regional Water Exposure Assessment Locations
The selection of a specific location for regional drinking water
assessments involves several steps. First, OPP identified the high OP usage
areas and high agricultural intensities within each region; these are shown on
a national scale in Figure I.E-4. Next, in each high usage area within the .
region, OPP determined the types and locations of drinking water sources.
The final step in choosing a location is to assess the vulnerability of drinking
water sources within the high usage area within the region. OPP adapted
vulnerability schemes proposed by Kellogg and others at USDA for this
purpose. Locations of surface drinking water intakes overlain on runoff
vulnerability maps (Figure I.E-5) were compared with the OP use areas to
determine whether potentially vulnerable surface water sources of drinking
water coincided with high use areas. For ground water, OPP compared OP
use areas with a pesticide leaching vulnerability map (Figure I.E-6).
| Figure I.E-4. Total organophosphorous (OP) pesticide usage on an area-weighted
I basis, showing high-use areas in each region.
I.E Page 23
-------
v-v-X
I
| Figure I.E-5. Runoff vulnerability (in/year), adapted from USDA (Kellog, 1998}
Legend
Pesticide leaching (IWacre)
i 1 Zero or no data
;:;.;.;!;! 0.0001 -0.0054
H|:O.Q055-0.0140
^0.0141 -0.0341
- 0.0753
JltS-0754- 0.1785
Figure I.E-6. Pesticide leaching vulnerability, adapted from USDA (Kellogg, 1998)
I.E Page 24
-------
v.-1-.v.v'
c
I Figure I.E-7. National map w/ Percent Crop Areas by 8-digit HUC
Details of this process are provided in each regional assessment. The
Northwest region (Region A) illustrates this process. Three OP-use areas
stand out in the region (Figure l-E-4): Yakima County and eastern
Washington are the highest OP use area (predominantly on orchards) and
highest percent crop area (Figure l-E-7). The Snake River Valley in
Southeast Idaho is the second highest use area (predominantly on potatoes,
sugar beets). The Willamette Valley, Oregon, is the third high-use area, with
a mix of OP uses. Ground water is the predominant source of drinking water
in Idaho and eastern Washington, with vulnerability to leaching potentially
higher in eastern Washington. A few surface-water intakes occur in the
Yakima County area; the Willamette Valley has more surface water intakes
and is more vulnerable to runoff. Available monitoring from NAWQA study
units in Willamette Valley, Snake River Basin, and Pugett Sound suggest that
Willamette Valley will be more vulnerable to OP contamination with a higher
potential for co-occurrence of multiple pesticides.
OPP based its surface water assessment for the Northwest Fruitful region
in the Willamette Valley in Oregon. We also looked at potential impacts of
OP pesticides on ground water resources in eastern Washington and
southeast Idaho, relying largely on ground-water monitoring available through
the USGS NAWQA program and state monitoring programs.
In the preliminary cumulative risk assessment, OPP selected eleven
vulnerable drinking water sources for the drinking water exposure
assessment (Table I.E.2). Each of the 12 USDA regions had a representative
vulnerable site except for the Prairie Gateway and Texas Fruitful Rim, which
shared the same Central TX Hills site, and the Basin and Range, where no
I.E Page 25
-------
vulnerable sites were identified. In the Central Valley (CA), two sites were
identified: (a) Fresno County and south, where OP usage is among the
highest in the country, and (b) north of Fresno County, where total OP usage
is lower, but surface water sources are more vulnerable to runoff, particularly
during the-dormant season.
This revised cumulative risk assessment combines several of the regions
._,_ (Table I.E.2). However, only two combined regions include more than one of
the original vulnerable sites. The Northeast/ Northcentral Region (D) includes
the original Northern Great Plains (Red River Valley), Heartland (Central IL),
and Northern Crescent (Southcentral PA) sites. The Humid Southeast
includes the original Southern Seaboard (Eastern NC) and Eastern Uplands
*^ (Western NC) sites. OPP compared the estimated cumulative distributions,
NAWQA monitoring results, and OP usage to select a single representative
site for each of these new regions. Because of the influence of the relative
potency factors (RPF) in the cumulative OP loads in water, sites with
significant usage and monitoring detections of the higher-RPF pesticides,
such as terbufos and phorate, were selected over sites which had higher
uses and monitoring detections of lower-RPF pesticides such as chlorpyrifos,
diazinon, and malathion. These comparisons are discussed in the regional
assessments in Part II.
d. Estimate of Pesticide Concentrations in Drinking Water Sources
Within Each Region
After considering several predictive tools, the Agency adapted its paired
PRZM and EXAMS models for the Index Reservoir (PRZM-EXAMS IR) to
estimate a distribution of daily drinking water concentrations that could be
used for multiple chemicals over several years of predictions across the
country. PRZM-EXAMS IR has been modified to calculate concentrations in
a small drinking water reservoir in a primarily agricultural watershed. PRZM-
EXAMS has the capability of predicting water concentrations over a number
of years based on collected historical weather data for the sites which are
being modeled.
The PRZM component of the model fs designed to predict the pesticide
concentration dissolved in runoff waters and carried on entrained sediments
from the field where a pesticide has been applied into an adjoining edge-of-
field surface water body. The model can simulate specific site, pesticide, and
management properties including soil properties (organic matter, water
holding capacity, bulk density), site characteristics (slope, surface roughness,
field geometry), pesticide application parameters (application rate, frequency,
spray drift, application depth, application efficiency, application methods),
agricultural management practices (tillage practices, irrigation, crop rotation
sequences), and pesticide environmental fate and transport properties
(aerobic soil metabolism half-life, soil:water partitioning coefficients, foliar
degradation and dissipation, and volatilization). OPP selects a combination
I.E Page 26
-------
of these different properties to represent a site-specific scenario for a
particular pesticide-crop regime.
The EXAMS component of the model is used to simulate environmental
fate and transport processes of pesticides in surface water, including: abiotic
and biotic degradation, sediment:water partitioning, and volatilization.
Currently, OPP is using an index reservoir as the benchmark surface water
body for drinking water exposure assessments.
For each component, the values used are derived from real world data.
Pesticide environmental fate properties used in the modeling come from
registrant-submitted data used for pesticide registration or reregistration. The
values used for soil properties and site characteristics are chosen from real
world databases appropriate for the sites on which the pesticide may be
used. For example, if the pesticide is approved for use on cotton, OPP uses
data reflecting the soil types in the Cotton Belt. The index reservoir being
modeled is based on and represents an actual, small flow-though reservoir
used for drinking water. Finally, the weather inputs for the model are taken
from regional specific weather data, based on the USDA Major Land
Resource Areas. PRZM modeling is generally simulated for 20 to 36 years in
order to calculate a return frequency of concentration in surface water body.
Further information on how the Index Reservoir model is used for screening-
level drinking water assessments of individual pesticides can be found in the
EPA Environmental Fate and Effects Division's pesticide science policy
paper, "Guidance for Use of the Index Reservoir Guidance for Use of the
Index Reservoir in Drinking Water Exposure in Drinking Water Exposure
Assessments."
Running the assessment with historical data for several years provides
more confidence that variations due to weather have been considered in the
assessment. Having the historical weather data, pertinent site information
and reported use histories allows the Agency to factor regional variations into
the assessment. With this method, multiple chemicals which have varying
uses and application factors are assessed and co-occurrence is realistically
accounted. Since the day by day component is retained, this distribution can
easily be paired with residues resulting from residential applications.
The PRZM-EXAMS/IR tool has been used in many of the individual
assessments to predict a reasonable high end screening concentration to
factor into the aggregate assessment. However, the cumulative assessment
focuses on the probability or likelihood a person will be concurrently exposed
to multiple pesticides from food, water, and residential use. The method
which was used in the aggregate assessments has been modified in several
ways to focus on the probability of co-occurrence from the various routes.
I.EPage27
-------
I The most significant change in terms of predicted exposure is th'at the
I entire range of PRZM-EXAMS/IR output is used for the probabilistic
| distribution. In other words, instead of choosing a single value at the upper
i end of the distribution to represent the exposure, all daily concentration
I values are used in the CALENDEX runs.
| • The cumulative assessment modeling used "typical" application rates with
[ typical numbers of applications instead of labeled maximum rates and
| maximum numbers of applications which were used in the individual chemical
';-- | assessments. While this is reflective of the "typical" condition, it does not
I reflect potential concentrations that may occur when the pesticide is used at
I maximum rates because of pest pressure.
- | The drinking water assessments for cumulative are regional in nature.
^ | This allows EPA to make informed judgements about when compounds co-
I occur and when they compete. Overall, the assessment is much more
^ | realistic on a regional basis. Scenarios chosen for regional assessments are
| reflective of regional differences in cropping and pesticide use as well as
's^ [ differences in run-off and leaching vulnerability.
\\ i • •
I i. Cumulative Adjustment Factors for Crop Area and OP Use
PRZM is a field-scale model, while the OP cumulative water
assessment focuses on watershed-scale impacts (i.e., the contributions of
multiple OP uses on multiple crops occurring in multiple fields in a
watershed). In individual chemical assessments, PRZM is used to
simulate a watershed. In the OP cumulative assessment, the Agency
used PRZM to model multiple fields in a watershed. While this approach
provides a more realistic depiction of multiple chemical usage in a
watershed, it still has limitations. PRZM can simulate multiple fields, but
provides no spatial context for those fields. It also assumes that the runoff
from each of those fields goes into the reservoir. In other words, each field
is assumed to be uniformly distributed in the watershed, with no distinction
made between those fields located in the upper reaches of the watershed
and those near the reservoir.
To adapt PRZM for this watershed approach, OPP must adjust the
estimated pesticide concentrations to account for the portion of the
watershed that is treated by a particular OP. This was done using a
Cumulative Adjustment Factor (CAP), which accounts for the percentage
of the watershed that is planted to a particular crop and the fraction of
those acres which receive OP applications.
The CAP accounts for the percent of the location area planted to crops
and treated with the corresponding OP pesticides. The CAP is based on
several different data sources. The Agency used the USGS 8-digit
Hydrologic Unit Codes (HUCs) to delineate watersheds, and the National
I.E Page 28
-------
Agricultural Census for 1997, reported on a county basis, to identify areas
planted to agriculture. This procedure was presented to OPP Science
Advisory Panel (available through the Agency web site at
http://www.epa.goV/scipoly/sap/1999/may/pca_sap.pdf) and is described
in an OPP science policy paper (available through the Agency web site at
http://www.epa.gov/oppfead1/trac/science/reservoir.pdf). Percent crop
area values were calculated for each region. To determine the total acres
planted for each crop within, the selected location, the Agency used the
most recent county level production statistics, generally taken from USDA
publications. And finally, to calculate the area treated, by the various OPs,
the most recent percent of crop treated estimates, generally taken from
USDA\NASS publications were applied.
In addition to primary USDA publications, various other data sources
(California Department of Pesticide Regulation, Pesticide Use Reporting
Data, academia publications) were used to obtain acres planted and acres
treated estimates.
The following example (Table I.E-3) illustrates how CAFs are
calculated and applied. Suppose, that after reviewing the various data
(drinking water source, vulnerability, crop production, pesticide use, and
monitoring data), a location (one or several counties) is identified around
which the drinking water assessment is conducted. The total area for this
location is 800,000 acres; agricultural cropland accounts for 600,000
acres of this total area, and 320,000 acres of the agricultural cropland are
planted to four crops (corn, alfalfa, beans and apples) that are treated with
OP pesticides:
Table I.E-3. Cumulative Adjustment Factor Illustration: Deriving Cumulative OP
Percent Crop Area
Total Area
Crop Area, All Agricultural Uses:
OP Uses in Region:
Corn
Alfalfa
Beans/legumes
Apples/pome fruit
Total OP Use Area
Acres
800,000
600,000
200,000
80,000
16,000
24,000
320,000
Percent of area
25%
10%
2%
3%
PCA
75%
40%
I.E Page 29
-------
o
Further, suppose that 4 different pesticides are used on each of the 4
crops (some pesticides are used on more than one crop). Acres treated
represent the total number of acres of the crop that were treated with
each pesticide (may represent more than one application). Following the
numerical example above, if 60,000 acres of field corn were treated with
pesticide A, then the CAF for this particular use (field corn-pesticide A) is
0.075, or:
CAF Corn.op(A) = (Total Acres Planted AI|OPcrops / Total Acres)
x (Acres Treated Com-op(A) /Acres Planted All OP Crops)
= (320,000 / 800,000) x (60,000 / 320,000) = 0.075
Table I.E-4. Cumulative Adjustment Factor Illustration: Individual Crop-Pesticide
Factors Used for Conversions.
Crop
Corn
Corn
Com
Corn
Alfalfa
Alfalfa
Alfalfa
Alfalfa
Apples
Apples
Apples
Apples
Beans
Beans
Beans
Beans
Pesticide
A
B
C
D
A
B
E
F
A
F
G
H
B
E
1
J
Acres Treated
60,000
1,000
500
40,000
16,000
4,000
10,000
8,000
10,000
15,000
6,000
6,000
16,000
1,000
16,000
2,000
Cumulative
Adjustment Factor
.075
.00125
.000625
.05
.02
.005
.0125
.01
.0125
.01875
.0075
.0075
.02
.00125
.02
.0025
Again, these CAF are applied to the model is run for a particular
chemical:crop combination. In this manner, since the use statistics come
from reported data, competing and compatible uses are accounted for by
summing the appropriate distributions across days after the RPFs are
applied.
I.E Page 30
-------
Cs
~
5
ii. Relative Potency and Safety Factor Adjustments
The resulting CAF-adjusted concentrations for each OP-crop
combination must be converted to a concentration equivalent for an index
chemical. Once this is done, the concentrations can be combined into a
single set of daily distributions (spanning multiple years) for each region.
The concentrations were normalized to methamidophos equivalents using
the relative potency factor (RPF) and safety factor. This normalized output
for each chemical:crop combination was summed day by day to give a
single distribution of potential combined water residues for the region.
Factors to convert from individual to cumulative distributions:
where
Cv,
(OPx,CROPz)
RPF(OPx) x SF(OPX)
is the converted concentration for OPx on CROPz
c
-------
approach was undertaken to provide transparent modeling scenarios using
,the best currently available data.
i. Typical Pesticide Use (Rate and Frequency)
For regions exclusive of the Arid/Semiarid West, the primary sources
of information for percent crop treated, number of applications, and
. amount of active ingredient applied are MASS Agricultural Chemical
Usage summaries. These documents provide data for selected crops in
selected states; they are published annually for field crops and biennially
for vegetables and for fruits and nuts. Vegetable chemical usage
summaries are reported for even years; fruit and nut chemical summaries
are reported for odd years. The years 1997-2000 were reviewed for field
crops, 1998 and 2000 for vegetables, and 1997 and 1999 for fruits/nuts.
The most recent summary data is cited for state/crop combinations
appearing in the cumulative surface water assessments. Citations follow
the format: "MASS, 2000 Vegetable Summary."
In a given MASS summary, specific OP pesticides may be noted, by
use of an asterisk, as being applied to a crop but no usage data is
provided. This situation arises where the number of individuals reporting
use of the specific OP is so small (i.e., fewer than five) that respondent
confidentiality could be compromised through data disclosure. In such
instances, an earlier summary has been consulted.
MASS data were not available for all specific chemical/state/crop
combinations examined. In some cases, additional survey instruments
were consulted. All usage data sources are documented at their
occurrence in the regional summaries.
OPP used the average application rate reported in the MASS
summaries to represent the typical application rate for each OP-crop
combination in a region. Likewise, OPP used the average number of
applications to determine how many times the OP pesticide would be
applied to the crop in a particular year. These rates were frequently less
than the maximum allowable application rates and frequencies specified
on the label. A comparison of OP cumulative distributions estimated by
typical and maximum label rates in three regions found that use of all
maximum rates generates concentrations that are one to four times
greater than those estimated using typical rates (see I.G. and Appendix
III.E. 11 for detailed analysis). In reality, it is unlikely that aN OP pesticides
would be used on all crops at maximum rates in the same year. Thus, the
difference between OP cumulative loads in a "typical" year and in a year
when intense pest pressure requires maximum label rates for one or more
OP pesticide on one or more crops is likely to be less than the one- to
four-times estimated.
I.E Page 32
-------
T
ii. Timing of Pesticide Application(s)
An application window has been established for each of the OP
pesticide crop uses reported in each region. This window represents an
approximate beginning and ending date for the use of the pesticide on a
particular crop. Delineation of these windows was based on review of
crop profiles and other relevant crop production publications; surveys
such as the Doane Marketing Research, Inc. Agrotrak1"1 reports on
agronomic, row and specialty crops; and on consultations with field
experts. Unless otherwise noted, the default planting and harvesting
dates for crops were taken from the following USDA documents:
Q United States Department of Agriculture, Crop Reporting Board,
Statistical reporting Service. 1977. Usual Planting and Harvesting
Dates for Fresh Market and Processing Vegetables. Agriculture
Handbook No. 507.
Q United States Department of Agriculture, National Agricultural
Statistics Service. 1997. Usual Planting and Harvesting Dates for U.S.
Field Crops. Agricultural Handbook No. 628.
These USDA handbooks also provide "most active" periods during the
planting and/or harvesting windows. The mid-point of the most-active
period was selected as the application date for a pesticide applied at the
"planting" stage of crop production. A case in point is the data input for
terbufos on corn in North Carolina:
Pesticide
Terbufos
Stage
Planting
Application Date
April 17
Range
April 1 - May 20
Most Active
April 10- April 25
When most active periods are not provided, the single application date
for a pesticide is set at the beginning of the crop stage window. Multiple
applications, such as OP cover sprays for tree fruits, are placed at the
beginning and equidistant within the application window. The following
example is for three cover sprays of phosmet on apples in the Northeast
(Pennsylvania):
Pesticide
Phosmet
Stage
Foliar
Application Dates
May 1
June 18
August 5
Range
May 1 - Sep 21
May 1 - Sep 21
May 1 - Sep 21
Because the application dates are held constant through a series of
years of weather patterns, variations in the selected date may affect the
estimated peak concentrations. Relatively high pulse loads from runoff
may occur if application events are closely followed by runoff-producing
I.E Page 33
-------
rains. However, a comparison of OP cumulative distributions resulting
from varying the application dates found that notable differences only
occur at the very highest concentrations that distributions at the 99th
percentile only vary by a factor of 2 or less (see I.G. and Appendix III.E.11
for discussion).
OPP assumed that the entire application of a given pesticide on a
given crop occurred on the same day. Except in Region C, where detailed
pesticide use reports from California Department of Pesticide Regulations
were available, sufficient usage information was not available to split .
applications. While this is likely to result in conservative (health-protective)
estimates, the assumption is not unreasonable in the smaller, more
vulnerable watersheds represented by the index reservoir. A comparison
of estimated OP cumulative distributions using split- and single-
applications in Region C found a difference of less than a factor of two
across the distribution profile (see I.G and Appendix III.E. 11 for
discussion).
A most likely, or predominant, application method is also designated
for each pesticide. The choice is simply "air" or "ground." Review of
MASS and proprietary data bases, crop production profiles, as well as
consultation with field experts, informed these application method
determinations.
iii. Use of CDPR Use Information in Region C
For the Central Valley (CA), used in Region C, the California
Department of Pesticide Regulation, Pesticide Use Reporting (PUR) data
was used to determine both the acres treated and the application dates.
The PUR contains detailed information on every commercial pesticide
application made within the State of California. Since the two locations
identified and assessed in this region were located in the State of
California, the Agency used the PUR data base to calculate the total area
treated by each pesticide, on each crop for each date. For some uses,
growers reporting making applications on numerous dates (>50 days)
throughout the Calendar year. For data management purposes, five
application dates were selected for each crop-OP use to be used in the
assessment; each application date represents 20% of the total acre
treatments made for that particular use.
Evaluations of CDPR and NASS usage information in California found
no routine under- or over-estimation of pesticide usage from the survey
methods used by NASS. A comparison of OP cumulative distributions
generated using both data sources found that the distributions generated
with the more complete CDPR information were greater than those
generated with the NASS survey data by a factor of 3 (see Appendix
III.E.11 for comparison).
I.E Page 34
-------
»
*«**
f. Incorporate the Drinking Water Exposure Estimate into the
Cumulative Assessment
In summary, within each region, a residue file was generated by PRZM-
EXAMS/IR for each pesticide:crop combination which was reported in the
county or counties selected for assessment. This day-by-day residue file was
modified by the CAP specific to that pesticide:crop combination and the
relative potency factor for that pesticide. Then, the modified residue files for
all pesticide:crop combinations for that location were summed across days to
give a distribution of combined daily residues in drinking water.
This distribution of combined daily residues can then be used as an input
file for the CALENDEX model which is discussed elsewhere in this document.
CALENDEX allows the Agency to combine OP concentrations from water and
residential exposures which are time and location dependent with food
exposures which are not time and location dependent.
The distribution of daily residues can also be compared to any water
monitoring data available for the chemicals and region being examined. Plots
of the daily distributions can be analyzed to ascertain which uses may be
expected to contribute significant exposures. The comparison of monitoring
data and the understanding of which uses contribute to exposure are
important aspects of risk characterization of the water portion of the OP
cumulative risk assessment.
For each vulnerable site, OPP developed a site-specific scenario for each
crop group with reported OP usage (see Appendix III.E.7.a and b for a
description of scenario development and documentation). Thus, the site and
soil characteristics are representative of those that actually occur in the
region and support that particular crop growth.
I.E Page 35
-------
XOMMM
AViW
.IW.W*
I I. Revised OP Cumulative Risk Assessment
5
F. Cumulative Assessment
1. Introduction
The previous four sections of this document have described the development
of the major components of the risk assessment. They describe a highly
complex process of combining multiple data sets to develop a description of the
possible risks from OP pesticides by each of the pathways described. OPP has
had to develop new methods for each component of the assessment in order to
produce an assessment which presents as realistically as possible the potential
exposure to OP pesticides. The purpose of this section is to explain the
concepts used to accumulate risk from each pathway into a total risk estimate,
and to provide a basis for understanding the presentations that are provided in
Section III for each of the regional assessments.
2. Basic Concepts
The definition of cumulative risk developed as a result of the passage of
FQPA requires OPP to conduct a risk assessment for a group of pesticides with
a common mechanism of toxicity that is multi-pathway, multi-route, and multi-
chemical in scope. As described in section I.B above, the RPF method was
used to address the issue of combining toxic responses from OPs with varying
propensities to inhibit acetylcholinesterase. Exposure to each OP was
normalized to equivalent exposure to the index compound, methamidophos.
The toxicity data currently available for conducting this analysis are estimates of
response by route-specific dosing, and do not support estimating delivered dose
to the target tissue. OPP decided to address this problem by comparing route-
specific exposures to route-specific points of departure to produce unitless
margins of exposure for each route. In this case, the POD was a BMD10. MOEs
were combined by taking the inverse for each route, adding them together, and
then taking the inverse of that sum. This process was used to produce a
distribution of daily estimates for the subpopulation of concern that reflects
regional and seasonal variation in the patterns of exposure that are likely to
occur throughout the US across the year. OPP used a probabilistic assessment
to capture the full range of exposure possibilities from all sources analyzed. The
intent was to produce an estimated range of risk that is as realistic as possible.
The OP cumulative risk assessment is not a high end risk assessment. It
attempts to reflect the full range of likely exposures for consideration in a
regulatory context. However, at the same time it is designed to avoid developing
extreme exposure estimates based upon the combination of exposure scenarios
and assumptions that are not reasonable.
I.F Page 1
-------
3. Framing the Population-Based Assessment
OPP focuses its risk assessment on exposure and resulting risk to the
population, not to risk to an individual. This distinction is an important one.with
regard to defining how the components of the assessment will be combined.
The current assessment focuses on highlighting inter-individual patterns of
exposure instead of attempting to define intra-individual patterns of exposure.
OPP made this choice because of the lack of acceptable longitudinal data
defining intra-individual behavior for any component of the risk assessment. This
issue has been repeatedly discussed at SAP meetings reviewing the conduct of
dietary risk assessment methodology. Longitudinal data permitting modeling of
the consumption of food and water by the US population is not available. The
data describing the use of pesticides in a residential setting is even more
uncertain. Although ranges of use parameters are available and have been
used in this assessment, they are only adequate to define the behavior of the
population across time, and cannot accurately reflect the day to day variability in
behavior of an individual. Therefore, OPP decided to develop a series of daily
exposure distributions and array them as a distribution across time.
The distribution of daily exposures and resulting MOEs are developed such
that the exposures from OPs in foods, drinking water, and from residential uses
are all calculated simultaneously for each hypothetical individual in the
subpopulation. OPP uses the Calendex software to develop the distributions
and resulting MOEs. Calendex permits incorporation of time course information
with regard to residential uses of pesticides, but does not permit specific
allowance for regional variability. OPP addressed this issue by running separate
risk assessments for each of seven regions of the US. The regions correspond
to agronomic cropping areas and reflect climatic and soil conditions that are
likely to affect pest pressure and resulting pesticide use. Regional differences in
pesticide use are major considerations in appropriately estimating exposure from
pesticides in drinking water and residential uses.
i
To generate a daily distribution of exposure, consumption records are
selected from the CSFII for each individual in the survey. Calendex uses this
consumption record to estimate OP exposure from food by randomly assigning a
residue value for each food reported consumed by that individual. After
multiplying each amount of food consumed by its selected residue value, the •
total exposure from food for this individual is calculated by summing the
exposures from the individual foods which were reported consumed. At the
same time, all appropriate residential scenarios that may be encountered for the
calendar day 1 (January 1) are reviewed. A probability-based decision is made
as to whether or not that scenario will be encountered (e.g., a lawn treatment
would probably not occur in January in the Northeast/North Central region). If
the scenario is assigned a "yes" answer (i.e., treatment does occur), then the
appropriate values defining the exposure are selected from the many
distributions of input parameters for residential exposure scenarios. The .
exposures for the dermal, oral and inhalation pathways are calculated for all
I.F Page 2
-------
i selected residential scenarios. A drinking water value taken from the
I PRZM/EXAMS output for January 1 is selected and paired with the water
I consumption reported in the CSFII consumption record. These values are used
| to calculated exposure from drinking water for that date. All of the exposures are
| converted to route-specific MOEs to define the total exposure to the hypothetical
| individual on January 1. The process is repeated for each consumption record
| for the age group in the CSFII ten times (i.e., ten iterations) to build a distribution
| of exposures for January 1. This process is repeated for January 2, January 3
| and so forth across the year.
I
| The 365 daily exposure distributions are arrayed together in order to provide
I a profile of possible exposures by each route and in total as MOEs. An example
f of such a distribution of distributions is presented in Figure I.F.1. In this figure,
| each daily distribution is arrayed on the yz plane of the plot. Day 365 can be
I clearly seen on the right side of the plot. This distribution of total risk is
I expressed as a cumulative distribution function of MOEs versus percentile of
I exposure. Percentile of exposure refers to that portion of the population output
= distribution that has less than or equal exposure. For example, at the 80th
I percentile of exposure, 20% of the output distribution has an MOE lower than the
I one at the 80th percentile point on the distribution.
E
I The distribution of daily distributions is used to estimate'the potential risk,
| with accompanying distributions generated for each pathway and route. OPP
I acknowledges that this approach does not describe intra-individual risk. In all
I likelihood, the variability in an individual exposure would be much greater than in
I a population-based approach because of the limited likelihood of repetitive
| events such as residential pesticide applications. However, the population at
| large will experience some degree of exposure each day. This factor is a likely
1 source of conservatism in the current assessment.
I 4. Interpreting the Outputs
I The results of the final assessment are presented in tabular (Calendex
I output) form in the appendices. The reflect year-long slices across the 3-
| dimensional plot in Figure I.F.1. In that plot, dark lines can be seen across the
1 total MOE surface. For instance, the top line in the 3-dimensional plot
I represents the 99.9th percentile of exposure for the population. A slice through
| the surface parallel to the xy plane at the 99.9th percentile would look like the plot
I presented in Figure I.F.2. This plot presents the potential total MOE for the
I exposure scenarios included in this assessment. In addition, the contributions
| from various pathways and routes of exposure are arrayed separately to assist
I the risk manager in identifying contributors to risk for further evaluation. Other
| age groups (or percentiles) of exposure may also be of interest. For example,
| Figure I.F.3 presents the results of the 99.9 percentile assessment for the age
I group Adults, 20-49.
I.F Page 3
-------
I 5. The Rolling Time Frame Approach
:
:
| One important aspect of the revised cumulative risk assessment for the
I organophosphate pesticides (OPs) is the manner in which estimated exposure is
| compared with toxicity endpoints. The above paragraphs detail and describe
= one "mode" or option of analysis (termed the single consecutive day option) in
[ which separate, independent estimates are made for each day of the year
{ (January 1, January 2, etc.). As discussed above, these can be arrayed into an
I exposure timeline for any selected percentile (and graphed, if desired). That is,
I for example, the estimated 99.9th (or any other percentile) percentile exposure
I value is calculated by DEEM/Calendex for each day of the year from January 1
I through December 31. These represent independent daily estimates of the 99th
| percentile exposures on each day of the year and do not necessarily represent
5 = the same individual on consecutive days1. Thus, it is NOT possible (with this
mode of analysis) to interpret an extended period (or series) of elevated
exposures over time as necessarily representing extended exposures to the
same individual, and comparison of any estimated exposure to multi-day
endpoints (e.g., a multi-day BMD10 would be expected to provide a very
conservative estimate of risk to the extent that exposures on consecutive days at
any given percentile are unlikely to be the same individual.
[ 1 For example, biomonitoring data from CDC and others indicate that a sizeable
I percentage of the U. S. population has measurable levels of OP metabolites in their
I urine or blood.
I
| I.F Page 4
-------
<
100
1000
> Figure I.F-1. Three-dimensional plot of the total MOE by day of the year and
* %»»>• •
**£f | percentile of exposure
fMKf
I.F Page 5
-------
Figure I.F-2. Cumulative Assessment - 99.9th Percentile Estimate for Children Ages 1-2 Years for AM Routes and
Pathways
Cumulative MOEs for Children 1 -2 Region A One Day Analysis
Julian Days
..
:";'' '•';
1000000 i
-FoodMOE PRZM-EXAMS Water MOE Total MOE Inhalation MOE Dermal MOE Oral (non-dietary) MOE
I.F Page 6
-------
Figure I.F-3. Cumulative Assessment - 99,9th Percentile Estimate for Aduits 20-49 Years for AH Routes and
Pathways
Cumulative MOEs for Adults 20-49 Region A Single Day Analysis
Julian Days
CMCN(CM(N
-------
The DEEM/Calendex program can perform analyses under a second option.
Under this second option (termed the multiple sequential day option), a rolling (or
sliding) time frame is used and multi-day average exposures are calculated for
each individual (e.g, average exposures for each individual for January 1 through
January 7, January 2 through January 8, etc.). Under this mode, average
exposures over multiple consecutive days (e.g., January 1 through 7, January 2
through 8, etc.) are assessed for the same individual. It is then this distribution
of multi-day average exposures at any given percentile which serves as a basis
of comparison with the (multi-day) BMD10. An example graph of this is presented
in Figure I.F-4 which shows a seven day rolling average exposure profile for
Children 1-2.
In the Preliminary Cumulative Risk Assessment, exposures were estimated
I on a single-day basis (the first option) and a comparison made of each
| independent DEEM-estimated single-day exposure with the steady-state (21
I day) equilibrium BMD10 value That is, separate exposure estimates were made
| for January 1, January 2, etc. for each individual in the CSFII survey for each
{ (single) day of the yearwith exposure at a given percentile (e.g., 99th) calculated
I and compared to a multi-day BMD10. In viewing these results, and despite their
I one-day exposure basis, OPP is NOT concerned with exposure spikes lasting
| only one or perhaps a few days since the MOE's associated with these "spikes"
| are based on multi-day toxicity endpoints. Rather, OPP is interested in extended
I periods of high exposure (or, equivalently, low MOEs) which indicate not that an
I individual is being exposed to high levels of OP pesticides over a multi-day time
| period, but instead that the overall level of exposure to the sub-population in the
I tails of the distribution has increased. This is an important distinction which
| brings up two issues:
I Q comparing a series of elevated single-day exposures to multi-day endpoint
I may have less relevance than comparing a multi-day average exposure (at
I any given percentile) to a multi-day endpoint.
rj Consecutive single-day estimates of exposure are likely to significantly
overestimate multi-day exposures to an individual (at higher percentiles) e.g.,
the 99.9th percentile individuals are unlikely to be the same individual on
consecutive days.
An alternative option - which was explored and incorporated into this revised
CRA and supplements the single consecutive day option - is to estimate multi-
day rolling average exposures in which average exposures over multiple
consecutive days (e.g., January 1 through 7, January 2 through 8, etc.) are
assessed for the same individual. It is this multi-day average exposure which
then serves as a basis of comparison with the (multi-day) BMD10. There are a
number of advantages to this alternative. In addition to providing a means of
estimating exposure which is more directly comparable to a multi-day endpoint,
the multiple sequential day mode of analysis better incorporates variability in
exposure for an individual across multiple days and is likely to provide a more
I.F Page 8
-------
1 realistic estimate of exposures for individuals across multiple days. It is also
I flexible with respect to matching time-frame associated with BMD10 in that multi-
| day averages can be calculated over 7, 14, 21, or 28 days. However, as
| discussed in the February 2002 Scientific Advisory Panel meeting associated
| cholinesterase inhibition level will be underestimated if one fails to allow for the
i residual (or lingering) cholinesterase inhibition effect from those previous days in
O f cases where a day's exposure is preceded by nonnegligible exposures on
i previous days.
=
jj
a
<••
^""5
*"*&
I.F Page 9
-------
Figure I.F-4. Cumulative Assessment - Seven Day Rolling Average 99.9th Percentile Estimate for Children Ages 1-
2 Years for All Routes and Pathways
Cumulative MOEs for Children 1-2 Region A Seven Day Rolling Average Analysis
Julian Days
lOOtXXW
: Si a i-i^SSSSSsSSsiSij; SSSSSSSSSS
... ' . . . . '
; ;: ' • ' :-••:•;. • . ' ' '
: ; ' . . ' 'SffiF •' "• "" ''!':""f
1 ' • ' ' . . :i$i:i:!:!:i::S:j:^ " . J.v::>:^:^
•;.•; • .' . ''''''''•'; ' . ;' . • ":'-..'. •"• -' '."
'''Ill II j\, :|||''
%Psi;";
' ,.• "... :;-r: .". '," : -•'-:•• '•^.•-:-
'•:?••&'$$$$$$. is \ . • " 111 ;. •' '.-•'•:• ' ::
• . •'•>:
1
•: ;•; :• •;•: ;•;•:•:• -'•
•FoodMOE PRZM-EXAMS Water MOE • Total MOE Inhalation MOE Dermal MOE Oral (non-dietary) MOE
I.F Page 10
-------
I. Revised OP Cumulative Risk Assessment
G. FQPA Safety Factor
1. Introduction
There is currently a significant focus on the potential susceptibility and
increased sensitivity of infants and children to toxic effects of chemicals (see
National Resource Council's 1993 report, Pesticides in the Diets of Infants and
Children). The Food Quality Protection Act of 1996 (FQPA) instructs the U.S.
Environmental Protection Agency (EPA or the Agency), in making its "reasonable
certainty of no harm" finding, that in "the case of threshold effects, ... an
additional tenfold margin of safety for the pesticide chemical residue and
other sources of exposure shall be applied for infants and children to take into
account potential pre- and postnatal toxicity and completeness of data with
respect to exposure and toxicity to infants and children." Section 408
(b)(2)(C) further states that "the Administrator may use a different margin of
safety for the pesticide chemical residue only if, on the basis of reliable data,
such margin will be safe for infants and children."
a. Guidance Used for Consideration of the FQPA Safety Factor
EPA's Office of Pesticide Programs (OPP) has recently released revised
guidance addressing application of the FQPA safety factor provision in risk
assessments for individual pesticide chemicals (USEPA, 2002a).
Additionally, OPP has prepared a separate guidance document addressing
the application of the FQPA safety factor provision in the context of
cumulative risk assessments for two or more pesticides sharing a common
mechanism of toxicity (USEPA, 2002b; released February 28, 2002 for a 60-
day comment period). Both FQPA safety factor guidance documents
(USEPA, 2002a,b) were used to provide general guidance on applying
traditional uncertainty factors and on implementing the FQPA safety factor
provision for the cumulative risk assessment of organophosphorus (OP)
pesticides. In implementing the FQPA safety factor provision, key
considerations in a cumulative risk assessment are:
Q Determining the completeness of the data with respect to effects that may
occur in the young due to the common mechanism of toxicity;
Q Evaluating the degree of concern regarding the potential for pre- and
postnatal effects associated with the common mechanism of toxicity and
determining the residual uncertainties not addressed by application of
traditional uncertainty factors to account for deficiencies in the toxicity
data; and
Determining the completeness of the exposure database for all pertinent
pathways, of exposure to OP pesticides.
I.G Page 1
-------
I b. Scope of Analysis on Sensitivity and Susceptibility
| Single-chemical risk assessments should generally be conducted for each
| member of a common mechanism group before a cumulative assessment is
I attempted. Thus, previous determinations have been made whether to retain
I or replace the FQPA 10X safety factor for the individual pesticide members of
| the OP cumulative risk assessment group. These FQPA safety factor
I decisions should be revisited, however, in the cumulative risk assessment
| process because they are based on broader considerations of potential toxic
| effects in the young (e.g., teratogenicity, carcinogenic effects) that may not
I relate to the common mechanism of toxicity. A cumulative risk assessment
differs from the single-chemical risk assessment both in focus and purpose.
i The cumulative risk assessment of the OP pesticides is based on their ability
to target and inhibit the enzyme acetylcholinesterase (AChE) in nerve tissue,
in other words, the common mechanism of toxicity for which these pesticides
are grouped. Thus, decisions on the FQPA safety factor for the cumulative
assessment group (CAG) reflect considerations that pertain to the common
effect and the common mechanism of toxicity.
Several years ago, the International Life Sciences Institute/Risk Sciences
Institute (RSI) convened an expert panel to address whether the OP
pesticides act by a common mechanism of toxicity (Mileson et a/., 1998).
Although some OP pesticides may act by several different neurotoxic
mechanisms through interaction with other esterases and nonesterase
targets (for review see Pope and Liu, 2001), there are insufficient data to
support subgrouping of the OP pesticides based on other actions operating
instead of, or in addition to, the inhibition of AChE. It should be pointed out
that these other mechanisms are considered in the individual risk
assessments of the OP pesticides when there is sufficient available
information. For example, in evaluating the susceptibility of the young to
chlorpyrifos, OPP considered data that showed effects on the developing rat
brain such as structural defects and changes in macromolecular synthesis,
neurotransmitter levels, and cell signaling. Although these other
neurodevelopmental mechanisms are considered in the single chemical
assessment, they will only be considered in the cumulative analysis as they
relate to AChE inhibition. Because AChE inhibition is the mechanism of
toxicity and precursor event to toxicity, functional effects in the young that
result from the inhibition of AChE activity should not occur at doses lower
than those causing AChE inhibition.
I.G Page 2
-------
1 2. Hazard Assessment: Sensitivity and Susceptibility1
| The hazard assessment, below, considers the potential pre- and postnatal
| developmental effects that may be associated with the inhibition of AChE, the
| comparative AChE inhibition between adults versus the immature animal, and
| the completeness of toxicity data on AChE inhibition in young animals.
| a. Role of Acetylcholinesterase in Neurodevelopment
T
AChE is the enzyme that hydrolyzes the neurotransmitter acetylcholine at
cholinergic synapses and neuromuscular junctions. The inhibition of AChE
leads to accumulation of synaptic acetylcholine, overstimulatioh of
postsynaptic cholinergic receptors and consequent signs of neurotoxicity or
cholinergic toxicity. It has been suspected, however, for more than 25 years
that AChE may have an extrasynaptic, noncholinergic role during
development (e.g., Karczmaref a/., 1973; Drews, 1975). Recent research
indicates that the roles of AChE during development center around
neurogenesis, cell adhesion and possibly stress response (e.g., Layer and
Willbold, 1995; Grisaru etal., 1999; Bigbee era/., 1999; Brimijoin and
Koenigsberger, 1999). Moreover, the widespread expression of AChE is
often mirrored by the expression of acetylcholine, which is involved with basic
developmental processes such as mitosis, cell-to-cell contact, cell adhesion,
cell differentiation, and organization of the cytoskeleton (reviewed in Wessler
era/., 1999; Lauder and Schambra, 1999).
Both AChE and acetylcholine are highly conserved molecules which have
multiple roles in the developing nervous system as well as extraneuronal
functions. Because AChE controls acetylcholine levels in neuronal as well as
extraneuronal tissues and blood (e.g., Wessler etal., 1998; Fujii and
Kawashima, 2001; Kirkpatrick etal., 2001) and because AChE activity is
more commonly measured as compared to acetylcholine levels, most of the
work reviewed below concentrates on changes in AChE activity rather than
acetylcholine levels. One may assume, however, that as mentioned above, a
decrease in AChE activity should also increase acetylcholine concentration.
Changes in the structure, activity or level of these neuromodulators, AChE or
acetylcholine, may elicit novel effects on the developing brain. It is not known
to what extent neuronal AChE needs to be altered to have adverse effects on
the developing brain, nor is it known what adverse effects on
neurodevelopment may result from AChE inhibition. Nevertheless, because
of the potential developmental role of AChE, it is reasonable to consider the
evidence for whether inhibition of AChE in the developing nervous system
may affect neural development.
1The term susceptibility is used qualitatively to indicate unique effects (e.g., a different pattern of effects
of concern) in the young. The term sensitivity is used to refer to quantitative susceptibility, or to
quantitatively indicate effects of a type similar to those seen in adults, but which occur at doses lower
than those causing effects in adults.
I.G Page 3
-------
In vitro work has shown that some OP compounds can inhibit neurite
outgrowth, but enzyme inhibition does not appear to correlate completely with
inhibition of outgrowth (Dupree and Bigbee, 1994; Layer era/., 1993, Bigbee
era/., 1999). Inhibition of neurite outgrowth is compound-specific, as some
compounds inhibit AChE activity but do not inhibit neurite outgrowth. It is
now accepted that the cell adhesive function of AChE is mediated by a
peripheral anionic site located at the rim of the 20 A gorge, a site distinct from
the catalytic site located at the bottom of that gorge (Johnson and Moore,
1999; Sternfeld et a/., 1998). OP inhibitors bind to the catalytic site; little is
known about prerequisites for binding to the peripheral anionic site mediating
cell adhesiveness. Perhaps some OPs bind specifically to that site or
perhaps some OPs can perturb the function of that site when bound to the
catalytic site (e.g., Bigbee era/., 1999).
In any event, AChE inhibition does not necessarily predict perturbations of
neuronal differentiation. It is possible to create fruit flies (Greenspan et a/.,
1980) or mice (Xie et a/., 2000) that do not produce AChE because they have
no gene for AChE. In fruit flies, this is a lethal mutation, but in mice the
absence of AChE is only lethal to approximately 25% of the homozygous
fetuses in utero. At birth, the surviving homozygous animals appear overtly
normal, but fail to develop normally and usually die by day 21 unless care is
taken to provide their nutritional needs, in which case they may live to
adulthood. The authors speculate that the animals survive because
butyrylcholinesterase assumes many of the biochemical functions of the
absent AChE. As with any study with knockout mice, the phenotype must be
interpreted with caution as compensation may occur during development that
would not mimic AChE inhibition during development.
Is there evidence that exposure to OP pesticides pre- or postnatally
perturbs neurodevelopment? Some animal studies using prenatal exposures
show effects on neurodevelopment, while other studies do not show any
effect. In general, the literature shows that high levels of dosing of an OP
during gestation (e.g., affecting maternal weight gain) will tend to be
embryotoxic (i.e., lethal). More subtle effects may be noted at lower doses if
other neurodevelopmental specific tests are employed. For example, the
offspring of mice receiving diazinon during gestation showed developmental
delays and abnormal endurance and coordination at doses of 0.18 or 9
mg/kg/day (Spyker and Avery, 1977). Malathion or dicrotophos showed dose
and age-related abnormalities (assessed histologically) of nervous and
extranervous system development in one-, two-, and three-day-old chick
embryos (Wyttenbach and Thompson, 1985; Garrison and Wyttenbach,
1985) An in vivo study of malathion, however, showed no teratological
effects in rabbits dosed from day 7 to day 12 of gestation (100 mg/kg; Machin
and McBride, 1989); note that this study did not include any detailed
assessment of nervous system tissues. Fetal brains of rats given chlorpyrifos
repeatedly during late gestation show abnormalities in neuronal migration and
other biochemical endpoints (Lassiter et a/., 2002; Qiao et a/., 2002).
I.G Page 4
-------
I Gestational exposure (day 6 to day 15) to tribufos, oxydemeton-methyl,
I azinphos-methyl, fenamiphos, isofenphos or fenthion at doses that produced
I 20-50% maternal brain cholinesterase (ChE) (ChE is used when there was
| no distinction between butyryl- or acetyl-cholinesterase in the experimental
| procedure) inhibition showed no embryotoxicity or teratogenicity;
| neurodevelopment was not assessed (Astroff and Young, 1998). Although
O \ the authors conclude that gestational dosing with these compounds caused
I "no effect on fetal ChE," this activity was not assessed until five days after the
I last dose, a time that is not optimal for assessing AChE inhibition in fetal
I tissues (Lassiterera/., 1998; Michalek era/., 1985).
* I Rats given OP pesticides postnatally may show abnormal nervous system
^ | development. In a series of papers exploring the neurotoxicity of postnatally
^- I administered chlorpyrifos, many changes were noted (e.g., RNA levels,
9^ I transcription factor expression, disruption of catecholaminergic and
<3 I cholinergic pathways) (Johnson et a/., 1998; Crumpton et a/., 2000; Dam et
(/) | a/., 1999), resulting in persistent biochemical and behavioral changes long
0) | after the dosing ceased (Dam et a/., 2000; Slotkin ef a/., 2001 a,b; Levin et a/.,
CD I 2001; Slotkin et a/., 2002). Other studies in which chlorpyrifos was
f$ I administered to the dam so that the pups received their dosage only through
the milk were largely negative (Breslin et a/., 1996; Deacon ef a/., 1980;
Maurissen et a/., 2000), although the endpoints examined were not as
targeted and discriminating as those used by the Slotkin laboratory. The
relationship of these neurodevelopmental changes to ChE inhibition is
unclear because many studies are lacking correlative ChE activity, thus
making it difficult to draw firm conclusions. In the few prenatal studies where
ChE activity was assessed, however, few of these effects occur at dose
levels that do not inhibit ChE activity in the fetal brain, and probably none of
these effects occur in the absence of ChE inhibition in maternal tissues. In
both the studies assessing prenatal effects of chlorpyrifos, effects on brain
development were noted at dosages (1 mg/kg/day) that did not inhibit fetal
brain ChE (Lassiter ef a/., 2002; Qiao et a/., 2002), but would be predicted to
show inhibition of maternal blood and brain ChE activity (Maurissen era/.,
2000). In postnatal studies, there are no reports of effects in the absence of
ChE inhibition. In some cases, this assertion is made by the authors, but the
authors fail to ascertain that the ChE measurements were taken at the time of
peak effect. Often the measurements are taken 24 hours after the last dose,
rather than assessing ChE activity during the entire dosing period. Thus, it is
reasonable to assume that adverse neurodevelopmental outcomes that are a
result of the inhibition of ChE should not occur at doses that do not inhibit
ChE. Because, however, the cumulative risk assessment is based on adult
brain ChE data, it is important to address the age-related sensitivity of ChE
inhibition in the adult versus the young animal. The available studies are
reviewed below.
I.G Page 5
-------
b. Differential Sensitivity of the Young Compared to the Adult
Although reports of increased sensitivity of the young following exposure
to OP pesticides date back over two decades, it is the work that has emerged
recently that provides a better basis for understanding the issues concerning
the sensitivity of the young to ChE inhibition. This understanding comes from
recently generated chemical-specific data in young animals on ChE activity,
as well as generic human and animal studies on the biological and
biochemical parameters involved in age-dependent sensitivity. The current
state of the knowledge is summarized and discussed below.
i. Human Incident Information
There are reports of symptoms associated with cholinergic toxicity due
to accidental acute exposures. A 1999 review based on pesticide-related
exposures (excluding cases of exposure to multiple products, attempted
suicides, malicious intent, and confirmed non-exposure) examined Poison
Control Centers Data from 1993 through 1996 (USEPA, 1999). Of the
exposures that occurred in a residential setting 16% were due to OP
pesticides. The review of the residential pesticide exposure concluded:
i Organophosphate pesticides pose a greater hazard from accidental
I acute exposure than do other pesticides, especially for children under
f six years-of-age. Children were three times more likely to be
I hospitalized, five times more likely to be admitted for critical care, and
| four times more likely to have experienced a major medical outcome or
| death than if exposed to some other pesticide.
I In this review of residential exposures, there were 24,889 exposures
| reported in children under the age of six, 5,080 exposures among children
| six to 19 years-old, and 32,087 exposures among adults. Of those cases
I with medical outcomes determined, children under age six were 22%
I more likely to experience a life-threatening or fatal outcome as a result of
1 their exposure than adults or children six to 19 years-old. Additionally,
I . based on the Centers for Disease Control mortality data (see
1 I http://wonder.cdc.gov/mortsqu.shtml). the ratio for death in young children
[ exposed to OP pesticides was 3.3 times higher than in adults.
These data show that there is more potential for harmful exposures in
young children than in older age groups, but they do not necessarily
demonstrate an increase in the sensitivity of young children. There is a
possibility that young children may be exposed to higher doses on a body
weight basis compared to adults (from spills, ingestion, inhalation)
because they are ignorant of the hazard, and not because of differences
in sensitivity based on age to the effects of these pesticides.
Furthermore, the human data on children come from accidental
exposures to these pesticides that are associated with acute poisoning
I.G Page 6
-------
resulting in significantly higher blood, tissue, and urine concentrations of
these chemicals compared to exposures that humans would normally
encounter in food or the environment.
Because of the reasons stated above, it is difficult to draw conclusions
from human incident data on the sensitivity of the young compared to
adults. The animal literature below allows for evaluations of age-
dependent sensitivity.
ii. Laboratory Animal Studies
Some studies are available in the open literature that have evaluated
ChE inhibition following in utero or lactational exposures to OP pesticides,
as well as dosing of young animals. EPA issued a Data Call-In (DCI) on
September 10, 1999 for adult and developmental neurotoxicity (DNT)
studies on the OP pesticides, and as part of the DNT protocol, measures
of brain, red blood cell (RBC), and plasma ChE activity in dams and pups
were required to characterize comparative levels of inhibition at the time
of peak effect. However, very few DNT rat studies have been submitted
to the Agency.2 In addition to studies on OP pesticides that allows a
comparison of the differential in response to ChE inhibition between adult
and immature rats, several recent published studies provide an important
perspective on the underlying basis for observed increased sensitivity.
The analyses below will focus on differences in ChE inhibition between
fetal, neonates, and juvenile rats compared to adults.
Differential Sensitivity Following Gestational/Lactational
Exposure
Fenamiphos, tribufos, trichlorfon, and oxydemeton-methyl
were evaluated for ChE inhibition in a rat multigeneration reproductive
feeding study (Astroff etal., 1998; discussed in Sheets, 2000). Dams
were treated with these OP pesticides via the diet during gestation and
continuing throughout the lactation period. Pups are assumed to be
exposed due to consumption of feed at about 14-21 days-old.
Plasma and RBC ChE activity were measured in the adults during
the premating phases of both generations following eight weeks of
exposure to each of the OP pesticides, and again at termination when
Out of the 30 OP pesticides included in the December 2001 preliminary cumulative assessment, DNT studies have only been
submitted for chlorpyrifos, dimethoate, malathion, methyl parathion, methamidophos, and tribufos. The DNT studies submitted for
dimethoate, malathion, and methyl parathion also included comparative ChE activity. These studies investigated ChE activity in
adult and immature rats following either acute or repeated dosing. A review of the chlorpyrifos DNT study was completed in 1999,
and reviews of the dimethoate and malathion DNT studies have recently been completed. The DNT studies for tribufos,
methamidophos, and methyl parathion are currently under review, although a review has been completed on the ChE data for
malathion and methyl parathion. It should be pointed out that the DNT studies on methamidophos and tribufos are feeding studies
in which the pups were not directly dosed, and thus the pups were presumed to be exposed only in utero and during lactation; no
comparative ChE data have been submitted for methamidophos and tribufos.
I.G Page 7
-------
brain ChE activities were measured. Separate contingents of
postnatal rats were evaluated for plasma, RBC, and brain ChE activity
on lactation day (LD) 4 and on LD21. The effects found on LD4 could
be due to gestational and lactational exposure, whereas the results on
LD21 may reflect exposure through the milk and some exposure
through the diet as pups begin consuming feed in the late lactational
period, in other words, postnatal days (PND)14-21. Each study
consisted of a control and three dose groups. The highest dose level
was selected based on parental toxicity.
Toxicity (reduced body weights or viability) in the young was not
apparent until there were significant maternal effects (decreased body
weights and food consumption) and substantial ChE inhibition in the
blood and brain of the parental animals. In fact, the adult animals
were more affected than the young in this study. Although young rats,
when exposed to these OPs in utero and via lactation, do not appear
to exhibit more ChE inhibition than is found in maternal tissues, the
dose that may be absorbed by the fetus and adult is unknown. Thus,.
conclusions can not be reached about the relative sensitivity of fetuses
versus dams to ChE inhibition.
Maternal and fetal ChE inhibition were evaluated following maternal
exposure to azinphos-methyl, fenamiphos, fenthion, isofenphos,
tribufos, and oxydemeton-methyl in a prenatal developmental
toxicity study in rats (Astroff and Young, 1998). These pesticides were
administered to the dams by gavage on gestation days (GD) 6-15.
Maternal ChE activity (brain, RBC, plasma) was measured onGD16
and 20, and fetal brain ChE activity was measured on GD20. The
dose levels for these studies were selected such that maternal ChE
inhibition at the highest dose tested was greater than 20%. At the
highest dose tested on GD16 (in plasma [except for azinphos-methyl],
RBC [except for fenamiphos], and brain [except for fenamiphos], and
on GD20 (in plasma [only for fenthion], RBC [except for aniphos
methyl], and brain [except for fenamiphos]), maternal ChE was
significantly inhibited. However, no remarkable brain ChE inhibition
was observed in fetuses at any dose on GD20.
The effect of treatment with chlorpyrifos on ChE activity was
compared in dams and fetuses by Mattsson et al. (1998; 2000).
Pregnant Sprague-Dawley rats were administered chlorpyrifos by
gavage at doses of 0, 0.3, 1.0, or 5.0 mg/kg/day on GD6-20. The
magnitude of brain, plasma, and RBC ChE inhibition in the fetus on
GD20 was found to be less than or equal to that observed in dams. At
5.0 mg/kg/day, ChE activity in fore- and hindbrain of the dams on
GD20 was inhibited by 76.0 and 86.7%, respectively, and by 58.8% in
fetuses. At 1.0 mg/kg/day, brain ChE activity in fore- and hindbrain
was inhibited in dams by 7.8 and 8.0% (statistically significant at
I.G Page 8
-------
I p<;0.05 or 0.01), respectively; there was no statistically significant
I depression of brain ChE activity in fetuses. In another study of the
1 comparative ChE inhibition between dam and fetus with chlorpyrifos,
| Lassiter et a/. (1998) concluded that the fetal brain ChE inhibition was
| less than the maternal brain ChE inhibition during repeated dosing
;\l | primarily because the fetal brain tended to recover more completely
Z> f between doses than the maternal brain ChE. When dams were given
pT I a single dose, both maternal and fetal brain ChE appeared to be
L» | depressed to the same degree, but when subjected to a repeated
^ | dosing regimen, the fetal brain showed less inhibition probably
X) | because of the higher rates of new synthesis or more rapid turnover of
i I inhibited molecules of ChE in the fetuses compared to the adult. In
^ | two different studies which compared the tissue burden of chlorpyrifos
^ | and metabolites in dam and fetus, one group (Mattsson et a/., 1998,
§5 j MRID 44648102, Mattsson et a/., 2000) found lower blood
f^ I concentrations of chlorpyrifos in the fetus as compared to the dam,
0) | whereas another group (Hunter et a/., 1999) found three times more
trichloropyridinol (a metabolite of chlorpyrifos) in the fetal brain as
compared to the maternal brain. Trichloropyridinol (TCP) can either be
produced as a by-product of a toxic action (i.e., TCP is the leaving
group wh chlorpyrofos-oxon binds to ChE) or as a detoxification action
(e.g., TCP can be produced when chlorpyrifos-oxon is catalyzed by
PON1).
Results of a recently submitted DNT study with dimethoate
indicated that treatment by gavage of dams with the pesticide induces
equal or less inhibition of ChE in the fetus compared with the dams
(Meyers, 2001; MRID 45529702). Treatment of dams with 0.5
mg/kg/day of dimethoate during GD6-20 induced statistically significant
but marginal ChE inhibition (10%) in brain tissue of both adult and fetal
rats. The responses at 3 mg/kg/day (the highest dose tested)
indicated less brain and RBC ChE inhibition in fetuses (33% and 31%,
respectively) compared with dams (60% and 58%, respectively).
Measurements of ChE inhibition were also conducted on four-day-old
pups that were exposed to dimethoate in utero from GD6 to GD20, but
not directly exposed postnatally. At 3.0 mg/kg/day, brain and RBC
ChE activity was inhibited by 13% and 17% in the PND4 pups.
CD i
v?<> :
Cl) I
XW* -
{/> 1
A DNT study performed with malathion (Fulcher, 2001; MRID
45566201) also showed that there was less effect on ChE activity
(measured at GD20) in fetuses than in dams that had been treated by
gavage with the pesticide during GD6-GD20. At the highest dose
examined (150 mg/kg/day), RBC ChE was inhibited by 19% in fetuses
and by 51% in the dams. No effects on brain ChE activity were
observed in either the fetuses or dams at that dosage. At PND4, at
which time the only exposure to malathion could be through milk, ChE
activities in treated pups were comparable to controls.
I.G Page 9
-------
I ChE data that were recently submitted to the Agency (Beyrouty,
I 2002b; MRID 45656501), supplemental to a DNT study on methyl
1 parathion, demonstrated that treatment of dams by gavage from GD6-
[ 20 induced more ChE inhibition in the brain of dams than in the
[ fetuses. Analyses of brain tissue at GD20 showed that ChE activity
| was inhibited by 31% in dams at a dose of 0.60 mg/kg/day (the highest
O | . dose tested), while there was no brain ChE inhibition in their fetuses.
Vvv •• / '
I At the same dose, RBC ChE inhibition was 58% in dams and 22% and
| 18% in male and female fetuses, respectively. In PND4 pups, ChE
I was not inhibited in any compartment.
I In summary, results of studies with fenamiphos, tribufos, trichlorfon,
i oxydemeton-methyl, chlorpyrifos, methyl parathion, dimethoate,
I malathion, and azinphos-methyl show that treatment of pregnant dams
i with an OP pesticide during gestation induces more ChE inhibition in
i the dams than in the fetus. Data from these studies also show that the
| newborn (one- to four-day-old pups), when only exposed in utero or
| possibly through the milk, also has less inhibition of ChE than the
[ maternal, adult rat. The lack of similar levels of ChE inhibition in
| fetuses or neonates relative to adults may be due to the fetuses
I receiving a lower dose of these OP pesticides compared to their dams
I because of pharmacokinetic differences, such as a lower dose being
I transferred to the fetus through the placenta or to the neonate through
CO I the milk than is received by the dam directly in the diet. A lower
V' I response in the immature animal may also be due to the increased
"*"* I synthesis or more rapid turnover of inhibited molecules of ChE in the
Cp I fetal brain compared to the adult (Lassiter et a/., 1998; Mortensen et
> I a/., 1998).
Differential Sensitivity Following Direct Postnatal Exposures
Neonatal, juvenile, and adult rats show differential sensitivity to
ChE inhibition following an acute gavage treatment with chlorpyrifos
(Pope, 2001 a). When rats from each age group were administered
chlorpyrifos at 0.5 times the LD10(7.5 mg/kg, neonates; 23.5 mg/kg,
juveniles; 68 mg/kg, adults), peak ChE inhibition in the cortex
(estimated from Figure 11 of the report) was 70% (neonates), 65%
(juveniles), and 68% (adults). Thus, based on similar magnitudes of
peak ChE inhibition at 0.5 of a LD10 dose and considering the
differentials in the 0.5 LD10 doses, neonates were shown to be about
threefold more sensitive than juveniles and about ninefold more
sensitive than adults. In another study by Moser et a/. (1998) a single
gavage dose of 20 mg/kg produced 89% and 91% (males and
females) brain ChE inhibition in PND17 pups, compared to 39% and
36% (males and females) inhibition in adults (i.e., about a twofold
difference in relative sensitivity).
I.GPagelO
-------
Chlorpyrifos produces a minimal difference in ChE inhibition in
neonatal rats compared to adult rats following repeated dosing (14
treatments by gavage during PND7 to PND21). Based on ED50 levels
(adults, 3.3 mg/kg/day and neonates, 2.2 mg/kg/day), the difference in
the response of neonates to brain ChE inhibition compared to adult
males is a 1.5-fold increase (Zheng et a/., 2000).
Similar results were reported in a recent study in which neonatal
and adult rats were administered chlorpyrifos by subcutaneous
injection (Liu etal., 1999). Neonatal (seven-day-old) pups and adults
were administered 0, 5, or 10 mg/kg/day for seven or 14 days and
sacrificed for ChE measurements one day after the final dose. At
seven days, inhibition of ChE activity in the cortex and striatum of the
neonates was 62 and 65%, respectively, compared with 50 and 55% in
adult animals. Following 14 days of treatment, ChE inhibition in the
cortex and striatum of neonates was 60 and 65%, respectively and
65% in both of these tissues from adult animals.
i Diazinon. A recent abstract by Moser and coworkers (Padilla et
I . a/., 2002) reported an increased sensitivity to ChE inhibition for
| diazinon when PND17 pups were given a single oral dose (via gavage)
| of 75 mg/kg (75% brain ChE inhibition) compared to adult rats (38%
| brain ChE inhibition). This observation was correlated with
I detoxification by carboxylesterases and A-esterases (as discussed
I later in Section H.C.). There are no data available on the effects on
I ChE activity following repeated dosing of neonates or young adults
i with diazinon.
Dimethoate. In a recent DNT study (Meyers, 2001; MRID
45529702), dimethoate was evaluated for ChE activity in plasma,
RBC, and brain following acute exposures and repeated dosing
(subacute exposures) to 0.1, 0.5, and 3 mg/kg/day of dimethoate.
Dimethoate was given by gavage to pregnant rats GD6 through LD10
(LD10; equivalent to PND10); their pups were treated by gavage from
PND11 through PND21. Plasma, RBC, and brain ChE was measured
at GD20 (dams and fetuses), PND4 (pups only), and PND21 (pups
only). ChE activity was also measured following an acute dose of
dimethoate to additional groups of young adult and PND11 rats. In
general, there was no striking difference in sensitivity to dimethoate-
induced brain or plasma ChE inhibition between males and females of
either adults or pups following acute or repeated treatment.
Acute (single dose) treatment with dimethoate of adult male and
female rats and 11-day-old offspring with 3 mg/kg/day induced
statistically significant, treatment related ChE inhibition in brain or
RBC. At that dose, brain ChE inhibition in adult male and female rats
was 12% and 14%, respectively, and in day 11 male and female
I.GPage11
-------
"-•"•.•<„
offspring 17% and 18%, respectively. At 3 mg/kg/day, RBC ChE
inhibition was greater in adult females than in adult males (26% versus
17%) and there was no statistically significant depression of RBC ChE
activity in PND11 offspring.
Repeated dosing (11 doses) with 0.5 mg/kg/day dimethoate
induced statistically significant but marginal brain ChE inhibition (10-
13%) in both sexes of adult and 21-day-old rats. The response is
likely due to treatment with the chemical because of the positive
finding in both sexes of both age groups and because data from GD20
dams also showed an effect at 0.5 mg/kg/day. At the 3 mg/kg/day
dose level, brain ChE inhibition was substantial in both adults (up to
58%) and 21-day-old offspring (up to 45%). In the repeated dosing
study, a small, but statistically significant, difference in brain ChE
inhibition was found at the low dose (0.1 mg/kg) between adults and
pups (GD20, PND4 and PND21). As the dose was increased, this
differential was not found at the high dose 3.0 mg/kg (see Table 1).
Because dimethoate does not show age-dependent sensitivity
(discussed above), it is reasonable to assume that its •
oxon-omethoate-will also not show a differential toxicity in adults
versus pups. Unlike acephate (discussed later), the parent
compound-dimethoate-has been characterized for ChE inhibition in
the young animals compared to adults.
Malathion. Recently submitted ChE data (supplemental to a DNT
study) (Fulcher, 2001; MRID 45566201) clearly demonstrate a
differential sensitivity to inhibition of the ChE enzyme in immature
animals compared to adult rats treated with acute or repeated
exposure to malathion. In this study, pregnant rats were administered
malathion by gavage from GD6 through LD10; gavage dosing of their
pups was then continued from PND11 through 21. Plasma, RBC, and
brain ChE was measured at GD20 (dams and fetuses), PND4 (pups
only), and PND21 (pups only). ChE activity was also measured in
additional groups of young adult and immature (PND11) rats that had
been administered a single (acute) dose of malathion. The dose levels
were 5, 50, 150 mg/kg/day in the repeated dosing studies, and 5, 50,
150, and 450 mg/kg in the acute studies.
Following an acute dose of malathion, brain ChE was inhibited in
PND11 pups at 150 mg/kg (44% in males and 48% in females) and at
450 mg/kg (84% in males and 81% in females), while brain ChE was
not affected in young adults at either of those doses. At 450 mg/kg,
however, RBC ChE was inhibited in both young adults (25% in males
and 17% in females) and PND11 pups (72% in males and 61% in
females).
I.GPage12
-------
Repeated dosing of malathion at 150 mg/kg/day from PND11
through PND21 (11 days of treatment) produced a marked inhibition of
plasma (24-32%), RBC (67-68%), and brain (16%) ChE compared to
controls. For dams, 14 days of treatment at 150 mg/kg/day resulted in
RBC ChE inhibition (51%), but no inhibition of plasma or brain ChE.
Similarly, for young adult rats that were treated for 11 days with 150
mg/kg/day malathion, RBC ChE was inhibited 43% in males and 48%
in females, while plasma and brain ChE were not affected. At 50
mg/kg/day, plasma (19%) and RBC (34-39%), but not brain, ChE
activity was inhibited in the PND21 offspring, while in dams and young
adults, only RBC ChE (19-20%) was inhibited. No effects on ChE
activity were seen at 5 mg/kg/day for dams or young adults. In PND21
offspring, however, RBC but not plasma or brain ChE was inhibited
(17% in males, 15% in females).
Methamidophos was evaluated for age-related differences in ChE
inhibition by Moser (1999). Comparisons for brain and blood ChE
activity were made between PND17 and adult rats following acute oral
doses of 1, 4, or 8 mg/kg. The dose response curves for ChE
inhibition were quite similar between pups and adult rats. ED50 values
for brain ChE inhibition in PND17 and adult rats were approximately
3.3 and 3.0 mg/kg/day, respectively. The ED50 values for blood ChE
inhibition were 2.5 (PND17) and 2.2 (adults) mg/kg/day.
Although acephate is metabolized to methamidiphos, it is not
possible to determine, based on available data, whether acephate
would show comparable responses in adult and young rats. This is
because acephate, the parent compound, has not been evaluated for
comparative ChE activity in young versus adult animals. In rats, only a
small portion of acephate is metabolized to methamidiphos (5%)
(Warnock, 1973; MRID 00014219). Furthermore, it is unknown to
what extent the parent chemical may induce ChE inhibition or to what
extent the parent chemical may alter the effects of methamidiphos on
ChE activity in the adult or young rat.
Treatment of neonatal, juvenile, and adult rats with a single gavage
dose of methyl parathion induces a differential response among the
age groups in ChE inhibition in the brain (cortex) (Pope, 2001 a).
Treatment with methyl parathion at doses of 1.0 mg/kg (neonates),
2.05 mg/kg (juveniles), or 7.3 mg/kg (adults) induced similar
magnitudes of peak ChE inhibition (60%-70%, estimated from Figure
14 of the report). Based on a comparison of the doses that induced
similar levels of ChE inhibition, neonates are 2.5-fold more sensitive
than juvenile rats and about sevenfold more sensitive than adult rats to
methyl parathion induced ChE inhibition.
I.G Page 13
-------
Methyl parathion was investigated by Liu et a/., (1999) for ChE
activity and other neurochemical effects after repeated dosing in
postnatal and adult male rats. Adult and postnatal rats (eight-day-old)
were treated with methyl parathion subcutaneously at 1.5 mg/kg/day or
3.0 mg/kg/day for either seven or 14 consecutive days. Brain ChE
activity was measured in the cortex and in the striatum one day after
seven days of dosing or eight days after 14 days of dosing. Brain ChE
activity was more reduced in postnatal pups compared to adults.
Following seven days of dosing at 1.5 mg/kg/day, neonates showed 62
and 75% ChE inhibition in the cortex and striatum, respectively,
compared to 25 and 30% in the adult male rats. In neonates treated
subcutaneously with methyl parathion for 14 days, ChE activity was
inhibited in the cortex by 65% and in the striatum by 80%. ChE activity
was inhibited in adult rats by 40% (cortex) and by 50% (striatum).
A recently submitted DNT study on methyl parathion with
supplemental ChE data (Beyrouty, 2002b; MRID 45656501)
demonstrated a differential sensitivity of immature versus adult rats to
ChE inhibition following acute or repeated exposure. The protocol for
this study was similar to that used for the DNT ChE studies conducted
for dimethoate and malathion. Methyl parathion was administered by
gavage to pregnant rats from GD6 through LD10 at doses of 0.03,
0.11, 0.3, and 0.6 mg/kg/day. Pups from these litters were then
administered methyl parathion by gavage from PND11-21. Plasma,
RBC, and brain ChE was measured at GD20 (dams and fetuses),
PND4 (pups only), and PND21 (pups only). Additional groups of
young adult and PND11 rats were dosed acutely with methyl parathion
(at doses of 0.03, 0.11, 0.3, and 0.6 for adults and doses of 0.03, 0.11,
0.3, and 1.0 for pups), and ChE activity was measured. Following
acute exposures of 0.3 mg/kg, ChE was inhibited in brain (15-18%),
RBC (20-31%), and plasma (25%) in PND11 pups; no inhibition was
observed in any compartment for adults.
Repeated dosing of PND11 to PND21 pups (which had also been
exposed in utero and via lactation) also showed an increased
sensitivity of neonates compared with adult rats to treatment with
methyl parathion. At 0.3 mg/kg/day, ChE activity was inhibited in
PND21 pups (brain 26-29%, RBC 62-65%, and plasma 24-31%). In
dams treated with the same dosage from GD6 to GD20 (i.e., 14 days
of treatment), ChE inhibition was seen in brain (9%) and RBC (35%),
but plasma ChE was not affected. In adult rats treated with 11
repeated doses of 0.3 mg/kg/day methyl parathion, ChE inhibition was
seen in RBC (30% in males and 35% in females) and plasma (25% in
males), but there was no inhibition of brain ChE. At 0.6 mg/kg/day,
brain ChE was inhibited 60-62% in PND21 pups, 31% in GD20 dams,
and 6-13% in adults; RBC ChE was inhibited 85-86% in PND21 pups,
58% in GD20 dams, and 40-58% in adults; and plasma ChE was
I.G Page 14
-------
inhibited 56-61% in PND21 pups, 29% in GD20 dams, and 28-34% in
adults.
Summary of Differential Sensitivity
• Table 1 shows results of ChE measurements performed in acute
and repeat dosing studies with OP pesticides using neonatal, juvenile,
or adult rats. The information provided in Table 1 is confined to data
that could be used to estimate the relative sensitivities of different age
groups based on the amount of ChE inhibition reported following
treatment with an OP pesticide. Estimates of relative sensitivities (4th
column) were derived by either: (1) the ratio of the response (i.e.,
percent ChE inhibition) for adults: pups at the same dose of chemical,
or (2) the ratio of doses in adults: pups that induce a comparable
amount of ChE inhibition. The different approaches to estimating the
relative sensitivities to a ChE-inhibiting chemical were necessary
because of the differences in study designs and results among the
studies evaluated. For example, some studies used single doses such
as a proportion of an LD whereas other studies used multiple doses
that allowed calculations of an ED50. It should be noted that estimates
of relative sensitivities are a function of the doses or percentages of
ChE reported in studies and that, depending on dose-response
characteristics of ChE inhibition among different age groups, actual
sensitivities may be different at doses other than those used in Table
I.G-1.
[Since this section was written, preliminary BMD10's were derived for
the dose-response ChE data from repeated dosing studies on pups
and adults in RBC and brain for malathion (Fulcher 2001), methyl
parathion (Beyrouty 2002)and chlorpyrifos (Zheng et al., 2000). These
data were modeled using the EPA Benchmark Dose Software version
1.3.1 - Hill model (available at website:
http://cfpub.epa.aov/ncea/cfm/bmds.cfm?ActType=default). The
| modeling confirmed that there was less than two-fold difference in
| response between adults and pups following repeated dosing with
| chlorypyrifos. For malathion, a difference between adults and pups up
| . to approximately 3-fold was found for RBC ChE inhibition based on the
I BMDIOs. For brain ChE .there was 16% inhibition in pups at the
f highest dose tested (150 mg/kg/day) but no inhibition in adults. Thus,
| relative sensitivity could be determined because of the lack of
| comparable dose response data in pups and adults. Although Table 1
| reports a differential in brain ChE inhibition for methyl parathion up to
I 3.-fold (comparing percent inhibition), modeling the Beyrouty data
| showed differences up to approximately 4-fold based on BMDIOs. ]
I.G Page 15
-------
Table I.G-1. Summary of Sensitivity to ChE Inhibition in Neonatal or Juvenile Rats Treated with
Organophosphorus Pesticides
Pesticide &
Reference
Chlorpyrifos
(Pope, 2001 a)
Moser etal.,
(1998)
(Zheng ef a/.,
2000)
Diazinon
(Padilla etal.,
2002)
Dimethoate
(Myers, 2001.;
MRID
45529702,
unpublished)
TfeaimentOfoupis; Doses
(fpg/fcg/dayfcFteuteof
Administration
.
Neonates, juveniles,
and adults
7.5 neonates, 23.5 juveniles, 68
adults
single gavage dose
Adults and PND1 7 pups -
20 mg/kg
single gavage oral doses
Repeated gavage doses of
0.15,0.45,0.75, 1.50,4.50,
7.50, or 15.0
Adults and PND17 pups
single gavage dose of 75
Adults and PND1 1 pups
single gavage doses of 0.1, 0.5,
or 3.0
RSSUJfe
Neonates: 70% ChEl in cortex at 7.5 mg/kg
Juveniles: 65% ChEl in cortex at 23.5 mg/kg
Adults: 60% ChEl in cortex at 68 mg/kg
Pups Brain ChEl: 89% (
-------
Table I.G-1. Summary of Sensitivity to ChE Inhibition in Neonatal or Juvenile Rats Treated with
Organophosphorus Pesticides
Pesticide &
Reference
T*ea&Bettt<3f0Mp$; Doses
{fngfaQf&xy'f, Routed
Administration
Adults and PN01 1-21
repeated gavage doses of 0.1,
0.5, or 3.0
Results
Pups: 45% brain ChEl at 3.0 mg/kg/day;
13% ChEl at 0.5 mg/kg/day; 65% RBC ChEl at 3.0 mg/kg/day;
no RBC ChEl at 0.5 mg/kg/day
Adults: 60% brain ChEl at 3.0 mg/kg/day;
1 3% brain ChEl at 0.5 mg/kg/day; 63% RBC ChEl at 3.0 .
mg/kg/day; no RBC ChEl
at 0.5 mg/kg/day
Relative Sensitivity
^otd Difference)
REPEATED
Pups/Adults: At 3 mg/kg/day
45% ChEI/60% ChEI=0.8 (no difference)
At 0.5 mg/kg/day 13% ChEI/13% ChEI=1
(no difference)
At 3 mg/kg/day 65% ChEI/63% ChEI=1
(no difference)
Malathion
Fulcher, 2001,
MRID
45566201,
unpublished)
Adults and PND1 1 pups
single gavage doses of
5,50, 150, or 450
Adults and PND11-21 pups
repeated gavage doses of 5, 50.
or 150
Pups: 84% brain ChEl at 450 mg/kg; 48% brain ChEl at 150
mg/kg; 72% RBC ChEl at
450 mg/kg; 55% RBC ChEl at 150 mg/kg
Adults: No brain ChEl at 1 50 or 450 mg/kg;
25% RBC ChEl at 450 mg/kg; no RBC ChEl
at 1 50 mg/kg
Pups: 16% brain ChEl at 150 mg/kg/day; no brain ChEl at 50
mg/kg/day; 68% RBC ChEl at 150 mg/kg/day; 39% RBC at 50
mg/kg/day
17% RBC ChEl at 5 mg/kg/day
Adults: No brain ChEl and 51% RBC ChEl at 150 mg/kg/day;
20% RBC ChEl at 50 mg/kg/day
ACUTE
Pups/Adults: 84% brain ChEI/no brain ChEl at 450
mg/kg; fold difference uncertain
Pups/Adults: 72% RBC ChEI/25% ChEl at 450 mg/kg
= 2.9
REPEATED
Pups/Adults: 16% brain CHEI/no CHEI at 150
mg/kg/day and no brain at 50 mg/kg/day in pups or
adults; fold difference uncertain
Pups/Adults: 68% RBC ChEI/51% RBC ChEl at 150
mg/kg/day =1 .3; 39% RBC ChEI/20% RBC ChEl at 50 '
mg/kg/day =2.0
Methamidophos
(Moser, 1999)
Adults and PND17 pups
single gavage dose of
1,4, or 8
Pups: ED50 for brain ChEl 3.0 mg/kg;
EDM for blood ChEl 2.3 mg/kg
Adults: ED5Qfor brain ChEl 3.0 mg/kg;
EDM for blood ChEl 2.0 mg/kg
ACUTE
Pups/Adults: 3 mg/kg/3 mg/kg=1
(no difference); 2.3 mg/kg/2 mg/kg=1 .2
(no difference)
Methyl Parathion
I.G Page 17
-------
Table I.G-1. Summary of Sensitivity to ChE Inhibition in Neonatal or Juvenile Rats Treated with
Organophosphorus Pesticides
Pesticide &
Reference
; Route o?
Relatwe Sensitivity
^otd Difference)
(Pope, 2001a)
Neonates, juveniles, and adults
treated with a single gavage
dose
neonates 1.0, juveniles 2.05,
adults 7.3
Neonates: 60% ChEl in cortex at 1.0 mg/kg
Juveniles: 60% ChEl in cortex at 2.05 mg/kg
Adults: 70% ChEl in cortex at 7.3 mg/kg
ACUTE
Juvenile/neonate: 2.05 mg/kg/
1.0 mg/kg = 2.05
Adult/neonate: 7.3 mg/kg/1.0 mg/kg = 7.3
(Beyrouty.
2002a; MRID
45656501,
unpublished)
Adults and PND11 pups
single gavage doses of 0.03,
0.11, 0.3, or 1.0 pups; 0.03,
0.11, 0.3, or 0.6 adults
Pups: 18% brain ChEl at 0.3 mg/kg; 31% RBC ChEl at 0.3
mg/kg
Adults: No brain or RBC ChEl at 0.3 mg/kg
ACUTE
Pups/Adults: 18% brain ChEI/no brain ChEl at 0.3
mg/kg; fold difference uncertain
31% RBC ChEI/no RBC ChEl at 0.3 mg/kg; fold
difference uncertain
Adults and PND11-21 pups
repeated gavage doses of 0.03,
0.11,0.3, or 0.6
Pups: 62% brain ChEl at 0.6 mg/kg/day; 29% brain ChEl at
0.3 mg/kg/day; 86% RBC ChEl at 0.6 mg/kgVday; 65% RBC
ChEl at 0.3 mg/kg/day
Adults: 31 % brain ChEl at 0.6 mg/kg/day;
9% brain ChEl at 0.3 mg/kg/day; 58% RBC ChEl at 0.6
mg/kg/day; 35% RBC ChEl at
0.3 mg/kg/day
REPEATED
Pups/Adults: 62% brain ChEI/31% brain ChEl at 0.6 =
2.0
29% brain ChEI/9% brain ChEl at 0.3 =3.2; 86% RBC
ChEl /58% RBC ChEl at 0.6=1.5;
65% RBC ChEI/35% RBC ChEl at 0.3 =1.9
I.G Page 18
-------
' E
^0444* S
iii. Recovery from ChE Inhibition in Young Rats Treated with
Organophosphorus Pesticides
Studies that included analyses of recovery from ChE inhibition in
young rats have been performed on chlorpyrifos, methamidiphos, methyl
parathion, and parathion.
PND17 pups were reported to recover from ChE inhibition in one week
after cessation of dosing with a single maximum tolerated dose (MTD) of
chlorpyrifos of 15 mg/kg compared with a greater than two-week
recovery period in adults administered a MTD dose of 80 mg/kg/day
(Moser and Padilla, 1998). Pope et al. (1991) also reported a faster
recovery in ChE activity in neonates compared to adults when treated with
a MTD of chlorpyrifos. Adults treated with an acute, subcutaneous, dose
of 279 mg/kg chlorpyrifos showed about a 90% inhibition of brain ChE
activity seven days after dosing compared to approximately 40% inhibition
in the brains of seven-day-old neonates treated with 45 mg/kg
chlorpyrifos.
One day following oral treatment every other day with three doses of 3
mg/kg/day and eight doses of 6 mg/kg/day from PND1-21, brain ChE
activity was inhibited by 57% (Tang era/., 1999). Following a 19-day
recovery period, brain ChE activity (about 20% inhibition relative to
controls) was still depressed. Thus, although ChE levels in juvenile rats
return to control levels after an acute treatment with chlorpyrifos, repeated
treatments can lead to prolonged ChE inhibition.
PND17 and adult rats orally administered 8 mg/kg methamidiphos
each showed about 80-85% brain ChE inhibition 1.5 hours after dosing
(Moser, 1999). Twenty-four hours after dosing, more recovery was noted
in the pups than adults (30-35% brain ChEl in pups versus 55% in adults).
At 72 hours post dosing, ChE activity in pups had returned to normal but
there was still brain ChE inhibition in adults (10-15%)
Neonatal (seven-day-old) pups and adults were found to have similar
brain ChE activities (about 20% activity compared to controls) when
administered a MTD of methyl parathion (7.8 mg/kg: neonates; 18
mg/kg: adults) but a more rapid recovery was reported for the neonates
•0 | (Pope et al., 1991). By seven days after treatment, neonatal ChE activity
-------
I Repeated treatments with methyl parathion of PND7 or 14 neonates
I and adults showed more inhibition initially but a faster recovery in the
I young rats (Liu et a/., 1999). On day 8, one day after seven days of
i subcutaneous treatment of neonates and adults, inhibition in the cortex of
i neonates administered 1.5 mg/kg/day was 73%,(neonate) and 32%
C\! = (adults). At a dose of 3.0 mg/kg/day, more inhibition was found in striatal
O | than cortex tissues: Striatal inhibition at that dose was reported as 86%
(neonate) and 64% (adult) one day following seven days treatment; seven
days after cessation of dosing, brain ChE inhibition was about 45% in both
age groups, indicating that more recovery had occurred in neonates (41%)
than in adults (only 19%).
Liu era/. (1999) also investigated the effects on ChE inhibition in
§ neonatal rats and adults following administration of MTD doses of
i parathion. As with methyl parathion, maximum brain ChE inhibition was
I similar (>85% on the day of dosing in neonates and 90% in adults four
| days after dosing), but recovery in the neonates was more rapid. Seven
I days after cessation of dosing, brain ChE activity had essentially returned
I to normal in neonates but brain ChE was inhibited by 80% in the adults.
I c. Mechanisms Underlying the Differential Age-Related Sensitivity For
ChE Inhibition
:
Age-related differences in sensitivity to pesticides can occur for a number
of reasons (Pope, 2001b). Exposures to pesticides are age-related
(discussed in Section D) where children may be more exposed than adults'
based on their diet and behaviors. Toxicodynamic and toxicokinetic
differences may-also contribute to the young being at a different risk to
pesticide exposure. As summarized below, there are several reports in the
literature investigating the basis underlying the differential sensitivity found for
certain OP pesticides.
Toxicodynamic Considerations: There may be different mechanisms
underlying age-related sensitivity to ChE inhibition for different OP pesticides.
The exact mechanisms are not clearly understood in laboratory animals. For
obvious reasons, no data are available in humans. There are studies,
however, in laboratory animals that provide such information. Intrinsic
differences in neuronal AChE (i.e., differential inhibition of the target enzyme
itself) do not appear to account for the observed age-related sensitivity found
in young animals as suggested by in vitro studies (Benke and Murphy, 1975;
Chanda etal., 1995; Mortensen et a/., 1996; Atterberry efa/., 1997). Another
toxicodynamic factor, the ability to restore function following exposure, does
not appear to be the basis for age-related sensitivity to the OP pesticides
because more rapid recovery of AChE activity in younger animals is found
compared to adults (Chakraborti era/., 1993; Moser, 1999; Pope etal., 1991;
Pope and Liu, 1997).
I.G Page 20
-------
Other toxicodynamic differences that could affect age-related sensitivity to
AChE inhibition concern the regulation of acetylcholine receptor number as
well as acetylcholine release. Inhibition of ChE activity in the nervous system
results in the accumulation of acetylcholine in the synapses causing
hyperstimulation of the cholinergic receptors on postsynaptic cells. It is this
hyperstimulation that leads to cholinergic toxicity. This hyperactivity may also
lead to a decrease in the number of muscarinic receptors (i.e., down-
regulation). As a measure of toxicodynamic response to OP dosing, some
studies have compared the degree of muscarinic down-regulation in adult
| and young rats. In a study of repeated, subcutaneous dosing with methyl
1 parathion or chlorpyrifos, Liu et a/. (1999) found that muscarinic receptor
i | number was markedly reduced in pups compared to adult rats following
•*%* | repeated dosing with methyl parathion, suggesting age-dependent
£™ | differences in mucarinic receptor adaptation. Interestingly, the chlorpyrifos
9^ I exposure also produced more receptor down-regulation in the pup as
^ I compared to the adult, but the effect was not as pronounced as the methyl
(/j | parathion effects. Moreover, using a different route, the same group showed
Cp § that repeated oral dosing with chlorpyrifos caused equal down-regulation of
CD I muscarinic receptors in neonatal and adult brain (Zheng et a/., 2000). The
^ |. effect on receptor down-regulation appears to be compound-specific, and
possibly, route-specific. In the normal cholinergic synapse, it is known that
feedback inhibition of acetylcholine release occurs through activation of the
muscarinic acetylcholine receptors located on the presynaptic nerve terminals
(see Pope, 2001 b). Activation of the muscarinic acetylcholine receptors
would decrease further acetylcholine release, thereby reducing the excessive
stimulation of the postsynaptic receptors following AChE inhibition. A limited
ability or adaptability of this presynaptic regulatory process in the young could
lead to increased sensitivity to OP pesticides. There are only a few reports
exploring age-related differences in muscarinic presynaptic acetylcholine
receptor activity: evoked acetylcholine release was lower in brain tissues
from newborn animals and aged animals compared to rats aged one to six
months (Pedata et a/., 1983; Meyer and Crews, 1984). There is no
information on the receptor response (either total muscarinic receptor number
or feedback inhibition of acetylcholine release) in the developing human
brain.
Toxicokinetic Considerations: Toxicokinetic differences among age
groups can contribute to age-related differences in response, with the
interplay of metabolic activation and detoxification processes being an
important major factor, particularly in the first few months after birth (e.g., see
Ginsberg et a/., 2002). It appears from the literature that toxicokinetic
differences play an important role in the differential sensitivity of the young to
ChE inhibition following treatment with OP pesticides (e.g., Brodeur and
DuBois, 1963; Benke and Murphy, 1975; Scheldt et a/., 1987, reviewed in
Pope, 2001 b). In addition to inhibiting AChE, OP pesticides also interact with
other esterases, i.e., carboxylesterases and/or A-esterases, an by doing so
become inactivated or detoxified. A-esterases (e.g., chlorpyrifos oxonase,
I.G Page 21
-------
paraoxonase, or PON1) detoxify some OP pesticides by hydrolysis, whereas
some OPs bind to carboxylesterases, a reaction which effectively lessens the
amount of pesticide available for inhibiting AChE. Many investigators have
noted the decreased capability of the young animal to detoxify OP pesticides
by A-esterase or carboxylesterase esterases compared to adults (Mortensen
etal., 1996; Atterberry etal., 1997; Costa etai, 1990; Padilla etai, 2000;
Padilla et ai, 2002; Karanth and Pope, 2000).
Laboratory Animal Literature: The importance of A-esterase protection
against the toxic effects of the anticholinesterase activity of OP pesticides has
been demonstrated in several studies in which exogenous administration of
A-esterase can lessen OP toxicity in rodents (Costa et ai, 1990; Li et ai,
1993; Main, 1956). Studies with an A-esterase knockout mouse reinforced
the important role that A-esterases play in the detoxication of OP pesticides:
knockout mice were much more sensitive to chlorpyrifos oxon or diazoxon
(the active metabolites of chlorpyrifos or diazinon, respectively) than their
wildtype litter mates (reviewed in Furlong ef ai, 2000). In rats, A-esterase
activity is virtually nonexistent in the fetus (Lassiter et ai, 1998) and
increases from birth to reach adult levels around PND21 (Mortensen ef ai,
1996; Li etal., 1997). The animal data regarding the role of carboxylesterase
in mediating OP toxicity are also quite extensive (e.g., Clement, 1984;
Fonnum et ai, 1985; Maxwell, 1992 a,b), but there are sparse data on the
role of carboxylesterase activity mediating age-related toxicity to OP
pesticides. Fetal rats possess very little carboxylesterase activity (Lassiter ef
ai, 1998) with increasing activity as the postnatal rat matures, reaching adult
values after puberty (50 days-of-age) (Morgan ef ai, 1994; Moser ef ai,
1998; Karanth and Pope, 2000).
The temporal pattern of A-esterase activity (and carboxylesterases)
correlates reasonably well with studies on OP sensitivity (see summary in
Table 2). Several studies have shown an increased sensitivity of newborn
rats to OP compounds which are detoxified via the A-esterase and/or
carboxylesterase pathways (Gagne and Brodeur, 1972; Benke and Murphy,
1975; Pope ef ai, 1991; Chambers and Carr, 1993; Padilla ef ai, 2000; 2002;
Karanth and Pope, 2000). For example, Padilla ef ai (2002) and Karanth
and Pope (2000) have correlated age-related sensitivity with the maturational
profiles of these esterases. Using an in vitro assay, Padilla ef ai (2000)
showed that methamidophos, a pesticide which is not more toxic to the young
rat, is not detoxified by A-esterases or carboxylesterases. These
observations have been extended in a recent abstract to other OP pesticides
using this in vitro model which measures the detoxification potential via these
esterases in various tissues (e.g., liver, plasma) (Padilla ef ai, 2002). It was
reported that the differential sensitivities of paraoxon (the active metabolite of
parathion), malaoxon (the active metabolite of malathion), and diazoxon (the
active metabolite of diazinon) were also correlated with the less efficacious
detoxification by these esterases in young animals (Table 2). Karanth and
Pope (2000) noted that the lower levels of esterases in neonatal and juvenile
I.GPage22
-------
rats correlated with the increased in vivo sensitivity to ChE inhibition found for
chlorpyrifos and parathion.
Table t.G-2. Summary of General Results for Age-Related Detoxification and
Sensitivity in Rat Studies
Pesi)$id£
Chlorpyrifos
Diazinon
Dimethoate
Malathion
Methamidophos
Methyl Parathion
Parathion
(not included in
cumulative
assessment)
Hydralyzad by
A-Esterasfes?
Yes
Yes
Not tested
No
No
No
No
Bind to
Cafbdxy!-
esterases?
Yes
Not Much
Not tested
Hydrolyzed
No
Yes
(limited)
Yes
Age-Related
Detoxification
7
Yes
Yes
Not tested
Yes
Not tested
Yes
Yes
Wore Sensitive
1$ Young?
Yes (acute
dose of PND
No (repeated
dosing of)
Yes
No
Yes
No
Yes
Yes
Ref$fen<$$
Karanth and Pope,
2000;Padillaefa/.,2002
Padilla et at., 2002
Meyers, 2001
Fulcher, 2001; Padilla ef
a/., 2002
Moser, 1999;
Padilla era/., 2000
Pope, 2002a; Chambers
and Carr, 1 993
Karanth and Pope, 2000;
Padilla ef a/., 2002
Human Literature: There are only a few studies in the older literature that
have assessed A-esterase activity in children. Based on these studies, it
appears that serum A-esterase levels are very low in human infants
compared to adults (Augustinsson and Barr, 1963; Mueller et a/., 1983;
Ecobichon and Stephens, 1972). After birth, there is a steady increase of this
activity during the first six months to about one year (Augustinsson and Barr,
1963). In a related study of the age-dependence of total serum arylesterase
activity (of which a large component is A-esterase activity), adult levels were
achieved by two years-of-age (Burlina et a/., 1977). Although serum A-
esterases are reported to achieve adult levels around six months to one year-
of-age, there is uncertainty surrounding those values for the one-year-old due
to the variability in the rate of maturation expected as these enzyme systems
mature at different rates in a cross-section of one-year-old children.
Suggestive evidence of this is the large degree of variability seen in the six-
month and one-year age groups in the limited serum esterase data available
for children (Augustinsson and Barr, 1983). This source of variability
(maturational rate) is unique to children and is in addition to the host of
factors that contribute to interindividual variability in the rest of the population
and normally considered in noncancer risk assessments. Given the small
number of children studied for this parameter, population distributions that
reflect the central tendency and lower percentile value for A-esterase function
I.G Page 23
-------
in one-year-olds relative to adults cannot be discerned from the data (for
example see Ecobichon and Stephens, 1972; Figure 1). Moreover, these
studies have only examined a few children, and given the high interindividual
variability, it is very difficult to discern with confidence the maturation profile
for serum A-esterases in young children. In ongoing studies in C. Furlong's
laboratory, the same child is being evaluated for the appearance of serum A-
esterase overtime (i.e., so that the high natural variation does not obscure
developmental patterns) to better define the developmental profiles for serum
A-esterases (See Figure 1 below). Preliminary results indicate that children
reach adult levels of A-esterases around 12 to 15 months-of-age. Note that
this age of maturation corresponds reasonably well with the maturation of
human serum arylesterases mentioned above (Burlina ef a/., 1977). It should
also be pointed out that there is no information on the maturational profile of
A-esterase in the human liver (an organ very important for detoxification), and
there appears to be no information about the maturational profile of
carboxylesterases in humans.
Figure I.G-1. Maturation Profile of Serum A-Esterase (Paraoxonase)
Appearance in Infants and Children (Costa ef a/., 2002).
% 250 r
S ^ '
Baby 2
•*:
;,*-*:.- '
0. & 10 15
Month:?, of Age
3 1?
li
800 ;
400
:C
:**&
* «
'••- ,
) 10:: 20 30. .40
Months of Age
Any anticholinesterase pesticide that is a substrate for A-esterase, the
lower A-esterase levels in the blood of very young would result in more
inhibitor available to reach target neuronal tissues. It should be noted that in
addition to age-dependent differences in A-esterase activity, a human and
animal genetic polymorphism has been well established (e.g., Mackness et
a/., 1998). Differences in observed rates of hydrolysis of paraoxon between
individuals can vary by at least 20-fold (Furlong ef a/., 2000). This large
difference in A-esterase activity does not necessarily translate into
equivalently large differential sensitivities to OP pesticides. There is also
some recent evidence in the literature that low A-esterase activity may
I.G Page 24
-------
-
predispose adult humans to a greater toxic response (Haley era/., 1999;
Cherry et al., 2002) to nerve agents and/or pesticides.
Not only is limited detoxification a factor in increased sensitivity of the
young, but another potential factor is the age-dependent ability to activate OP
pesticides via oxidation by cytochrome P450s to their oxon form (i.e., the
active anticholinesterase metabolite). For example, oxidation by CYP3A4
plays a key role in the oxidation of OP pesticides in humans (Butler and
Murray, 1997). Ginsberg et al. (2002) using the children's pharmacokinetic
data from the therapeutic drug literature showed that compared to adults,
oxidation by CYP3A4 tends to be more active in children beginning as early
as two to six months-of-age with this difference lasting until at least two
years-of-age. While this may increase concern for greater oxidative
bioactivation in the young, the CYP-mediated oxidative dearylation pathway,
which may also be more active at these ages, is involved in the detoxification
of these pesticides. Therefore, it is important to compare the maturation
profiles for these two GYP pathways. Based on data from rat liver
microsomes (Ma and Chambers, 1994) and as modeled by Timchalk et al.
(2002) for humans and rats, the activation step is 2.5-fold faster (based upon
Vmax/Km ratios) and importantly, the activation step has a 8.4-fold lower Km
than the dearylation step. The significance of this is that at relatively low,
environmental exposures, OP molecules reaching the liver may be much
more likely to be oxidized by the activation pathway than detoxified by the
dearylation pathway. This evidence supports the potential concern that
greater oxidative capacity in one- to two-year-olds may lead to more OP
activation than seen in adults. The enhanced ability of the young to
bioactivate OP pesticides to their oxon form, however, has not been
correlated with an increased sensitivity to ChE inhibition. Nonetheless, when
coupled with the potential limited ability of young children to detoxify these
pesticides via the A-esterase and carboxylesterase pathways, this produces a
source of uncertainty in the pesticide risk assessment for children.
b;: | d. Hazard Characterization Summary
/""H ^
*^ | There have been reports of signs and symptoms associated with
cholinergic toxicity following high exposures to OP pesticides of adults and of
young children. Common signs and symptoms of cholinergic toxicity in
humans range from changes in heart rate and blood pressure, miosis,
diarrhea, headaches, nausea, muscle weakness to unconsciousness,
convulsions, and death. Not only can cholinergic toxicity occur in children
following exposure to OP pesticides, but emerging investigations have raised
concern about the effects of antiChE activity on neurodevelopment which
may be a sensitive process susceptible to adverse perturbations.
As discussed in Section A, there is evidence that ChE and acetylcholine
act as important neuromodulators in the developing brain. Because
neurogenesis is not limited to the intrauterine period and may continue
I.G Page 25
-------
throughout childhood, all stages of brain development are considered to be
potentially susceptible to disruption by ChE inhibition. During the first few
years of life, brain development is a tightly orchestrated process of migration
and "pruning," which is under the influence of neuromodulators (ChE,
acetylcholine, and other neurotransmitters), genetic controls, and the
experiences of the child. Although OP pesticides may influence the migration
of cells and the connectivity of the central nervous system (CMS) and result in
consequences that could last into adulthood, it is not known how much of a
perturbation (i.e., degree of ChE inhibition) is needed, or how long this
perturbation must be sustained, to disrupt normal development. The majority
of OP pesticides included in the cumulative risk assessment have not been
evaluated for neurodevelopmental effects (e.g., functional, behavioral, or
neuropathogical effects) or for ChE activity in immature animals.
In light of this uncertainty, it should be assumed that small perturbations
resulting from either a single exposure or repeated exposure could potentially
disrupt neurodevelopment. Therefore, it is important to insure that the adult
brain ChE endpoints used in the cumulative risk assessment for OP
pesticides are adequately protective of the young. Thus, a key issue in this
assessment is whether ChE inhibition in the young will be caused at lower
doses of these pesticides compared to adults or whether the young will show
a higher level of ChE inhibition at comparable doses. It is the integration of
the chemical-specific information along with the basic biological
understanding of sensitivity and susceptibility that informs the need for the
application of additional safety or uncertainty factors in the cumulative risk
assessment to protect fetuses, infants, and children.
Because in humans, the process of brain development begins during
gestation and continues postnatally through adolescence, it is important to
identify the developmental windows of age-dependent sensitivity to ChE
inhibition. In laboratory animals, ChE inhibition can be found to occur in all
developmental stages of the young (i.e., in fetal, neonatal, juvenile, and
young adult rat tissues). In general, oral dosing of pregnant rats with OPs
causes ChE inhibition in the fetus and/or neonate, but fetuses/neonates that
are exposed in utero (and via early lactation) generally do not exhibit more
ChE inhibition than is found in maternal tissues. These studies need to be
interpreted with caution with respect to comparative sensitivity because the
absorbed dose to the dam and fetus is typically not known. Also, the fetal rat
appears to be less affected from repeated exposures to OP pesticides
presumably because of the rapid recovery and resynthesis of the AChE in
fetal tissue compared to the dam, making it difficult to compare relative
responses in the fetus versus dam. It should be noted that rat fetal tissues
and the placenta are deficient in key detoxification systems, including A-
esterases and carboxylesterases. Overall, there is limited pharmacokinetic
information available in fetal versus maternal tissues for OP pesticides.
|.G Page 26
-------
Continued treatment following birth is important to ensure that critical
periods of sensitivity are evaluated. Direct dosing of the postnatal rat may be
necessary, however, because of the possibility of limited exposure through
the milk via lactational transfer of OP pesticides. Although direct dosing of
the pups (typically via oral gavage) maximizes and allows for quantification of
exposure to the pups, it does not necessarily mimic the dietary intake
exposure patterns in children. Furthermore, certain stages of brain
development in the early postnatal rat are equivalent to the third trimester
human fetus, and thus direct dosing of very young postnatal rats would not
represent the pharmacokinetics of the chemical in the mother. Nonetheless,
direct dosing experiments do provide a better basis to determine the
comparative sensitivity of the pups and adult animals. Some direct dosing
studies of postnatal rats are available on OP pesticides; however, these few
studies have shown that acute postnatal exposures via direct dosing to young
rats results in an increased sensitivity to ChE inhibition for certain OP
pesticides (e.g., malathion, methyl parathion, chlorpyrifos, diazinon), but not
all (e.g., methamidophos, dimethoate and by extension, its metabolite
omethoate).
Age-dependent sensitivity to ChE inhibition by OP pesticides can
sometimes also be found following repeated dosing studies in laboratory
animals. An important issue with repeated dosing is the more rapid recovery
(synthesis of new ChE enzyme) in postnatal (and fetal) rat tissues. In most
repeated dosing studies comparing the responses of adults to postnatal
animals dosed at the same frequency, this faster recovery in the young
animals may result in less inhibition as compared to the adults, which is
interpreted by some as lower sensitivity of the young. The results of such
studies are critically dependent on the time interval between the doses and
also the time (in relation to the last dose) at which the ChE inhibition is
sampled in both age groups. As acute studies have shown, age-related
sensitivity differences in rodents depend on the age at dosing, since the
detoxification pathways are rapidly maturing. Therefore, in repeated dose
studies, .the fact that the animals are probably becoming less sensitive over
time by virtue of this changing toxicokinetic pattern is an additional
confounding factor. For all these reasons, a smaller differential for ChE
inhibition has often been found between the pups and adults following
repeated dosing when compared to acute exposure.
Although age-dependent sensitivity is found in some animal experiments,
a key question is whether this sensitivity will occur in children. Children may
respond to toxicity at lower doses than adults because infants and very young
children may be less able than adults to metabolize and excrete toxic
substances (Ginsberg et a/., 2002). Animal studies have shown a correlation
of age-dependent sensitivity to certain OP pesticides with the developmental
profiles of the A-esterases and/or carboxylesterases (enzymes that detoxify
OP pesticides). As described in Section C, based on limited data, young
children may have lower levels of these detoxification pathways. The most
I.G Page 27
-------
O
V-
highly exposed age group in the OP cumulative risk assessment was
identified as the one- to two-year-olds. Although after birth there is a steady
increase of A-esterase activity during the first six months to one year, these
maturation profiles may vary among children (due to interindividual variability)
and may vary among different tissues. Maturation profiles are not available
for the carboxyesterases, and the developmental profile for either A-
esterases or carboxylesterase has not been delineated in liver (a major
detoxication organ). Furthermore, young children may also have an
increased ability to activate OP pesticides to the oxon form as compared to
the adult. Therefore, given the uncertainty surrounding the maturation
profiles of young children for A-esterases and carboxylesterase, their
potential to be more active than adults at bioactivating OP pesticides to their
oxon form, as well as their rapidly developing nervous system, infants and
very young children (including children in the one- to two-year age group)
would potentially be vulnerable to chemical interference due to OP pesticide
exposure.
Because some OP pesticides do show age-dependent sensitivity, and
there are missing ChE data in young animals for many of the OP pesticides in
this cumulative risk assessment, there is a degree of uncertainty regarding
the estimation of risk. Under the children's safety factor provision a default
safety factor of 10X is required to address this database deficiency unless
there are reliable data to support a conclusion that a different safety factor
would be safe for infants and children. As the following discussion indicates,
OPP has concluded that reliable data do exist to support use of a database
uncertainty factor to address this data deficiency. To determine whether a
database uncertainty factor could protect infants and children, the degree of
difference between the doses needed to cause a certain level of ChE
inhibition between the young and adult was evaluated. As shown in Table 1 ,
the differential between adults and immature animals following repeated
dosing (typically 11 consecutive days) is at most approximately threefold. A
single acute dose is found to cause differences ranging from about twofold up
to approximately ninefold.
The relative sensitivities of immature animals found in repeated dosing
studies are considered more appropriate than the results of the acute dosing
studies for the cumulative risk assessment of OP pesticides for several
reasons. Acute dosing studies were done with PND11 pups, which are more
like the human newborn with limited detoxification ability. Repeated dosing
studies of OP pesticides usually started treatment at PND1 1 and continued to
PND21. As the immature animal ages, it rapidly reaches adult levels of A-
esterases around PND21 . Thus, evaluation of ChE activity in repeated
dosing studies more closely mimics the maturation of A-esterase activity in
children around one year-of-age when children are reaching adult levels of A-
esterases. Thus, the use of repeated dosing studies better approximates the
maturation profile of the age group that is significantly exposed to OP
pesticides in the cumulative risk assessment. Children generally do not begin
I.G Page 28
-------
•SXX-W
to consume fresh (uncooked) fruits and vegetables until after six months-of-
age or more. The highly exposed group in the cumulative risk assessment is
the one- and two-year-olds, not the infants. Repeated dosing studies were
also used to determine relative sensitivity because people are exposed every
day to an OP pesticide through food, and thus an animal study using repeat
exposures is considered appropriate. Also, following exposure to an OP,
regeneration of ChEs to preexposure levels does not occur for days or
weeks, making the exposed individual potentially more vulnerable to
subsequent exposures during that period.
Repeated dosing studies are now available on six of the 22 OPs in the
cumulative risk assessment. For three of these OP pesticides, the repeated
dosing studies showed no increased sensitivity in the y9ung, whereas
increased sensitivity was seen in the other three. The differential sensitivity
between adult and immature animals ranged from 1X-(i.e., no differential) up
to a 3X difference. These studies are considered to provide a reasonable
basis on which to establish the size of a database uncertainty factor for the
following reasons. Although these six OP pesticides do not represent every
structural and pharmacokinetic characteristic of the large class of OP
pesticides included in the cumulative risk assessment, they are nonetheless a
reasonable subset of different structural and pharmacokinetic characteristics.
For example, methamidophos is a phosphoramidothioate of small molecular
weight with no ring structure, does not require metabolic activation to
generate an oxon form, and is not detoxified by esterases. On the other
hand, methyl parathion is a phosphorothioate of larger molecular weight with
a ring structure, hepatically bioactivated to its oxon form, and detoxified by
esterases. In addition to the observed differential between adult and young
animals following repeated dosing, it must also be kept in mind that the
differential between the adult and young animal decreases as the animal
ages and reaches adult levels of the detoxification enzymes. For these
reasons, there are sufficient data to conclude that a 3X database uncertainty
factor should be applied, and that the 3X UFDb should be sufficient to account
for potential age-dependent sensitivity to ChE inhibition. It should be noted
that the application of a 3X UFDb is in addition to the application of the
customary intra- and interspecies uncertainty factors, which takes into
account variability among the human population.
The question remains as to how such a database uncertainty factor
should be incorporated into the cumulative risk assessment. In the
cumulative risk assessment process, uncertainty or safety factors can be
either applied to estimates of individual members toxic potencies (i.e.,
relative potency factors or RPFs) or applied as a group factor on the index
chemical's point of departure.3 Because age-dependent sensitivity to ChE
i 3ln the cumulative risk assessment, the RPF approach is used to determine the joint risk of the OP
1 pesticides, which applies dose addition. The RPF approach uses an index chemical as the point of
| reference for standardizing the common toxicity of the chemical members of the (CAG). Relative
5
S
| I.G Page 29
-------
inhibition is not common to all OP pesticides, application of a database
uncertainty factor would be more appropriately applied as chemical-specific
adjustments to the RPFs to account for ChE inhibition potentially occurring at
lower doses in the young than in the adult or resulting in a more potent
response at the same dose compared to the adult. These chemical-specific
adjustments should be made on the RPFs for those OP pesticides that lack
ChE data in the young. There are ChE data for a few OP pesticides that
show age-dependent sensitivity. However, RPFs are based on a uniform
measure of toxic potency using the same species, sex, endpoint, and age
T^v, group from studies of comparable methodology. Given that there are too few
data in young animals to determine RPFs for the OP Cumulative Assessment
Group (CAG), the RPFs for those chemicals showing age-dependent
sensitivity should also be adjusted to account for sensitive effects in the
young. The RPFs of those OP pesticides that do not cause age-dependent
sensitivity after brief periods of repeated exposure (dimethoate and by
extension omethoate, methamidophos, chlorpyrifos) should not be adjusted.
In conclusion, the limited animal data on the relative sensitivity of young
animals to cholinesterase inhibition (ChEl) caused by OP pesticides has
raised uncertainty about the adequacy of the adult RPFs to be protective of
the young and should be addressed by application of the traditional database
uncertainty factor (UFDb). Application of this UFDb should be protective of
potential age-dependent sensitivity to ChE inhibition and of potential adverse
neurodevelopmental outcomes that are a result of the inhibition of ChE.
Thus, there are reliable data to assign an additional factor (a database
uncertainty factor of 3X) other than the default 10X additional safety factory.
Further, because the database uncertainty factor addresses potential age-
dependent sensitivity there is no a need to retain an additional special FQPA
safety factor for potential pre- or postnatal toxicities associated with inhibition
of ChE.
3. Cumulative Exposure Assessment
Another important consideration for the FQPA safety factor is the
completeness of the exposure database. Whenever appropriate data are
available, OPP estimates exposure using reliable empirical data on specific
pesticides. In other cases, exposure estimates may be based on models and
assumptions (which in themselves are based on other reliable empirical data).
This section explains how the safety of the exposure estimates to infants and
children were estimated.
potency factors (i.e., the ratio of the toxic potency of a given chemical to that of the index chemical) are
then used to convert exposures of all chemicals in the CAG into exposure equivalents of the index
chemical.
I.G Page 30
-------
EPA identified and included three exposure pathways for the OP Pesticide
cumulative risk assessment: food, drinking water, and
residential/nonoccupational. Each of these pathways was evaluated separately,
and, in doing this step of the analysis, EPA determined which of the OP
pesticides were appropriately included for a particular pathway. The cumulative
assessment of potential exposure to OP pesticides in food includes: 22 OP
pesticides that are currently registered in the U.S. or have import tolerances;
residential or nonoccupational pesticide uses included 11 OP pesticides (Note:
many residential uses have been canceled as a result of risk mitigation efforts or
are not expected to result in any significant exposure); and 24 OP pesticides (as
well as several toxic transformation products) were considered in the cumulative
I water exposure assessment. Calendex™ software was used to determine the
I distribution of exposures and resulting Margins-of-Exposure (MOEs) from OPs in
| foods, drinking water and from residential uses.
Up until this time, OPP has performed its risk assessments using several
different age groups for children including nursing infants less than one year,
non-nursing infants less than one year, children one to six years-old, children
seven to 12 years-old, and children 13 to 19 years-old. Because of the
availability of more extensive data on children's food consumption, EPA was able
to subdivide the children's age group one through six years-of-age into two
different age groups: children one through two years-old and children aged three
through five years-old. EPA also made some other slight adjustments to the age
breaks defining groups for older children. As explained below, EPA analyzed all
of the different children's age groups, but did not analyze every age group for
every scenario. The children's age groups that were analyzed for all of the
exposure scenarios in the revised OP cumulative risk assessment were one
through two years-of-age and three through five years-of-age. EPA selected
these two age groups because in single chemical risk assessments (including for
the individual OPs) they most frequently reflect the highest levels of exposure.
Thus, a narrow range of ages were used to capture the finer details associated
with major contributors to risk under the premise that they reflect the exposure
scenarios most likely to be emphasized in risk management activities.
In addition, EPA produced exposure estimates for all of the other children's
age groups (children less than one year, children six through 10 years and
children 11 through 19 years) for the Florida region. Florida was selected
because it appears to have the highest level of exposure from all sources of
pesticides combined. As expected, the exposures estimated for children less
than one year-old or six and older were consistently lower than the exposures
estimated for one- to two- and three- to five-year-old children. Based on this
analysis, EPA concludes that, by focusing on two age groups of children (one- to
two-year-olds and three- to five-year-olds), its risk assessment does not
underestimate potential exposure to any age group of children.
I.G Page 31
-------
a. Food Pathway
Exposure to foodborne pesticides is an important factor in evaluating the
susceptibility of the young. Young children tend to eat more food in
proportion to their body size and they tend to eat more frequently than adults.
As discussed below, these characteristics are incorporated in the assessment
of exposure to OP pesticides via food.
The food component of the cumulative risk assessment has been highly
refined to reduce OPP's Tier 1 default assumptions (all foods contain
residues at the maximum amount allowed under tolerance) to more realistic
estimates of actual human exposure. It is based on residue monitoring data
-•^f from U.S. Department of Agriculture (USDA) Pesticide Data Program (PDP),
supplemented with information from the U.S. Food and Drug Administration
(FDA) Surveillance Monitoring Programs and Total Diet Study. The PDP data
provide a very reliable estimate of pesticide residues in the major children's
foods. They also provide direct measures of co-occurrence of OP pesticide
residues in the same sample, alleviating much of the uncertainty about co-
occurrence in foods that are monitored in the program.
Another important aspect of the food exposure assessment is that it is
based on actual consumption data from the USDA's Continuing Survey of
Food Intakes by Individuals (CSFII). The CSFII provides a detailed
representation of the food consumption patterns of the U.S. public across all
age groups, during all times of the year, and across the 48 contiguous states.
Additionally, OPP used a more recent CSFII in the OP cumulative
assessment (the 1994-96 CSFII) that was supplemented in 1998 by the
Supplemental Children's Survey. This 1998 survey focused on children from
birth to nine years-old and greatly expanded (by several fold) the number of
birth to four-year-old children in the survey database. OPP believes that the
food consumption information used provides a very realistic estimate of
potential risk concerns because it reflects the current eating habits of the U.S.
public, including those of children. The use of the newer CSFII and the finer
age breakouts should increase the accuracy and utility of the risk assessment
overall by making it more descriptive of the anticipated exposures and risks
for children.
A large percentage of the foods consumed in children's diet is addressed
in this assessment. Only about 3% of the foods consumed by children still
sremained unaccounted for after using PDP and the FDA Total Diet Study and
FDA monitoring data.
OPP is aware that some or all baby food manufacturers have adopted
policies that restrict the use of OPs on fruits and vegetables that will be used
Ui™ in their products. As a result, children consuming commercially prepared
baby food may not be exposed to OPs in their diet. OPP has investigated the
impact of this assumption for children one through two years-of-age, and for
I.G Page 32
-------
children less than one year-old. The residues in commercially prepared baby
foods were assumed in the first case to be equivalent to those found in an
adult diet. They were also set to zero to bound the lower limit and determine
the extent of the impact on any risk assessment. Setting all residues in
commercial baby food to zero had little impact on the magnitude of risk
estimated for children one through two years-of-age. This observation is
consistent with the very small amounts of baby food consumed by this age
group. However, a substantial impact was observed for the age group of
children less than one year-of-age because of the relatively large proportion
of baby food in their diets. OPP believes that estimating exposure to
pesticides from baby food as containing residues comparable to those in
adult diets will not impact regulatory decision-making because the overall
exposure to children less than one year-of-age is less than exposure to
children one through two years-of-age.
Two exposure issues unique to children are not directly addressed in the
current assessment. OP exposure from breast milk is not incorporated
quantitatively. A review of the literature to identify any potential pesticide
transfer from breast milk to children indicated no evidence that this pathway
would represent a significant source of exposure (ILSI, 1998). However,
further analysis has identified a study that demonstrated the presence of
chlorpyrifos and chlorpyrifos-oxon in the milk of rats (Mattsson et a/., 1998,
2000). The results of studies generated by the regulated community in
support of pesticide registration indicate no significant transfer of OPs to milk.
OP pesticides are not found to transfer into cow's milk when cattle are fed a
diet containing OPs. This finding is uniform across the entire class of OP
pesticides. As a result, OPP believes that breast milk is not likely to be an
important contributor of OPs to the diets of infants and children, especially at
environmentally relevant levels of exposure. Baby formula is included in the
current assessment with its consumption reflected in the FCID (Food
Commodity Intake Database) translation of CSFII food consumption survey,
and residue data available for all of its components.
OPP's dietary assessment also captures the metabolites of OPs that are
known to occur at significant levels in food commodities (e.g.,
omethoate-metabolite of dimethoate; methamidophos-metabolite of
acephate; and, dichlorvos-metabolite of naled and trichlorfon). Although
there is not extensive analytical data on other OP metabolites, there is
adequate data (e.g., from metabolism studies, FDA monitoring data) to
indicate that the food assessment is not missing significant residues in food
(such as for malaoxon- metabolite of malathion).
In summary, given the comprehensive data on potential exposure to OP
pesticide residues through the food, OPP is confident that exposure to all age
groups, including children via the food dietary pathway has been well
characterized.
I.G Page 33
-------
b. Drinking Water Pathway
Daily drinking water exposure estimates were generated using the
simulation models PRZM and EXAMS (a description of the use of these
models can be found http://www.epa.qov/oppefed1/modeis/water/index.htm).
The use of these models allows estimation of concentrations of OP
O pesticides. OPP used these models to provide daily distributions of OP
pesticide levels in water for incorporation into the probabilistic cumulative
exposure assessment. Twelve regional water exposure assessments were
conducted that were designed to represent exposures from typical OP usage
conditions at one of the more vulnerable surface watersheds in the region.
Each regional assessment focused on areas where combined OP pesticide
exposure is likely to be among the highest within the region as a result of total
OP usage and vulnerability of the drinking water sources. These methods
have provided OP pesticide distributions that are, in many cases, reasonably
comparable with available monitoring data in the same or nearby locations.
There are too little data to quantify OP degradates that may result in drinking
water. These metabolites, however, have been qualitatively assessed in the
revised cumulative risk assessment by assuming complete oxon conversion
with a 10-fold increase in toxicity compared to the parent compound: it was
found that this assumption did not have an impact on the upper percentile
distributions. In summary, OPP believes that the current cumulative
assessment represents a reasonable, health protective estimate of likely
exposure to OP pesticides from water to all age groups, including children.
c. Residential or Nonoccupational Exposure Pathway
The residential/nonoccupational exposure analysis includes the exposure
from home lawn and garden treatments, pesticides used on golf courses, and
applications made by governmental entities for the control of public health
pests such as wide area mosquito sprays. The oral, dermal, and inhalation
routes are considered. This analysis has incorporated activity patterns of
children and the major sources of exposure to young children (e.g.,
nondietary ingestion and hand-to-mouth behavior as established by video
tapes of children). Furthermore, pet uses have been incorporated in the
revised assessment. For the first time, the residential analysis used
distributions of data and exposure elements instead of point values. In most
cases, these data and exposure elements were chemical-specific. The
analysis reflects all remaining residential uses of OP pesticides, consideration
of both homeowner and professional applications, and postapplication
exposures resulting from these applications. The analysis also employed the
most recent survey data of residential uses and use information. Exposure
due to activity in and around schools and parks is not addressed directly,
because there does not appear to be any remaining OP pesticide uses for
school structures and grounds. Nonetheless, the possibility of exposures
encountered away from the home has been indirectly built into the
I.G Page 34
-------
assessment by conservatively extending the duration of residential exposure
beyond the two hours spent on grass to 3.5 hours spent outdoors.
The calendar-based model (Calendex™) that was used in the preliminary
OP cumulative risk assessment allowed for the temporal aspects of the
residential use of pesticides in twelve distinct geographic regions to be
accounted for; these regions not only represent major crop growing areas
and their influence on residues of OP pesticides in surface and ground water,
but also present an opportunity to consider the unique climate patterns, pest
patterns and potential socioeconomic patterns that influence residential
pesticide use and expected exposure to OP insecticides. Furthermore,
Calendex™ allows one to delineate the critical timing aspects of seasonal
uses of OP insecticides that result in exposure, as well as to identify potential
co-occurrences from multiple sources. Again, it cannot be emphasized too
strongly that the exposure, monitoring, and residue studies that were used as
input parameters in the modeling of residential/nonoccupational exposures
represent the best available data on these pesticides (i.e., the input
parameters for residue levels and dissipation rates based on actual
measurements).
d. Biological Monitoring Studies of Children
Biomarkers can serve as a useful measure of direct exposure aggregated
over all sources and pathways by measuring integrated exposure from all
routes. Biomarkers can be used to characterize the relative magnitude of
exposure within population subgroups. In addition, biomonitoring can be
used to verify predictions of exposure models.
Urinary biomarkers of OP pesticides and their metabolites have been
used to characterize reference body burden levels for adult and children
^
^ I populations in the U.S. and Europe (Murphy et a/., 1983; Kutz et at., 1992;
£ f Hill et a/., 1995; Aprea et a/., 1996, 1999, 2000; Macintosh et a/., 1999;
ir | Fenske et at., 2000; Quackenboss et a/., 2000; Adgate et a/., 2001; Heudorf
""* and Angerer2001; Kriegeref a/., 2001). Most of this research has been
designed to determine if children have higher exposures to OP pesticides
than adults, and, if so, what are the differences in these exposures and what
are the factors that influence these higher exposures.
Several researchers have conducted monitoring studies that have
collected environmental and/or biological samples to assess the potential
aggregate (inhalation, dermal, ingestion (dietary and indirect)) exposure to
OPs by adults in their everyday environments. Hill et al. (1995) analyzed
single spot urine samples from approximately 1000 adults (20-59 years-of-
| age) living in the U.S. to establish reference range concentrations for OP
| pesticide residues. Chlorpyrifos exposure was indicated by TCPY
I concentrations of 13 ug/L (95th percentile value) and 77 ug/L (maximum value
I observed). Macintosh etal. (1999) reported on the relationship between
I.G Page 35
-------
short-term and long-term average levels of OP biomarkers for 80 adults living
in Maryland. First-morning void urine samples were collected at up to six
different time periods equally spaced over a one-year period, with the range
of urinary OP metabolite values being similar to the levels reported by Hill et
al. (1995).
Only a handful of studies have been published in the literature that were
specifically focused on biomonitoring of children for OP pesticides and their
metabolites. These researchers noted that young children may be more
highly exposed and are more susceptible to health risks from exposures to
OP pesticides than adults because they are undergoing rapid physiological
and behavioral development. Furthermore, in comparison to adults, young
children: have a larger surface area to volume ratio; have a relatively large
brain size as compared to total body mass; take in more air, food, and water
on a per unit body weight; and, absorb, distribute, metabolize, and eliminate
pesticides differently than adults (Guzelian et al,, 1992). Children also
engage in specific activities in which they may more likely come into direct
contact with contaminated surfaces and objects (Hubal et al., 2000). These
child-type activities include: sitting, playing on the floor; eating while roaming
around the house; putting hands, objects, toys into the mouth; licking the
furniture, pet, floor; etc.
The Minnesota Children's Pesticide Exposure Study (MNCPES) was the
first published study to report multipathway pesticide exposures in a
population-based sample children (Quackenboss et al., 2000; Lioy et al.,
2000; Adgate et al., 2001). Personal (hand rinse, duplicate diet, time activity
diaries and questionnaires, videotape segment), biological (urine), and
environmental (indoor/outdoor air, residential surfaces, soil, drinking water)
samples were collected to assess children's aggregate pesticide exposure
and attempt to identify the critical pathways of exposure. Three first-morning
void urine samples were collected on three separate days during the study.
The urine samples were then analyzed for metabolites of commonly used OP
pesticides (Adgate et al., 2001). Analysis of these urine samples for OP
pesticides and their metabolites have shown that children do have a body
burden level of OP metabolites (Quackenboss et al., 2000; Lioy et al., 2000;
Adgate et al., 2001). While the MNCPES study didn't assess adult
exposures, Adgate et al. (2001) compared these urinary biomarker levels of
OPs for children with the reference levels reported by Hill era/. (1995) and
found similar ranges for both children and adults.
Fenske et al. (2000) collected and analyzed single void urine samples for
OP metabolites from 109 children (up to six years-of-age) who lived in an
agricultural community in the State of Washington. The children's urine
samples were collected at the convenience of the child and parent. From the
children's OP pesticide doses derived from this biologic monitoring study, the
authors suggested that residents of agricultural communities may be more
exposed to pesticides than the general population.
I.G Page 36
-------
Two studies have compared urinary metabolite levels for all members of a
household (Heudorf and Angerer 2001; Krieger et al., 2001). Heudorf and
Angerer (2001) examined urinary metabolite concentrations for children and
adults living in dwellings that had not been recently treated with OPs (most
recent treatment was four years prior). These investigators suggested that
urinary OP metabolite concentrations in children and adults were not
different. Krieger et al. (2001) assessed the extent of exposures for family
members (adults and children) residing in homes where pesticides have been
used. Chlorpyrifos was applied in this study by three different methods:
fogger, broadcast, and crack and crevice. Analysis of the family urine
samples for OP metabolites showed no significant difference between
children and adult exposures for those family members living in the same
households. However, both studies only examined the sample results
without considering the factors associated with the physiological and
behavioral differences between adults and children, a step needed to better
describe and understand the real potential for exposure.
Interpreting the results of these published studies presents several
challenges. First, only a few studies have been conducted and the results
published in the literature. Secondly, the methodologies employed in each
study have varied. Only spot urine samples have been collected, and, more
importantly, the sample collection times for these spot urine samples have
differed for many of the studies, ranging from first-morning voids to
convenience samples collected during the morning hours. In the few studies
that have collected both the environmental and biological samples, the levels
of the OP pesticides and urinary metabolites have not correlated with any of
the OP concentrations in the other environmental samples analyzed (Lioy et
al., 2000). Some investigators have tried to compare the children's and
adult's OP pesticide and metabolite levels without correcting these data for
the differences found in children associated with their differences in metabolic
rates, muscle mass, creatinine production, and urinary output.
Although the available biological monitoring studies generally indicate
children do have a body burden level of OP pesticides, based on the limited
number of published studies and the inconsistencies noted above, it is
difficult to make any general statements concerning the study population,
much less the general population. Equally important, from these limited sets
of published data, it is difficult to assess whether children's exposures to OP
pesticides are the same, higher, or lower than corresponding adult
exposures.
Several relatively large-scale children's aggregate pesticide exposure
studies which include OP pesticides are ongoing or near completion by the
U.S. EPA National Exposure Research Laboratory (NERL) scientists and
academicians. However, the analyzed and published results of this research
will not be available for several years. Without these additional data,
questions regarding whether children's exposures to pesticides are higher
I.G Page 37
-------
than adults or the validity of the cumulative exposure estimates can not be
readily answered.
e. Exposure Characterization Summary
The cumulative exposure assessment of OP pesticides represents the
first probabilistic assessment of multichemical and multipathway exposures to
pesticides. Estimates of residues in food are based on actual monitoring and
consumption data that capture the major food groups consumed by children.
Several age groups are defined such that they reflect an adequate number of
individuals in each age group and are based on real differences in age-
related eating patterns. Estimates from food dietary intake are considered to
confidently approximate dietary food exposure of children to OP pesticides.
There is also confidence that the dietary drinking water exposure assessment
for OP pesticides does not understate potential exposure to children (or any
age group) given that regional water exposure assessments were conducted
that were designed to focus on areas where combined OP exposure is likely
to be among the highest within the, region as a result of total OP usage and
vulnerability of the drinking water sources. Furthermore, the cross check of
PRZM and EXAMS predicted estimates with actual drinking water monitoring
data gives confidence in the drinking exposure assessment. Finally, the
residential and nonoccupational exposure estimates are also considered to
provide protective estimates of children's exposures given the quality of the
data and the conservative assumptions used. In summary, there is a high
degree of confidence in the exposure data and methodologies used when
assessing cumulative risk to children, that are considered to be protective of
children without understating risk. Thus, for the exposure assessment,
reliable data show that it is not necessary to retain the default 10X special
FQPA safety factor.
4. Integrative Analysis of Hazard and Exposure
A weight-of-evidence analysis has been conducted to determine the
completeness of the toxicity and exposure databases, and the degree of concern
for pre- and postnatal toxicity associated with the common mechanism of toxicity,
AChE inhibition. It was determined in the hazard assessment that there are
reliable data to support application of a 3X database uncertainty factor (which is
used to address the FQPA safety factor provision's expressed concern as to the
"completeness of the data with respect to... toxicity to infants and children....") to
address the limited data on ChE inhibition in immature animals and evidence
that supports the potential for OP pesticides to show ChE inhibitory effects at
lower doses in young animal compared to adults (i.e., the age group on which
the relative potencies values are based). There is no need for an additional
special FQPA safety factor to address potential pre- and postnatal toxicity
associated with ChE inhibition because application of the database uncertainty
factor to the RPFs for the OP accounts for age-dependent sensitivity in the
young and potential neurodevelopmental effects associated with ChE inhibition.
I.G Page 38
-------
The revised cumulative risk assessment for OP pesticides is based upon the
most comprehensive and data-specific exposure assessment ever performed by
OPP. This statement is true for all aspects of the exposure estimates including
pesticide sources from food, drinking water and residential uses. Each aspect of
the assessment relied upon the use of the best available data for input
parameters. The data were introduced into the assessment in large part in the
form of distributions, permitting the assessment to reflect the full range of
variability in each input parameter. This approach deviates from the past
practice used particularly for drinking water and residential exposure estimates
that relied upon high endpoint estimates. In this assessment, drinking water and
| residential estimates have been refined in much the same manner previously
I established for food assessments. The comprehensiveness and thoroughness
of this exposure assessment allows OPP to conclude that an additional safety
factor is not needed to address the completeness of the exposure database.
While the available data and methodologies used by OPP to estimate
exposures cannot be used to precisely define an exact exposure for any given
percentile of the population, OPP can bracket or otherwise define the range
within which exposures are expected to fall. Specifically, OPP believes that the
traditional single-day analysis in which individual days are assessed in isolation
I reflects a likely upper-bound of exposures. OPP also believes that the actual
| upper-bound of exposure is lower than the high-end estimated by the rolling
| average exposure (discussed in the Risk Characterization Section of the OP
| cumulative risk assessment). Additionally, the cumulative assessment was
| conducted in a way that does not intrinsically bias the analysis toward over
I estimation or under estimation of exposures, but instead reflects exposures
I anticipated to be experienced by the public. Accordingly, OPP believes that the
analysis captures the highly exposed groups (including children) and represents
exposures reasonably likely to occur and that the above-mentioned "bracketing"
represents realistic expected upper and lower bounds on the estimated
exposure. Final determinations regarding which predicted exposures will be
considered in making a regulatory decision will depend on sensitivity analyses of
predicted high-end exposures. These determinations could also play a role in a
final conclusion about whether OPP remains confident that the analysis
adequately captures the upper-bound of estimated exposures and, therefore,
whether there is continuing support for the conclusion that an additional FQPA
safety factor is not needed to address the completeness of the exposure
database.
s
In summary, given the highly refined nature of estimates for all pathways of
exposure, the use of bounding estimates to reflect the potential issues
associated with timing and repeated exposures, and the application of the
database uncertainty factor of 3X, the presumptive 10X safety factor can be
removed.
I.G Page 39
-------
I I. Revised OP Cumulative Risk Assessment
H
H. Risk Characterization
1. Introduction
Risk characterization is the interpretation phase of the assessment process.
Appropriate interpretation of results is particularly important for an assessment as
complex as the OP cumulative risk assessment. Many types of data, derived
from a variety of sources, have been combined to produce highly detailed
estimates of risk from multiple OPs in food, drinking water or from residential use.
The outputs of the assessment should be evaluated in a variety of ways.
Potential biases in input parameters, the direction of the bias, and the uncertainty
surrounding the inputs and the exposure model must be considered with regard
_ to their potential impact on the results of the assessment.
WVWW
OPP has attempted to reflect the completed risk mitigation measures from the
single-chemical assessments. OPP will continue to make risk decisions about
individual pesticides over the next months. Changes resulting from risk
mitigation measures completed through March 2002 have been included in this
assessment. Modifications in OP use patterns made after that date can be
incorporated after they occur. The current document presents the estimates of
risk associated with exposures to OPs in food, drinking water and from
residential uses of OPs. The current assessment has used two modes of
analysis (single-day and 7-day)to provide bounding estimates of potential
exposures, and also reflects the risk estimates at a variety of percentiles of
exposure. In addition, analyses were performed for periods of 14 days and 21
days to demonstrate that extending the averaging time for the risk assessment
has little impact on the results obtained. The detailed results of this assessment
are presented as a plot of MOEs over a period of 365 days. Contributions from
various pathways and routes of exposure are arrayed separately. The results are
presented graphically for the seven Regions, for the 1-2 year old and 3-5 year
old age groups (Appendix III.J.2 to III.P.2). Data output tables for the 20 to 49
year old and 50+ groups are presented as spreadsheets (Appendix III.J.3) The
results presented are based on a one day and seven day rolling average. For
Region A the assessment also presents the 14 and 21 day rolling average results
for the 1-2 and 3-5 year old groups. No single value in the assessment
should be used to independently arrive at the interpretation of the results.
As discussed below, interpretation of the assessment depends upon the
| synthesis of a vast body of information about the input data and the processing of
1 that data to determine whether an acceptable risk has been achieved. A number
| of crop/chemical combinations in the food assessment and one chemical/use
| combination in the residential risk assessment warrant additional scrutiny in
| determining any future activities arising from this assessment.
I.H Page 1
-------
2. Hazard and Dose-Response Assessment
The hazard and dose-response assessment is presented in detail in section
I.B. That section outlines the steps in developing the dose-response
| relationships for each pesticide and its capacity to inhibit acetylcholinesterase in
CM ! the brain of female rats. It includes a description of all of the data used in the
O | dose-response analyses. Reasons for the selection of methamidophos as the
index chemical for the OP cumulative risk assessment are also discussed.
Finally, section I.B. describes the exponential dose-response model used to
develop the dose response curves that provided the basis for developing the
relative potency factors (RPF) for each chemical and the points of departure
(POD) for the index chemical for each route of exposure (i.e., oral, dermal, and
inhalation).
a. Acetylcholinesterase Inhibition: Data Quality & Common Effect
The first step in deciding that a cumulative risk assessment was needed
was the determination that the OPs were toxic by a common mechanism, i.e.,
cholinesterase inhibition. This determination was made and subjected to peer
review by the Scientific Advisory Panel in 1998
(http://www.epa.gov/scipoly/sap/1998/march/comec.htm). Once a common
mechanism was identified, the next step in the process was to select an
appropriate method for combining the risks from exposures to several
pesticides from more than one source/route. A large body of data describing
the inhibition of acetylcholinesterase in plasma, red blood cells and brain has
been generated for each registered OP. OPP has elected to use the brain
acetylcholinesterase data from female rats as the basis for developing RPFs
and PODs for use in the assessment. The choice addressed a number of the
concerns raised by the SAP and the public. Brain acetylcholinesterase
inhibition is the endpoint used because it reflects a response in a target
tissue of concern that is relevant to humans. Although RBC and plasma
cholinesterase inhibition do reflect exposure to OPs and, therefore, the
potential for adverse effects, brain acetylcholinesterase inhibition is an
indication of direct effects upon the brain itself. Error due to the extrapolation
between the response in a surrogate tissue (i.e., red blood cell and plasma)
and a target tissue itself (brain) is eliminated. In addition, the data for the
brain compartment have very narrow confidence limits when compared to
those from the plasma and RBC compartments, suggesting that there is much
less variability in this compartment across the data base.
This assessment uses the Relative Potency Factor (RPF) approach which
applies dose addition. Briefly, the RPF approach uses an index chemical as
the point of reference for standardizing the common toxicity of the chemical
members of the cumulative assessment group (CAG). Relative potency
factors (i.e., the ratio of the toxic potency of a given chemical to that of the
index chemical) are then used to convert exposures of all chemicals in the
CAG into exposure equivalents of the index chemical. The RPF approach
I.HPage2
-------
utilizes dose-response information to provide an estimate of each OP's
potency for the common toxicity, and thus allows for the quantification of
exposure as it relates to the joint risk of the GAG. OPP selected the relative
potency factor approach based upon the relatively rich oral toxicity data base
on cholinesterase inhibition available for the OPs which permitted
consideration of the entire dose-response curve for each pesticide rather than
only focusing on NOAELs that are a function of study design. Although a
biological or pharmacokinetic modeling approach would have advantages in
determining the cumulative risk for these OPs, the input parameters for such
an approach are not available. Thus, the pharmacokinetic (PK)
characteristics of the OPs could not be incorporated in the dose-response
assessment which would allow for a more refined estimate of the combined
risk to humans. Therefore, OPP has applied simple dose addition and used
an empirical curve fitting model (i.e., the exponential model described below)
to determine RPFs and PODs.
b. Exponential Dose-Response Model
OPP, in collaboration with ORD, developed an exponential model to
describe the oral dose response curves for each OP that permitted fitting of a
combination of cholinesterase (ChE) activity data from different studies. This
model has been subjected to public comment and peer review by the SAP
(http://www.epa.gov/scipoly/sap/2001/september/finalreport.htm). Although a
PK model is the ideal approach, the SAP regarded the exponential model
(with their recommended improvements) to be appropriate for the data being
analyzed for derivation of relative potency factors and points of departure.
OPP has responded to the SAP recommendations on the exponential model
by making modifications to address the issues raised. One issue was that the
original model did not appropriately reflect cholinesterase inhibition at very
low doses. The revised statistical model now incorporates, to the extent
"3 supported by the data, a flat region at the low dose portion of the dose
response curve. Another issue raised by the SAP concerned the derivation of
the factor "B". The B value is the limiting value for the maximum
cholinesterase inhibition (called the horizontal asymptote). The SAP raised
the issue that the weighting strategy used for calculating the "B" which
assumed 100% cholinesterase inhibition (i.e., 0% ChE activity) did not
adequately reflect the actual B value for each OP (the B value was often less
than 100% inhibition at the asymptote). The revised approach has been
modified in order to generate B values for each OP reflective of its
dose-response data.
OPP assumed dose additivity by application of a single model to all of the
OP's dose-response curves. There is some uncertainty surrounding the
assumption of dose additivity given that the B values (horizontal asymptotes)
are heterogeneous among the OPs analyzed. This heterogeneity is indicative
that the dose-response curves are not parallel and, therefore the application
of simple dose addition is only an approximation of joint risk and may not be
I.H Page 3
-------
I precise. Dose additivity assumes that the common mechanism chemicals
I behave in a similar fashion (i.e., same pharmacokinetics and
| pharmacodynamics) and that their dose response curves will be parallel (i.e.,
| the ratio of their relative toxic potencies remain the same throughout their
| dose range). The underlying biological processes that determine the toxic
| potency of each OP are extremely complex and involve several metabolic
I • systems in different organs as well as re-synthesis rates of the different
I cholinesterases. The activation and/or deactivation rates differ for some of
I these OPs. However, because of insufficient data, these pesticides can not
| be separated into subgroups based on pharmacokinetic and
1 pharmacodynamic characteristics. Thus, current information on OP
1 pharmacokinetics and pharmacodynamics cannot provide a sufficient basis to
| depart from the assumption of dose additivity. Also, available studies on OP
| mixture interactions do not provide a sufficient basis for departure from dose
I additivity.
In summary, OPP believes that the model fitting procedure used in this
assessment provides reliable estimates of relative potency and points of
departure. The cholinesterase data used for the oral route of exposure was
quite extensive and, in general, of good quality for dose-response modeling.
The data for the inhalation and dermal routes tended to be less extensive and
not as robust for dose response modeling. OPP has refined the dose
response modeling for the oral dose by incorporating the SAP
recommendations in its dose response assessment of these OPs. OPP has
attempted to address uncertainty in extrapolating to lower human exposures
by the revised model and by using a low model estimate (BMD10) to develop
OP relative potency factors. OPP acknowledges that there is uncertainty that
dose addition applies to all of these OPs at human exposure levels. In the
absence of data to the contrary, however, dose additivity is assumed. OPP
realizes that the assumptions of additivity and the dose response modeling
used in this assessment may slightly overestimate response due to the
assumption that response will be uniform regardless of the underlying
background exposure level.
A BMD10 was selected as the basis for comparison of the dose-response
curves for the OPs. OPP's goals in selecting a point of comparison were to
choose a point in the observed response range, but low enough on the curve
to reduce the impact of any lack of proportionality between response that
might result from deviation from the assumption of proportionate dose
response between OPs. In addition, OPP was concerned that the magnitude
of the response (cholinesterase inhibition) be sufficient to ensure that it was
reliably distinguishable from background. A power analysis of the data used
in deriving the 21-day steady state determination indicated that there was
insufficient power to distinguish the change in cholinesterase inhibition reliably
below 10% inhibition. In addition, OPP has used the central estimate of the
BMD10 instead of the BMDL generally used for single chemical assessment.
This decision reflects the complexity brought to the analysis by the joint
I.HPage4
-------
I consideration of multiple studies for multiple chemicals. The use of the BMDL
I has been suggested for those instances in which single studies are modeled
f to provide an indication of a reasonable lower limit on response. In the OP
| cumulative risk assessment, the results of several studies were combined,
| introducing the potential for inappropriately broadening the confidence limits
I about the BMD10 and making the BMDL a likely underestimate of the POD
I for each chemical. These considerations strengthened the case for selection
| of methamidaphos as the index chemical because the BMD10 and the BMDL
| were very similar suggesting a very good fit of the data to the model.
s
s
I c. Selecting the Index Chemical
| OPP selected methamidophos as the index chemical for the current
| assessment. Methamidophos has sufficient data for cholinesterase inhibition
I to support modeling of a BMD10 by all three routes of exposure. The high
I quality dose response data for methamidophos permits reliable estimates of
I PODs for all routes without resorting to the use of the less precise NOAELs.
| Certainty in the PODs was considered to be of great importance in as much
I as they will impact the outcome of the assessment to a greater extent than
I any other aspect of the toxicity data base.
f d. Use of Steady State
| During the data evaluation phase, OPP elected to use only those data
| points that resulted from exposure of rats for 21 days or longer. This choice
| was made for a number of reasons. First, because of the many agricultural
| uses of OPs and the resulting residues that occur in food and water, and also
I the application of OPs in homes across the US (as reflected in the
I assessment), the likelihood of encountering an exposure to OPs with no prior
I recent exposure was considered to be small. Therefore, the use of single-day
"3 I toxicity data was considered inappropriate. Further, following exposure to an
C I OP, regeneration of cholinesterases to pre-exposure levels occurs in the time
h™ | scale of days to weeks, not a single day, making the exposed individual
""""* I potentially more vulnerable to subsequent exposures during that period.
I Examination of the rat data suggested that for most pesticides, cholinesterase
§ inhibition reached steady state approximately by 21 to 30 days after the start
i of dosing. After that point, little change occurred in the degree of inhibition
I resulting from continued administration of the dose for a longer period. OPP
| selected 21 days as a reasonable time point to assume that steady state had
1 been achieved. For the purposes of this discussion, steady state is defined
1 as the point at which further inhibition of the enzyme is offset by regeneration
1 of the enzyme and equilibrium has been achieved. All of the pesticides
1 considered have very stable, reproducible levels of cholinesterase inhibition in
I all compartments measured.
The selection of the data set to support the steady state decision hinged
upon two determinations about the data available. The first was the
I.H Page 5
-------
evaluation of rate of change in cholinesterase inhibition as discussed above.
The second was the decision that a 10% inhibition of brain cholinesterase
was a tolerable level of inhibition that was unlikely to result in an unacceptable
adverse outcome for the exposed individual. This decision is the more critical
in determining the application of the hazard data to the running time frame
approach than the actual selection of the time point of determining the extent
of inhibition. The application of a steady state approach is predicated on the
assumption that the extent of cholinesterase inhibition on any given day
reflects the balance between prior exposures and the extent of recovery
experienced. The processes of inhibition and recovery are balanced in the rat
data as they are in exposed human populations. The major distinction
between the steady state data from the rat studies and the likely inhibition in
the exposed population is that the actual dose to the rat on any day and on
preceding days is known. In the human population, the prior exposures can
not be known with certainty. However, as demonstrated by the current
exposure assessment, the prior exposures may be either higher or lower than
for the current day. OPP believes that the use of steady state data is
consistent with the results of biomonitoring data available in the literature.
There is a body of evidence that indicates a sizeable proportion of the US
population has a fairly constant background exposure to OPs. This is evident
from the results of the NHANES III in which 82% of people who provided
urine samples for analysis were found to be positive for trichloropyridinol, a
metabolite of the OPs chlorpyrifos and chlorpyrifos-methyl (Hill et al., 1995).
Further examination of the NHANES III data indicate that a sizeable
proportion of the population have metabolites in their urine that are not
compound specific, but are associated with other OPs. Preliminary analyses
of data collected under the auspices of NHEXAS also indicate that
metabolites from a variety of OPs are found in urine from populations of
adults and children sampled.around the US.
The use of 21-day steady state data in rats may over- or understate the
potential for cholinesterase inhibition based upon exposure in the current
day combined with residual effects from the preceding day(s). The extent
and direction of the error can not be known, however, data pertaining to
prior exposure to OPs such as those cited above reflect a pattern of
exposure to humans that is qualitatively different from the repeated daily
dosing used in rat feeding studies and therefore there is a possibility that
risk is overstated. The use of the 21-day steady state data fixes the
estimate of dose relating to a 10% cholinesterase decrement and permits
a reasonable estimate of risk from exposure to OPs.
This finding was important in determining the appropriate manner in which
to incorporate the available acetylcholinesterase inhibition data into the
hazard assessment. In conjunction with the understanding that the period of
reversibility for OP-induced cholinesterase inhibition is on the order of several
days to weeks, it provides a reasonable basis for the decision to use steady
state measures of cholinesterase inhibition as the basis for OPs RPFs and
I.HPage6
-------
I the PODs for methamidophos. It also supports the need to consider multiple
1 modes of exposure, focusing on both more extreme episodic exposures as
I well as longer term average exposures. These two modes of analysis when
I used together acknowledge the potential for continuous exposure over an
| extended period of time while allowing an evaluation of the potential impact of
| periods of elevated OP exposure.
I In summary, OPP has taken steps to address the most significant
I methodological issues raised concerning the dose response assessment
| developed in support of the OP cumulative risk assessment. OPP is
I confident that the assessment as performed is scientifically and statistically
I sound and based upon a reliable data set.
| 3. Use of Calendex and the Mode of Analysis
I The use of Calendex in conducting the current assessment is described in
= section I.F of this document. Calendex permits the simultaneous evaluation of
| more than one pathway of exposure. It also permits the evaluation of exposure
| on a calendar basis, considering changes in exposure patterns with season as
I pest pressures change. In the December 3 Preliminary OP Cumulative Risk
I Assessment, OPP demonstrated the use of Calendex to develop a distribution of
| linked single-day exposures. This approach to estimating an annualized
| distribution of exposures for use in risk assessment received numerous public
| comments and reviews from the FIFRA Scientific Advisory Panel.
5
| The single day analysis approximates an analysis similar to that performed for
1 acute dietary evaluations in single-chemical assessments. As in the acute
1 dietary analyses, a full range of inputs for dietary residues is paired with the
I individual consumption records from the CSFII. In addition, a water
I concentration value is paired with the water consumption value developed as
I described in section I.E. Finally, an estimate of exposure from residential
| sources is calculated and combined to develop an estimate of the total MOE. A
| distribution of total MOEs is generated for each day of the year. A series of
I percentiles are then selected for each day to evaluate potential risk concern for
I the combined uses of the OPs. This analysis assumes that the potential exists to
I experience a high-end exposure on every day of the year. In the current
| assessment, values are presented for the 99.9th, 99.5th, 99th, and 95th percentiles
E of exposure.
I
| OPP developed a second assessment considering the potential risk from a
| series of 7-day rolling averages across the year. This process is described in
| section I.F. This analysis is an attempt to better match the time frames for the
1 toxicity data with the consumption data which are not directly comparable. The
| toxicity data used in the assessment are based upon 21 days of exposure to rats
| in feed. This data reflects steady state measures of cholinesterase inhibition and
| is readily available for all OPs. Food consumption data are available only on a
| single-day basis. The 7-day average allows for consideration of the variable
E
I I.H Page 7
-------
nature of likely exposures to an individual across time. It permits consideration of
the impact of varying exposure in the diet and from residential uses of OPs from
day to day. OPP also investigated the potential impact of longer averaging times
(14 and 21 days) on the results of the risk assessment. The longer averaging
times resulted in incrementally small decreases in the estimated risk, with the
effect of duration decreasing with increasing time. This behavior is not
unexpected in that with longer averaging time, the exposure will approach the
,r_ mean exposure for the output distribution. However, the longer time frames will
^™ further obscure any time-related variability in exposure. The use of the 7-day
rolling average provides a more realistic profile of exposure across a series of
multi-day exposures while maintaining some sense of anticipated variability.
OPP believes that the results of the single day and 7-day analyses
successfully bound the anticipated exposures resulting from residues of multiple
OPs in food. The one-day analysis assumes that an individual is exposed to OP
exposures from the tail of the distribution every day. This assumption
overestimates risk. The seven-day analysis incorporates day-to-day variability in
exposure and is more representative of anticipated exposures. The major
sources of difference in the results of the two analyses arise from differences in
how the data is incorporated into the analyses and the ability to reflect day to
day variability. The differences are: the selection process for the two days of
CSFII data, the assumption of independence of residue data inherent in the use
of the PDP data in the assessment, and the inability of the approaches to allow
for carryover cholinesterase inhibition from prior exposures.
a. Use of CSFII Data
In the single day analysis, one diary for each individual in the CSFII is
selected to be paired with a randomly selected set of residue values for each
food consumed. A set of exposures from OPs in foods is developed and
arrayed as a distribution from high to low exposures. This type of
assessment assumes that the consumption of foods from one day to another
is independent, with no consideration of the potential for eating leftovers or
consuming foods purchased in bulk such as juices or potatoes. As a result,
the assessment over emphasizes variability in the diet. This factor may bias
the exposure assessment up or down depending the food/pesticide
combination that is not repeated appropriately. The magnitude of the effect
also will vary depending upon the specific food/pesticide combination.
\
The use of the CSFII data in the 7-day rolling average consists of a
random redraw of the two available days of consumption data for each person
in the data base. This process is intended to maintain the integrity of the data
for individuals, including to the extent possible, any information defining
patterns of diet peculiar to them. However, the redraw process results in the
implicit assumption that every individual in the CSFII consumes a diet that is
limited to the records in the diaries repeated randomly across the year. As a
result, the variability likely to occur in the diet is not fully expressed in the
I.H Page 8
-------
I current risk assessment. This factor is expected to reduce the range of
1 exposures to which any particular individual can be exposed by limiting the
| number of commodities and pesticides possible to those reported in the two
I daily diaries. This factor is anticipated to introduce a downward bias into the
| exposure assessment. However, the impact on the assessment is anticipated
[ to be small because all possible combinations of exposures are still available.
I The shape of the final distribution may be modestly affected by the difference
f in averaging that occurs due to the reduction in combinations available, but
§ the exposure estimates at the upper percentiles of exposure should not be
I significantly affected.
l
I b. Use of PDF Data
| POP data are used for most of the pesticide residues in food assessment.
| These data are introduced into the assessment in a manner that imposes the
I assumption that all eating days are independent with regard to the source of
= food consumed. In fact, consumption on any given day may not be
| independent of preceding days to the extent that individuals consume bulk
I items such as juice, bunches of grapes, or bags of produce or left-overs that
| have the same level of residues on multiple days. As a result, exposure from
I items of these types may be under represented in the single day and 7-day
| rolling analyses to the extent that a high end residue may be selected on one
I day, but not resampled on the subsequent days. As result, these
| assessments may be biased downward with respect to the exposure
I estimates developed, although the magnitude of the error is not known.
c. Impact of Residual Cholinesterase Inhibition
Cholinesterase inhibition resulting from OP exposure is not immediately
reversible. OPs bind covalently to the active site of the enzyme. Recovery is
largely due to regeneration of the enzyme rather than dissociation. As a
result, the recovery time (time required for Cholinesterase levels to return to
pre-exposure levels) is extended, requiring on average 1 to 2 weeks in
humans. As a result, the Cholinesterase inhibition experienced on any given
day is the sum of inhibition from that day's exposure combined with residual
inhibition due to exposure on the preceding several days. As each day
passes, the importance of inhibition from any given preceding day declines
until it is fully reversed. The single day analysis does not incorporate an
estimate of the phenomenon. This results in a downward bias of unknown
magnitude. The magnitude of the bias will be dependent on the likelihood of
exposures on previous days. The Calendex model attempts to incorporate
some aspect of the prior exposure in the 7-day rolling average approach by
allowing for the combined exposure over a 7-day period of time. To this
extent, Calendex captures the carryover aspect of exposure to pesticides.
However, this approach can not account for the biological aspect of declining
importance of an exposure with the passage of time. It also de-emphasizes
the impact of intermittent high exposures as they are averaged into the
I.H Page 9
-------
background. To the extent that Calendex can not reflect this aspect of the
impact of exposure to OPs, the 7-day rolling average will tend to be biased
downward with regard to the estimate of risk from exposure to OPs.
4. Food Assessment
The food component of the OP cumulative risk assessment is based primarily
v^ upon two extensive, reliable data sets: 1) USDA's Pesticide Data Program, and
2) USDA's Continuing Survey of Food Intakes by Individuals, 1994 -1996 + 1998
(CSFII). The PDP data provide a very reliable estimate of pesticide residues in
the major children's foods. They also provide an indication of the co-occurrence
of OPs in the same sample, alleviating much of the uncertainty about
^f co-occurrence in foods that are monitored in the program. The CSFII provides a
detailed representation of the food consumption patterns of the US public across
all age groups, during all times of the year and across the 48 contiguous states.
r~ These two data components provide a firm foundation upon which to assemble
,;<; other data to develop the OP cumulative risk assessment.
/ -\
a. Consumption Data
Up until this time, OPP has performed its risk assessments using the
1989-91 Continuing Survey of Food Intakes by Individuals (CSFII). This
survey was conducted by USDA and was based on responses over three
consecutive days. A more recent CSFII was performed (the 1994-96 CSFII)
which was supplemented in 1998 by the Supplemental Children's Survey.
This 1998 survey focused on children from birth to 9 years old and greatly
expanded the number of birth to 4 year old children in the survey data base.
Importantly, the Supplemental survey was designed in a manner such that the
results from the 1998 CSFII survey could be combined with the 1994-96
survey. OPP believes that the newer survey information provides a more
realistic estimate of potential risk concerns because it reflects the current
eating habits of the US public. Based in part on past recommendations of the
FIFRA Scientific Advisory Panel and other advisory bodies, based in part on
OPP analyses of dietary and behavioral patterns, and based in part on a
minimum number of individuals needed to provide a good representation of
eating patterns, OPP has determined that the following age groupings are
appropriate for the cumulative assessment: birth to 1 year of age (i.e., 0-11
months); 1 to 2 year of age (i.e., 12-36 months); 3 through 5 years of age; 6
through 12 years of age; 13 through 19 years of age; 20 through 49 years of
age; and 50 years of age and greater.
For this assessment, the following age groups were analyzed for all
regions: 1 to 2 years of age; 3 through 5 years of age; 20 through 49 years of
age; and 50 years of age and greater. These age groups were selected
because the other age groups are rarely the most highly exposed in the
single-chemical assessments. For Region A, all subpopulations were
evaluated to confirm this assertion. Region A was selected as an appropriate
I.H Page 10
-------
analysis to demonstrate the impact of a variety of parameters within the
assessment because it consistently demonstrates the highest exposures and
risks estimated for regions across the US. The change to the more refined
age groupings should improve our ability to identify age-related differences in
food consumption (especially among young children). The use of the newer
CSFII and the finer age breakouts should increase the accuracy and utility of
the risk assessment overall by making it more descriptive of the anticipated
exposures and risks for each age group.
OPP is confident that the consumption data available from the CSFII
permit a reasonable basis for estimating exposure to OPs in foods. The data
are used empirically in combination with residue values to estimate exposure.
As a result, no issues relating to the appropriateness of curve fitting
procedures have been introduced into the assessment. OPP also believes
that an adequate number of samples are available to support estimation of
exposure. Approximately 4000 consumption days for 2000 individuals are
available for each subpopulation. This body of data is sufficient to support
simulation well out into the tails of the exposure distribution with little concern
for overestimation of consumption. However, as is the case with all sampling
protocols, the proportion of samples available declines toward the extremes
of the output distribution. As a result, extreme output distribution values are
less well represented than those reflecting the central tendency for the output
distribution. OPP acknowledges that the use of CSFII in this assessment
| may not fully reflect the eating habits of high end eaters, introducing some
I uncertainty with respect to the tails of the distribution of estimated exposures
i in the assessment.
jj b. PDF Monitoring Data in the Assessment
i
I The use of POP as a source of residue data has a number of inherent
f benefits that preclude the need for the use of conservative assumptions in the
| assessment. POP provides a direct measure of the occurrence of more than
I one OP in any sample analyzed. OPP can use these data as an indication of
I pesticide co-occurrence likely to be encountered in foods. OPP assumes that
I co-occurrence mirrors the POP values; in fact POP composites contain
= multiple individual units which may have different profiles of co-occurrence.
| Therefore, use of POP data in this manner may overstate potential risk. POP
| implicitly reflects the percent of a crop that has been treated with any given
I OP by measurement of the residues.
| Samples with non-detectable residues are assumed to be "zero" values in
I this assessment. The impact of this assumption was tested in the OP
I Cumulative Risk Case Study (USEPA, 2000c) that was presented to the SAP
I in December 2000. In the Case Study, a similar use of PDP data as the
i residue data source in this assessment was demonstrated for 24 OPs. The
I resulting data set had characteristics very similar to the one used in the
| current assessment. The analysis performed demonstrated that the use of
I
I I.H Page 11
-------
the "zero" values had only negligible impact on the MOEs developed at the
upper percentiles of exposure.
Although the result of replacing all non-detectable residues with "zero"
values would intuitively suggest an under-estimation bias, OPP has
demonstrated through its case study that this change has little impact at all on
the portion of the exposure curve likely to be used for regulatory purposes.
This result is not surprising for a multiple chemical assessment addressing
the number of chemicals under evaluation here. This assessment combines
many data elements, with no.single chemical or commodity dominating the
exposure. The residue data used in this assessment include highly
consumed foods, and several of these have large numbers of detects as well
as a few high detects of OPs. There are detectable residues of at least one
OP on 25% of the samples used in this assessment with a high of 66% on
one commodity. Generally, the LODs for POP data are very low (the average
LOD for the entire data base is about 0.01 ppm). Therefore, it seems
reasonable that the effect of assumptions related to estimation of values"
below the LOD would not significantly influence exposures at the highest
percentiles of exposure. This result may not be the case for other
assessments containing fewer foods or lower levels of detectable residues
and should be evaluated for each subsequent case.
c. Data Translation from PDF
Not all foods to which OPs are applied are monitored in POP. OPP has
developed a scheme by which commodities that are measured by POP serve
as surrogate data sources for commodities that are not. This approach is
outlined in OPP/HED SOP 99.3 (USEPA, 1999b). It is based upon the
concept that families of commodities with similar cultural practices and insect
pests are likely to have similar pesticide use patterns. Although this approach
is generally sound, it introduces uncertainty with regard to how similar the use
patterns for a given pesticide are to those for even closely related
commodities.
For example, the same OP may be applied on a similar time schedule.
However, the rates of application may differ between the crops treated. The
number of treatments may also differ for application to the two crops. This
issue is of importance to consider when conducting sensitivity analyses of the
results of the risk assessment. When the data are adapted for the use of
several chemicals simultaneously, and estimates of co-occurrence are
derived from that data, the likelihood of an inappropriately assigned residue
becomes greater. Although the commodities may have similar cultural
practices, they may differ in the number of OPs registered for these uses. In
addition, the translation from one commodity to another implicitly assigns the
inherent percent crop treated information from one commodity to another.
The direction and magnitude of this error will differ from one commodity to
another.
I.H Page 12
-------
OPP believes that this potential source of error in its assessment will most
likely result in over-estimation bias. However, the magnitude of the error is
probably not great in that the commodities for which PDP data was translated
represent only ~1 % of a child's diet.
d. Other Sources of Residue Data
PDP data and surrogate PDP data do not cover all commodities of
interest. For meats, seafood and eggs, FDA's Total Diet Study and FDA
Monitoring data provided residue estimates. These data suggest that eggs
and seafood contain negligible residues. For most meats (beef, pork, sheep
and goat), the maximum residue from the Total Diet Study was used.
Although the use of the maximum residue as a single data point for meats is
an overestimate, OPP has conducted a sensitivity analysis making all
residues for meats zero and found that there was no change in the outcome
of the risk assessment at the upper percentiles of exposure. This is not
surprising in that the highest residue observed is itself very low. Therefore,
OPP considers these factors neutral with regard to their impact on the results
of the assessment.
Approximately 3% of the foods consumed by children 1-2 years of age
still remained unaccounted for after using FDA Total Diet Study and FDA
Monitoring data. Sugar, molasses and syrups were assigned a residue value
of zero. These products are highly processed commodities that are unlikely
to retain any significant residues following the processing steps. The limited
data from the Total Diet Study found no residues in pancake syrup or sugar.
No data are available for field corn or dried beans. However, these
commodities are also blended and highly processed before consumption.
OPP believes that omission of these foods from the assessment will not result
in any change in the results of the assessment.
e. Impact of Regulatory Actions
Inherent in the use of monitoring data to estimate future residues is the
concern that any changes in use patterns will not be reflected in the data.
The OPs are currently undergoing use changes as a result of the individual .
chemical decisions. In most cases for which legal agreements have been
signed, the uses have been removed from the assessment. For other
pesticides, pre-harvest intervals have been extended or rates have already
been reduced. These changes are not reflected in the assessment as they
are not yet apparent in the monitoring data available. A specific example of
this issue is the rate reductions agreed upon for azinphos-methyl on apples in
1999. Although the rate reductions have been implemented, they will not be
reflected in monitoring data until the 2002 PDP data become available. This
delay reflects the year lag in affecting a new growing season following
implementation as well as the long period during which treated apples can
remain in the chain of commerce following harvest.
I.H Page 13
-------
Decisions have not been completed for all OPs included in this
assessment. Completion of the regulatory process for these pesticides could
result in additional exposure and risk reduction measures. These changes
could result in further reductions in exposure in the food portion of the
assessment. The magnitude of that change is uncertain.
f. Model Outputs
•"•vv.
The single-day food component of the OP cumulative risk assessment
was conducted using the DEEM software. This program evaluates the full
range of dietary exposures across a single day. It permits a detailed
evaluation of the source of exposures with regard to which foods and
pesticides are the likely sources of the exposure. This analysis served as the
basis for determining which commodity/pesticide combinations warrant further
scrutiny in the event that further regulatory action is determined to be needed.
The use of the single day assessment was considered to be appropriate
because exposure to OPs in foods is uniform nationally, and it has no
significant seasonal variations. OPP has extensive experience with the two
data bases that confirm this assumption as reasonable. OPP has conducted
a large number of seasonal assessments of exposure to individual pesticides
in foods. These assessments show virtually no differences in exposures
across seasons. This finding is not surprising in light of the widespread
distribution of foods across the United States, and the proportion of foods that
are imported. Lack of seasonal consumption patterns is also not unexpected
given the ability to preserve and store foods for delayed consumption, and the
import of seasonal foods to bridge gaps in domestic production periods.
Similarly, POP does not suggest any significant alteration in the types of
pesticides encountered or the magnitude of residues across the year. The
assumption of nationally uniform distribution of foods does not reflect highly
localized consumption events that may be encountered by individuals who
obtain foods at road side stands and consume it closer to the time of harvest
than the foods available in larger grocery stores. OPP does not have reliable
data on either consumption or anticipated pesticide residues to permit
evaluation of this type of exposure, however we anticipate that only a small
percentage of food consumed would be affected.
The results of the food portion of the revised OP cumulative risk
assessment are summarized in Table I.H-1 . The results are presented in the
form of MOE for children 1 - 2 years of age, 3-5, adults 20 - 49, and adults
50+, at the 95th, 99th, 99.5th and 99.9th percentiles of exposure. The percentile
of exposure as used in this document indicates the percent of the output
distribution that is predicted to experience exposure less than or equal to the
exposure at that point on the exposure distribution curve. In other words, at
the 95th percentile of exposure, 95% of the output distribution is likely to
have the exposure indicated or less. Five percent are likely to be exposed to
higher amounts of OPs. The 1 - 2 year age group is consistently the most
highly exposed subgroup in the analysis. This is due to a higher consumption
I. H Page 14
-------
to body weight ratio for this age group. Results are presented for both single-
day and 7-day analyses for all regions, with 14-day and 21-day analyses
included for Region A. The FQPA Safety Factor was incorporated into the
RPFs to permit modification of the assessment on a chemical-specific basis
as appropriate based upon our current understanding of age-related
sensitivity. The toxicity endpoints for this assessment were developed in
consideration of a 10X uncertainty factor to account for interspecies variability
and a 10X uncertainty factor to account for intraspecies variability. As
discussed fully in section I. G., because some OP pesticides show
age-dependent sensitivity and there are missing comparative ChE inhibition
data in young animals for many of the OP's, the magnitude of the FQPA
Safety Factor was set at 3X for most of the OP pesticides. Age-dependent
susceptibility data are available for seven of the OP's. The data for
dimethoate, omethoate (a metabolite of dimethoate), chlorpyrifos, and
methamidophos support reducing the FQPA Safety Factor to 1X.
MOEs from the 7-day analysis exceeded 100 with all remaining uses
i (Table I.H-1). The MOE for children 1 - 2 years was 119 at the 99.9th
| percentile of exposure. As discussed above, OPP believes that this estimate
| is a reasonable approximation of the risk anticipated from consumption of
I OPs in foods.
| MOEs for the single-day assessment do not reach the target value of 100
I at the 99.9th percentile (Table I.H-1). The MOE for children 1 - 2 years was
I 45 at the 99.9 percentile of exposure. An MOE of 100 was reached at the
I 99.4th percentile of exposure. OPP believes that the 99.9th percentile of
| exposure in the single-day assessment is an overestimate of anticipated
| exposure, especially when considered as occurring over more than one day
| at a time. In addition, there is an overestimation of exposure resulting from
I the inability to reflect changes in residues due to recently implemented
I mitigation activities such as application rate changes and extended pre-
[ harvest intervals increases. This value may be biased toward overstating the
I risk from OPs in food. However, bias reflected in this particular point estimate
| is anticipated to diminish at lower percentiles of exposure. OPP can not
| determine at what point in the exposure distribution the exposure estimate
| begins to be biased toward understating the exposure anticipated.
| The decision as to whether additional mitigation activities are needed can
I not be made by looking at any single value in the results. Many factors must
| be weighed in determining the extent to which any particular value over- or
I understates the need for additional action. OPP believes that the actual
| exposures to the US public fall somewhere between the results of the two
I analyses presented. In addition, the application of hazard values results in
I offsetting issues with regard to the direction of change in the MOEs
i calculated.
I.H Page 15
-------
OPP has identified commodity/pesticide combinations that appear at the
upper end of the distribution and may warrant further study. These include:
acephate on green peppers and succulent beans; azinphos methyl on apples
and pears; dimethoate on apples, grapes, green peppers, pears, spinach,
succulent beans, and tomatoes,; methamidaphos on potatoes and tomatoes;
mevinphos on grapes and spinach; phorate on potatoes; and phosmet on
apples, grapes, and pears. Until the individual assessments for DDVP and
_,,,., dimethoate are complete, it is premature to attach significance to these
commodity/pesticide combinations. The significance of these
•-•;-vv. commodity/pesticide combinations cannot be fully understood without taking
into account all other relevant information, such as the results of the
sensitivity analyses.
)
OPP has evaluated the consumption records occurring in the tail of the
distribution to ensure that they reflect reasonable consumption patterns.
L- Analysis of the tail of the distribution (>99th percentile) indicates that no small
subset of consumption records dominates the outcome. This observation
increases OPP's confidence that the food and water components of the
assessment are not unduly influenced by unusual consumption patterns and
reflect the consumption habits of the public at large.
I.H Page 16
-------
Table I.H-1. Summary of the OP Cumulative Food Assessment
Cnilelrenl-2
Route:
Food*
i Percent! le
95
99
99.5
99.9
Exposure PeTrloo* i
Single Day Analysis1
MOE** :
353
128
91
45
Exposure Period
7-day Analysis :
Mean MOE
475
249
197
119
Exj;k>swre Period
14-day Analysis
Mem MOE
517
295
239
151
Exposure Pftpiott
21-day Analysis
Mean MOE
539
320
262
166
The additional FQPA Safety Factor is included as an adjustment to the chemical-specific Relative Potency Factors
"For the single day analysis for food, MOEs were calculated using DEEM software rather than Calendex software and thus no mean is applicable
Children 3-5
Router
Food* .
Percent! le
95
99
99.5
99.9
Exposure Period i
Single Day Analyst
MOE**
437
158
111
53
Exposure Period
7*day Analysis
Mean MOE
570
290
225
131
Exposure Period
14-day AnaJyste
Mean MOE
616
340
271
165
Exposure Period
21-day Analysis
Mean MOE
634
364
295
184
The additional FQPA Safety Factor is included as an adjustment to the chemical-specific Relative Potency Factors
"For the single day analysis for food, MOEs were calculated using DEEM software rather than Calendex software and thus no mean is applicable
Adults 29-49
Route:
Food*
: Pereentile
95
99
99.5
99.9
Exposure Period j
Single Day Analysis
MOE** \
1286
439
304
146
Exposure Period :
7-day Analysis
Mean MOE
826
784
622
364
The additional FQPA Safety Factor is included as an adjustment to the chemical-specific Relative Potency Factors
"For the single day analysis for food, MOEs were calculated using DEEM software rather than Calendex software and thus no mean is applicable
Adults 50+
Route:
Food*
Pereentile
95
99
99.5
99.9
Exposure Period
SingJe Pay Analysis
MOE**
1136
403
282
139
Exposure Period
7-day Analysis
Maart MOE
824
735
537
340
The additional FQPA Safety Factor is included as an adjustment to the chemical-specific Relative Potency Factors
"For the single day analysis for food, MOEs were calculated using DEEM software rather than Calendex software and thus no mean is applicable
I.H Page 17
-------
5. Residential Assessment
The residential component of the preliminary OP cumulative risk assessment
is the most sophisticated analysis of its type that OPP has ever conducted. It is
the first application of distributional analysis to residential exposure assessments.
It also factors in the seasonal and regional aspects of pesticide use. Three types
of data are used in the residential assessment: pesticide use; pesticide residue
dissipation; and exposure contact and exposure factors. Pesticide use data are
utilized to determine the percent of households using a pesticide, the timing of
the pesticide treatments, frequency and duration of exposure. Use data are also
important in identifying geographic regions where the pesticide will be applied. In
the current assessment, use data are specific to the region under evaluation and
vary according to the specific aspects of that region. Pesticide residue
dissipation data address the fate of the pesticides once applied to an
environment (e.g., lawns). Exposure contact data are exposure-specific metrics
that relate human exposure to pesticide residues. Humans come in contact with
the residues by contacting the product directly or by contacting the residues left
after the pesticide applications are made. Distributions of human exposure
factors, such as breathing rates, body weight and surface areas used in this
assessment come from the Agency Exposure Factors Handbook. These will not
be discussed in the risk characterization of the document because the values are
established and used throughout the Agency.
OPP has used all of the known available data to assess the significant
residential uses of the OP pesticides. The residential uses not covered by this
assessment are pet collars (DDVP and tetrachlorvinphos), crack and crevice
uses (DDVP) and pest strips (DDVP) used in attics, basements and other areas
with limited human access. Use of DDVP pest strips in closets and cupboards
were included. It should be noted that the DDVP pet collars are currently not
being marketed. While the tetrachlorvinphos pet collars have not been
assessed, the CRA does address the use of tetrachlorvinphos pet shampoos ,
sponge-on treatments and powders. Exposure from the shampoo, sponge-on
and powder treatments is likely to be higher than from pet collar use. This is
because greater amounts of active ingredient are applied and larger areas of the
pet are being treated. Although tetrachlorvinphos treated pet collars represent
the largest usage of the product, the number of people treating pets with the
liquid and powder products were adjusted upwards to reflect the collar use in
addition to the use of the other products. The usage data was taken from
NHGPUS.
Each data set used in the assessment introduces some potential bias in the
outcome of the exposure assessment. A summary of these biases, their
direction and magnitude, is presented in Table I.H-2.
I.H Page 18
-------
I EPA recently funded a study assessing adult and children's exposure to
I insecticides in flea collars. Preliminary results show that the use of pet collars
I does not result in significant exposure to pesticides (Boone et al., 2001). Spot
f urine analysis of 110 pre-school children in the Seattle Metropolitan area
i monitored for dialkylphosphate (DAP) metabolites suggested that DAP
jj concentrations were not significantly higher in children whose parents reportedly
I used pet care products (Lu et al., 2002).
i a. Exposure Contact and Pesticide Residue Dissipation Data
I Exposure contact data used to assess exposures experienced by the
I applicator of consumer oriented pesticides are by far the most robust
I information used in the residential portion of this assessment. In addition, the
I application of pesticides is one of the more straight-forward activity patterns to
I measure since it represents easily defined activities. Recent data generated
I by the Outdoor Residential Exposure Task Force (ORETF) have been used to
= assess the use of hose-end sprayers (lawn care products), rotary granular
| spreaders (lawn care products), hand-pump sprayers (home gardens and
| orchards) and hand held dusters (home vegetable gardens). Another study,
I submitted by a registrant, was also used to assess residential applicator
1 exposure using granular shaker cans to apply disulfoton. All studies meet or
I exceed current Agency guideline requirements (in particular regarding the
I number of replicates) and can be extrapolated to include clothing scenarios
I ranging from short-sleeved shirts and short pants to long-sleeved shirts and
I long pants. OPP has high confidence in the use of these data. Exposure
I contact data used to address the pet scenarios include chemical specific
| handler exposure
:
There are two post-application dermal exposure scenarios addressed in
this assessment. These are: post application dermal exposure to lawn care
products, and post-application exposure to vegetable and home orchard
pesticide applications. Like the applicator scenarios, the post application
garden and home orchard exposure scenarios are easily defined activities.
For harvesting vegetables or weeding, there is a substantial amount of data
I based on farm worker exposure performing similar activities in crops requiring
I substantial hand labor. These contact values have the potential to
I overestimate exposure since they are based on individuals working for profit
I based largely on their productivity. Such workers are likely to be more
| efficient and therefore exposed to a larger amount of treated surface than
i most home gardeners. A uniform distribution of values representing hoeing
I and harvesting may overestimate early season activities that consist of
1 potential exposure to small plants.
I Dermal exposure from post-application contact with the lawn chemicals is
I equally varied. Contact data, representative of the range of human activities
I has been difficult to model. Dermal contact exposure values were identified
1 in data described in Vaccaro et al., 1996, for adults who performed scripted
I I.HPage19
-------
activities and contact values for children performing non scripted activities on
lawns treated with a non-toxic substance were described by Black in 1993.
Rates of pesticide transfer in the studies with surrogate compounds were
similar to those observed in the chemical specific dissipation data available to
OPP.
Turf transferable residue data are available for all turf chemicals. For
malathion, these studies were conducted at multiple locations. Studies
conducted in Missouri, North Carolina and Pennsylvania were used for the
eastern regions and the study conducted in California was used for the
western regions. Similar regional residue data were available for the use of
malathion on home gardens and orchards and were used accordingly in this
assessment. These data are of good quality and provide accurate estimates
for this parameter.
There are no chemical specific data that measure the influence of wet
hands and the mouthing behavior of young children on the efficiency of
residue transfer. OPP considered a study performed by Clothier et al. (2000)
in which he observed an increase in transfer efficiency (1.5- to 3-fold) when
comparing a turf residue collection method to volunteers pressing dry hands .
or hands wetted with saliva. He observed a higher transfer rate for the
compound with the lowest application rate. This may suggest that the hand
surface becomes saturated and thus results in a lower transfer rate at higher
application rates. The factor of from 1.5- to 3-fold was used in the
assessment. The factor may overestimate the transfer of residues at higher
application rates.
Estimates of exposure to pet care products were developed using an
approach similar to the one taken with the turf care products. For applicator
exposure, the Agency used dermal and inhalation unit exposures coupled
with important statistics that influence exposure such as animal weights and
number of animals treated. The latter two variables were gleaned from
proprietary sources and an EPA funded study (Boone et al., 2001). For post
application exposure, surrogate dermal exposure data of individuals exposed
to treated animals were used to generate transfer coefficients, based on the
transfer efficiency of the available dislodgeable residue data for
tetrachlorvinphos on fur.
Tetrachlorvinphos specific data addressing exposure of individuals while
treating pets and post application pet fur measurements were recently
submitted to the Agency. The unit exposures from pet care product applicator
data (n-15) were expressed as an empirical distribution. Dog weights (n-176)
were expressed as a cumulative distribution. To assess post application
dermal exposure, an exposure study of 16 pet groomers, each exposed to 8
dogs treated with carbaryl, was used. Dermal transfer coefficients were
generated based on the transfer efficiencies of the tetrachlorvinphos pet fur
data and the measured exposure of the groomers. These data were also
I.H Page 20
-------
:
:
treated as an empirical distribution. Duration of exposure was based on video
I analysis of children (n-3) playing with pets (Freeman et al., 2001). At this
time the method OPP is using in this assessment is the best available as it
uses chemical specific data (applicator and fur residue), real world contact
data (groomers and video analysis of children).
b. Pesticide Use Data
| Accurate pesticide use data are key to the residential risk assessment.
| Useful information include regional site/pest markets, timing of application
| and the percent of households using their products. In the absence of
I specific pesticide use information, OPP developed exposure scenarios based
| on timing aspects found in regional Cooperative Extension Service
§ publications and surveys such as the National Home and Garden Pesticide
| Use Survey (NHGPUS), the National Garden Survey, and Doane's GolfTrak.
I The Cooperative Extension Service publications were useful for establishing
i the timing of various turf chemicals. The survey data were used to establish
| the number of households that may use a given pesticide. For some regions,
| these application windows were expanded to account for the differences in
| length of growing season. This is particularly important when regions consist
of several USDA Plant Hardiness Zones (e.g., Region 8). The NHGPUS
delineates percent of households using pesticides based on a large national
survey. These values consider users and non-users as well as homes having
lawns and those that do not. The use of this survey introduces uncertainty
into the analysis because of the age of the survey (1989-90). The data may
not reflect reductions in current OP use patterns and therefore overestimate
exposure. Doane's GolfTrak was used to identify the percent of golf courses
treated with pesticides and is more timely (1998-99). OPP believes this is a
robust data source. The National Garden Survey has been tracking percent
of households employing lawn care applicators and is considered very robust.
In addition, variables such as vegetable garden size are well characterized
since these gardens are easy for survey respondents to define.
c. Use of Calendex
OPP believes using a calendar-based model is justified in order to
manage the timing of pesticide applications and delineating subsequent
exposures in the general population. Models that can employ distributions of
the available residue and contact exposure data are needed to capture the
inherent variability in the exposed population and can be used to provide
justification regarding co-occurrence of pesticide exposure events. This
method is preferable to relying solely on point estimates and combining "what
if scenarios which only adds uncertainty, while providing little information to
risk managers regarding the potential numbers of exposed individuals and
their ranges of exposure. Calendex provides the ability to evaluate route
specific pathways which are defined by the model user so that appropriate
residue and residue contact data can be used.
I.H Page 21
-------
d. Non-dietary ingestion
Non-dietary ingestion is an important exposure pathway in the residential
assessment in the southern regions. Frequencies of hand to mouth events
used in the assessment are based on real world observations of children in
homes and day care centers enumerated on video tape. However, a number
of issues surround the estimation of the impact of this activity. The number of
hand-to-mouth events occurring in a given time frame was developed by
observing children's behavior during quiet play. Video tape data are based on
children situated indoors and not outdoors. Hand to mouth frequency may be
higher when children are engaged in "quiet play" (e.g., listening to stories)
than when engaged in active play (running/tag, etc.). Children playing on
lawns are likely to be engaged in active play. Therefore, the frequency of.
hand-to-mouth events used in the current assessment may be an
overestimate.
The variety of hand-to-mouth events (such as the hand being near the
mouth rather than in it) makes the enumeration of events difficult. Further,
video tape values provide no information on rate of transfer from treated
surfaces to hands. Transfer estimates in the assessment were based on
studies measuring wet hand transfer efficiency with wet hands using
surrogate compounds. No chemical specific data are available. For each
hand-to-mouth event, the hand is assumed to have residue when data
indicates a child may touch other things (e.g., clothing, non-treated surfaces
or nothing).
e. Results
The results of the residential portion of the cumulative risk assessment are
relatively straight-forward to interpret. The results of the individual regional
assessments can be found in section II of this document. Inhalation
exposures to DDVP from No-Pest strips are the major contributor to
residential exposures. This determination is relatively obvious because this is
the only remaining indoor use for OPs. Removal of DDVP from the
assessment resulted in MOEs that were essentially the same as those
deriving from food alone. Some of the regional assessments from the
southern regions also indicate hand-to-mouth activities by children in
conjunction with lawn scenarios as an important contributor to exposure.
Some uncertainty surrounds the estimate of exposure from hand-to-mouth
behaviors in the assessment Any bias from this uncertainty is anticipated to
overestimate exposure. The magnitude of overestimation is uncertain. OPP
| believes that the current OP cumulative risk assessment represents a
| reasonable, health protective estimate of likely exposure to OPs from
I residential uses.
I.H Page 22
-------
Table I.H-2. input Parameters Used in the Ex
)osure Models: Bias, Assumptions, Uncertainties, and Strengths
Model
Input Parameter
Bias*
Assumptions, Uncertainties, or Strengths
and Other Comments
Exposure Mode!: for
fte$«J
(Rex)
*• ~
- » neuiraf
- ~ dowrowarcf
Lawn Exposure
Unit Exposure:push-type
rotary spreader (mg
exposure per amount of
active ingredient applied)
Assumptions/Uncertainties
1. This unit exposure is based on 30 replicates consisting of
individuals using a push-type rotary spreader. A number of
clothing scenarios are possible to be generated from these
data. In this assessment short-sleeved shirt and short pants
were assumed. This may overestimate exposure as large
portion of exposure is to the lower legs. Although a surrogate
compound was used, exposure is believed to be more
influenced by the type of equipment used rather being
chemical specific. OPP has high confidence in these data.
2. A lognormal distribution was selected.
3. Assumed gloves are not worn. Survey data do indicate that
some residential handlers use gloves. Because consumers
are unlikely to use, remove and care for PPE in the manner
of professionals, it is unclear what impact this may have on
actual use.
4. The surrogate compound (dacthal) used in the exposure
study may be dustier than the granular formulations of the
OP compounds assessed. This factor increases confidence
that this variable will not underestimate exposure.
Area treated (square feet)
-to-
Assumptions/Uncertainties
5. A difficult variable to estimate. However, the assumption is
reasonable given the application equipment used.
Although, may underestimate areas that have larger lawns
(midwest), margins of exposure are large.
I.H Page 23
-------
Model
Exposure Mocfe) for
Rssidentta* Pathway
(Rex)
Public Health
Input Parameter
Huma?» AcMty Palter rt
Dermal Contact Transfer
Turf Residues: dermal
Turf Residues: hand-to-
mouth
Frequency of hand-to-
mouth events
Duration on lawn
Drift
Population Exposed
Bias*
* * upward
-•«rieu&ai
-«<&Wftward
~to +
~
-to +
~to +
-to +
~
~to +
Assumptions, Uncertainties, or Strengths
and Other Comments
6. Adults: activities performed with tank tops and short pants,
lognormal distributions may be reflective of study design
rather than actual activities (choreographed)
7. Children: Includes above scripted activities and a range of
non scripted activities. Non-scripted activities lognormal
distribution may be influenced by use of a non-toxic
substance (not a pesticide)
8. Assumes all adults and children living in households being
treated with lawn care products are exposed (enter treated
area).
9. Chemical specific data reflect a range of high values (e.g.,
immediately after application) and influenced byVatering-in
and rainfall.
10. Based on surrogate data. Lone OP in surrogate data had
the lowest transfer.
11. Based on video-observations of children situated indoors.
Active play outdoors may result in lower hand-to-mouth
frequencies.
1.2. For children, the value is time spent outdoors in addition to
time spent on lawns. Does not account for survey responses
of individuals that did not play on lawns or go outside.
13. Distribution of aerial and ground equipment values
14. Assumes a large percentage of the population being
exposed (based on those having lawns).
I.H Page 24
-------
Model
Exposure Model for
Rfcskteftt&l Palfsvay i
(Rex)
Home Garden
Input Parameter
Nymar* Activity Pattern
Applicator: Small Tank
Sprayer
Applicator: Granular
Area treated: ornamentals
Area treated:
vegetables/fruits
Postapplication:
vegetables/fruits
Frequency of applications
Bias*
* v ypward
~ ~mn&t&\
-«
-------
Model
Esdpowre Modeler
R
-------
6. Regional Water Exposure Assessments
The regional water exposure assessments are designed to represent
exposures from typical OP usage conditions at one of the more vulnerable
surface watersheds in the region. Regions were selected to reflect the climate
and soil conditions causing increased pest pressure and resulting OP use. Each
regional assessment focuses on areas where combined OP residues in drinking
water is likely to be among the highest within the region as a result of total OP
usage and vulnerability of the drinking water sources. In this manner, OPP is
confident that if the regional cumulative risk assessment finds that exposure in
water is not a significant contributor to the overall OP exposure from a vulnerable
area, it will not be a significant contributor in other areas in the region. However,
because the assessment is based on typical usage, it is not a high-end estimate
of pesticide exposure via drinking water at that vulnerable site. A comparison of
the estimated concentrations from individual OPs with available monitoring
indicates that this assessment is by no means worst case or unrealistic. In each
region, levels of one or more OP pesticides detected in monitoring studies exist
that are greater than that estimated by the cumulative water assessment; in
some cases, the estimates are off by an order of magnitude or more. However,
in that same region, estimates of other OP pesticides are similar to or greater
than detections found in monitoring studies (see Appendices III.E.1 and III.E.3,
as well as the regional assessments in 11.A through II.G, for detailed
comparisons). Although the potential exists that peak water concentrations for
one or more OP pesticides may not be captured in this approach, the impact on
the contribution from water to the overall risk assessment is anticipated to be
small.
The discussion that follows characterizes the results of the regional water
exposure distributions, and identifies assumptions and approaches to the
assessment that might impact the level of certainty in the results.
a. What Each Regional Assessment Represents
Each region in the assessment is represented by a geographic area with
the highest apparent potential for cumulative exposure to OPs in drinking
water. The vulnerable drinking water source in each geographic area
represents an area with relatively high usage of multiple OP pesticides in
relation to other parts of the region and coincides with surface water sources
of drinking water that are vulnerable to potential contamination by these OPs.
The focus on surface water sources of drinking water is a likely source of
overestimation bias inasmuch as ground water sources generally have lower
OP residues than are found in surface water.
Because OP usage varies within the region, the initial evaluation focused
on the areas of highest use, based upon the crops grown, which OP(s) are
used on these crops, how much OP pesticides are applied and when they are
used. Because the relative potency factors (RPF) have a large impact on the
I.H Page 27
-------
overall OP cumulative distribution, site selection tended to favor high-RPF
OPs such as disulfoton, dicrotophos, and terbufos. Since the purpose of the
assessment is to identify the impact from multiple OPs occurring in water in
the same area, the area(s) selected for the assessment do not necessarily
represent the highest exposure of a single chemical, but rather the highest
multiple OP exposure within the region. Since OP use may vary from year to
year and cropping and usage patterns may change, some areas in other parts
of the region may have greater water exposure in a given year.
Because OPP considers both total OP usage and vulnerability of the
drinking water sources, the site selected may not necessarily coincide with
the highest OP use area in the region or the area where runoff alone is
greatest. For instance, the highest OP use areas in the Northwest region
(Region B) are in Yakima County and eastern Washington and in southeast
Idaho. However, because of low rainfall, few surface-water intakes, and
irrigation-dominated agriculture, OP use in this area did not necessarily pose
the greatest risk to drinking water sources. Instead, the surface-water
sources of drinking water in the Willamette Valley were potentially more
vulnerable, despite lower OP usage.
Comparisons of the estimated pesticide concentrations with available
monitoring in each region indicate that, in almost every region, a few known
detections of one or more OP pesticides occur at higher levels than are being
predicted for the cumulative assessment. As noted, because the estimate
focuses on the cumulative impact from multiple OP pesticides, it doesn't
necessarily focus on the conditions that lead to the highest concentration of
one particular OP. In addition, some of the monitoring data may come from
water bodies that are not representative of drinking water sources. In some
instances, the higher monitoring levels may reflect uses that are being
cancelled, such as the residential uses of chlorpyrifos and diazinon. In the
;™j; case of azinphos methyl, in which upper percentile regional distributions were
consistently one to three orders of magnitude less than monitoring detections,
the underestimates may be due to inadequate or missing data on pesticide
fate and transport properties or usage.
b. What PRZM-EXAMS and the Index Reservoir Represent
OPP adapted available tools to provide daily distributions of OP levels in
water for incorporation into the probabilistic cumulative exposure assessment.
While these tools have provided OP distributions that are, in many cases,
comparable with available monitoring data in the same or nearby locations,
assumptions regarding the nature of the drinking water source and watershed
influence the estimated distributions.
I.HPage28
-------
I i. Nature of the Drinking Water Source
| The Index Reservoir is based on the specific geometry (watershed and
I reservoir size) of an actual reservoir (Shipman City) in the midwest. As
i such, it may best represent potential transport to similar drinking-water
| sources in high rainfall areas such as the eastern US. It may not so well
I represent reservoirs in drier parts of the west, where inflow and outflow
| are artificially managed. In addition, while the Index Reservoir scenario
| will not necessarily reflect short pulses of higher concentrations found in
| flowing rivers and streams, long-term average concentrations in a
I reservoir may be greater than in streams because of differences in the
[ residence time for water in these water bodies.
| The Index Reservoir is adapted to the runoff and stream inflow
i calculated from local soil and weather data. OPP used the PRZM runoff
I data for the cropping scenario that generated the lowest total runoff
1 volume in the region to derive the inflow and outflow of the Index
| Reservoir. This introduces a small additional error into the concentrations
| calculated for the other chemical-prop simulations in each region.
ii. Nature of the Watershed
PRZM is not a basin-scale model, but a field-scale model which
provides an edge-of-field estimate of pesticide loads in runoff to the
5.3-hectare reservoir simulated by EXAMS from a 172.8-hectare
watershed. PRZM does not explicitly account for the relative contributions
of each field to the Index Reservoir. OPP used a cumulative adjustment
factor (a combination of the regional percentage of the total watershed
area in crops with OP uses and the percentage of acres treated by each
OP on each crop) to adjust the resulting reservoir concentrations
calculated by EXAMS. Further information on the assumptions involved in
applying Percent Crop Area (PCA) factors for drinking water assessments
of individual pesticides can be found in the science policy paper, "Applying
a Percent Crop Area Adjustment to Tier 2 Surface Water Model Estimates
I for Pesticide Drinking Water Exposure Estimates" (USEPA, 2000e).
i
I PRZM does not account for location in the watershed: all fields are
I assumed to be uniformly distributed within the watershed, with runoff
I going directly into the reservoir. The simulation of multiple chemicals to
I multiple crops grown in different soils represents a significant adaptation
| of PRZM-EXAMS. Ideally, the cumulative drinking-water exposure
| assessment for a region would allow separating the different crop-soil
I regions within a watershed, and could simulate the different path lengths
I through runoff and stream-flow to the Index Reservoir. However, since
PRZM is an edge-of-field model, runoff from fields representing the
application of each OP to a different crop follows the same path length in
the treated field and empties directly to the reservoir. In other words, this
I.H Page 29
-------
w.vJ-*
simulation assumes that the treated fields with their individual soils are
uniformly distributed throughout the watershed and essentially ring the
index reservoir for direct deposition of the edge-of-field load.
Each crop use simulated in PRZM assumes that the entire area of the
watershed planted in the crop consists of a single soil. In each of the
regions, OPP used actual soil data from local soils on which the crops are
grown. When possible, the soil selected for each scenario was a
benchmark soil that was prone to runoff (classified as hydrologic group "C"
or "D" soils). While OPP attempted to simulate soils that might be prone
to runoff, the emphasis in developing the scenarios was to choose
important local soils for which sufficient data are available, and which are
known to be used to grow the crops of interest. These soils may not
represent those most prone to runoff, but afford reasonable certainty that
the simulation represents local soil conditions. While an assessment
using a single soil assumes that each part of the watershed will be equally
vulnerable to runoff, areas of higher and lower runoff vulnerability will exist
in an actual watershed.
iii. Multiple Years of Local Weather Data
Because the application rates, frequencies, and timing are held
constant, the PRZM/ EXAMS Index Reservoir simulations over multiple
years evaluate the impact of the variability in precipitation on the amount
of pesticide that reaches surface water. Because weather data spanning
24 to 36 years is available for many locations across the country, PRZM
and EXAMS can account for OP runoff from a wide range of weather
patterns not otherwise possible with monitoring studies that span relatively
few years. The age of the data (collected through 1983) limits OPP's
ability to compare of the modeling output to more recent monitoring data.
Weather data files for PRZM are available for weather stations across
the country. The weather station nearest to the county or counties used
for the simulations was chosen for the cumulative assessment. To the
extent that precipitation in these counties over the period of record might
have been greater or less than that recorded at the nearest weather
station, runoff for that area may have been over- or underestimated by
PRZM.
Additional uncertainty in the modeling results is associated with
application of OPs to irrigated crops, PRZM has a relatively simple
irrigation subroutine, applying a user-specified amount of irrigation to the
simulated field when the moisture content of the top soil layer drops to
some fraction of field capacity. Actual irrigation in the field follows a more
complicated formula, with irrigation timing dependent on the grower's
professional judgement of crop needs. In addition, PRZM has a limited
ability to distinguish between various irrigation methods.
I.H Page 30
-------
•.^w.%
c. What the Usage, Cropping Areas, and Acre Treatments Represent
Typical application rates and frequencies for each OP pesticide on each
crop were generated by taking the average reported in the USDA NASS
(National Agricultural Statistics Service) Agricultural Chemical Usage
summaries. This assumes that all applications were made at this typical or
average rate and that frequencies of applications were constant year to year.
The assessment considered only yearly variations in weather, and not
variations in application rates. Thus, using these typical application rates and
frequencies may underestimate water concentrations in years when pest
pressure is higher than in our reported years and may overestimate in years
when lower amounts of pesticide is used. The usage data was generally not
sufficient to conduct a probabilistic assessment over a distribution of actual
application rates.
In some instances, the typical and maximum label application rates were
identical. For instance, the typical rate for phorate application on sugarcane
in Florida was at the maximum label rate. In many cases the maximum label
rates were one to eight times greater than the typical rates (see Appendix
III.E.11). The extent to which the differences in rates would be reflected in
the OP cumulative distribution depends on a number of factors, including
timing of application relative to runoff events and relative potency of the
pesticide. As a result, the net difference in estimated cumulative distributions
between all typical and all maximum rates ranges from no difference in all but
the lowest percentiles in Region A to a factor of 2 to 4 times greater at the
higher percentiles (95th and above) in Regions E and G (Appendix III.E. 11).
Those comparisons reflect the maximum potential difference between
typical and maximum application rates by assuming that all OP pesticides
would be applied at the maximum rates to all crops. In reality, given the
.range in crops and pests treated by OP pesticides, it is more likely that only
some of the OP pesticides might be applied at maximum rates in a given year
and, thus, the difference would be less than that found in the comparison.
The regional percent crop area (PCA) factors are based on a large area:
the size of the hydrologic units (average > 1000 square miles) used generally
span multiple counties and may contain several watersheds that supply
drinking water intakes. These regional PCAs represent the aggregation of
crop areas from county-level NASS data and assume that the cropping area
is uniformly distributed. However, cropping intensity is variable and smaller
watersheds, including those capable of supporting drinking water supplies,
may have a much different (higher or lower) percentage of crop land than the
rest of the large basin. An example is Zollner Creek in the Willamette River
Valley. This watershed had the highest concentrations and frequencies of
detection of OPs among all of the NAWQA monitoring sites in the Willamette
Valley. This stream drained a watershed that was 99% agriculture, much
greater than the regional PCA of 60%.
I.H Page 31
-------
The regional assessment areas coincided with the area with the highest
PCA in all of the regions except the Northwest. In the Northwest, the regional
assessment focused in the Willamette River Valley, a generally lower-intensity
agricultural area which was otherwise more vulnerable because of OP usage
and/or the nature of the drinking water source. However, as already noted,
portions of the Willamette Valley had higher percentages in agriculture than
reflected by the PCAs generated using the larger hydrologic units.
The typical application rates and percent acres treated are derived from
state-level data (or NASS reporting districts) and assume uniform use
practices across the state. Indeed, an uneven distribution of application rates
and percent acres treated is expected in response to differing pest pressures.
This assumption will underestimate areas where pest pressures may dictate a
higher percentage of acres treated in a given year; similarly, it will
overestimate areas where low pest pressures will require fewer acre
treatments. In the Red River Valley (Northeast/North Central region),
differences in percent acres treated and application rates between the
Minnesota counties and the North Dakota counties located within the Red
River Valley are more likely due to differences in the state-level data than in
actual differences between the adjacent counties.
d. Timing of Application
OPP used crop profiles and other relative crop production publications to
establish a time frame for making the applications of the pesticide on a
particular crop (application window). The length of the window doesn't
necessarily reflect the range over which a pesticide will be applied in a
particular year, but the year-to-year variation in the application dates over
time. Thus, in any given year, the timing of application may be clustered
within a shorter time-frame than suggested by the application window.
However, because of weather and other environmental factors, the timing of
intensive pest pressure and/or OP application may vary across the window.
The date of application can have an effect on the predicted concentrations
generated by PRZM/EXAMS, depending on how close the pesticide
application coincides with rainfall events in any given year. To evaluate how
this may impact on the OP cumulative distribution, where multiple pesticides
are applied at different dates, OPP varied dates of application across the
active window for each OP-crop combination in Regions A and D (see
Appendix 11 I.E. 11 for details). The impact of varying dates of application was
most evident at the extremes in the distributions. The ratio in maximum
concentrations between the lowest and highest estimates was a factor of 5 to
6. For 99th and lower percentiles of exposure to OPs, the differences were
not as dramatic, with the ratio between lowest and highest values generally
two or less. This analysis only looked at the cumulative OP distribution and
did not evaluate variations in individual chemical distributions. In both
regions, the cumulative distribution generated at the beginning of the
I.H Page 32
-------
application window and used for the regional assessment was less than the
maximum estimated distribution. The ratio between the highest estimated
concentration distributions and that used for the regional assessment was
between 2 to 4 for the maximum estimated concentrations, but less than 2 for
99th and lower distributions.
In the absence of data to show otherwise, OPP assumed that all of the
pesticide applied on a particular crop is done on the same date. While this
may be an unreasonable assumption for a large watershed, it is not
unrealistic for the size of the watershed used in this assessment. This
assumption may result in higher peaks, but similar overall average
I concentrations than if applications are spread out over time. The resulting
I estimate of exposure may result in a small overestimation bias in the results
= that will be greater in large than in small watersheds.
In California (Region C - Arid/Semiarid West), OPP used California
Department of Pesticide Regulations (CDPR) census data for its regional
assessment. This information provided a distribution of applications by actual
date of application. For that regional assessment, OPP split the total
application into 5 applications, with each application representing 20% of the
total amount applied on that particular chemical. The absence of information
on application dates in MASS precludes OPP from taking a similar approach
in other regions. OPP also generated an estimated cumulative OP
distribution by using a single application at the beginning of the application
window, as was done in other regions. The cumulative OP concentration
distribution estimated using a single application was greater than that
estimated using 5 split applications by a factor of two or less (see Appendix
III.E.11 for details). While splitting the application over multiple days is
expected to result in lower peaks than a single application, the degree to
which a difference is seen depends on a number of factors, including the
mobility and persistence of the pesticide and the timing of applications in
relation to runoff-producing rainfalls.
e. Water Treatment Effects
Although not extensive, scientific evidence suggests that many of the
parent OP pesticide residues in water are likely to be transformation by
oxidation during water treatment, through chlorination or similar disinfection
treatments. These oxidative transformation products, such as sulfones,
sulfoxides, and oxons, are still of toxicological concern, have been detected in
treated water from water treatment plants. Limited data suggest that these
treatment by-products may be stable for sufficient periods of time (for least 24
to 96 hours) to move through the distribution system.
I.H Page 33
-------
The information is not sufficient to make quantitative adjustments to the
cumulative exposure estimates. To estimate potential impacts and to
determine whether additional information is needed, OPP assumed that any
transformation due to chlorination results in the conversion to a product of
toxicological concern. Thus, all OP parents that form oxons, sulfoxides, or
sulfones (see Table I.E-1) were assumed to be transformed into those
products as a result of oxidation. Where the transformation is less than
complete, and where non-toxic products are also formed, the such an
assumption will overestimate the ultimate drinking water exposure. While
limited information suggests that the other OP parent would be transformed
and removed from treated drinking water, sufficient information is not
available to quantify this for all OP pesticides. Thus, OPP did not assume
that any of the other OP parent pesticides would be removed. OPP assumed
that the sulfoxide and sulfone products are equal in toxicity to the parent and
that the oxon products are ten times more toxic than the parent. A
comparison of the RPFs for dimethoate (0.32) and omethoate (0.96), the
oxon of dimethoate, suggests that this assumption would be protective.
Table I.H-3 compares the cumulative OP distribution used in the risk
assessment (labeled "No treatment effects") with an estimated distribution
using the assumptions of treatment impacts described above (labeled "oxon
conversion w/ 10X increase in RPF"). In each region, the main cumulative
OP "pulse" in any given year is dominated by an OP which transforms into
sulfoxide and sulfone products (terbufos, phorate, or disulfoton). Since the
estimated distributions for those OP pesticides reflect the combined parent
plus sulfoxide/sulfone residues, any potential treatment effects from oxidation
of these chemicals is covered in the assessment. Conversion of OP parents
to oxons would not add significantly to the cumulative OP load in these
regions because (a) those OP pesticides which form oxons are not
contributing significantly to the cumulative "pulse" for the region, and/or (b)
those oxon-forming OP pesticides that are frequently detected in water
(chlorpyrifos, diazinon, malathion) have very low RPFs in.comparison to other
OP pesticides (such as dicrotophos, terbufos, and phorate).
I.H Page 34
-------
Table I.H-3. Comparison of OP cumulative distribution (ppm, methamidophos
equivalents) assuming no drinking water treatment effects to distribution
assuming oxon conversion with increase in toxicity
Region
Vlax
)9th
)5th
)0th
JOth
fSth
>0th
>5th
10th
Vtin
Mean
Contributors
o cumul.
DP pulses
Dxon-
ormers
legion
Max
39th
35th
30th
50th
f5th
50th
25th
10th
\Am
Mean
Contributors
o cumul.
DP pulses
Dxon-
brmers
Cumulative OP
Distribution, ppm
s»o
reatment
jffects
A
1.4E-02
9.0E-04
7.8E-05
3.6E-05
2.0E-05
1.7E-05
8.1E-06
3.4E-06
1.5E-06
4.1E-07
4.6E-05
Convert to
OXOI1W/10X
Increase in
m , ,
*atle
r»o
10X
9XO«
(Florida)
1.4E-02
9.0E-04
1 .OE-04
5.8E-05
3.5E-05
2.9E-05
1.6E-05
8.3E-06
4.5E-06
1.1E-06
5.5E-05
1.0
1.0
1.3
1.6
1.7
1.7
2.0
2.4
3.1
2.6
1.2
3horate + sulfoxide/ sulfone;
ethoprop
Chlorpyrifos, diazinon
D (Northeast/ North Central)
4.9E-03
1.5E-03
4.8E-04
2.0E-04
5.5E-05
3.1E-05
5.5E-06
1.5E-06
5.8E-07
2.0E-08
9.2E-05
4.9E-03
1 .5E-03
4.9E-04
2.2E-04
8.5E-05
6.0E-05
1.2E-05
3.8E-06
1.8E-06
2.0E-07
1. OE-04
1.0
1.0
1.0
1.1
1.5
2.0
2.3
2.5
3.2
10.0
1.1
ferbufos, phorate with sulfoxide/
sulfone transformation products
teM, chlorpyrfos, dimethoate
Cumulative OP -
Distribution, ppm
!4o
treatment
affects
£(
1.4E-04
1.2E-04
9.2E-05
7.5E-05
5.1E-05
4.6E-05
3.0E-05
2.0E-05
1.5E-05
8.3E-06
3.7E-05
Convert to
oxon w/ 1QX
increase in
ItPF
Ratio
no
10X \
oxon
Northwest)
2.6E-04
1.4E-04
1. OE-04
8.1E-05
5.7E-05
5.3E-05
3.6E-05
2.6E-05
2.0E-05
9.5E-06
4.4E-05
1.8
1.1
1.1
1.1
1.1
1.1
1.2
1.3
1.3
1.1
1.2
Ethoprop; azinphos methyl;
ihlorpyrifos
<\zM, bensulide, Chlorpyrifos,
jiazinon, dimethoate, malathion,
MePara, Phosmet
E (Humid Southeast)
3.7E-03
1.1E-03
3.6E-04
1.6E-04
6.5E-05
4.9E-05
1.8E-05
9.6E-06
6.2E-06
3.9E-07
7.9E-05
3.8E-03
1.1E-03
3.9E-04
1.9E-04
8.3E-05
6.4E-05
2.2E-05
1.1E-05
7.2E-06
7.3E-07
8.9E-05
1.0
1.0
1.1
1.2
1.3
1.3
1.2
1.2
1.2
1.9
1.1
Ferbufos, phorate, & disulfoton
with sulfoxide/ sulfone; acephate
Chlorpyrifos, dimethoate
Cumulative OP
Distribution, ppm
to
reatment
affects
;-.C (Arid
7.6E-04
2.2E-04
1 .6E-04
1.4E-04
1.2E-04
1.1E-04
7.6E-05
4.6E-05
3.0E-05
1.7E-05
8.3E-05
Convert to
oxon w/1 OX
Increase In
RPF
Ratio
no
treat
tax
oxon
Semiarid West)
9.9E-04
2.7E-04
2.0E-04
1.7E-04
1.4E-04
1.3E-04
1.1E-04
7.8E-05
5.4E-05
2.4E-05
1.1E-04
1.3
1.2
1.2
1.2
1.2
1.:
1.4
1.7
1.(
1.4
1.4
Disulfoton + sulfoxide/ sulfone,
3horate + sulfoxide/ sulfone
teM, Chlorpyrifos, diazinon,
dimethoate, malathion, MePara,
3hosmet
F (Lower Midwest)
3.7E-03
1.3E-03
4.7E-04
2.3E-04
5.7E-05
3.0E-05
4.6E-06
1.8E-06
9.7E-07
1.5E-07
8.2E-05
3.9E-03
1.4E-03
5.7E-04
3.1E-04
1.2E-04
8.2E-05
2.3E-05
8.4E-06
3.4E-06
4.7E-07
1.2E-04
1.1
1.1
1.2
1.4
2.1
2.7
5.C
4.7
3.5
3.2
1.4
Terbufos + sulfoxide/ sulfone;
Dhostebupirim
Chlorpyrifos, dimethoate,
nalathion MePara ohostebuoirim
I.H Page 35
-------
Region
Max
)9th
)5th
)0th
)0th
fSth
>0th
!5th
10th
Vlin
Mean
Contributors
o cumul.
}P pulses
Dxon-
ormers
Cumulative OP
Distribution, ppm
Mq
treatment
jffects
6
8.7E-03
4.3E-03
1.9E-03
1.0E-03
4.4E-04
3.1E-04
4.1E-05
8.4E-06
4.2E-06
1..4E-06
3.6E-04
Convert to
oxen wl 10X
increase in
RPF
Ratio
no
10X
PXOfl
Mid-south)
9.0E-03
4.4E-03
2.0E-03
1.1E-03
5.2E-04
3.8E-04
7.4E-05
1.5E-05
6.8E-06
1.8E-06
4.1E-04
1.0
1.0
1.0
1.1
1.2
1.2
1.8
1.8
1.6
1.3
1.1
Dicrotofos; acephate; terbufos +
>ulfoxide/ sulfone
Chlorpyrifos, dimethoate,
nalathion. MePara. Dhostebuoirim
Cumulative OP
£>j$trj button, ppro
No
treatment
effect*
Convert to
»0n Vft 10X
increase in
f*PF
Ratio
no
tox
?xon
Cumulative OP
Distribution, ppro
No
treatment
effects
Convert to :
oxonwMOXi
increase in
RPF
Ratio
no
treat;
WX
pxon
The assumption of oxon conversion with a 10X increase in toxicity had no
impact on the upper percentile of the concentration distributions for OPs in
water in the two regions with the highest estimated cumulative OP load in
drinking water - Region A or Region G . In Region B, the assumptions
regarding oxon transformation increased the maximum estimated cumulative
OP concentration by a factor of 2, but had little effect on the 99th or lower
percentiles of the water concentration distribution. This resulted in two spikes
off the peak OP pulses in two years of simulations, but a lower impact during
other times.
7. Conclusions
A multi-route, calendar-based risk assessment for a single chemical requires
the assessor to consider a variety of new issues in designing and interpreting a
risk assessment. The issues are more complex when the analyses address the
simultaneous exposures to more than one pesticide. OPP advanced its risk
assessment methods, across the board, as it developed this specific OP
cumulative risk assessment.
Many questions arise when interpreting results generated in a complex, highly
refined assessment. The detailed outputs allow in-depth analysis of interactions
of data sets to estimate the possible risk concerns and identify the sources of
exposures. In this assessment, assumptions are replaced with data from
surveys and monitoring studies and, as a result, the assessment provides a more
refined picture of what is likely to be encountered in the real world. In most
cases the assessment uses distributions of data. This practice permits
expression of the full range of values for each parameter.
I.H Page 36
-------
This revised assessment presents results as a range of estimated Margins of
Exposure (MOEs) using one-day and seven-day rolling averages at different
percentiles of exposure distribution. After careful analysis, the Agency believes
that the real world exposure is somewhere between the one-day and seven-day
rolling average, and generally these MOEs do not represent a concern. OPP is
analyzing the sources of exposure that are significant at the lower end of the
MOE range at the high percentiles of exposure distribution.
One of the major factors influencing the results at the highest portion of the
range (99.9th percentile) of exposure is the fact that for a few individual OPs, risk
assessments and mitigation actions have not been finalized. This is particularly
true for DDVP and dimethoate. This conclusion is supported by the results of the
analysis that removed pest strips containing DDVP from the assessment. The
resulting total (food, water and residential) MOE is essentially identical to that for
a food-only MOE analysis.
OPP has identified that a few uses of OP pesticides on food crops generally
play a larger role in the results of the food risk assessment. Overall evaluation of
the risk from exposure to OPs in foods suggests that, with the exception of
completion of outstanding single chemical assessments, the cumulative MOEs
from exposure to OPs in foods do not raise a concern.
The results of the residential risk assessment indicate that remaining uses of
OPs in a residential setting are anticipated to provide only minimal contributions
to the cumulative risks from OPs with the exception of pest strips containing
DDVP. The single chemical risk mitigation activities for DDVP have not been
completed. The impact of these activities may significantly reduce the
contribution of DDVP to the cumulative risk assessment.
OP cumulative risk from drinking water is generally at least one order of
magnitude lower than the contribution from OPs in food at percentiles of
exposure above 95th for all subpopulations evaluated. As the percentile of
exposure increases, the difference between the food and water contributions
increase. Below the 95th percentile of exposure, the water risk comes within one
order of magnitude of the food contribution. This pattern is consistent for all
regions in the current risk assessment. OP exposure from drinking water does
not play a significant role in the cumulative risk from OP use in the US
I.H Page 37
-------
I. Revised OP Cumulative Risk Assessment
I. Future Work
The Revised OP Cumulative Risk Assessment provides a detailed picture of
possible exposure to 32 OPs. Details retained in the assessment are sufficient to
evaluate the impact of the methods and assumptions on the results of the
assessment. This process is particularly important for a cumulative OP assessment
because of its complexity and much additional data compared to single-chemical
assessments. It uses distributions of data in place of point estimates to the extent
possible. Appropriate information submitted during the comment period has been
incorporated as appropriate as have comments from the March 2002 SAP review.
This revised assessment utilizes the same innovative methodologies as that in the
preliminary document. Since the issuance of the preliminary assessment OPP has
analyzed the results presented therein and revised the outputs as necessary. Also
further risk mitigation on individual chemicals which has occurred since December
2001 is incorporated in this revised assessment. This document also contains a
section discussing the FQPA safety factor as it was applied in this cumulative
assessment. That Section of the assessment will be reviewed by the SAP in June
of 2002. This revised assessment will undergo an additional comment period during
June and July. EPA will evaluate the SAP's comments as well as other comments
or data that it receives and will modify this assessment as appropriate. In addition,
as existing analyses are revised or new information is obtained, EPA will review this
assessment and make further changes as appropriate.
I With respect to the next steps discussed in the preliminary assessment, the
| revised document reflects essentially all of the proposed short term actions. In that
| document some of the activities were flagged as long term activities. These
I activities are not necessary for completion of the OP cumulative risk assessment,
| but are actively being pursued by OPP at present in the interest of improving OPP's
"3 | risk assessment process. These long term steps are listed below categorized by
C | discipline. Please note that no long term steps were discussed in the food and risk
:r 1 assessment methodology sections of the preliminary cumulative risk assessment
"""" | and therefore there are no listings in these sections below.
E
| As with the preliminary assessment, new information submitted during the
| comment period that will serve to improve the accuracy of the assessment will be
I incorporated into the assessment as appropriate. Also, further risk mitigation on
| individual chemicals which may have occurred since December of 2001 is
I incorporated in this revised assessment.
I.H Page 1
-------
1. Hazard Assessment
CD Long term: Research to develop and implement physiologically based
pharmacokinetic [PBPK] models, which describe the time course disposition
of chemicals and their metabolites, are well suited to provide more refined
estimates of relative toxic potencies and points of departure for future
cumulative risk assessment. OPP is currently working with the EPA's Office
of Research and Development on the development and testing of such
models for common mechanism pesticides.
® Long term: Pursue with ORD investigations on the interactions among simple
mixtures of common mechanism pesticides to better understand the concept
and application of dose additivity.
2. Food Exposure Assessment
To be determined.
3. Drinking Water Exposure Assessment
® Long term: What aspects of the modifications to the water residue modeling
process can be applied to the conduct of single chemical aggregate
assessments? What differences in assumptions may be needed for
implementation of that process for single chemical assessments?
4. Residential Exposure Assessment
(j) Long term: Develop a science-based process for incorporation of spray drift
and other sources of exposure into residential exposure assessment.
® Long term: What aspects of the modifications to the residential exposure
estimation process can be applied to the conduct of single chemical
~,,; aggregate assessments? What differences in assumptions may be needed
for implementation of that process for single chemical assessments?
© Long term: Develop better data defining the hand to mouth behavior of
children in a variety of settings and for active and quiet play.
5. Risk Assessment Methodology
To be determined.
I.H Page 2
-------
1 I. Draft OP Cumulative Risk Assessment
I (Please note: This section is still undergoing editing)
:
J. References
Abu-Qare AW, Abdel-Rahman AA, Ahmad H, Kishk AM, and Abou-Donia MB.
2001 a. "Absorption, distribution, metabolism and excretion of daily oral doses of
[14C]methyl parathion in hens;" Toxicol Lett. 2001. Nov 30;125(1-3):1-10.
Abu-Qare AW, Abdel-Rahman A, Brownie C, Kishk AM, and Abou-Donia MB.
2001 b. "Inhibition of cholinesterase enzymes following a single dermal dose of
chlorpyrifos and methyl parathion, alone and in combination, in pregnant rats;" J
Toxicol Environ Health A. 2001. Jun 8;63(3): 173-89.
Adgate JL, Barr DB, Clayton CA, Eberly LE, Freeman NC, Lioy PJ, Needham LL,
Pellizzari ED, Quackenboss JJ, Roy A, Sexton K. 2001. "Measurement of
children's exposure to pesticides: analysis of urinary metabolite levels in a
probability-based sample." Environmental Health Perspectives 109(6):583-90.
Aerts M, Cockrell P, Botts D, Lamberts M, Pernezny K, and Shuler K. 1999. "Crop
Profile for Beans (Snap) in Florida." USDA Office of Pest Management Policy and
Pesticide Impact Assessment Program. October 8,1999.
Aerts M, Cockrell P, Neussly G, Raid R, Schueneman T, and Seal D. 1999. "Crop
Profile for Corn (Sweet) in Florida." USDA Office of Pest Management Policy and
Pesticide Impact Assessment Program. Updated August, 1999.
Aprea C, Strambi M, Novell! MT, Lunghini L, Bozzi N. 2000. "Biologic monitoring of
exposure to organophosphorus pesticides in 195 Italian children." Environmental
Health Perspectives. Jun;108(6):521-5.
Aprea C, Betta A, Catenacci G, Lotti A, Magnaghi S, Barisano A, Passini V, Pavan I,
Sciarra G, Vitalone V, Minoia C. 1999. "Reference values of urinary
3,5,6-trichloro-2-pyridinol in the Italian population-validation of analytical method
I and preliminary results (multicentric study)." Journal of AO AC International. 9
I Mar-Apr;82(2):305-12.
2
Aprea C, Sciarra G, Orsi D, Boccalon P, Sartorelli P, Sartorelli E. 1996. "Urinary
excretion of alkylphosphates in the general population (Italy)." The Science of the
Total Environment. Jan 5;177(1-3):37-41.
Astroff, AB, Freshwater KJ, Eigenberg D. 1998. "Comparative organohposphate-
induced effects in adult and neonatal sprague-dawley rats during the conduct of
multigeneration toxicity studies." Reproductive Toxicology. 12(6):619-45.
I.J Page 1
-------
Astroff AB, Young AD. 1998. "The relationship between maternal and fetal effects
following maternal organophosphate exposure during gestation in the rat."
Toxicology and Industrial Health. Nov-Dec; 14(6):869-89.
Atterberry TT, Burnett WT, Chambers JE. 1997. "Age-related differences in
parathion and chlorpyrifos toxicity in male rats: target and nontarget esterase
"r"' sensitivity and cytochrome P450-mediated metabolism." Toxicology and Applied
Pharmacology. 1997 Dec;147(2):411-8.
Augustinsson KB, Barr M. 1963. "Age variation in plasma arylesterase activity in
children." Clin. Chem. Acta. 8:568-573.
Bacheler JS, Edmisten KL, and Koenning SR. 1999. "Crop Profile for Cotton in
North Carolina." USDA Office of Pest Management Policy and Pesticide Impact
Assessment Program. Updated November, 1999.
Bailee, D. 1990. "A Golfer Exposure Study with Chlorothalonil Used for Golf
Course Maintenance-1985." Lab. Proj. No. 1148-85-0059. Unpublished study
prepared by Ricerca, Inc. 264 p. MRID 424338-11.
BeckleyP. "Crop Profile for Sugarcane in Louisiana." 1999. USDA Office of Pest
Management Policy and Pesticide Impact Assessment Program. April 26, 1999.
Belcher T, and Owen N. 1996. "Transferable Residue Study and Postapplication
Exposure Study-Malathion Residues in Turf after Handspray Applications to Turf."
Lab. Proj. No. 95463. 95470. Unpublished study prepared by ABC Labs California.
397 p. MRID 439450-01.
Benke GM, Murphy SD. 1975. "The influence of age on the toxicity and metabolism
of methyl parathion and parathion in male and female rats." Toxicology and Applied
Pharmacology. Feb;31(2):254-69.
(
Beyrouty, P. 2002a. "A developmental neurotoxicity study of orally administered
methyl parathion in the rat." ClinTrials BioResearch Ltd., Senneville, Quebec. Lab
Project Number: 97574, March 1, 2002, MRID 45630301, unpublished.
Beyrouty, P. 2002b. "A study on the effects of orally administered methyl parathion
on cholinesterase levels in adult, juvenile, and neonatal rats." ClinTrials
BioResearch Ltd., Senneville, Quebec. Lab Project Number: 97558, February 26,
2002, MRID 45656501, unpublished.
Bigbee JW, Sharma KV, Gupta JJ, Dupree JL. 1999. "Morphogenic role for
acetylcholinesterase in axonal outgrowth during neural development."
Environmental Health Perspectives. Feb; 107 Suppl 1:81 -7.
J Page 2
-------
I Black KG. 1993. "Assessment of Children's Exposure to Chlorpyrifos from Contact
I with a Treated Lawn." A dissertation submitted to the Graduate School-
i NewBrunswick Rutgers. UMI Dissertation Services.
E
| Blancato JN, Knaak J, Dary C, and Power F. 2000. "Multi-Route Pesticide
I Exposures from a PBPK Model for Three Pesticides: Chlorpyrifos, Isofenphos, and
3TI I Parathion." Presented at Annual International Meeting of ISEA, Monterey, CA,
| October, 2000; paper submitted for review.
I
| Brandenburg RL, Bailey JE, and Jordan D. 2000. "Crop Profile for Peanuts in
| North Carolina." USDA Office of Pest Management Policy and Pesticide Impact
I Assessment Program. Updated March, 2000.
=
| Breiman L, Friedman JH, Olshen RA, and Stone CA. 1984. Classification and
| Regression Trees. Wadsworth: New York.
Breslin WJ, Liberacki AB, Dittenber DA, Quast JF. 1996. "Evaluation of the
developmental and reproductive toxicity of Chlorpyrifos in the rat." Fundamental and
Applied Toxicology. Jan;29(1): 119-30.
Brimijoin S, Koenigsberger C. 1999. "Cholinesterases in neural development: new
findings and toxicologic implications." Environmental Health Perspectives. Feb;107
Suppl 1:59-64.
Brodeur J, DuBois KP. 1963. "Comparison of acute toxicity anticholinesterase
insecticides to weanling and adult male rats." Proc. Soc. Exp. Biol. Med. 114:509-
511.
Burlina A, Michielin E, Galzigna L. 1977. "Characteristics and behaviour of
arylesterase in human serum and liver." European Journal of Clinical Investigation.
Feb;7(1):1.7-20.
Butler AM, Murray M. 1997. "Biotransformation of parathion in human liver:
participation of CYP3A4 and its inactivation during microsomal parathion oxidation."
The Journal of Pharmacology and Experimental Therapeutics. Feb;280(2):966-73.
Butterfield BW, NGA Research Director. 1997. "National Gardening Survey
1996-97." The National Gardening Association, Inc.
Calabrese EJ. 1991. Multiple Chemical Interactions. Lewis Publishers, Inc:
Chelsea, Michigan; pp. 3-115 and 355-375.
California Environmental Protection Agency Department of Pesticide Regulation.
"Pesticide Use Reporting." Online. Available:
http://www.cdpr.ca.gov/docs/pur/purmain.htm
I.J Page 3
-------
Casida JE, Baron RL, Eto M, and Engel JL. 1963. "Potentiation and neurotoxicity
induced by certain organophosphates;" Biochem Pharmacol. 12: 73-83.
Chakraborti TK, Farrar JD, Pope CN. 1993. "Comparative neurochemical and
neurobehavioral effects of repeated chlorpyrifos exposures in young and adult rats."
Pharmacology, Biochemistry, and Behavior. Sep;46(1):219-24.
Chambers JE, Carr RL. 1993. "Inhibition patterns of brain acetylcholinesterase and
hepatic and plasma aliesterases following exposures to three phosphorothionate
insecticides and their oxons in rats." Fundamental And Applied Toxicology.
>w>w
Chanda SM, Harp P, Liu J, Pope CN. 1995. "Comparative developmental and
maternal neurotoxicity following acute gestational exposure to chlorpyrifos in rats."
Journal of Toxicology and Environmental Health. Feb;44(2):189-202.
Cherry N, Mackness M, Durrington P, Povey A, Dippnall M, Smith T, Mackness B.
2002. "Paraoxonase (PON1) polymorphisms in farmers attributing ill health to
sheep dip." Lancet. Mar 2;359(9308):763-4.
Cherry RH and Schueneman TJ. 1998. Insect Management in Sugarcane.
University of Florida Department of Entomology, Florida Cooperative Extension
Service, Institute of Food and Agricultural Sciences. ENY-406. Available:
http://edis.ifas.ufl.edu
Clement JG. 1984. "Role of aliesterase in organophosphate poisoning."
Fundamental and Applied Toxicology. Apr;4(2 Pt 2):S96-105.
Clothier JM, and Lewis RG. 1999. "Dermal Transfer Efficiency of Pesticides from
Turf Grass to Dry and Wetted Palms." Prepared for U.S. Environmental Protection
Agency, National Exposure Research Laboratory, Research Triangle Park, NC.
Cohen SD. 1984. "Mechanisms of Toxicological Interactions Involving
Organophosphate Insecticides;" Fundam Appl Toxicol. 4:315-324.
Collins RD, and DeVries DM. 1973. "Air Concentrations and Food Residues from
Use of Shell's No Pest Insecticide Strips;" Bull Environ Contam Toxicol. 9(4): 227-
233.
Costa, LG, W-F Li, RJ Richter, DM Shih, AJ Lusis, Furlong CE. 2002. PON1 and
organophosphate toxicity. Submitted for publication in: Paraoxonase (PON1) in
Health and Disease: Basic and Clinical Aspects. L.G. Costa and C.E. Furlong, Eds.
Kluwer Academic Press.
I.J Page 4
-------
X-JiOCI-?
X-K-M.
Costa LG, McDonald BE, Murphy SD, Omenn GS, Richter RJ, Motulsky AG, Furlong
CE. 1990. "Serum paraoxonase and its influence on paraoxon and
chlorpyrifos-oxon toxicity in rats." Toxicology and Applied Pharmacology. Mar
15;103(1):66-76.
Cronholm G, Knutson A, Merchant M, and Teetes G. 1993. "Managing Insect and
Mite pests of Texas Sorghum." Texas Agricultural Extension Service, The Texas A
& M University System. B-1220.
Crowe F, Gingrich G, Lundy R, Mellbye M, and Ostlund B. 1999. "Crop Profile for
1 Mint in Oregon." USDA Office of Pest Management Policy and Pesticide Impact
iw | Assessment Program. Revised September 2, 1999.
S«ww 3
&M4*
Crumpton TL, Seidler FJ, Slotkin TA. 2000. "Developmental neurotoxicity of
| chlorpyrifos in vivo and in vitro: effects on nuclear transcription factors involved in
I cell replication and differentiation." Brain Research. Feb 28;857(1-2):87-98.
I Dam K, Seidler FJ, Slotkin TA. 2000. "Chlorpyrifos exposure during a critical
I neonatal period elicits gender-selective deficits in the development of coordination
I skills and locomotor activity." Brain Research. Developmental Brain Research. Jun
I 30;121(2):179-87.
j:
I Dam K, Garcia SJ, Seidler FJ, Slotkin TA. 1999. "Neonatal chlorpyrifos exposure
I alters synaptic development and neuronal activity in cholinergic and
I catecholaminergic pathways." Brain Research. Developmental Brain Research.
I Aug5;116(1):9-20.
I Dam K, Seidler FJ, Slotkin TA. 1998. "Developmental neurotoxicity of chlorpyrifos:
I Delayed targeting of DNA synthesis after repeated administration." Brain Research.
I Developmental Brain Research. 108:39-45.
| Darnell T, Ewart W, Mielke E, Nelson T, Olsen J, Niederholzer F, Riedl H, and van
I Buskirk P. 1999. "Crop Profile for Apples in Oregon." USDA Office of Pest
I Management Policy and Pesticide Impact Assessment program. December 19,
I 1999.
| Davidian M, and Giltinan DM. 1995. Nonlinear Models for Repeated Measurement
I Data. Chapman and Hall: New York.
f Deacon MM, Murray JS, Pilny MK, Rao KS, Dittenber DA, Hanley TR Jr, John JA.
I 1980. "Embryotoxicity and fetotoxicity of orally administered chlorpyrifos in mice."
I Toxicology and Applied Pharmacology. Jun 15;54(1):31-40.
I Doane and GolfTrak. DOANE Marketing Research, Inc. GolfTrak, 1998-1999.
| Drews U. 1975. "Cholinesterase in Embryonic Development." Progress in
I Histochemistry and Cytochemistry. 7(3): 1 -52.
s
1 I.J Page 5
-------
I DuBois KP. 1969. "Combined Effects of Pesticides." Canad Med Assoc J.
I 100:173-179.
i •
I DuBois KP. 1961. "Potentiation of the Toxicity of Organophosphorus Compounds;"
i Adv Pest Control Res. 4:117-151.
| Dupree JL, Bigbee JW. 1994. "Retardation of neuritic outgrowth and cytoskeletal
I changes accompany acetylcholinesterase inhibitor treatment in cultured rat dorsal
| root ganglion neurons." Journal of Neuroscience Research. Dec 1 ;39(5):567-75.
j Ecobichon DJ, Stephens DS. 1972. "Perinatal development of human blood
I esterases." Clinical Pharmacology and Therapeutics. 14:41-47
s
I Edmisten KL, York AC, Yelverton FH, Spears JF, Bowman DT, Bacheler J,
I Koenning SR, Hodges SC, Naderman GC, Brown AB, and Culpepper AS. 2001.
i "2001 Cotton Information." North Carolina Cooperative Extension Service. AG-417.
I Available: http://ipmwww.ncsu.edu/Production Guides/Cotton/contents.htmi
Eigenberg, DA, Sangha GK, Thyssen JH. 1996. "Organophosphate-induced
maternal and pup cholinesterase inhibition in two-generation reproduction studies
with rats." Toxicologist. 30:310.
I Eto M. 1974. Organophosphorus pesticides: organic and biological chemistry.
I CRC Press, Cleveland. 387 pp.
I Felland CM. 2000. "Crop Profile for Peaches in Pennsylvania." USDA Office of
| Pest Management Policy and Pesticide Impact Assessment Program. May, 2000.
I Fenske RA, Kissel JC, Lu C, Kalman DA, Simcox NJ, Allen EH, Keifer MC. 2000.
I "Biologically based pesticide dose estimates for children in an agricultural
I community."
I Environmental Health Perspectives. Jun;108(6):515-20.
5
I Fenske RA, and Lu C. 1994. "Determination of Handwash Removal Efficiency:
| Incomplete Removal of the Pesticide Chlorpyrifos from Skin by Standard Handwash
I Techniques." Am Ind Hyg Assoc J. 55:425-432.
I Fenske RA, Wong SM, Leffingwell JT, and Spear RC. 1986. "A Video Imaging
1 Technique for Assessing Dermal Exposure II. Fluorescent Tracer Testing." Am Ind
I Hyg Assoc. 47:771-775.
I.J Page 6
-------
SVAW.VL
FIFRA SAP. 2001 a. "End Point Selection and Determination of Relative Potency in
Cumulative Hazard Assessment: A Pilot Study of Organophosphorus Pesticide
Chemicals." Report from the FIFRA Scientific Advisory Panel Meeting of September
27, 2000. FIFRA Scientific Advisory Panel, Office of Science Coordination and
Policy, Office of Prevention, Pesticides and Toxic Substances, U.S. Environmental
Protection Agency. Washington, DC. SAP Report 2000-OX. Available:
http://www.epa.qov/scipoiv/sap/2000/September/
FIFRA SAP. 2001 b. "Case Study of the Cumulative Risk of 24 Organophosphate
Pesticides; Cumulative Risk Assessment Method for Dietary Food Exposure;
Cumulative Risk Assessment for Residential Exposure; Cumulative Risk
Assessment for Drinking Water; Integrated Cumulative Risk Assessment." Report
from Session II of the FIFRA Scientific Advisory Panel Meeting of December 7-8,
2000. FIFRA Scientific Advisory Panel, Office of Science Coordination and Policy,
Office of Prevention, Pesticides and Toxic Substances, U.S. Environmental
Protection Agency. Washington, DC. SAP Report 2001-06. Available:
http://www.epa.aov/scipolv/sap/2001/December/
FIFRA SAP. 2001 c. "Preliminary Cumulative Hazard and Dose Response
Assessment for Organophosphorus Pesticides: Determination of Relative Potency
and Points of Departure for Cholinesterase Inhibition." Report from the FIFRA
Scientific Advisory Panel Meeting of September 5-6, 2001 (Report dated September
11, 2001). FIFRA Scientific Advisory Panel, Office of Science Coordination and
Policy, Office of Prevention, Pesticides and Toxic Substances, U.S. Environmental
Protection Agency. Washington, DC. Available:
http://www.epa.Qov/scipoly/sap/2000/September/
FIFRA SAP. 2000a. "Proposed Guidance for Conducting Cumulative Hazard
Assessments for Pesticides that have a Common Mechanism of Toxicity." Report
from Session II of the FIFRA Scientific Advisory Panel Meeting of September 23,
1999. FIFRA Scientific Advisory Panel, Office of Science Coordination and Policy,
Office of Prevention, Pesticides and Toxic Substances, U.S. Environmental
Protection Agency. Washington, DC. SAP Report 99-05D. Available:
http://www.epa.gov/scipoly/sap/2000/September/
FIFRA SAP. 2000b. "Cumulative Risk Assessment Methodology Issues of
Pesticide Substances that Have a Common Mechanism of Toxicity." Report from
Session II of the FIFRA Scientific Advisory Panel Meeting of December 8-9, 1999
(Report dated February 4, 2000). FIFRA Scientific Advisory Panel, Office of
Science Coordination and Policy, Office of Prevention, Pesticides and Toxic
Substances, U.S. Environmental Protection Agency. Washington, DC. SAP Report
99-06B. Available: http://www.epa.gov/scipolv/sap/2000/December
I.J Page 7
-------
FIFRA SAP. 1999. "Overview of Issues Related to the Standard Operating
Procedures for Residential Exposure Assessment." Report from Session I of the
FIFRA Scientific Advisory Panel Meeting of September 21, 1999. FIFRA Scientific
Advisory Panel, Office of Science Coordination and Policy, Office of Prevention,
Pesticides and Toxic Substances, U.S. Environmental Protection Agency. •
Washington, DC. SAP Report 99-05. Available:
http://www.epa.qov/scipolv/sap/1999/September/
Florida Agricultural Statistics Service. 2000. "Citrus Chemical Usage." Florida
Agricultural Statistics Service.
Florida Agricultural Statistics Service. 1999. "Vegetable Chemical Use." Florida
Agricultural Statistics Service.
Fonnum F, Sterri SH, Aas P, Johnsen H. 1985. "Carboxylesterases, importance for
detoxification of organophosphorus anticholinesterases and trichothecenes."
Fundamental and Applied Toxicology. Dec;5(6 Pt 2):S29-38.
Frawley JP, Weir R, Tusing T, DuBois KP, and Calandra JC. 1963. "Toxicologic
Investigations on Delnav." Toxicol Appl Pharmacol. 5:605-624.
Frawley JP, Fuyat HN, Hagan EC, Blake JR, and Fitzhugh OG. 1957. "Marked
Potentiation in Mammalian Toxicity From Simultaneous Administration of Two
Anticholinesterase Compounds;" J Pharmacol Exp Therap. 121:96-106.
Fujii T, Kawashima K. 2001. An independent non-neuronal cholinergic system in
lymphocytes. Japanese Journal of Pharmacology. 85:11-15.
Fulcher, S.M. 2001. "Malathion - Effects on cholinesterase in the CD rat (adult and
juvenile) by oral gavage administration." Huntingdon Life Sciences, Ltd.,
Cambridgeshire, England. Laboratory study no. CHV067/012452, November 30,
2001, MRID 45566201, unpublished.
Furlong CE, Li W-F, Richter RJ, Shih DM, Lusis AJ, Alleva E, Costa LG. 2000.
"Genetic and temporal determinants of pesticide sensitivity: Role of paraoxonase
(PON1)." Neurotoxicology. 21(1-2):91-100.
Gagne J, Brodeur J. 1972. "Metabolic studies on the mechanisms of increased
susceptibility of weaning rats to parathion." Canadian Journal of Physiology and
Pharmacology. Sep;50(9):902-15.
Garrison JC, Wyttenbach CR. 1985. "Teratogenic effects of the organophosphate
insecticide dicrotophos (Bidrin): histological characterization of defects." The
Anatomical Record. Nov;213(3):464-72.
I.J Page 8
-------
:
Gearhart JM, Jepson GW, Clewell HJ, et al. 1994. "Physiologically Based
I Pharmacokinetic Model for the Inhibition of Acetylcholinesterase by
Organophosphate Esters." Environ Health Perspect. 102 (Suppl 11),
51-60.
Geno PW, Camann DE, Harding, HJ, Villalobos K, Lewis RG. 1995. "Handwipe
Sampling and Analysis Procedure for the Measurement of Dermal Contact with
Pesticides." Arch Environ Contam Toxicol. 30:132-138.
Ginsberg G, Hattis D, Sonawane B, Russ A, Banati P, Kozlak M, Smolenski S,
Goble R. 2002. "Evaluation of child/adult pharmacokinetic differences from a
database derived from the therapeutic drug literature." Toxicological Sciences.
Apr;66(2): 185-200.
Glogoza P, McMullen M, Zollinger R, Peel M, and Fisher N. 2000. "Crop Profile for
Hard Red Spring and Durum Wheats in North Dakota." USDA Office of Pest
Management Policy and Pesticide Impact Assessment Program. June, 2000.
Gold RE, and Holcslaw T. 1985. "Dermal and Respiratory Exposure of Applicators
and Residents to Dichlorvos-Treated Residences. 0097-6156/85/0273-0353.
Gouker, e. 1999. "Determination of Transferable and Total Turf Residues on Turf
Treated with Bensulide. Lab. Pro). No. 98703. 44679. Unpublished study prepared
by ABC Laboratories, Inc. 265 p. MRID 447990-01.
Gray M, and Steffey K. 2000. "Insect Pest Management for Field and Forage
Crops." 2000 Illinois Agricultural Pest Management Handbook. Cooperative
Extension Service, Agricultural Experiment Station and College of Agricultural,
Consumer and Environmental Sciences, University of Illinois.
Greenspan RJ, Finn JA Jr, Hall JC. 1980. "Acetylcholinesterase mutants in
Drosophila and their effects on the structure and function of the central nervous
system." The Journal of Comparative Neurology. Feb 15; 189(4):741 -74.
Grisaru D, Sternfeld M, Eldor A, Click D, Soreq H. 1999. "Structural roles of
acetylcholinesterase variants in biology and pathology." European Journal of
Biochemistry. Sep;264(3):672-86.
Groot ME, Lekkerkert MC, and Steenbekkers LPA. 1998. "Mouthing Behaviour of
Young Children-An Observational Study." Agricultural University Wageningen,
Household and Consumer Studies, Wageningen, Netherlands.
Guzelian, PS, Henry CJ , Olin SS. 1992. Similarities and differences between
children and adults: Implications for risk assessment. Washington, D.C.: ILSI
Press. 285 pages.
I.J Page 9
-------
Haley RW, Billecke S, La Du BN. 1999. "Association of low PON1 type Q (type A)
arylesterase activity with neurologic symptom complexes in Gulf War veterans."
Toxicology and Applied Pharmacology. Jun 15;157(3):227-33.
Hanan D, Heer T, Kiyokawa B, Long L, Mielke E, Facteau T, Nelson T, and Olsen J.
1999. "Crop Profile for Cherries (Sweet) in Oregon." USDA Office of Pest
Management Policy and Pesticide Impact Assessment Program. September 7,
1999.
Harrington E, and Good G. 2000. "Crop Profile for Pears in New York." USDA
Office of Pest Management Policy and Pesticide Impact Assessment Program.
January 2000.
Hawkins W, Matthews C, Cockrell P, and Aerts M. 1999. "Crop Profile for
Tomatoes in Florida." USDA Office of Pest Management Policy and Pesticide
Impact Assessment Program. April 5,1999.
Heimlich, R. 2000. Farm Resource Regions. USDA Economic Research Service No.
760. Available: http://www.econ.ag.gov/whatsnew/issues/regions
Hertzberg, RC, Rice G, and Teuschler LK. 1999. "Methods for Health Risk
Assessment of Combustion Mixtures." Hazardous Waste Incineration: Evaluating
the Human Health and Environmental Risks. Roberts S, Team C, and Bean J,
Editors. CRC Press LC. pp. 105-148.
Heudorf U, Angerer J. 2001. "Metabolites of organophosphorous insecticides in
urine specimens from inhabitants of a residential area." Environmental Research.
May;86(1):80-7.
Hill RH Jr, Head SL, Baker S, Gregg M, Shealy DB, Bailey SL, Williams CC,
Sampson EJ, Needham LL. 1995. "Pesticide residues in urine of adults living in the
United States: reference range concentrations;" Environ Res. 71(2) 1995, pp.99-
108.
Hofen J. 2000. "Determination of Transferable Turf Residues on Turf Treated with
Trichlorfon." Lab. Proj. No. SARS-98-71. 109529. Unpublished study prepared by
Stewart Agricultural Research Services, Inc. 177 p. MRID 450672-01
Hogmire HW, and Biggs AR. 1999. "Crop Profile for Apples in West Virginia."
USDA Office of Pest Management Policy and Pesticide Impact Assessment
Program. February 1999.
Hubal EA, Sheldon LS, Zufall MJ, Burke JM, Thomas KW. 2000. "The challenge of
assessing children's residential exposure to pesticides." Journal of Exposure
Analysis and Environmental Epidemiology. Nov-Dec; 10(6 Pt 2):638-49.
I.J Page 10
-------
;vv i
I
Hunter DL, Lassiter TL, Padilla S. 1999. "Gestational exposure to chlorpyrifos:
comparative distribution of trichloropyridinol in the fetus and dam.
Toxicology and Applied Pharmacology. Jul 1;158(1):16-23.
lhaka R, and Gentleman R. 1996. "R: A language for data analysis and graphics."
J Comput Graph Stat. 5 (3): 299-314.
International Life Science Institute. 1998. ILSI Aggregate Exposure Subcommittee
Report. "Status Report on Biological Monitoring Research Relevant to Aggregate
Exposure Assessment under the Food Quality Protection Act." October 12, 1998.
Jenkins J, and Thomson P. 1998. "Pesticide Use in Oregon's Drainage Basins."
(!) I Agriculture Chemistry Extension, Department of Environmental and Molecular
fZ [ Toxicology, Oregon State University.
&«•» -
CA i
l/l I Johnson D, Thompson R, Butterfieid B. 1999. "Outdoor Residential Pesticide Use
0 i and Usage survey and National Gardening Association Survey." Unpublished study
0) = prepared by DOANE Marketing Research, Inc., and Gallup Organization, Inc. 761 p.
C/} I MRID 449722-02
Johnson DE, Seidler FJ, Slotkin TA. 1998. "Early biochemical detection of delayed
neurotoxicity resulting from developmental exposure to chloropyrifos." Brain
Research Bulletin. 45(2): 143-7.
Johnson G, Moore SW. 1999. The adhesion function on acetylcholinesterase is
located at the peripheral anionic site. Biochemical and Biophysical Research
Communications. May 19;258(3):758-62.
Jordan DL, Spears JF, York AC, Brandenburg RL, Brown AB, Bailey JE, and
Roberson GT. 2001. "2001 Peanut Information." North Carolina Cooperative
Extension Service. AG-331.
Karanth S, Olivier K, Liu J, and Pope C. 2001. "In vivo interaction between
chlorpyrifos and parathion in adult rats: sequence of administration can markedly
influence toxic outcome." In press: Toxicol Appl Pharmacol.
Karanth S, Pope C. 2000. "Carboxylesterase and A-esterase activities during
maturation and aging: relationship to the toxicity of chlorpyrifos and parathion in
rats."
Toxicological Sciences. Dec;58(2):282-9.
Karczmar AG, Srinivasan R, J Bernsohn. 1973. "Cholinergic function in the
developing fetus." In Fetal Pharmacology (L. Boreus, Ed), Raven Press, NY.
I.J Page 11
-------
I Kellogg RL, Nehring R, Grube A, Plotkin S, Goss DW, and Wallace S. 1999.
I "Trends in the Potential for Environmental Risk from Pesticide Loss from Farm
I Fields." USDA Natural Resources Conservation Service. Available:
| http://www.nhq.nrcs.usda.gov/land/pubs/pesttrend.html
j Kellogg RL, Goss DW, et al. 1997 (revised maps, 1998). "Potential Priority
I Watersheds for Protection of Water Quality from Nonpoint Sources Related to
i Agriculture.: USDA Natural Resources Conservation Service. State of the Land
I website, http://www.nhq.nrcs.usda.gov/land/pubs/wqpost2.html
| Keplinger ML and Deichmann WB. 1967. "Acute Toxicity of Combinations of
[ Pesticides." Toxicol Appl Pharmacol. 10: 586-595.
S
I Kirkpatrick CJ, Bittinger F, Unger RE, Kriegsmann J, Kilbinger H and Wessler I.
| 2001
i "The non-neuronal cholinergic system in the endothelium: Evidence and possible
[ pathobiological significance." Japanese Journal of Pharmacology. 85:24-28.
I KisselJC, Shirai JH, Richter KY, Fenske RA. 1998. "Empirical Investigation of
I Hand-to-Mouth Transfer of Soil." Bull Environ Contam Toxicol. 60:379-386.
=
| Kline and Co. Professional Markets for Pesticides and Fertilizers, 1998 and 1999.
I Kline and Co. Professional Market Data. 1997-1998.
| Kline and Co. Consusmer Markets for Pesticides and Fertilizers, 1995 and 1997.
| Klonne D. 1999. "Integrated Report for Evaluation of Potential Exposures to
I Homeowners and Professional Lawn Care Operators Mixing, Loading, and Applying
I Granular and Liquid Pesticides to Residential Lawns." Lab. Proj. no. OMA550,
I OMA001, OMA002. Unpublished study prepared by Ricerca, Inc., and Morse
I Laboratories. 2213 p. MRID 449722-01.
E
1 Korpalski SJ, and Bruce ED. "2000. Agronomic and Statistical Clustering of
I Agricultural Reentry Transfer Coefficients." Agricultural Reentry Task Force (ARTF).
MRID 448026-01.
Krieger Rl, Bernard CE, Dinoff TM, Ross JH, Williams RL. 2001. "Biomonitoring of
persons exposed to insecticides used in residences." The Annals of Occupational
Hygiene. Apr;45 Suppl 1:S143-53.
Kutz FW, Cook BT, Carter-Pokras OD, Brody D, Murphy RS. 1992. "Selected
pesticide residues and metabolites in urine from a survey of the U.S. general
population." Journal of Toxicology and Environmental Health. Oct;37(2):277-91.
I.J Page 12
-------
:
s
Lai.J. 1999. "Determination of Transferable Turf Residues on Grass Treated with
Acephate." Lab. Proj. No. V11983. 9900130. Unpublished study prepared by
Valent U.S.A. Corporation and Weed Systems, Inc. 267 p. MRID 448064-01.
Lassiter TL, White LD, Padilla S, Barone S, Jr. 2002. "Gestational exposure to
chlorpyrifos: Qualitative and quantitvtive neuropathological changes in the fetal
neocortex." Presented at the 41st Annual meeting of the Society of Toxicology,
March 2002.
Lassiter TL, Padilla S, Mortensen SR, Chanda SM, Moser VC, Barone S Jr. 1998.
"Gestational exposure to chlorpyrifos: apparent protection of the fetus"?
Toxicology and Applied Pharmacology. Sep;152(1 ):56-65.
Lauder JM, Schambra UB. 1999. "Morphogenetic roles of acetylcholine."
Environmental Health Perspectives. Feb;107 Suppl 1:65-9.
Lavigne A, Matthews C, Cockrell P, and Aerts M. 1999. "Crop Profile for Citrus in
Florida." USDA Office of Pest Management Policy and Pesticide Impact
Assessment Program. February 26, 1999.
Layer PG, Willbold E. 1995. "Novel functions of cholinesterases in development,
physiology and disease." Progress in Histochemistry and Cytochemistry.
29(3): 1-94.
Layer PG, Weikert T, Alber R. 1993. "Cholinesterases regulate neurite growth of
chick nerve cells in vitro by means of a non-enzymatic mechanism." Cell and Tissue
Research. Aug;273(2):219-26.
Lessard S, Preston D, Glogoza P, Olson D, Lamey A, Gudmestad N, Secor G and
Zollinger R. 2000. "Crop Profile for Potatoes in North Dakota." USDA Office of
Pest Management Policy and Pesticide Impact Assessment Program. December,
2000.
Levin ED, Addy N, Nakajima A, Christopher NC, Seidler FJ, Slotkin TA. 2001.
"Persistent behavioral consequences of neonatal chlorpyrifos exposure in rats."
Brain Research. Developmental Brain Research. Sep 23;130(1):83-9.
Li WF, Matthews C, Disteche CM, Costa LG, Furlong CE. 1997. "Paraoxonase
| (PON 1) gene in mice: sequencing, chromosomal localization and developmental
| expression." Pharmacogenetics. Apr;7(2):137-44.
I Li WF, Costa LG, Furlong CE. 1993. "Serum paraoxonase status: a major factor in
I determining resistance to organophosphates." Journal of Toxicology and
| Environmental Health. Oct-Nov;40(2-3):337-46.
I
I Lindstrom MJ, and Bates DM. 1990. "Nonlinear mixed effects models for repeated
| measures data." Biometrics 46: 673-687.
I
| I.J Page 13
-------
Lioy, PJ, Edwards RD, Freeman N, Gurunathan S, Pellizzari E, Adgate JL,
Quackenboss J, Sexton K. 2000. "House dust levels of selected insecticides and a
herbicide measured by the EL and LWW samplers and comparisons to hand rinses
and urine metabolites." Journal of Exposure Analysis and Environmental
Epidemiology. Jul-Aug;10(4):327-40.
"*'"" Liu J, Olivier K, Pope CN. 1999. "Comparative neurochemical effects of repeated
methyl parathion or chlorpyrifos exposures in neonatal and adult rats." Toxicology
and Applied Pharmacology. Jul 15; 158(2): 186-96.
Lucas RM, Boyle KE, Dever JA, George BJ, and Jeffries CJ. 1995. Final Report.
"Volume 1 Results of the 1993 Certified/Commercial Pesticide Applicator Survey."
Research Triangle Institute. August 22,1995.
Ma T, Chambers JE. 1994. "Kinetic parameters of desulfuration and dearylation of
parathion and chlorpyrifos by rat liver microsomes." Food and Chemical Toxicology:
An International Journal Published for the British Industrial Biological Research
Association. Aug;32(8):763-7.
Machin MG, McBride WG. 1989. "Teratological study of malathion in the rabbit."
Journal of Toxicology and Environmental Health. 26(3):249-53.
Macintosh DL, Needham LL, Hammerstrom KA, Ryan PB. 1999. "A longitudinal
investigation of selected pesticide metabolites in urine." Journal of Exposure
Analysis and Environmental Epidemiology. Sep-Oct;9(5):494-501.
Mackness B, Durrington PN, Mackness Ml. 1998. "Human serum paraoxonase."
> General Pharmacology. Sep;31(3):329-36.
Magara Y, Aizawa T, Matumoto N, and Souna F. 1994. "Degradation of Pesticides
by Chlorination During Water Purification. Groundwater Contamination,
Environmental Restoration, and Diffuse Source Pollution." Water Sci Tech.
30(7):1 19-128.
Mahajna M, Quistad GB, and Casida JE. 1997. "Acephate insecticide toxicity:
safety conferred by inhibition of the bioactivating carboxyamidase by the metabolite
methamidophos." Chem Res Toxicol. 10:64-69.
Main AR. 1956. The role of A-esterases in the acute toxicity of paraoxon, TEPP
and parathion. Canadian Journal Biochem. 34:197-216.
Matthews C, Cockrell P, and Aerts M. 1999. "Crop Profile for Peppers (Bell) in
Florida." USDA Office of Pest Management Policy and Pesticide Impact
Assessment Program. April 5, 1999.
I.J Page 14
-------
[ Mattsson JL, Maurissen JP, Nolan RJ, Brzak KA. 2000. "Lack of differential
I sensitivity to cholinesterase inhibition in fetuses and neonates compared to dams
I treated perinatally with chlorpyrifos." Toxicological Sciences. 2000
| Feb;53(2):438-46.
| Mattsson, JL et a/. 1998. "Effects of chlorpyrifos administered via gavage to CD
| rats during gestation and lactation on plasma, erythrocyte, heart, and brain
I cholinesterase, and analytical determination of chlorpyrifos and metabolites."
I Health and Environmental Research Laboratories, The Dow Chemical Company,
I Midland, Ml, Laboratory Study No. 971162. August 31, 1998. 322 p. MRID
I 44648102.
s
=
Maurissen JP, Hoberman AM, Garman RH, Hanley TR Jr. 2000. "Lack of selective
developmental neurotoxicity in rat pups from dams treated by gavage with
chlorpyrifos." Toxicological sciences. 2000 Oct;57(2):250-63.
Maxwell DM. 1992a. "The specificity of carboxylesterase protection against the
toxicity of organophosphorus compounds." Toxicology and Applied Pharmacology.
Jun;114(2):306-12.
I Maxwell DM. 1992b. Detoxication of organophosphorus compounds by
| carboxylesterases. In Organophosphates Chemistry, Fate and Effects (J.E.
Chambers and P.E. Levi, eds) pp. 183-199. Academic Press, New York.
| McGrath D, Antonelli A, and Bechinski E. 2001. "Pacific Northwest Insect
I Management Handbook." Oregon State University.
McGrath D, Burt J, Ocamb CM, and William R. 2001. "Crop Profile for Cauliflower
in Oregon." USDA Office of Pest Management Policy and Pesticide Impact
Assessment Program. February, 2001. Available:
http://cipm.ncsu.edu/cropprofiies/docs/ORCauliflower.html
Merricks D. 1997. "Carbaryl Mixer/Loader/Applicator Exposure Study During
Application of RP-2 Liquid (21%), Sevin Ready to Use Insect Spray or Sevin 10 Dust
to Home Garden Vegetables." Lab. Proj. No. 1519. 10564. ML97-0676-RHP.
Unpublished study prepared by Agrisearch Inc., Rhone-Poulenc Ag Co. and Morse
Labs., Inc. 358 p. MRID 445980-01.
Merricks L. 2001. "Determination of Dermal (Hand and Forearm) and Inhalation
Exposure to Disulfoton Resulting from Residential Application of Bayer Advanced
Garden 2-in-1 Systematic Rose and Flower Care to Shrubs and Flower Beds." Lab
Prj. No. 4201. Unpublished study prepared by Agrisearch Inc. 178 p. MRID
453334-01.
Meyers, D. 2001. "Dimethoate effects on cholinesterase in the CD rat (adult and
juvenile) by oral gavage administration." Huntingdon Life Sciences, Ltd., Suffolk,
England, Lab Project Number: CHV/070: 012226, MRID 45529702, unpublished.
I.J Page 15
-------
Meyer EM, St Onge E, Crews FT. 1984. "Effects of aging on rat cortical
presynaptic cholinergic processes." Neurobiology of Aging 5(4):315-7.
Michalek H, Pintor A, Fortuna S, Bisso GM. 1985. "Effects of
diisopropylfluorophosphate on brain cholinergic systems of rats at early
developmental stages." Fundamental and Applied Toxicology. Dec;5(6 Pt
2):S204-12.
Mileson BE, Chambers JE, Chen WL, et al. 1998. "Common Mechanism of
Toxicity: A Case Study of Organophosphorus Pesticides." Toxicol Sci. 41:8-20.
Moran JW, Saunders, DG, Carter, JE, and Emery KAB. 1987. "Exposure of
Workers and Golfers to Flurprimidol from use of Cutlass SOW on Golf Course Turf."
Lab Proj. I.D. AAC8606. Unpublished study prepared by Lilly Research
Laboratories. 211 p. MRID 401844-14.
Morgan EW, Van B, Greenway D, Parkinson A. 1994. "Regulation of two rat liver
microsomal carboxylesterase isozymes: species differences, tissue distribution, and
the effects of age, sex, and xenobiotic treatment of rats." Archives of Biochemistry
and Biophysics. Dec;315(2):513-26.
Morrison WP, Cronholm GB, Parker RD, Baugh B, Patrick CD, and Archer TL.
1995. "Managing Insect and Mite Pests of Texas Corn." Texas Agricultural
Extension Service, The Texas A&M University System. B-1366.
Mortensen SR, Hooper MJ, Padilla S. 1998. "Rat brain acetylcholinesterase
activity: developmental profile and maturational sensitivity to carbamate and
organophosphorus inhibitors." Toxicology. Jan 16;125(1):13-9.
X •«•»» Z
CO I Mortensen SR, Chanda SM, Hooper MJ, Padilla S. 1996. "Maturational differences
in chlorpyrifos-oxonase activity may contribute to age-related sensitivity to
p* I- chlorpyrifos."
£- I Journal of Biochemical Toxicology. 11(6):279-87.
I Moser VC. 1999. "Comparison of aldicarb and methamidophos neurotoxicity at
I different ages in the rat: behavioral and biochemical parameters." Toxicology and
I Applied Pharmacology. Jun 1 ;157(2):94-106.
| Moser VC, Padilla S. 1998. Age- and gender-related differences in the time course
| of behavioral and biochemical effects produced by oral chlorpyrifos in rats.
I Toxicology and Applied Pharmacology 149:107-119.
I
^} I Moser VC, Chanda SM, Mortensen SR, Padilla S. 1998. "Age- and gender-related
°:^; i differences in sensitivity to chlorpyrifos in the rat reflect developmental profiles of
$% I esterase activities." Toxicological Sciences. Dec;46(2):211-22.
I.J Page 16
-------
Mueller RF, Hornung S, Furlong CE, Anderson J, Giblett ER, Motulsky AG. 1983.
"Plasma paraoxonase polymorphism: a new enzyme assay, population, family,
biochemical, and linkage studies." American Journal of Human Genetics.
May;35(3):393-408.
Murphy RS, Kutz FW, Strassman SC. 1983. "Selected pesticide residues or
metabolites in blood and urine specimens from a general population survey."
Environmental Health Perspectives. Feb;48:81-6.
f National Center for Food and Agricultural Policy (NCFAP). 2000. "Pesticide Use in
1 U.S. Crop Production: 1997 National Summary Report." November 2000. Available:
C3 I http://www.ncfap.org/commissi.htm
E
s
I National Research Council. 1993. Pesticides in the Diets of Infants and Children,
| Committee on Pesticides in the Diets of Infants and Children, Board on Agriculture
I and Board on Environmental Studies and Toxicology, Commission on Life Sciences,
I National Research Council, National Academy Press, Washington, DC. Available:
I http://www.nap.edu/
: '
1 Nigg HN, and Knaak JB. 2000. "Blood Cholinesterase as human biomarkers of
I organophosphorus pesticide exposures." Rev Environ Contam Toxicol. 163:29-
I 112.
I North Carolina State University College of Agriculture and Life Sciences. 2001.
I "2001 North Carolina Agricultural Chemicals Manual."
| North Dakota State University Extension Service. 1997. "Corn Production Guide."
| North Dakota State University Extension Service in cooperation with the North
| Dakota Corn Utilization Council: The North Dakota Corn Growers Association. May
I 1997. A-1130.
| Ol-Sebae AH, Ahmed NS, and Soliman SA. 1978. "Effect of pre-exposure on
I acute toxicity of organophosphorus insecticides to white mice." J Environ Sci Health
f B13(1): 11-24.
I
I Oregon State University Cooperative Extension Service. 2001. "Cherry. 2001 Pest
| Management Guide for the Willamette Valley." Revised February 2001. EM 8329.
f Oregon State University Cooperative Extension Service. 2001. "Pear. 2001 Pest
j Management Guide for the Willamette Valley." February 2001. EM 8420.
| Oregon State University Cooperative Extension Service. 2001. "Hazelnut. 2001
1 Pest Management Guide for the Willamette Valley." Revised February, 2001. EM
I 8328.
I.J Page 17
-------
I Oregon State University Cooperative Extension Service. 2001. "Hazelnut. 2001
I Pest Management Guide for the Willamette Valley." Revised February 2001. EM
I 8328.
:
| Oregon State. University Cooperative Extension Service, Willamette Research and
I Extension Center. 2001. "Commercial Vegetable Production Guides. Peas-
1 Western Oregon. Pisum sativum." April 4, 2001.
| Oregon State University Cooperative Extension Service, Willamette Research and
t I Extension Center. 2001. "Commercial Vegetable Production Guides. Peas for
*&«* I Processing - Eastern Oregon. Pisum sativum" April 4, 2001.
•!%V>.ytf 2
<|} | Oregon State University Cooperative Extension Service, Willamette Research and
|Z | Extension Center. 2001. "Commercial Vegetable Production Guides. Dry Bulb
^ | Onions - Western Oregon. Allium cepa". April 4, 2001.
\£ v1 I:
(|> I Oregon State University Cooperative Extension Service, Willamette Research and
C.I) I Extension Center. 2001. "Commercial Vegetable Production Guides. Cole Crop
I Insect Control." April 3, 2001.
I
| Oregon State University Cooperative Extension Service, Willamette Research and
| Extension Center. 2001. "Commercial Vegetable Production Guides. Cauliflower.
| Brassica oleracea (Botrytis Group)." April 3, 2001.
| Oregon State University Cooperative Extension Service, Willamette Research and
| Extension Center. 2001. "Commercial Vegetable Production Guides. Sweet Corn
| for Processing. Zea mays." Aprils, 2001.
Oregon State University Cooperative Extension Service, Willamette Research and
Extension Center. 2001. "Commercial Vegetable Production Guides. Snap
I Beans- Green, Romano, Yellow Wax. Phaseolus vulgaris." April 3, 2001.
Oregon State University Cooperative Extension Service, Willamette Research and
Extension Center. 2001. "Commercial Vegetable Production Guides. Broccoli.
Brassica oleracea (Italica Group)." April 3, 2001.
Oregon State University Cooperative Extension Service, Willamette Research and
Extension Center. 2001. "Commercial Vegetable Production Guides. Cabbage.
Brassica oleracea (Capitata Group)." April 3, 2001.
Oregon State University Cooperative Extension Service, Willamette Research and
Extension Center. 2001. "Commercial Vegetable Production Guides. Pumpkin and
Winter Squash." August 13, 2001.
Oregon State University Cooperative Extension Service, Willamette Research and
Extension Center. 2000. "Commercial Vegetable Production Guides. Slicing
(Fresh Market) Cucumbers. Cucumis sativus." Revised March 29, 2000.
I.J Page 18
-------
I Oregon State University Cooperative Extension Service, Willamette Research and
I Extension Center. 2000. "Commercial Vegetable Production Guides. Zucchini and
I Summer Squash. Cucurbita pepo" Revised March 29, 2000.
I
i Oregon State University Cooperative Extension Service, Willamette Research and
i Extension Center. 2000. "Commercial Vegetable Production Guides. Pickling
I Cucumbers. Cucumis sativus." Revised September 18, 2000.
| Oregon State University Extension Service "Oregon Agricultural Information
| Network." Online. Available: http://ludwig.arec.orst.edu/Econlnfo/
C ! Orzolek MD, Greaser GL, and Harper JK. 2001. "Cantaloupes. Penn State
(0 | Cooperative Extension Agricultural Alternatives." The Pennsylvania State
fZ | University.
Orzolek MD, Fleischer SJ, and MacNab, AA. 1998. "Crop Profile for Pumpkins in
Pennsylvania." USDA Office of Pest Management Policy and Pesticide Impact
Assessment Program. Prepared September 14, 1998.
Padilla S, Sung H-J, Jackson L, Moser V. 2002. "Development of an in vitro assay
which may identify which organophosphorus pesticides are more toxic to the young."
Presented at the Society of Toxicology meeting, March 2002.
|
I Padilla S, Buzzard J, Moser VC. 2000. "Comparison of the role of esterases in the
i differential age-related sensitivity to chlorpyrifos and methamidophos."
| Neurotoxicology. Feb-Apr;21(1-2):49-56.
I Pedata F, Slavikova J, Kotas A, Pepeu G. 1983. "Acetylcholine release from rat
= cortical slices during postnatal development and aging." Neurobiology of Aging.
I Spring;4(1):31-5.
sss%w S
E~" | Peel MD, and Riveland N. 1997. "Winter Wheat Production in North Dakota."
1 North Dakota State University Extension Service. Revised September, 1997.
Extension Bulletin 33.
Pike D, Steffey K, and Babadoost M. 2000. "Crop Profile for Corn in Illinois."
USDA Office of Pest Management Policy and Pesticide Impact Assessment
Program. Prepared July, 2000.
Pinheiro J, and Bates DM. 2000. Mixed Effects Models in S and S-Plus. Springer.
Berlin.
Pope C . 2001 a. "Age and Interactive Toxicity of Organophosphorus Insecticides."
U.S. Environmental Protection Agency. NCERQA Grant Project Number R 825811,
February 28, 2002.
I.J Page 19
-------
Pope C. 2001 b. "The influence of age on pesticide toxicity." In Handbook of
Pesticide Toxicology (ed. R. I. Krieger) Volume 1, Principles Chapter 41, Academic
Press, pages 873-885.
Pope C, Liu J. 2001. "Nonesterase Actions of Anticholinesterase Insecticides" in
Handbook of Neurotoxicology. Volume 1, Chapter 3, edited by E.J. Massaro
(Totowa, NJ: Humana Press Inc), pages 29-43.
Pope CN, Liu J. 1997. "Age-related differences in sensitivity to organophosphorus
pesticides." Environmental Toxicology and Pharmacology. 4:309-314.
Pope CN, Chakraborti TK, Chapman ML, Farrar JD, Arthun D. 1991. "Comparison
of in vivo cholinesterase inhibition in neonatal and adult rats by three
organophosphorothioate insecticides." Toxicology. 68(1):51-61.
Pope CN and Padilla S. 1990. "Potentiation of organophosphorus-induced delayed
neurotoxicity by phenylmethylsulfonyl fluoride." J Toxicol Environ Health. 31: 261-
273.
Qiao D, Seidler FJ, Padilla S, Slotkin TA. 2002. "Developmental neurotoxicity of
chlorpyrifos: What is the vulnerable period"? Journal of Toxicology and
Environmental Health, in press.
Quackenboss JJ, Pellizzari ED, Shubat P, Whitmore RW, Adgate JL, Thomas KW,
Freeman NC, Stroebel C, Lioy PJ, Clayton AC, Sexton K. 2000. "Design strategy
for assessing multi-pathway exposure for children: the Minnesota Children's
Pesticide Exposure Study (MNCPES)." Journal of Exposure Analysis and
Environmental Epidemiology.. Mar-Apr; 10(2): 145-58.
Raun E, Martin A, Mayo ZB, and Watkins J. 2000. "Crop Profile for Sorghum in
Nebraska;" USDA Office of Pest Management Policy and Pesticide Impact
Assessment Program. Prepared June, 2000.
Reiss R, and Griffin J. 2001 "Analysis of the National Pest Management Assoc.
Pest Control Operators (PCO) Product Use and Usage Information Survey."
Completion Date May 16, 2001
Richardson JR, Chambers HW, and Chambers JE. 2001. "Analysis of the additivity
of in vitro inhibition of cholinesterase by mixtures of chlorpyrifos-oxon and azinphos-
methyl-oxon." Toxicol Appl Pharm. 172: 128-139.
Riedl H, and Van Buskirk P. 1999. "Crop Profile for Pears in Oregon." USDA
Office of Pest Management Policy and Pesticide Impact Assessment Program.
Revised October 26, 1999.
I.J Page 20
-------
Rinehold J, and Jenkins JJ. 1994. "Pesticide Use Survey. Oregon Pesticide Use
Estimates for Seed and Specialty Crops, 1992." Oregon State University
Publication No. EM 8658.
Rinehold JW, Jenkins JJ, and Lundy R. 1999. "Pesticide Use in Oregon
Peppermint and Spearmint [DRAFT]." Prepared for the Mint Industry Research
Council, Stevenson, WA.
Rinehold JW. 1999. "Crop Profile for Christmas Trees in Oregon and Washington."
USDA Office of Pest Management Policy and Pesticide Impact Assessment
Program. Revised January, 1999.
Rinehold JW, and Jenkins JJ. 1994. "Oregon Pesticide Use Estimates for Seed
and Specialty Crops, 1992." Oregon State University Cooperative Extension
Service. EM 8568.
Robert M, and Wade S. (May 5, 1998). Carbaryl Mixer/Loader/Applicator Exposure
Study during Application of RP-2 Liquid (21%), Sevin® Ready to Use Insect Spray
or Sevin® 10 Dust to Home Garden Vegetables. Check PDMS
Rosenberg P, and Coon JM. 1958. "Potentiation Between Cholinesterase
Inhibitors." Proc Soc Exp Biol Med. 97: 836-839.
Scheidt AB, Long GG, Knox K, Hubbard SE. 1987. Toxicosis in newborn pigs
associated with cutaneous application of an aerosol spray containing chlorpyrifos.
Journal of the American Veterinary Medical Association. Dec 1;191(11):1410-2.
Serat WF, and Bailey JB. 1974. "Estimating the relative toxicologic potential of
each pesticide in a mixture of residues on foliage." Bull Environ Contam Toxicol.
12(6): 682-686.
Seume FW, and O'Brien RD. 1960. "Potentiation of the Toxicity to Insects and
Mice of Phosphorothionates Containing Carboxyester and Carboxyamide Groups."
Toxicol Appl Pharmacol. 2: 495-503.
Sheets LP. 2000. "A consideration of age-dependent differences in susceptibility to
organophosphorus and pyrethroid insecticides." Neurotoxicology. 21(1-2):57-63.
Singh AK. 1986. "Kinetic analysis of acetylcholinesterase inhibition by
combinations of acephate and methamidophos." Toxicology. 42(2-3): 143-56.
Slotkin TA, Tate CA,'Cousins MM, Seidler FJ. 2002. "Functional alterations in CMS
catecholamine systems in adolescence and adulthood after neonatal chlorpyrifos
| exposure. Brain Research. Developmental Brain Research. 133:163-173.
=
I.JPage21
-------
Slotkin TA, Cousins MM, Tate CA, Seidler FJ. 2001 a. "Persistent cholinergic
presynaptic deficits after neonatal chlorpyrifos exposure." Brain Research. Jun
1;902(2):229-43.
Slotkin TA, Tate CA, Cousins MM, Seidler FJ. 2001 b. "Functional alterations in
CNS catecholamine systems in adolescence and adulthood after neonatal
chlorpyrifos exposure." Brain Research. Developmental Brain Research. 133:163-
173.
Smith D, McCallum A, Bade D, Bean B, Grichar JC, Patrick C, and Stichler C. 2000.
"Crop Profile for Alfalfa in Texas." USDA Office of Pest Management Policy and
Pesticide Impact Assessment Program. Prepared May, 2000.
Smith D, and Moerbe T. 1999. "Crop Profile for Cotton in Texas." USDA Office of
Pest Management Policy and Pesticide Impact Assessment Program. Prepared
Septembers, 1999.
Smith WD, Peedin GF, Fisher LR, Southern PS, Melton TA, Brown AB, Moore CL,
Sr, Boyette MD, and Moore JM. 2001 . "2001 Flue-Cured Tobacco Information."
North Carolina Cooperative Extension Service.
Sommer JE, and Mines FK. 1991. "Diversity in U.S. Agriculture: A New Delination by
Farming Characteristics." USDA. Economic Research Service. Agricultural
Economic Report No. 646.
Southern PS, Fisher L, Melton T, Peedin G, and Smith WD. 1999. "Crop Profile for
Tobacco in North Carolina. USDA Office of Pest Management Policy and Pesticide
Impact Assessment Program. Updated January, 1999.
Spyker JM, Avery DL. 1977. "Neurobehavioral effects of prenatal exposure to the
organophosphate Diazinon in mice." Journal of Toxicology and Environmental
Health. Dec;3(5-6):989-1002.
Sternfeld M, Ming G, Song H, Sela K, Timberg R, Poo M, Soreq H. 1998.
"Acetylcholinesterase enhances neurite growth and synapse development through
alternative contributions of its hydrolytic capacity, core protein, and variable C
termini."
The Journal of Neuroscience. Feb 1 5 ; 1 8(4 ): 1 240-9 .
Su MQ, Kinoshita FK, Frawley JP, and DuBois KP. 1971. "Comparative inhibition
of aliesterases and cholinesterases in rats fed eighteen organophosphorus
insecticides." Toxicol Appl Pharmacol. 20(2): 241-249.
Sutton TB, Walgenbach J, Mitchem W, Unrath CR, Sullivan WT, and Parker M.
1999. "Crop Profile for Apples in North Carolina." USDA Office of Pest
Management Policy and Pesticide Impact Assessment Program. Prepared January,
1999.
IJ Page 22
-------
I Tang J, Carr RL, Chambers JE. 1999. "Changes in rat brain cholinesterase activity
I and muscarinic receptor density during and after repeated oral exposure to
I chlorpyrifos in early postnatal development." Toxicological Sciences.
I Oct;51(2):265-72.
Thomson P, Parrott W, and Jenkins J. 2001. "Crop Profile for Peas in Oregon."
USDA Office of Pest Management Policy and Pesticide Impact Assessment
Program. Prepared February, 2001.
Thomson P, Parrott W, and Jenkins J. 2000. "Crop Profile for Corn in Oregon."
USDA Office of Pest Management Policy and Pesticide Impact Assessment
Program. Prepared October, 2000.
Thomson P, Parrott W, and Jenkins J. 2000. "Crop Profile for Cucumbers in
Oregon." USDA Office of Pest Management Policy and Pesticide Impact
Assessment Program. Prepared November, 2000.
Thomson P, Parrott W, and Jenkins J. 2000. "Crop Profile for Broccoli in Oregon."
USDA Office of Pest Management Policy and Pesticide Impact Assessment
Program. Prepared November, 2000.
Thomson P, Parrott W, and Jenkins J. 1999. "Crop Profile for Beans in Oregon."
USDA Office of Pest Management Policy and Pesticide Impact Assessment
Program. Revised September 2, 1999.
Thomson P, Parrott W, and Jenkins J. 1999. "Crop Profile for Hazelnuts in
Oregon." USDA Office of Pest Management Policy and Pesticide Impact
Assessment Program. Revised September 2, 1999.
Thomson P, Parrott W, and Jenkins J. 1999. "Crop Profile for Raspberries in
Oregon." USDA Office of Pest Management Policy and Pesticide Impact
Assessment Program. Revised September 7, 1999.
Thomson P, Parrott W, and Jenkins J. 1999. "Crop Profile for Onions in Oregon."
USDA Office of Pest Management Policy and Pesticide Impact Assessment
Program. Prepared November 9, 1999.
Thomson P, Parrott W, and Jenkins J. 1999. "Crop Profile for Hops in Oregon."
USDA Office of Pest Management Policy and Pesticide Impact Assessment
Program. Revised November 23, 1999.
Thompson R. 1998. "Agricultural Worker Crop Contact from Reentry Activities
Performed in the United States and Canada: Growers Results. Unpublished study
prepared by DOANE Marketing Research, Inc. 7147 p. MRID 448026-01.
I.J Page 23
-------
Tierney DP, Christensen BR, and Culpepper VC. 2001 a. "Drinking water
monitoring study for six organophosphate insecticides and four oxons from 44
community water systems in the United States." Syngenta Crop Protection, Inc.
Study No. 1330-00.
Tierney DP, Christensen BR, and Culpepper VC. 2001 b. "Chlorine degradation of
six organophosphorus insecticides and four oxons in a drinking water matrix."
Syngenta Crop Protection, Inc. Study No. 1562-00.
Tieze NS, Hester PG, and Shaffer KR. 1994. "Mass Recovery of Malathion in
Simulated Open Field Mosquito Adulticide Tests." Arch Environ Contam Toxicol. 26:
473-477.
Timchalk C, Nolan RJ, Mendrala AL, Dittenber DA, Brzak KA, Mattsson JL. 2002.
"A Physiologically Based Pharmacokinetic and Pharmacodynamic (PBPK/PD) Model
for the Organophosphate Insecticide Chlorpyrifos in Rats and Humans."
Toxicological Sciences. Mar;66(1):34-53.
Travis JW. S. Tibbetts and N. Serotkin, editors. 2001. Pennsylvania Tree Fruit
Production Guide 2000-2001. College of Agricultural Sciences. The Pennsylvania
State University. Updated April 3, 2001. 2CDocuPS1/00;2M1/OOCP.
U.S. Department of Agriculture, Crop Reporting Board, Statistical Reporting Service.
1977. Usual Planting and Harvesting Dates for Fresh Market and Processing
Vegetables. Agriculture Handbook No. 507.
U.S. Department of Agriculture, National Agricultural Statistics Service. 1997.
Usual Planting and Harvesting Dates for U.S. Field Crops. Agricultural Handbook
No. 628.
| U.S. Department of Agriculture. Science and Technology Programs at AMS.
| Online. Available: http://www.ams.usda.gov/science/pdp
| U.S. Environmental Protection Agency. 2002a. "Determination of the Appropriate
| FQPA Safety Factor(s) for Use in the Tolerance-Setting Process;" February 28,
I 2002. Office of Pesticide Programs, Office of Prevention, Pesticides, and Toxic
1 Substances. Available: http://www.epa.gov/oppfead1/trac/science/tf10-fold
| U.S. Environmental Protection Agency. 2002b. Draft Document. "Consideration of
I the FQPA Safety Factor and Other Uncertainty Factors in Cumulative Risk
I Assessment of Chemicals Sharing a Common Mechanism of Toxicity;" February 28,
I 2002. Office of Pesticide Programs, Office of Prevention, Pesticides, and Toxic
I Substances. Washington, DC. Available:
I http://www.epa.gov/oppfead1/trac/science/tf10-fold
I.J Page 24
-------
U.S. Environmental Protection Agency. 2001 a. "Guidance on Cumulative Risk
Assessment of Pesticide Chemicals that Have a Common Mechanism of Toxicity."
Office of Pesticide Programs, Office of Prevention, Pesticides, and Toxic
Substances, U.S. Environmental Protection Agency. Washington, DC. Available:
http://www.epa.gov/oppfead1/trac/science/
U.S. Environmental Protection Agency. 2001 b. "Preliminary Cumulative Hazard
and Dose Response Assessment for Organophosphorus Pesticides: Determination
of Relative Potency and Points of Departure for Cholinesterase Inhibition." Office of
Pesticide Programs, US Environmental Protection Agency, Washington, DC. July
31,2001. http://www.epa.gov/scipolv/sap
U.S. Environmental Protection Agency. 2001c. "Supplementary Guidance for
Conducting Health Risk Assessment of Chemical Mixtures." Final Draft. Risk
Assessment Forum, Office of Research and Development, U.S. Environmental
Protection Agency. Washington, DC. NCEA-C-0148.
Available: www.epa.gov/ncea/new.htm
U.S. Environmental Protection Agency. 2001 d. "General Principles For Performing
Aggregate Exposure And Risk Assessments." Final. December 2, 2001 Office of
Pesticide Programs, Office of Prevention, Pesticides, and Toxic Substances, U.S.
Environmental Protection Agency. Washington, DC. Available:
http://www.epa.gov/oppfead1/trac/science/
U.S. Environmental Protection Agency. Super Sample PCO Study, Client
presentation, March 22, 2001.
U.S. Environmental Protection Agency. 2000a. "Proposed Guidance on Cumulative
Risk Assessment of Pesticide Chemicals that Have a Common Mechanism of
Toxicity." Public Comment Draft. Office of Pesticide Programs, Office of Prevention,
Pesticides, and Toxic Substances, U.S. Environmental Protection Agency.
Washington, DC. Available:
www.epa.gov/fedrgstr/EPA-PEST/2000/June/Dav-30/6049.pdf
U.S. Environmental Protection Agency. 2000c. "Cumulative Risk: A Case Study of
the Estimation of Risk from 24 Organophosphate Pesticides", November 2, 2000,
Office of Pesticide Programs, Office of Prevention, Pesticides, and Toxic
Substances, U.S. Environmental Protection Agency. Washington, DC. Available:
http://w.epa.gov/scipoly/sap/2000/december/sap-casestudv2.pdf
U.S. Environmental Protection Agency. 2000d. Use of Data on Cholinesterase
Inhibition for Risk Assessments of Organophosphorus and Carbamate Pesticides,
Office of Pesticide Programs, (issued in revised form in September 2000), Office of
Pesticide Programs, US Environmental Protection Agency, Washington DC.
Available: http://www.epa.gov/pesticides/trac/science/cholin.pdf
I.J Page 25
-------
SW.WIM
WWW
V.Vf?
U.S. Environmental Protection Agency. 2000e. "Drinking Water Screening Level
Assessment. Part B: Applying a Percent Crop Area Adjustment to Tier 2 Surface
Water Model Estimates for Pesticide Drinking Water Exposure Assessments;" Draft
Paper. September 1, 2000. Office of Pesticide Programs, Office of Prevention,
Pesticides, and Toxic Substances, U.S. Environmental Protection Agency.
Washington, DC. Available: http://www.epa.gov/pesticides/trac/science
U.S. Environmental Protection Agency. 1999a. "Guidance for Identifying Pesticide
Chemicals and Other Substances That Have A Common Mechanism of Toxicity."
Environmental Protection Agency, Office of Pesticide Programs. Fed. Reg.
64:5796-5799. Available: http://www.gov/fedrgstr/EPA-PEST/1999/Februarv/Dav-
05/6055.pdf
U.S. Environmental Protection Agency. 1999b. Memorandum from Margaret
Stasikowski, Health Effects Division to Staff. "Translation of Monitoring Data. HED
Standard Operating Procedure 99.3 (3/26/99);" March 26, 1999. Office of
Pesticide Programs, Office of Prevention, Pesticides, and Toxic Substances,
Washington, D.C.
U.S. Environmental Protection Agency. 1999c. "Guidance for Performing
Aggregate Exposure and Risk Assessments;" draft document. October 29, 1999.
Office of Pesticide Programs, Office of Prevention, Pesticides, and Toxic
Substances, Washington, D.C. 64_FR61343. Available:
http://www.epa.gov/fedrgstr/EPA-PEST/1999/November/Dav-107.
U.S. Environmental Protection Agency. 1999. Memorandum from Jerome Blondell,
Office of Pesticide Programs, Health Effects Division to Dennis Utterback, of the
Office of Pesticide Programs, Special Review and Reregistration Division. " Review
of Poison Control Center Data for Residential Exposures to Organophosphate
Pesticides, 1993-1996;" February 11,1999. Office of Pesticide Programs, Office of
Prevention, Pesticides, and Toxic Substances. Washington, DC.
U.S. Environmental Protection Agency. 1997a. "Exposure Factors Handbook.
13 | Volume 1/General Factors. Update to Exposure Factors Handbook; EPA/600/8/043
r ^ ^ - May 1989." Office of Research and Development, National Center for
Environmental Assessment, U.S. Environmental Protection Agency. EPA/600/P-95-
002Fa. Available: http://www.epa.gov/ncea/exposfac.htm
U.S. Environmental Protection Agency. 1997b. "Exposure Factors Handbook.
Volume 3/Activity Factors. Update to Exposure Factors Handbook; EPA/600/8/043 -
May 1989." Office of Research and Development, National Center for
Environmental Assessment, U.S. Environmental Protection Agency. EPA/600/P-95-
002Fa. Available: http://www.epa.gov/ncea/exposfac.htm
U.S. Environmental Protection Agency. 1992. National Home and Garden
Pesticide Use Survey, March 1992. Prepared by Research Triangle Institute.
I.J Page 26
-------
i U.S. Environmental Protection Agency. 1990. "Nonoccupational Pesticide
I Exposure Study (NOPES) Final Report." Atmospheric Research and Exposure
| Assessment Laboratory, Office of Research and Development, U.S. Environmental
I Protection Agency, Research Triangle Park, NC 27711
I U.S. Food and Drug Administration. Center for Food Safety and Applied Nutrition:
I "Pesticides, Metals, Chemical Contaminants & Natural Toxins." Online. Available:
(0 I http://vm.cfsan.fda.gov/~ird/pestadd.html
! | U.S. Geological Survey Hydrologic Investigations Atlas. Online. Available:
I http://capp.water.usgs.gov/gwa/gwa.html
I U.S. Geological Survey Circulars. Online. Available:
i http://pubs.usgs.aov/products/books/circular.html
^ I U.S. Geological Survey Fact Sheets. Online. Available:
/I) § http://pubs.usgs.gov/products/books/factsheet.html
CO |
CO f U.S. Geological Survey Professional Papers. Online. Available:
-------
• #A
v#^
°*>
Wessler I, Krikpatrick CJ, Racke K. 1998. "Non-neuronal acetylcholine, a locally
acting molecule, widely distributed in biological systems: Expression and function in
humans." Pharmacology & Therapeutics. 77:59-79.
Wester RC, and Maibach HI. 1989. "Dermal Decontamination and Percutaneous
Absorption." In: Percutaneous Absorption. 2nd ed. R. Bronaugh and H.I. Maibach,
editors. New York: Marcel Dekker, pp 335-342.
Wyttenbach CR, Thompson SC. 1985. "The effects of the organophosphate
insecticide malathion on very young chick embryos: malformations detected by
histological examination." The American Journal of Anatomy. 174(2): 187-202.
Xie W, Stribley JA, Chatonnet A, Wilder PJ, Rizzino A, McComb RD, Taylor P,
Hinrichs SH, Lockridge O. 2000. "Postnatal developmental delay and
supersensitivity to organophosphate in gene-targeted mice lacking
acetylcholinesterase." The Journal of Pharmacology and Experimental
Therapeutics. 293(3):896-902.
Young WC, III, Mellbye ME, and Gingrich GA. n.d. The Oregon Grass Seed
Industry. Oregon State University, Crop & Soil Science Dept.
Zheng Q, Olivier K, Won YK, Pope CN. 2000. "Comparative cholinergic
neurotoxicity of oral chlorpyrifos exposures in preweanling and adult rats."
Toxicological Sciences. 55(1): 124-32.
Zollinger RK, Dexter AG, Dahl GK, FittererSA, McMullen MP, Waldhaus GE,
Glogoza P, and Ignaszewski K. 1998. "Pesticide Use and Pest Management
Practices for Major Crops in North Dakota" 1996. North Dakota State University
Extension Service. Extension report no. 43.
I.J Page 28
------- |