Page
1
of
5
UNITED
STATES
ENVIRONMENTAL
PROTECTION
AGENCY
WASHINGTON,
D.
C.
20460
OFFICE
OF
PREVENTION,
PESTICIDES
AND
TOXIC
SUBSTANCES
November
8,
2004
MEMORANDUM
Subject:
Transmission
of
Background
Materials
and
Charge
to
the
Panel
for
the
Session
of
the
December
3,
2004
FIFRA
Scientific
Advisory
Panel
Entitled
"
The
N­
methyl
Carbamate
Cumulative
Risk
Assessment:
Strategies
and
Methodologies
for
Exposure
Assessment"

To:
Joseph
Bailey,
Designated
Federal
Official
FIFRA
SAP
Office
of
Science
Coordination
and
Policy
(
7101C)

From:
David
J.
Miller
and
Anna
Lowit
Office
of
Pesticide
Programs,
Health
Effects
Division
(
7509C)

Through:
George
Herndon,
Acting
Director
Office
of
Pesticide
Programs,
Health
Effects
Division
(
7509C)

The
December
3,
2004
FIFRA
SAP
meeting
entitled
"
The
N­
methyl
Carbamate
Cumulative
Risk
Assessment:
Strategies
and
Methodologies
for
Exposure
Assessment"
is
the
first
in
a
series
of
SAP
meetings
planned
by
EPA
to
discuss
various
aspects
of
the
N­
methyl
carbamate
cumulative
risk
assessment.
The
primary
purpose
of
this
meeting
is
to
discuss
the
concepts
introduced
in
the
white
paper
entitled
"
Designing
Exposure
Models
that
Support
PBPK/
PBPD
Models
of
Cumulative
Risk"
and
developed
by
the
LifeLife
Group
Inc.
In
addition,
this
SAP
meeting
is
meant
to
provide
the
members
of
the
FIFRA
Scientific
Advisory
Panel
(
SAP)
and
the
public
with
the
general
framework
and
next
steps
in
the
development
of
the
cumulative
risk
assessment
for
the
N­
methyl
carbamate
pesticides.
Page
2
of
5
In
addition
to
this
memo,
three
documents
are
provided
to
members
of
FIFRA
Scientific
Advisory
Panel
in
preparation
for
the
December
3,
2004
meeting:

1.
A
background
document
covering
EPA
activities
related
to
the
N­
Methyl
carbamate
cumulative
risk
assessment
2.
Attachment
1:
Overview
of
Topics
for
February
2005
FIFRA
SAP
meeting
3.
A
document
entitled
"
Designing
Exposure
Models
that
Support
PBPK/
PBPD
Models
of
Cumulative
Risk"
developed
by
the
LifeLife
Group
Inc.

Note:
EPA
is
soliciting
comment
(
Question
2)
from
the
SAP
on
topics
discussed
in:

Price,
P.
S.,
Conolly,
R.
B.,
Chaisson,
C.
F.,
Young,
J.
S.,
Mathis
E.
T.,
Tedder
D.
T.
2003.
Modeling
Inter­
individual
Variation
in
Physiological
Factors
Used
in
PBPK
Models
of
Humans,
Critical
Reviews
in
Toxicology
Vol.
33,
(
5):
469­
503.

EPA
is
currently
working
with
Taylor
and
Francis,
the
publisher
and
copyright
owner
of
this
article,
to
provide
this
paper
to
the
panel.

As
part
of
this
SAP
session,
we
are
asking
Panel
members
to
consider
the
following
charge
and
questions:

Charge
and
Questions
to
the
Panel:

1.
The
LLG's
white
paper
entitled
"
Designing
Exposure
Models
that
Support
PBPK/
PBPD
Models
of
Cumulative
Risk"
presents
an
outline
of
the
fundamental
procedures
and
logic
required
to
deliver
appropriate
exposure
metrics
to
the
Physiologically­
based
Pharmacokinetic/
Pharmacodynamic
(
PBPK/
PD)
model
for
the
N­
methyl
carbamate
group
of
pesticides.
Specifically,
the
new
exposure
assessment
requires
an
approach
that
will
modify
the
exposure
information
that
is
currently
produced,
extend
the
software
to
provide
additional
information
on
the
individuals
being
modeled,
and
define
the
technical
process
by
which
information
will
be
transferred
from
the
exposure
model
to
the
PBPK/
PD
model.
The
LLG
white
paper
also
describes
the
data
requirements
of
a
PBPK/
PD
model,
briefly
reviews
the
state
of
existing
exposure
assessment
models
and
their
outputs,
and
presents
a
both
a
general
approach
and
an
N­
methyl
carbamate­
specific
approach
of
how
exposure
simulation
models
can
be
adapted
to
meet
the
needs
of
a
PBPK/
PD
model
of
cumulative
risks.

Please
comment
on
the
detail
and
clarity
of
this
document.
Page
3
of
5
2.
A
central
tenet
underlying
aggregate
and
cumulative
risk
assessment
is
that
exposure
occurs
to
a
hypothetical
individual
whose
specific
demographic
characteristics
such
as
age
group,
region
of
residence,
race/
ethnicity,
sex,
etc.
help
define
exposure
scenarios.
The
exposure
pattern
and
other
data
concerning
this
individual
should
be
consistent
with
those
characteristics.

The
use
of
PBPK/
PD
models
in
cumulative
assessments
adds
another
layer
to
the
complexity
of
generating
and
maintaining
a
set
of
internally
consistent
individuals
comprising
a
hypothetical
population.
In
defining
individuals
for
use
in
PBPK/
PD
models,
it
is
necessary
to
maintain
logical
consistency
and
linkage
between
the
various
anatomical
and
physiological
parameters
that
describe
that
individual.
For
example,
given
a
bodyweight,
age,
and
sex
of
an
individual
from
a
reference
population
such
as
Lifeline's
Natality
data
set,
it
is
necessary
that
the
organ
sizes,
compartmental
blood
flows,
breathing
rates,
etc.
all
be
consistent.

A
recent
journal
article
by
P.
S.
Price
et
al.
(
2003)
appearing
in
Critical
Reviews
in
Toxicology
summarizes
much
of
the
literature
in
this
area1.
The
article
presents
a
number
of
regression
and
other
equations
which
can
be
used
to
generate
the
linked
anatomic
and
physiological
characteristics
of
those
individuals.
2
a)
Please
comment
on
the
degree
to
which
the
article
comprehensively
summarizes
the
available
literature
concerning
the
anatomic
and
physiological
relationships
that
exist
between
organ
sizes
and
volumes,
blood
and
other
flows,
breathing
parameters,
etc.?

b)
Are
there
additional
data
or
data
sources
for
these
relationships
that
would
be
useful
to
include
or
consider?

c)
Please
comment
on
algorithms
provided
and
their
potential
utility
in
use
by
PBPK/
PD
models.

1
EPA
has
not
as
of
yet
received
permission
from
Taylor
and
Francis,
the
publisher
and
copyright
owner
of
this
article,
to
distribute
the
following
study
to
the
FIFRA
SAP
members.
Price,
P.
S.,
Conolly,
R.
B.,
Chaisson,
C.
F.,
Young,
J.
S.,
Mathis
E.
T.,
Tedder
D.
T.
2003.
Modeling
Interindividual
Variation
in
Physiological
Factors
Used
in
PBPK
Models
of
Humans,
Critical
Reviews
in
Toxicology
Vol.
33,
(
5):
469­
503
2
The
algorithms
present
in
this
journal
article
are
those
that
are
used
in
a
model
called
Physiological
Parameters
for
PBPK
Modeling
(
P
³
M)
and
available
from
http://
www.
thelifelinegroup.
org/
P3M/
index.
html
.
The
model
serves
as
a
convenient
tool
to
parameterize
exposure
and
PBPK
models
The
software
can
be
downloaded
from
the
aboveindicated
site
and
is
available
without
charge.
Page
4
of
5
3.
Traditional
non­
cancer
probabilistic
risk
assessment
methods
perform
a
direct
conversion
of
exposure
(
expressed,
for
example,
in
ug/
kg
day)
into
risk
(
expressed,
for
example,
as
a
unitless
margin
of
exposure
or
percent
of
reference
dose).
By
incorporating
a
PBPK/
PD
component
into
risk
assessments
in
order
to
more
appropriately
account
for
temporal
and
other
aspects
of
toxicity,
output
from
the
exposure
component
of
the
model
must
serve
as
input
to
the
PBPK
component.
In
order
for
this
to
occur,
a
time
series
of
exposures
must
be
developed
for
each
individual
considered
in
the
assessment.
Each
exposure
event
associated
with
that
individual
that
occurs
during
a
given
time
step
must
act
as
a
separate
input
to
the
PBPK/
PD
model.

In
order
for
this
to
occur,
data
from
the
USDA's
CSFII
must
be
placed
into
the
exposure
component
of
a
model
in
a
such
a
way
that
separates
each
individual's
eating
occasions.
In
addition,
data
from
NHAPS
and
other
databases
will
need
to
be
entered
in
such
a
way
that
each
event
occurring
during
a
given
time
step
is
distinct
and
separate.
Furthermore,
the
output
from
this
exposure
model
must
appropriately
link
or
interface
with
a
PBPK/
PD
model.
The
LLG's
white
paper
proposes
that
Lifeline
be
modified
such
the
analyst
can
customize
the
outputs
of
the
model
for
the
specific
PBPK/
PD
analysis
to
be
run,
selecting
from
among
23
tissues,
organs,
and
compartments
listed.
The
analyst
will
then
define
the
duration
of
the
time
step
used
for
creating
the
exposure
history
and
the
duration
of
the
exposure
history
for
the
basis
of
the
LifeLine
 
exposure
analysis
metrics
and
output
file.
LifeLine
 
output
files
will
be
created
as
Access
 
files
consisting
of
separate
records
for
exposures
of
each
simulated
individual
within
the
defined
population
of
the
analysis.
Each
individual's
exposure
history
will
be
captured
in
a
record
that
consists
of
two
tables.
Examples
of
data
tables/
outputs
were
presented
in
the
LLG's
background
document.

a)
Please
comment
on
the
format
and
structure
of
the
MS
Access
file
containing
the
records
for
each
individual's
exposure
and
anatomical/
physiological
parameters
(
Table
2
and
Table
3a
of
the
LLG
white
paper)?

b)
Are
there
additional
parameters
or
options
that
should
be
included?
Page
5
of
5
4.
The
suggested
approach
addressed
in
Question
#
3
will
make
resourceintensive
computational
demands
making
computer
run­
times
impractical
for
regulatory
purposes.
The
LLG
white
paper
proposes
that
not
every
record
generated
or
processed
by
the
LifeLine
model
be
saved.
These
limitations
will
require
that
model
runs
be
limited
to
a
few
hundred
or
a
thousand
individuals
and
that
only
some
fraction
of
the
records
be
retained
by
software
and
used
as
input
to
the
PBPK/
PD
model.
The
process
of
selecting
the
records
to
convey
to
the
PBPK/
PD
model
will
require
special
attention
and
a
transparent
prioritization
scheme
based
on
explicit
criteria.
The
specific
nature
of
how
this
will
be
done
could
be
based
on
any
of
several
criteria.
For
example:
the
exposure
software
could
create
a
demographic,
physiological
and
exposure
history
for
each
individual
and
"
tag"
only
those
individuals
with
estimated
exposures
(
or
relative
potency
factor­
adjusted
exposures)
greater
than
either
1)
a
certain
user­
defined
cut­
off
value
(
e.
g,
>
BMD10)
or
2)
greater
than
a
user­
defined
percentile
(
e.
g.,
90th
percentile).
Only
those
records
that
were
tagged
in
this
way
would
be
included
in
the
interface
file
(
MS
Access
 
)
that
will
be
exported
to
the
PBPK/
PD
model.
In
this
way,
only
the
records
that
were
at
the
high
end
of
the
exposure
distribution
(
however
defined
by
the
user)
would
be
run
through
that
model.

a)
Please
comment
on
the
proposal
to
retain
only
a
fraction
of
the
records
generated
by
the
LL
model
for
interface/
export
to
the
PBPK/
PD
model
due
to
computational
demands.

b)
Does
the
panel
have
any
comments
or
suggestions
on
the
criteria
which
should
be
used
to
select
records
for
input
into
the
PBPK/
PD
model?
