THE
LIFELINE
GROUP
Page
1
of
57
Designing
Exposure
Models
that
Support
PBPK/
PBPD
Models
of
Cumulative
Risk
Prepared
for
USEPA
Office
Pesticide
Programs
Under
EPA
Contract
EP­
W­
04­
016
In
Partial
Fulfillment
Work
Assignment
1­
4
"
Technical
Support
and
Model
Enhancement
for
LifeLine
 
Software"

Prepared
by
Paul
S.
Price,
Christine
F.
Chaisson,
Claire
A.
Franklin,
Michael
A.
Jayjock
The
LifeLine
Group
Inc.

www.
TheLifeLinegroup.
org
November
8,
2004
THE
LIFELINE
GROUP
Page
2
of
57
Page
is
intentionally
left
blank.
THE
LIFELINE
GROUP
Page
3
of
57
TABLE
OF
CONTENTS
1.
Introduction........................................................................................................
6
2.
Accommodating
the
Pesticide's
Characteristics
in
the
Conceptual
Design
of
Exposure
Models...............................................................................................
9
3.
Meeting
the
Needs
of
PBPK/
PD
Based
Models
of
Cumulative
Risk................
11
3
a.
Relevant
Physiological
Characteristics
of
the
Exposed
Individuals
.
11
3
b.
Kinetic,
Metabolic
and
Capacity
Parameters
...................................
12
3
c.
Data
on
Exposure
and
Dose
...........................................................
13
i.
Matching
LifeLine
 
Outputs
to
PBPK/
PD
Model
Inputs
............
13
ii.
Establishing
the
Time
Step
and
Exposure
History
.....................
14
4.
Current
Exposure
Assessment
Models:
A
Brief
Overview...............................
16
4
a.
Capabilities
Required
for
Supporting
PBPK/
PD
Modeling...............
16
5.
Development
of
a
New
Version
of
the
LifeLine
 
Software
for
the
Support
of
a
PBPK/
PD
Model
of
Cumulative
Risk
...............................................................
18
5
a.
The
Current
Framework
of
LifeLine
 
Software
..............................
18
i.
POM
Models
..............................................................................
18
5
b.
Changing
the
Time
Step
from
One
Day
to
a
Shorter
Time
Interval
.
24
i.
Proposed
Approach
...................................................................
27
5
c.
Modeling
Multiple
Pesticides
...........................................................
30
i.
Considerations
of
Residue
Data
Appropriate
for
Use
in
Modeling
Concurrent
Exposures
of
Multiple
Pesticides
..................................
30
ii.
Modification
of
LifeLine
 
to
Separately
Calculate,
Track,
and
Save
Exposure
Information
for
Multiple
Pesticides
.........................
35
5
d.
Matching
the
LifeLine
 
Exposure
Outputs
to
the
Definitions
of
Dose
in
the
PBPK/
PD
Model..........................................................................
38
i.
Oral
Exposures
..........................................................................
38
ii.
Dermal
Exposures......................................................................
39
iii.
Inhalation
Exposures..................................................................
40
iv.
Summary....................................................................................
40
5
e.
Modeling
Physiological
and
Genetic
Variability
in
Partitioning,
Metabolism,
and
Dose­
Effect
Relationships
in
PBPK/
PD
Models.........
40
i.
The
Lifeline
 
Framework
..........................................................
41
ii.
Proposed
Modifications..............................................................
42
5
f.
Designing
the
Interface
between
LifeLine
 
Exposure
Model
and
the
PBPK/
PD
Model
of
Cumulative
Risks...................................................
46
5
g.
Processing
Demands
......................................................................
47
5
h.
Output
File
Structure
.......................................................................
52
6.
References
......................................................................................................
55
Appendix
A.
Modeling
Inter­
individual
Variation
in
Physiological
Factors
Used
in
PBPK/
PD
Models
of
Humans
(
Price
et
al.
2003b)
...........................................
57
THE
LIFELINE
GROUP
Page
4
of
57
Definitions
Absorbed
dose
The
amount
of
a
substance
that
enters
an
individual's
body
as
a
result
of
passing
an
exposure/
absorption
barrier
(
e.
g.
skin
surface,
alveolar
wall
or
the
surface
of
the
gastrointestinal
tract).
Also
referred
to
as
the
internal
dose.

Aggregate
exposure
The
total
exposure
from
all
sources
(
excluding
occupational
exposures)
to
a
single
pesticide
by
all
routes
of
exposure.

Contact
dose
The
amount
of
pesticide
presented
to
an
exposure/
absorption
barrier
and
available
for
absorption.

Cumulative
exposure
The
total
exposure
from
all
sources
(
excluding
occupational
exposures)
of
multiple
pesticides
operating
by
a
common
mechanism
of
action.

Dose
The
amount
of
a
substance
available
for
interaction
with
metabolic
processes
or
biologically
significant
receptors
after
crossing
the
outer
boundary
of
an
organism.
It
can
be
described
as
potential
dose,
contact
(
applied)
dose,
absorbed
dose,
internal
dose,
delivered
dose
and
target
(
biologically
effective
dose)

Eating
event
A
collection
of
food
items
recorded
by
an
individual
as
being
consumed
at
a
particular
time
of
the
day
or
at
a
named
eating
occasion
(
i.
e.
lunch).

Exposure
Contact
of
an
organism
with
a
chemical,
quantified
as
the
amount
of
chemical
available
at
exposure/
absorption
barriers
of
the
organism
and
available
for
absorption.

Exposure/
absorption
barrier
Any
of
the
exchange
barriers
of
the
body
that
allow
differential
diffusion
of
various
substances
across
a
boundary
i.
e.
skin
surface,
alveolar
wall
or
the
surface
of
the
gastrointestinal
tract.
THE
LIFELINE
GROUP
Page
5
of
57
Metrics
Quantitative
measures
of
exposure
or
dose
that
specify
the
compartment
or
location
of
the
pesticide
(
on
the
skin,
inhaled,
or
crossing
the
GI
tract,
or
etc.)
and
the
units
that
the
measure
is
expressed
(
mg,
mg/
kg,
ppm
in
air
or
water).

Exposure
history
A
description
of
an
exposures
and
resulting
doses
that
occur
over
a
specific
period
of
time.
Composed
of
multiple
time
steps.

Food
item
A
description
of
a
food
(
with
a
defined
recipe
of
ingredients)
in
its
edible
form.

Internal
dose
See
absorbed
dose.

PBPK/
PD
model
A
model
that
is
a
combination
of
PBPK
and
PBPD
models.

Pharmacodynamic
model
Quantitative
models
of
measurable
effects
of
pesticides
typically
at
the
tissue
or
cellular
level.

Pharmacokinetic
model
Quantitative
models
of
the
intake,
movement,
metabolism
and
excretion
of
pesticides
in
humans
and
other
animals.

Person
Oriented
Model
(
POM)
A
model
of
exposure
produced
using
a
modeling
architecture
that
places
the
person
central
to
model
design.
Contrast
to
source­
to­
dose
models.

Simulation
model
A
quantitative
model
that
characterizes
exposure
and
dose
by
modeling
the
uncertainty
and
interindividual
variation
of
those
parameters
that
determine
exposure
and
uptake.

Target
(
biologically
effective)
dose
The
amount
of
pesticide
that
reaches
the
tissue
or
compartment
where
an
effect
occurs.

Time
step
A
period
in
an
individual's
exposure
history
that
is
sufficiently
short
that
the
inputs
to
an
exposure
or
dose
equation
can
be
treated
as
constants.
THE
LIFELINE
GROUP
Page
6
of
57
1.
Introduction
The
Environmental
Protection
Agency
(
EPA)
is
charged
with
assuring
the
safety
of
pesticides
by
regulating
their
use
in
the
United
States
under
the
Federal
Insecticide,
Fungicide
and
Rodenticide
Act
and
by
setting
tolerances
for
all
crop
commodities
(
domestic
and
imported)
under
the
Federal
Food
and
Drug
and
Cosmetics
Act.
In
assessing
the
potential
health
risks
associated
with
exposure
to
pesticides,
attention
has
historically
focused
on
single
pathways
of
exposure
(
e.
g.,
from
pesticide
residues
in
food,
water,
or
residential/
non­
occupational
uses)
for
individual
pesticides,
and
not
on
the
potential
for
individuals
to
be
exposed
to
multiple
pesticides
by
all
pathways
concurrently.
This
changed
in
1996
with
the
passage
of
the
Food
Quality
Protection
Act
(
FQPA)
which
required
the
consideration
of
human
health
risks
resulting
from
concurrent
exposures
from
multiple
sources
and
routes
of
exposure
for
dietary
and
non­
dietary,
nonoccupational
exposures
(
aggregate
exposure)
and
by
the
concurrent
exposures
to
all
pesticides
acting
through
a
common
mechanism
of
toxicity
(
cumulative
exposure).
The
FQPA
also
required
attention
to
potentially
vulnerable
population
groups.

In
response
to
this
requirement,
the
Agency
developed
guidance
for
performing
aggregate
and
cumulative
risk
assessment
(
EPA,
2001;
2002a).
In
2002,
EPA
performed
the
first
cumulative
assessment
on
the
organophosphorus
(
OP)
pesticides
(
EPA,
2002b).
As
part
of
the
assessment,
The
LifeLine
Group
Inc.
(
LLG)
was
hired
to
develop
Version
2.0
of
LifeLine
 
and
to
use
the
program
to
perform
an
assessment
of
the
OP
pesticides
(
LifeLine
 
,
2002a).

Over
the
past
twenty
five
years,
exposure
assessments
conducted
by
the
Office
of
Pesticide
Programs
within
the
EPA
evolved
from
calculation
of
the
exposure
presented
to
the
"
average
American"
for
lifetime
average
exposures
to
assessments
of
full
distributions
of
possible
exposures
to
individuals
within
defined
population
groups
over
variable
time
periods.
The
cumulative
risk
assessment
of
the
OP
pesticides
(
EPA,
2002b)
stands
as
the
most
sophisticated
attempt
to
date
to
realize
the
goal
of
the
FQPA.
The
assessment
achieves
many
elements
of
an
aggregate,
cumulative
assessment
for
the
general
population
of
the
United
States.
It
assesses
exposure
experienced
by
individuals
of
all
ages
via
multiple
routes
of
exposure
on
a
daily
basis.
To
achieve
consideration
of
concurrent
exposure
to
multiple
organophosphorus
pesticides,
the
Agency
used
a
Relative
Potency
Factor
(
RPF)
approach,
as
explained
in
detail
in
its
guidance
documents.
This
approach
accounts
for
the
variance
in
the
potency
of
each
pesticide
by
mathematically
comparing
the
potency
of
the
pesticides
using
a
common
toxicology
endpoint.
The
RPF
is
defined
as
the
ratio
of
the
potency
of
each
pesticide
to
the
potency
of
an
"
index
chemical".
Summation
of
the
RPFs
that
are
produced
for
each
of
the
OP
pesticides
allows
the
estimation
of
an
equivalent
total
dose
of
the
index
chemical.
THE
LIFELINE
GROUP
Page
7
of
57
Exposures
to
multiple
OP
pesticides
can
now
be
evaluated
in
terms
of
a
dose
of
one
equivalent
pesticide;
thus,
the
cumulative
risks
estimated
in
the
OP
assessment
were
determined
by
an
aggregate
model
of
the
total
dose
of
the
index
chemical.
By
performing
this
conversion
at
the
beginning
of
the
exposure
process,
the
OP
assessment
allowed
EPA
to
use
existing
aggregate
exposure
software
to
perform
the
cumulative
assessment.

I
n
the
N­
methyl
carbamate
assessment,
EPA
is
proposing
to
take
the
next
step
in
technical
improvement
of
the
exposure
and
risk
assessment
science.
This
exposure
assessment
will
independently
determine
and
track
the
route
specific
cumulative
doses
of
each
pesticide
that
occur
from
exposures
to
dietary
and
nondietary
sources.
The
approach
will
not
require
the
collapsing
of
these
pesticide
specific
doses
into
a
single
dose
of
an
index
pesticide.
The
approach
will
instead
use
the
pesticide
specific
doses
as
inputs
to
a
physiologically
based
pharmacokinetic
model
PBPK
model
and
calculate
the
internal
or
target
doses
of
each
pesticide.
In
order
to
support
such
modeling,
the
exposure
assessment
must
be
extended
in
two
ways.
First,
the
exposure
assessment
is
modified
to
consider
time
units
of
hours,
or
minutes,
instead
of
days.
Second,
the
exposure
assessment
is
expanded
to
define
not
only
the
doses
received
by
the
simulated
individuals
but
also
the
physiological
parameters
of
each
modeled
individual.
In
this
way,
the
dose
and
physiological
information
are
based
on
individuals
who
have
a
consistent
age,
sex
and
activity
level.

Finally,
the
internal
doses
from
the
PBPK
models
are
provided
to
another
type
of
model,
the
physiologically
based
pharmacodynamic
model
(
PBPD).
This
model
applies
information
on
the
effects
of
the
pesticides
(
and
their
metabolites)
on
the
cellular
or
tissue
level.
Processes
included
such
as
binding
and
release
of
substances
to
receptors,
the
resulting
effects
and
recovery
from
the
effects.
The
outputs
of
these
models
provide
a
basis
of
the
determination
of
whether
an
"
injury"
occurred.
Because
of
the
close
linking
of
PBPK
and
PBPD
models,
they
are
referred
to
as
a
combined
(
PBPK/
PD)
model.
EPA
will
employ
PBPK/
PD
models
such
as
those
under
development
by
EPA's
Office
of
Research
and
Development.

EPA
tasked
The
LifeLine
Group
(
LLG)
to
develop
a
new
approach
to
deliver
appropriate
exposure
metrics
to
the
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.

This
report
presents
the
fundamental
approaches
and
logic
for
the
required
changes
to
the
existing
exposure
assessment
methodology
of
the
LifeLine
 
software.
The
concepts
are
generic,
but
the
technical
approaches
are
specific
to
the
LifeLine
 
software.
In
order
to
explain
the
approach
and
to
provide
a
THE
LIFELINE
GROUP
Page
8
of
57
framework
for
the
logic
employed
in
this
task,
this
report
begins
by
describing
the
data
requirements
of
a
PBPK/
PD
model;
the
report
then
briefly
reviews
the
state
of
existing
exposure
assessment
models
and
their
outputs
and
presents
a
general
approach
of
the
how
LifeLine
 
exposure
simulation
model
can
be
adapted
to
meet
the
needs
of
any
PBPK/
PD
model
of
cumulative
risks.
Finally,
the
general
approach
is
shown
to
be
applicable
to
the
specific
modeling
needs
for
the
N­
methyl
carbamate
pesticides.
THE
LIFELINE
GROUP
Page
9
of
57
2.
Accommodating
the
Pesticide's
Characteristics
in
the
Conceptual
Design
of
Exposure
Models
The
characteristics
of
the
pesticides
under
consideration
influence
the
calculation
of
exposure
and
of
dose
and
risk
in
two
ways.
The
first
way
is
the
direct
application
of
pesticide
specific
information
within
a
given
model.
For
example,
in
considering
exposure
to
a
pesticide
used
on
crops,
the
pesticide
specific
values
for
residue
concentrations
are
entered
by
the
user
and
applied
to
the
calculations
of
dietary
exposure
estimates.
Second,
the
overall
methodology
and
design
of
the
model
should
accommodate
pesticide
specific
characteristics
as
well.
For
example,
if
the
relevant
toxicological
endpoint
for
risk
assessment
is
exhibited
only
after
long
durations
of
consistent
exposure,
exposure
averaging
techniques
over
variable
time
intervals
should
be
a
possible
calculation
choice
for
the
model
user.
The
model
should
be
sufficiently
flexible
to
allow
the
user
to
incorporate
the
relevant
values
of
potency
and
relevant
exposure
durations
for
the
specific
pesticide.

Exposure
models
deal
with
temporal
patterns
of
exposure
using
the
concept
of
a
time
step.
Ideally,
a
time
step
is
the
period
of
time
sufficiently
short
that
exposures
can
be
represented
by
time
invariant
values.
However,
the
shorter
the
duration
of
a
time
step
in
the
model,
the
greater
are
the
computational
demands
in
that
model.
In
addition,
in
many
instances,
data
that
would
allow
the
definition
of
the
variation
of
dose
on
a
fine
time
scale
are
not
available.
As
a
result,
models
calculate
longitudinal
exposures
using
time
steps
of
a
day,
a
week
or
a
year.
In
these
instances,
the
exposure
is
expressed
in
terms
of
average
exposure
over
the
period.

N­
methyl
carbamate
pesticides
have
a
number
of
characteristics
that
influence
the
design
of
both
the
exposure
and
PBPK/
PD
models
of
cumulative
risk.
These
include
the
rapid
binding
and
release
of
the
compounds
as
well
as
reversibility
of
the
cholinesterase
inhibition
and
the
rapid
metabolism
of
the
active
moieties.
Because
of
these
characteristics,
it
is
likely
that
the
effects
from
a
dose
of
a
pesticide
in
this
class
will
persist
only
over
short
periods
of
time.
These
characteristics
suggest
that
the
toxicological
effects
that
occur
from
a
series
of
exposures
on
a
given
day
would
be
limited
to
that
day
and
would
not
persist
into
a
second
day.
This
in
turn
suggests
that
the
temporal
framework
for
the
assessment
of
cumulative
risk
for
this
group
of
pesticides
could
be
limited
to
a
single
day.
However,
without
specific
empirical
data
to
confirm
this,
it
is
prudent
to
build
the
model
so
that
it
can
handle
any
carry­
over
time
that
is
warranted.

There
is
also
data
indicating
that
the
time
to
peak
exposure
via
the
oral
route
is
considerably
shorter
that
from
the
dermal
route.
These
characteristics
require
that
the
exposure
software
define
exposures
on
a
finer
time
scale
than
a
single
THE
LIFELINE
GROUP
Page
10
of
57
day
and
has
the
ability
to
define
time­
correlated
exposures
across
sources
and
routes
of
exposure.

It
should
also
be
noted
that
other
pesticides
or
classes
of
pesticides
that
operate
by
different
mechanisms
might
require
modeling
over
longer
or
shorter
periods.
Therefore,
this
report
will
develop
an
approach
that
can
also
be
applied
to
longer
longitudinal
periods
than
a
single
day
and
where
the
duration
of
the
time
step
required
can
vary
from
durations
of
less
then
one
day
to
multiple
days.
THE
LIFELINE
GROUP
Page
11
of
57
3.
Meeting
the
Needs
of
PBPK/
PD
Based
Models
of
Cumulative
Risk
The
goal
for
the
LifeLine
 
exposure
assessment
model
is
to
provide
the
appropriate
support
to
the
PBPK/
PD
model
of
the
cumulative
toxicological
effects
posed
by
N­
methyl
carbamate
pesticides.
The
initial
portion
of
the
PBPK/
PD
model
is
a
PBPK
model
for
the
various
pesticides.
This
PBPK
model
translates
the
history
(
or
temporal
pattern)
of
the
route­
specific
exposures
of
each
pesticide
received
by
an
individual
into
a
history
of
the
concentrations
of
the
pesticide
(
and
relevant
metabolites)
in
the
relevant
compartments
of
the
body
(
target
doses).

PBPK
models
use
three
different
types
of
information.
The
first
type
of
data
is
the
relevant
physiological
characteristics
of
the
individuals
being
simulated
and
how
those
characteristics
are
influenced
by
the
activities
of
the
individual.
The
second
type
is
compound­
specific
data
on
the
absorption,
distribution,
metabolism
and
excretion
(
ADME)
kinetics
and
metabolism
of
the
individual
pesticides
and
how
they
vary
across
individuals
in
the
population
of
interest.
Finally,
PBPK/
PD
models
utilize
time­
specific,
route
specific
estimates
of
exposure.

The
PBPK/
PD
model
under
development
by
the
EPA's
Office
of
Research
and
Development
(
ORD)
is
not
yet
available
for
inspection
and
discussion.
LLG
has
collaborated
with
the
ORD
development
team
to
identify
common
approaches
and
definitions
and
to
assure
that
the
concepts
developed
for
the
exposure
model
amendments
will
accommodate
the
needs
of
the
PBPK/
PD
model.
The
discussions
about
PBPK/
PD
in
this
report
are
developed
by
the
LLG
and
are
not
meant
to
represent
the
technical
documentation
of
any
ORD
model.

3
a.
Relevant
Physiological
Characteristics
of
the
Exposed
Individuals
PBPK/
PD
models
require
information
on
the
physiology
of
the
individual
and
how
an
individual's
physiology
changes
over
time.
This
physiological
information
includes:

1.
The
volume
of
each
compartment
in
the
PBPK/
PD
model;
2.
The
cardiac
output;
3.
The
fraction
of
the
cardiac
output
for
each
compartment;
and
4.
The
alveolar
ventilation
rate.

The
values
of
these
parameters
vary
from
one
individual
to
another
and
within
individuals
as
a
function
of
their
level
of
physical
activity
at
a
given
point
in
time.
The
values
of
these
parameters
are
correlated
with
the
estimates
of
the
doses
that
the
individual
receives
because
age
and
gender
influence
both
exposure
THE
LIFELINE
GROUP
Page
12
of
57
opportunity
to
pesticides
and
the
physiology
of
the
individual.
To
address
this
correlation,
the
PBPK/
PD
models
must
be
provided
descriptions
of
the
individuals'
physiology
that
are
consistent
with
the
assumptions
used
in
the
exposure
calculations.
Thus
if
the
model
simulates
the
exposures
for
a
three
year
old
playing
on
the
lawn,
then
the
physiological
parameters
used
in
the
PBPK/
PD
modeling
must
be
consistent
with
those
of
an
active
three
year
old.

The
values
of
the
physiological
parameters
for
an
individual
are
also
correlated.
Values
for
one
compartment
will
be
correlated
with
values
for
another.
An
aboveaverage
body
weight
and
height
implies
that
the
compartment
volumes
of
the
individual
will
be
larger
than
average
and
that
the
alveolar
ventilation
rate
and
the
cardiac
output
will
also
be
larger
than
average.
This
kind
of
correlation
must
be
captured
in
the
inputs
to
the
PBPK/
PD
model.

While
the
volume
of
the
compartments
are
constant
over
time
periods
of
a
few
hours
or
days,
the
cardiac
output
and
alveolar
ventilation
rates
are
not
constant.
Both
the
cardiac
output
and
alveolar
ventilation
rate
increase
with
the
activity
level
of
the
individual.
As
a
result,
the
values
for
these
parameters
must
be
specified
over
time
as
a
function
of
the
activity
levels
of
the
individual
at
various
points
in
time
during
a
single
day.
To
a
lesser
extent,
the
compartment­
specific
fractions
of
cardiac
output
are
also
affected
by
level
of
activity
and
from
behaviors
such
as
eating.
The
temporal
patterns
of
these
time­
varying
parameters
are
also
highly
correlated
with
each
other.
Factors
that
influence
one
parameter
will
tend
to
influence
some
or
all
of
the
parameters.

3
b.
Kinetic,
Metabolic
and
Capacity
Parameters
The
second
type
of
information
used
in
PBPK/
PD
models
is
the
rates
of
metabolism
and
partitioning
of
the
specific
pesticides.
These
parameters
also
vary
across
individuals
as
a
function
of
age,
gender
and
genetic
variation.
LifeLine
 
software
defines
the
gender,
age,
race
and
ethnicity.
These
data
can
be
used
within
the
PBPK/
PD
model
to
select
the
values
for
these
parameters.

The
pharmacodynamic
model
for
the
N­
methyl
carbamate
pesticides
has
not
been
finalized
but
is
expected
to
include
the
concept
of
capacity
or
tolerance
to
the
effects
of
the
compounds.
Tolerance
to
the
pesticide's
effects
is
likely
to
vary
across
the
population
and
may
differ
across
ages
and
gender.
If
this
is
the
case
then
gender,
age,
race
and
ethnicity
may
be
useful
for
predicting
values
for
this
parameter
as
well.
THE
LIFELINE
GROUP
Page
13
of
57
3
c.
Data
on
Exposure
and
Dose
i.
Matching
LifeLine
 
Outputs
to
PBPK/
PD
Model
Inputs
Exposure
to
pesticides
is
a
multi­
step
process
that
takes
the
pesticides
from
the
environment
in
which
an
individual
breathes,
consumes
food
and
water
and
contact
surfaces,
ultimately,
to
the
target
organ
where
the
effects
of
interest
occur
(
EPA,
1992;
EPA,
1997).
Exposures
to
pesticides
in
the
environment
occur
by
dermal,
oral
or
inhalation
routes
of
exposure.
As
a
pesticide
passes
from
the
environment
onto
and
through
the
skin,
or
is
ingested
or
inspired
and
absorbed,
the
concept
of
exposure
becomes
a
concept
of
dose.
The
relevant
metrics
change
from
concentration
in
the
environmental
medium
to
route­
specific
contact
dose
(
on
the
skin,
in
the
gut
following
oral
exposure
or
in
the
lungs
following
inhalation
exposure)
to
absorbed
dose
(
in
the
blood,
in
individual
organs
or
at
the
tissue
level).

This
process
requires
a
clear
description
of
each
of
the
steps
on
the
process.
Such
description
requires
common
definitions
of
modeling
concepts,
terminology
and
measures
of
exposure
and
dose
(
metrics).
For
example,
dermal
exposures
can
be
described
in
terms
of:

 
Dermal
loading
(
the
average
mass
of
pesticide
in
dirt,
dust
and
oil
on
the
skin
and
the
area
over
which
it
occurs)
during
a
time
step;
 
The
average
mass
of
pesticide
in
direct
contact
with
the
skin
during
a
time
step;
 
The
mass
of
pesticide
that
is
absorbed
by
the
stratum
corneum
in
a
time
step;
 
The
mass
of
pesticide
that
is
absorbed
by
the
dermis
in
a
time
step;
or
 
The
mass
of
a
pesticide
that
reaches
the
circulating
blood
supply
in
a
time
step.

Doses
from
exposures
that
occur
by
the
oral
and
inhalation
routes
also
have
multiple
descriptive
options.
No
particular
descriptive
option
is
inherently
correct
or
incorrect.

As
the
LifeLine
 
and
PBPK/
PD
models
interface
and
the
output
of
LifeLine
 
becomes
the
input
to
the
PBPK/
PD
model,
the
values
produced
by
LifeLine
 
must
interface
correctly,
neither
presenting
a
gap
in
the
process
nor
duplicating
a
process
modeled
in
the
PBPK/
PD
model.
Thus,
for
example,
if
the
PBPK/
PD
model
includes
a
quantitative
model
of
the
process
of
dermal
absorption,
then
the
exposure
must
estimate
the
amount
of
pesticide
that
reaches
the
surface
of
the
skin
and
the
duration
of
time
over
which
the
pesticide
remains
on
the
skin.
If
the
PBPK/
PD
model
does
not
consider
the
dermal
absorption
process,
the
exposure
assessment
model
must
include
algorithms
representing
absorption
through
the
skin
to
provide
the
absorbed
dose
resulting
from
dermal
exposure.
It
THE
LIFELINE
GROUP
Page
14
of
57
is
critical
to
recognize
that
the
input
exposure
metric
may
vary
with
different
PBPK/
PD
models
and
to
have
the
capacity
to
provide
what
is
required.
More
discussion
of
the
impact
of
these
factors
on
the
interface
is
offered
in
the
last
sections
of
this
document.

This
"
matching"
of
the
metrics
and
exposure/
dose
definitions
between
linked
models
is
a
critical
issue
and
often
complicated
by
the
variable
use
of
these
terms
by
different
communities
of
scientists.
The
definitions
including
those
of
exposure
and
dose
used
in
this
report
are
presented
in
the
beginning
of
the
report
and
are
meant
to
clarify
the
concepts
discussed
in
this
report
and
be
consistent
with
definitions
used
by
OPP.
They
may
differ
from
those
used
by
other
scientific
groups.

PBPK/
PD
models
characterize
the
movement
and
transformation
of
pesticides
in
the
body.
Thus,
the
PBPK/
PD
models
require
information
on
the
mass
of
the
pesticide
that
enters
the
body.
This
information
delivered
from
exposure
models
should
include:

 
The
amount
that
enters
by
each
of
the
three
routes
of
exposure;
 
Exposure
or
dose
metric
appropriately
"
matched"
to
the
definitions
of
exposure/
dose
used
in
the
PBPK/
PD
model;
 
Exposure
histories
of
the
simulated
individuals
over
time;
 
Time
steps
appropriate
to
the
pesticide
family
(
for
N­
methyl
carbamates
this
would
be
units
smaller
than
daily);
and
 
Separate
and
correlated
estimations
of
route­
specific
exposure
for
each
pesticide
during
each
time
step.

ii.
Establishing
the
Time
Step
and
Exposure
History
The
timing
of
all
exposure
events
for
all
routes
for
all
pesticides
must
be
placed
in
a
consistent
timeframe.
The
timeframe
must
be
sufficiently
detailed
to
be
appropriate
for
the
mechanism
of
action
relevant
to
the
pesticides
under
consideration.
To
be
relevant
to
the
mechanism
of
action,
the
timeframe
must
accommodate
modeling
of
the
time
course
of
pesticide
delivery
to
the
target
tissue,
expression
of
the
toxicity
mechanism
and
mechanics
of
recovery.

The
level
of
detail
in
an
exposure
assessment
is
defined
in
terms
of
the
duration
of
the
averaging
period
of
exposure.
Historically
for
pesticide
exposure
assessment,
averaging
lifetime
exposure
periods
have
been
used
for
cancer
risk,
one
year
periods
for
chronic
effects
and
one
day
periods
for
acute
effects.

PBPK/
PD
models
also
deal
with
doses
as
a
sequence
of
exposure
events.
These
events
have
a
specific
duration
and
are
referred
to
as
"
time
steps".
This
allows
the
PBPK/
PD
models
to
predict
the
time
course
of
a
pesticide
in
the
body
as
a
result
of
ongoing
exposures.
These
sequential
measurements
represent
the
THE
LIFELINE
GROUP
Page
15
of
57
"
exposure
history"
of
the
individual.
As
discussed
above,
some
pesticides
such
as
the
N­
methyl
carbamates
exert
their
effects
via
mechanisms
requiring
short
periods
of
time.
As
a
result,
PBPK/
PD
models
require
an
exposure
history
with
time
steps
much
shorter
than
one
day.

The
optimal
duration
of
a
time
step
will
vary
with
the
mechanism
and
properties
of
the
group
of
pesticides.
To
accommodate
the
myriad
of
pesticides
and
toxicity
mechanisms
involved
in
pesticides,
models
must
at
least
be
able
to
accommodate
short
time
steps
and
be
able
to
average
those
time
steps
over
multiple
periods
to
produce
metrics
of
the
appropriate
duration.

In
each
time
step,
the
exposure
software
must
determine
the
route
specific
doses
for
each
of
the
pesticides
in
the
assessment.
If
there
were
10
pesticides
to
be
modeled,
then
each
time
step
would
have
10
doses
(
one
for
each
pesticide)
and
three
route
specific
doses
(
dermal,
oral
and
inhalation).
This
will
result
in
30
doses
per
time
step.
If
the
time
step
is
10
minutes,
then
there
will
be
144
time
steps
in
a
day
or
4,320
doses.
Since
the
software
used
in
the
OP
cumulative
assessment
only
modeled
three
one­
day
route
specific
doses
per
day,
this
shorter
time
step
approach
generates
more
than
a
thousand
fold
increase
in
the
exposure
assessment
metrics.
THE
LIFELINE
GROUP
Page
16
of
57
4.
Current
Exposure
Assessment
Models:
A
Brief
Overview
Current
exposure
software
programs
used
by
OPP/
EPA
programs
integrate
exposures
from
food,
drinking
water
and
residential
non­
dietary,
nonoccupational
sources
to
assess
acute,
short
term,
intermediate
term
and
longterm
lifetime
exposures.
All
of
the
models
produce
exposure
estimates
in
metrics
tailored
to
the
risk
characterization
methodologies
employed
by
EPA
(
such
as
the
MOE,
%
RfD
and
lifetime
cancer
risk
probabilities).
The
models
also
place
the
estimates
of
exposure
into
some
form
of
chronological
framework.
The
models
typically
model
time
steps
of
a
single
day
using
databases
conveniently
structured
over
exposure
events
expressed
in
terms
of
days
or
subunits
of
days.
This
approach
has
paralleled
the
toxicology
metrics
for
most
toxicological
endpoints
where
effect
has
been
expressed
as
a
function
of
exposure
or
dose
over
one
or
more
days.
All
of
the
programs
deal
only
with
a
single
pesticide
at
a
time.
As
previously
discussed,
this
is
true
even
for
the
programs
used
in
the
OP
cumulative
risk
assessment.
Cumulative
exposure
assessment
for
these
chemicals
was
achieved
by
first
converting
the
individual
pesticides
into
equivalent
doses
of
an
index
chemical.
The
models
provide
route
specific
exposure
metrics
that
are
used
to
support
the
route
specific
toxicology
and
risk
characterization
methodologies
such
as
the
MOE
and
%
RfD.
Finally,
in
various
ways,
the
exposure
assessment
programs
identify
some
physical
and
demographic
characteristics
of
the
simulated
individuals
in
the
defined
population
under
consideration.
Demographic
information
is
sometimes
limited
to
age
and
gender.

The
earliest
design
of
the
LifeLine
 
software
architecture
was
influenced
by
the
recognition
that
aggregate
and
cumulative
exposure
assessment
software
programs
should
be
able
to
provide
the
types
of
data
required
by
PBPK/
PD
models
including
the
essential
route
specific
exposure
estimates
since
this
application
would
be
required
in
the
future.
The
LifeLine
 
software
already
has
the
basic
modules
that
can
be
enhanced
to
provide
the
information
required
by
PBPK/
PD
models.

4
a.
Capabilities
Required
for
Supporting
PBPK/
PD
Modeling
Based
upon
the
above
sections
there
are
four
areas
where
exposure
models
must
be
enhanced
in
order
to
support
PBPK/
PD
modeling
in
the
assessment
of
cumulative
risk.

1.
The
models
must
be
restructured
to
provide
real
pesticide­
by­
pesticide
cumulative
exposure
assessment
rather
than
using
a
single
RPF
approach
to
represent
multiple
pesticides.
THE
LIFELINE
GROUP
Page
17
of
57
2.
Time
steps
must
be
redefined
to
present
exposure
metrics
as
a
function
of
hours
or
minutes
rather
than
days.

3.
The
physiological
characteristics
of
the
exposed
individual
must
accompany
the
exposure
values
for
each
time
step.

4.
Finally,
the
interface
between
the
exposure
model
and
the
PBPK/
PD
model
must
be
fashioned
whereby
the
information
is
transferred
without
losing
the
interconnections
of
multiple
pesticide,
multiple
route
exposure
values
for
a
coherent
series
of
time
steps
for
each
individual
with
the
relevant
physiological
and
demographic
identifiers.
This
interface
must
faithfully
maintain
the
continuum
from
the
media
concentration
values
to
the
target
tissue
doses
without
creating
gaps
or
overlaps,
as
previously
discussed.
These
are
additional
factors
that
must
be
addressed
in
the
design
of
the
exposure
database
if
an
acceptable
interface
is
to
be
built
between
the
models.
The
interface
is
discussed
in
more
detail
in
the
last
sections
of
this
document.

The
following
section
of
the
report
outlines
an
approach
that
achieves
these
four
goals.
THE
LIFELINE
GROUP
Page
18
of
57
5.
Development
of
a
New
Version
of
the
LifeLine
 
Software
for
the
Support
of
a
PBPK/
PD
Model
of
Cumulative
Risk
This
section
of
the
report
presents
a
detailed
description
of
how
the
current
version
of
LifeLine
 
software
(
Version
2.1)
could
be
enhanced
to
provide
the
inputs
required
by
PBPK/
PD
models
such
as
that
being
developed
by
ORD/
EPA.
LifeLine
 
software
for
exposure
and
risk
characterization
has
been
used
by
EPA,
other
international
federal
and
state
regulatory
agencies,
stakeholders,
academia,
and
research
scientists
since
its
first
release
in
late
2000.
Version
2.0
was
used
to
characterize
the
cumulative
risks
of
the
OP
pesticides
(
LifeLine
 
2002)
and
has
been
used
in
the
Agency's
Voluntary
Children's
Chemical
Evaluation
Program
(
ACC,
2003).
The
software
has
been
the
subject
of
Science
Advisory
Panel
(
SAP)
reviews
in
1999,
2000,
and
2001.1
This
section
of
the
report
begins
with
a
review
of
the
design
of
the
LifeLine
 
software's
architecture
 
the
critical
element
to
consider
for
the
modification
approach.
The
Technical
Manual
for
the
LifeLine
 
software,
available
from
the
LLG
web
site2,
provides
a
detailed
description
of
the
model's
architecture,
databases
and
operating
algorithms.
This
report
discusses
those
elements
necessary
to
consider
the
framework
for
desired
modifications
for
supporting
PBPK/
PD
models
of
cumulative
risk
and
details
of
the
modifications.
The
section
concludes
with
a
discussion
of
how
to
convey
the
data
from
LifeLine
 
to
the
PBPK/
PD
model.

5
a.
The
Current
Framework
of
LifeLine
 
Software
LifeLine
 
software
was
designed
from
its
very
beginning
to
support
assessments
of
cumulative
risk
(
LLG,
1999,
LLG
2004).
The
design
of
the
software
is
consistent
with
Agency
guidance
for
performing
aggregate
and
cumulative
assessment
(
EPA,
2001;
2002a).

i.
POM
Models
LifeLine
 
belongs
to
a
class
of
programs
called
Person
Oriented
Models
(
POM).
These
models
place
the
design
focus
on
the
individual
receiving
the
exposure
rather
than
the
exposure
sources
(
Price
et
al.
2003a).
Figure
1
is
a
flowchart
of
the
basic
components
of
a
POM.
These
models
begin
by
defining
the
individual's
characteristics.
These
characteristics
are
those
aspects
of
the
individual
that
1
All
versions
of
LifeLine
 
software
are
made
publicly
available
to
all
stakeholders
without
charge
by
LLG.
LLG
is
a
501
(
c)(
3)
not­
for­
profit
corporation
created
for
the
development
and
public
dissemination
of
risk
assessment
software
and
related
materials.
2
www.
TheLifeLineGroup.
org
THE
LIFELINE
GROUP
Page
19
of
57
influence
the
probability
of
occurrence
of
an
exposure
opportunity
and
the
magnitude
and
duration
of
the
exposure
resulting
from
that
occurrence.
The
characteristics
could
include
the
individual's
body
weight,
diet
within
a
given
time
period,
activity
patterns,
residence,
location
(
region
of
the
country)
and
season
of
the
year.
LifeLine
 
utilizes
a
library
of
person­
oriented
databases
to
create
the
distributions
of
parameter
values.
This
provides
a
framework
to
express
interindividual
variability
for
the
population
of
interest.
THE
LIFELINE
GROUP
Page
20
of
57
Assign
individual's
characteristics
Update
exposure
history
Determine
dose
Source­
to­
dose
model
Source
Exposed?

More
sources?
Yes
No
Yes
No
Save
exposure
history
Another
individuals?
Yes
No
Stop
Start
Exposure
Event
Loop
Individual
Loop
Figure
1.
Person­
Based
Modeling
of
Inter­
individual
Variation
in
Dose
THE
LIFELINE
GROUP
Page
21
of
57
Once
the
individual's
characteristics
are
specified,
the
model
enters
the
Exposure
Event
Loop.
In
this
loop,
the
POM
systematically
determines
if
the
individual
is
exposed
to
any
of
the
possible
sources.
Examples
of
possible
sources
are
the
diet
or
a
series
of
microenvironments
that
the
individual
encounters
within
the
time
step.
If
the
model
determines
that
an
exposure
opportunity
has
occurred,
the
model
runs
the
appropriate
source­
to­
dose
model
and
calculates
the
magnitude
of
the
exposure
or
dose.
The
series
of
exposure
opportunities
for
a
given
individual
are
consistent
with
the
characteristics
of
the
individual,
the
point
in
time
and
with
each
other.
This
is
achieved
by
defining
each
exposure
opportunity
based
on
the
characteristics
of
the
simulated
person,
established
at
the
beginning
of
the
process.
This
yields
temporal
consistency
in
the
estimate
of
aggregate
doses
for
a
pesticide.
Note
that
if
such
consistency
is
not
established
within
a
model,
the
inconsistencies
magnify
as
the
model
considers
multiple
pesticides
with
route­
specific
exposures.

Once
the
model
has
calculated
the
exposures
or
doses
that
result
from
each
source,
the
information
can
be
saved
as
part
of
an
exposure
history
for
that
individual
and
that
day.
This
exposure
history
can
be
used
in
a
variety
of
subsequent
analyses3.
The
POM
then
repeats
the
process
for
other
individuals,
assigning
different
characteristics
by
re­
sampling
the
distributions
of
interindividual
variation.
This
modeling
of
multiple
individuals
happens
in
the
Individual
Loop.

LifeLine
 
Version
2.1
currently
creates
a
longitudinal
model
of
individuals'
daily
exposures
over
an
85­
year
life
span.
The
model
provides
route
specific
estimates
of
daily
doses
from
diet,
drinking
water
and
residential
uses
of
pesticides4.
Figure
2
presents
the
basic
design
of
Version
2.1.

LifeLine
 
,
as
with
all
POM
models,
begins
with
a
definition
of
the
exposed
individual
and
his
or
her
characteristics
on
that
day.
As
part
of
this
definition,
LifeLine
 
assigns
the
individual
a
record
from
the
US
Department
of
Agriculture's
Continuing
Survey
of
Food
Intakes
by
Individuals
(
CSFII)
and
a
record
from
EPA's
National
Human
Activity
Pattern
Survey
(
NHAPS).
The
selection
of
the
records
is
based
on
a
consideration
of
factors
such
as
the
individual's
age,
the
season
of
the
year
and
day
of
the
week.
5
The
model
then
cycles
through
each
food
consumed
on
that
day
using
the
Dietary
Exposure
Event
Loop.
The
exposure
from
the
consumption
of
each
food
item
is
determined
based
on
the
amount
of
the
food
consumed
and
the
residues
possibly
in
or
on
the
food.
The
same
process
is
used
for
each
instance
when
drinking
water
is
consumed.
The
exposures
from
all
consumption
events
are
summed
to
yield
the
3
Such
as
determining
the
highest
day's
exposure
for
an
individual
in
a
season
or
a
year,
or
to
calculate
the
average
daily
dose
over
a
season,
year
or
other
specified
time
period.
4
Residential
includes
exposure
to
pesticides
during
use
and
from
post­
application
exposures.
Sources
of
pesticides
include
food,
drinking
water,
indoor
products,
outdoor
products,
public
health
exposures
(
vector
control)
and
residues
on
golf
courses.
5
Details
about
data
binning
and
construction
of
the
distributions
of
values
available
for
this
process
are
given
in
the
Technical
Manual
for
LifeLine
 
and
available
from
the
LLG
web
site.
THE
LIFELINE
GROUP
Page
22
of
57
total
dietary
exposure
presented
by
the
exposure
opportunities
(
eating
events)
on
that
day.

Once
the
model
has
completed
the
calculation
of
the
dietary
exposure,
it
moves
to
the
second
exposure
event
loop,
the
Non­
dietary
Exposure
Event
Loop.
In
this
loop,
the
model
cycles
through
each
microenvironment
and
activity
in
the
NHAPS
record.
If
a
pesticide
residue
is
present
in
the
microenvironment
and
the
individual
interacts
with
the
residue
(
as
represented
on
the
NHAPS
record),
then
the
model
determines
the
exposure
presented
and
the
route
(
or
routes)
by
which
the
exposures
are
presented.
After
the
evaluation
of
the
last
microenvironment
and
activity
portrayed
in
the
NHAPS
record,
LifeLine
 
totals
the
exposure
presented
from
each
microenvironment
by
each
of
the
three
routes
of
exposure
to
provide
daily
route
specific
estimates
and
the
daily
aggregate
exposure.
These
calculated
exposure
metrics
are
saved
for
use
in
subsequent
analyses.

Once
the
calculations
are
saved,
the
model
moves
to
the
next
day
in
the
individual's
life.
This
occurs
in
the
"
day
loop".
The
day
loop
begins
by
updating
the
characteristics
of
the
individual,
pesticide
usage
probabilities
and
the
residues
in
the
individual's
environments.
The
two
exposure
event
loops
are
then
repeated.
This
process
continues
until
the
exposure
period
of
interest
to
the
user
is
complete.
Version
2.1
will
simulate
an
individual
for
durations
of
one
year
to
85
years
The
entire
process
is
repeated
for
every
individual
assigned
to
the
population
of
interest.
LifeLine
 
accommodates
calculations
for
up
to
100,000
individuals
within
a
population
of
interest
(
the
general
US
population
or
a
defined
subgroup
of
that
population)
as
specified
by
the
user.

This
LifeLine
 
framework
provides
many
functional
capabilities
necessary
to
calculate
the
exposure
metrics
relevant
to
PBPK/
PD
models.
These
capabilities
include:

o
Definition
of
the
exposed
individual
in
terms
of
o
Race;
o
Ethnicity;
o
Physiology
o
Definition
of
the
characteristics
of
the
individual's
source
of
exposure
o
Sources
of
exposure
o
Route
specific
doses
o
Definition
of
exposure
history
and
o
Definition
of
temporal
changes
in
demographic
and
physiological
parameters
and
linkage
of
these
parameters
to
the
activity
profiles,
exposure
opportunities,
and
estimates
of
exposure
and
dose
(
magnitude,
route)
for
a
time
step
of
a
day.
These
capabilities
provide
a
robust
starting
point
for
the
design
of
a
new
version
of
LifeLine
 
that
can
meet
the
needs
of
the
PBPK
/
PD
models.
THE
LIFELINE
GROUP
Page
23
of
57
Update
exposure
History
Determine
oral
dose
Residue
data
Yes
No
Yes
No
Figure
2.
Flowchart
for
LifeLine
 
Software
Version
2.1
Yes
No
Yes
No
Determine
route
specific
doses
Start
Dietary
Non
Dietary
Save
exposure
Update
characteristics
of
source
No
Yes
Stop
Select
activity
pattern
record
Residue
in
first/
next
food?

Residue
in
first/
next
microenv.?

More
microenv?
Residue
data
data
Update
individual's
characteristics
Assign
birth
and
fixed
characteristics
Select
CSFII
record
Another
day?
Another
individual?
No
Yes
Day
Loop
Individual
Loop
Exposure
Event
Loop
Diet
Exposure
Event
Loop
Non­

Diet
Update
exposure
History
More
food
items?
THE
LIFELINE
GROUP
Page
24
of
57
5
b.
Changing
the
Time
Step
from
One
Day
to
a
Shorter
Time
Interval
As
discussed
in
the
prior
section,
while
LifeLine
 
calculates
the
exposure
from
multiple
exposure
events
over
the
course
of
a
day,
the
program
sums
the
doses
from
the
separate
events
to
yield
a
daily
exposure
value.
With
this
approach,
the
time
step
for
the
exposure
history
of
the
individual
is
one
day.
The
LifeLine
 
framework
can
also
support
estimates
of
exposure
for
periods
shorter
than
a
day.
As
discussed
below,
the
LifeLine
 
framework
can
accommodate
as
short
a
time
step
as
the
analyst
desires.
The
practical
lower
limit
on
the
interval
for
the
time
step
is
imposed
however,
by
the
structure
of
the
databases
on
which
the
exposure
model
operates.
Temporal
information
directly
or
indirectly
supplied
by
the
data
sets
accommodates
daily,
hourly
and
to
a
limited
degree
sub­
hourly
time
steps.
Rappaport
(
2003),
in
a
review
of
the
work
of
Roach
(
1977),
points
out
that
many
of
his
early
concepts
have
been
borne
out
by
more
recent
research.
In
the
area
of
air
sampling
in
the
workplace,
Roach
discussed
the
appropriate
duration
of
exposure
measurement
for
fast
acting
chemicals,
i.
e.
those
which
have
a
T1/
2
in
the
order
of
0.5
to
1
hr.
Too
long
a
sampling
time
(
exposure
metric)
would
not
accurately
reflect
the
maximum
tolerated
concentration.
He
showed
that
selection
of
an
averaging
of
0.3T1/
2
avoids
this
problem.
Since
the
T
1/
2
for
orally
dosed
carbamates,
as
represented
by
carbaryl,
is
in
the
range
of
0.5
hr,
and
the
data
support
peak
exposures
rather
than
AUC,
Roach's
approach
is
applicable,
and
a
sampling
time
of
10
minutes
would
be
appropriate.
Additionally,
there
are
data
in
CSFII
for
consumption
of
a
snack,
and
LLG
believes
that
inference
of
a
10
minute
time
interval
for
eating
duration
is
reasonable.

The
following
section
presents
a
discussion
about
the
compromises
and
derivations
associated
with
such
data
structuring.
Shorter
time
steps
appear
to
be
impractical
given
the
presently
available
databases.

The
tasks
necessary
to
create
shorter
time
steps
are
as
follows:

 
Create
a
user
interface
to
define
the
duration
of
the
exposure
history
and
the
duration
of
the
time
step;
 
Restructure
the
NHAPS
and
CSFII
data
to
appropriate
time­
related
data
values.
 
Reconcile
the
conflicts
between
the
NHAPS
and
CSFII
temporal
information
units;
 
Revise
the
day
loop
framework;
and
 
Track
and
manage
the
exposure
assessment
outputs.
THE
LIFELINE
GROUP
Page
25
of
57
Design
of
User
Interface
for
Exposure
History
and
Time
Step
The
toxicology
of
the
N­
methyl
carbamate
pesticides
presents
several
issues
that
influence
the
design
and
requirements
of
the
PBPK/
PD
models
of
cumulative
risk.
These
pesticides
rapidly
bind
to
receptors
in
the
target
tissues.
The
parent
pesticides
are
rapidly
metabolized
and
the
receptors
are
rapidly
cleared.
Thus,
the
mechanism
by
which
the
pesticides
cause
the
effects
on
which
the
cumulative
risk
is
based
and
the
mechanism
of
recovery
from
the
toxic
effects
occur
over
short
time
intervals.
If
an
adequate
number
of
contiguous
time
steps
occur
with
no
new
exposure
events,
complete
recovery
of
the
target
tissue
is
expected.
If
no
exposure
opportunity
is
introduced
during
a
night's
sleep,
the
effects
of
this
class
of
pesticides
will
not
persist
into
the
next
day.
This
implies
the
assessment
of
cumulative
risk
for
this
group
of
pesticides
can
be
limited
to
a
single
day.
However,
the
rapid
mechanisms
of
receptor
binding
and
clearance
will
be
better
reflected
in
a
model
employing
a
finer
time
step
scale
than
a
single
day.

As
previously
explained,
the
scenario
presented
by
these
N­
methyl
carbamate
pesticides
is
not
universally
applicable
to
all
classes
of
pesticides.
Therefore,
this
report
will
outline
an
approach
that
can
be
applied
to
longer
periods
than
a
single
day
and
where
the
duration
of
the
time
step
can
vary
from
durations
of
a
few
minutes
to
one
day.
Therefore,
the
new
version
of
LifeLine
 
will
allow
the
user
to
specify
the
duration
of
the
exposure
history
and
the
duration
of
the
time
step
in
the
exposure
history.

Restructure
the
NHAPS
and
CSFII
Data
to
Appropriate
Time­
Related
Data
Values
There
are
two
databases
in
LifeLine
 
that
provide
temporal
information
with
scales
of
less
than
a
day,
the
CSFII
and
the
NHAPS.
The
CSFII
provides
daily
dietary
records
for
each
person
with
detailed
temporal
information
linked
to
each
eating
event
of
the
day.
Each
eating
event
of
the
day
is
recorded
noting
the
hour
and
minute
of
the
eating
occasion,
the
foods
eaten
within
that
event,
the
location
of
the
eating
event,
the
amount
of
each
food
eaten
and
the
demographic
information
about
the
consumer.
In
addition,
the
consumer
can
classify
the
event
as
a
major
meal
(
breakfast,
lunch,
dinner)
or
as
a
snack
event.

The
CSFII
does
not
include
a
measure
of
the
duration
of
the
eating
event.
Using
the
consumer's
categorization
of
the
eating
events
and
a
system
of
classification
by
the
number
of
foods
and
food
forms
involved
with
each
event,
the
duration
of
the
eating
event
can
be
inferred.
For
example
a
snack
event
involving
few
foods
and
food
forms
can
be
reasonably
assigned
a
duration
of
10
minutes.
For
a
major
meal
event,
the
eating
occasion
can
be
assigned
a
duration
of
20
or
30
minutes.
As
a
result,
each
eating
event
can
be
assigned
to
1,
2,
or
3
10­
minute
time
steps.
THE
LIFELINE
GROUP
Page
26
of
57
CSFII
records
do
not
record
the
order
of
the
consumption
of
the
foods
involved
in
the
eating
event
or
the
time
that
each
item
is
consumed.
It
is
reasonable
to
assume
that
foods
are
not
eaten
sequentially
but
rather
in
a
pattern
that
is
consistent
with
personnel
preference.
We
consider
this
to
be
a
random
pattern
of
eating
food
items
within
the
eating
event
over
the
time
steps
assigned
for
each
eating
event.

The
activity
patterns
in
NHAPS6
are
defined
in
terms
of
single
minutes;
however,
the
precision
of
the
survey
instrument
and
accuracy
at
that
level
of
detail
are
questionable.
It
is
widely
recognized
that
survey
diaries
can
only
capture
the
gross
temporal
patterns
of
activities
and
locations
particularly
when
an
adult
fills
out
the
diary
for
a
child
(
Elgethun
et
al.
2003).
Minute­
by­
minute
precision
for
parameters
such
as
a
location
and
activity
are
feasible
with
the
survey
instruments
employed
in
such
surveys.
Error
increases
with
the
decrease
in
the
size
of
the
time
step
when
these
data
are
used.

Based
on
reasonable
inference
from
CSFII,
LLG
believes
that
it
is
not
advisable
to
develop
time
steps
shorter
than
10
minutes
for
the
exposure
models.
The
limitations
in
the
definition
of
the
duration
of
meals,
the
uncertainty
in
the
order
of
the
consumption
of
food
items
and
the
known
limitations
of
24
hour
recall
diaries
make
estimation
of
shorter
time
frames
highly
uncertain.

Reconcile
the
Temporal
Information
Structures
for
Dietary
and
Non
Dietary
Activities
As
presented
in
the
previous
discussions
and
discussed
in
detail
in
the
LifeLine
 
Software
Technical
Manual,
the
exposure
event
characteristics
for
a
simulated
individual
are
drawn
from
databases
such
as
the
CSFII,
NHAPS
and
others.
Thus,
for
a
given
simulated
individual,
the
exposure
information
will
be
drawn
from
records
on
individuals
that
have
the
same
key
characteristics
of
the
simulated
individual.
While
the
records
are
derived
from
similar
individuals,
they
still
remain
records
from
different
individuals.
As
a
result,
there
is
a
potential
for
a
disagreement
in
the
hourly
data
in
these
records.
For
example,
the
dietary
exposure
data
(
taken
from
the
CSFII
record)
may
indicate
an
eating
event
at
the
same
time
as
the
NHAPS
indicates
the
child
is
napping.

Previous
assessments
calculated
daily
aggregate
and
cumulative
exposures.
Such
estimates
were
not
seriously
impacted
by
this
contradiction
since
the
total
daily
dietary
and
non­
dietary
exposures
were
not
affected
by
the
exact
time
of
the
exposure
event.
However,
when
the
goal
is
to
calculate
the
cumulative
exposure
over
a
10­
minute
period,
these
chronological
differences
must
be
reconciled.

6
Data
from
NHAPS
has
been
combined
with
other
survey
data
to
form
the
Consolidated
Human
Activity
Database
(
CHAD).
The
revised
version
of
LifeLine
 
should
consider
using
this
larger
database;
however,
the
data
from
CHAD
suffers
from
the
same
lack
of
precision
as
NHAPS.
THE
LIFELINE
GROUP
Page
27
of
57
There
are
a
number
of
ways
to
handle
this
problem.
For
example,
the
eating
events
can
be
given
precedence
over
the
activities
in
the
NHAPS
records.
In
this
case,
preference
for
the
time
spent
eating
meals
or
snacks
in
the
time
step
records
would
be
given
to
the
CSFII
schedule.
A
second
longer
term
option
is
to
move
the
assessment
to
databases
that
contain
both
dietary
and
non­
dietary
activity
information,
such
as
the
recent
(
2001­
2002)
National
Health
and
Nutrition
Evaluation
Survey
and
the
Department
of
Health
and
Human
Service­
US
Department
of
Agriculture's
Dietary
Survey
Integration.
Other
options
are
also
possible.

The
selection
of
the
final
approach
will
depend
on
a
number
of
considerations.
That
discussion
and
this
particular
operation
is
not
the
subject
of
this
report
but
will
need
to
be
addressed
prior
to
the
development
of
the
revised
software.

i.
Proposed
Approach
LifeLine
 
will
be
modified
to
allow
the
user
to
define
the
populations
to
be
simulated,
the
duration
of
the
exposure
history
(
from
8
hours
to
one
year)
and
the
duration
of
the
time
step
from
10
minutes
to
24
hours.
The
user
can
specify
that
separate
exposure
histories
be
modeled
for
one
individual
at
different
ages.
Once
this
information
is
entered,
the
model
will
begin
the
first
exposure
history
for
an
individual.
Figure
3
presents
a
flow
chart
explaining
the
process
of
modeling
an
individual's
exposure
history.

The
process
starts
by
defining
the
individual's
characteristics
(
age,
gender
and
other
factors).
Based
on
these
characteristics
the
model
pulls
the
CSFII
and
NHAPS
records
for
the
period
of
time
covered
by
the
exposure
history.
In
the
case
of
N­
methyl
carbamate
pesticides
where
the
duration
of
the
exposure
histories
is
24
hours,
only
one
CSFII
and
one
NHAPS
record
will
be
pulled.
These
records
are
then
converted
into
a
series
of
time
step­
specific
records.
Each
record
is
assigned
a
specific
start
and
end
time.
If
the
duration
of
the
time
step
that
is
selected
is
10
minutes
there
will
be
144
time
steps
created.
The
record
will
define
the
microenvironment
and
activity
for
that
time
step
and
the
foods
consumed
during
that
period7.
When
producing
these
records,
the
software
will
reconcile
any
conflicts
between
the
eating
event
times,
activities
and
microenvironment
locations.

As
part
of
this
process,
the
drinking
water
consumption
will
be
assigned
to
various
time
steps.
The
CSFII
record
does
not
record
the
time
of
drinking
water
consumption.
However,
it
does
indicate
the
source
of
the
water.
Water
that
is
consumed
as
part
of
a
food
can
be
assigned
to
one
or
more
of
the
eating
events.

7
The
individual
will
be
assumed
to
spend
all
10
minutes
in
one
microenvironment
performing
one
macro
activity.
If
the
duration
of
the
time
set
is
longer
than
10
minutes
then
the
individual
will
be
assigned
multiple
activities
and
microenvironments.
THE
LIFELINE
GROUP
Page
28
of
57
Other
water
consumption
can
be
assigned
as
occurring
over
the
course
of
the
day.

Once
the
time
step
specific
records
are
created,
LifeLine
 
selects
the
first
time
step
record
in
the
individual's
exposure
history
and
uses
the
information
on
the
foods
consumed,
the
location
(
microenvironment)
and
activity
to
determine
the
route
specific
exposures
for
each
pesticide.
The
model
then
cycles
through
each
of
the
time
steps
using
the
Time
Step
Loop.
THE
LIFELINE
GROUP
Page
29
of
57
Non
Dietary
Yes
No
Yes
No
No
Yes
No
Determine
route
specific
doses
Dietary
No
Yes
Stop
Residue
in
first/
next
food?

Residue
in
microenv.?
Update
individual's
characteristics
Assign
individual's
characteristics
Another
exposure
history?
Exposure
History
Loop
Select
the
first/
next
time
step
record
More
food
items?
Update
characteristics
of
sources
Save
data
on
individual's
exposures
Update
exposure
history
Residue
data
Residue
data
More
time
steps?
Time
Step
Loop
Start
Select
activity
pattern(
s)
and
CSFII
record(
s)

Convert
the
records
into
144
10­
min
records
of
behavior
during
the
time
step
Update
exposure
history
Figure
3.
Flow
Chart
for
Modeling
Exposure
Histories
with
a
10
Minute
Time
Step
Determine
oral
dose
Yes
THE
LIFELINE
GROUP
Page
30
of
57
The
exposures
to
the
pesticides
that
are
presented
at
each
time
step
are
saved
as
a
series
of
exposures
and
are
linked
to
the
times
when
they
occurred.
The
data
are
saved
and
eventually
exported
to
the
PBPK/
PD
models
as
the
individual's
exposure
history.

LifeLine
 
models
individuals'
entire
lifetimes.
If
there
is
a
need
to
investigate
exposures
that
occur
across
age
ranges
(<
1,
1­
2,
3­
5,
etc),
the
software
will
create
an
exposure
history
for
an
individual
at
one
age,
then
model
the
individual's
growth
and
when
the
individual
reaches
the
next
age
range
of
interest,
model
an
additional
exposure
history
for
that
age
range.
In
this
way,
multiple
sets
of
age
specific
exposure
histories
for
a
population
can
be
created
in
one
model
run.
This
process
is
performed
in
the
Exposure
History
Loop.

5
c.
Modeling
Multiple
Pesticides
One
of
the
requirements
for
the
cumulative
assessment
is
the
ability
to
model
concurrent
exposures
from
different
pesticides
within
the
family
of
pesticides
in
the
cumulative
exposure
assessment.
Up
to
this
time,
neither
LifeLine
 
nor
any
other
pesticide
exposure
model
has
attempted
to
model
the
concurrent
exposures
of
multiple
pesticides.
Moving
from
a
model
of
an
exposure
assessment
for
a
single
pesticide
to
an
exposure
assessment
model
of
multiple
pesticides
introduces
three
new
requirements.

1.
The
first
is
obtaining
data
that
accounts
for
the
correlation
between
the
occurrences
of
multiple
pesticide
residues
in
the
exposure
sources
(
i.
e.,
co­
occurrence
of
multiple
pesticide
residues
on
the
same
apple).

2.
The
second
is
the
modification
of
the
software
to
separately
calculate
and
track
the
route
specific
exposures
for
each
of
the
pesticides.

3.
The
third
task
requirement
is
the
management
of
the
increased
number
of
outputs
generated.

i.
Considerations
of
Residue
Data
Appropriate
for
Use
in
Modeling
Concurrent
Exposures
of
Multiple
Pesticides
Residue
Data
for
Food
and
Drinking
Water
Sources
As
discussed
in
the
Agency's
assessment
of
OP
pesticides,
the
collection
of
residue
data
cannot
be
based
on
independent
studies
for
each
pesticide
(
EPA,
2002c).
The
probability
of
occurrence
of
a
given
pesticide
and
the
magnitude
of
that
pesticide
on
a
crop
commodity
is
correlated
to
the
co­
occurrence
of
residues
of
other
pesticides
on
that
same
crop
commodity.
Since
pesticide
use
in
agriculture
is
related
to
seasonal
pest
pressures,
crop
treatment
traditions,
THE
LIFELINE
GROUP
Page
31
of
57
employment
of
competitive
pest
control
methods
and
pesticide
cost
factors
(
as
well
as
other
market­
related
factors),
occurrence
of
multiple
pesticide
residues
on
given
crop
commodities
are
related
in
terms
of
probability
of
occurrence
and
in
terms
of
magnitude
of
the
residues.

In
the
current
versions
of
LifeLine
J
software,
residue
distributions
can
be
applied
either
at
the
crop
level
or
to
the
CSFII
food
level.
Residue
data
relevant
to
the
crop
commodities
are
applied
to
the
crop
level
interface
along
with
relevant
processing
factor
information.
Residue
data
relevant
to
the
foods
as
eaten
can
be
applied
to
a
listing
of
foods
utilized
in
the
food
consumption
database.

One
solution
is
to
rely
on
contemporary
survey
data
where
every
pesticide
included
in
a
cumulative
risk
assessment
is
measured
concurrently
in
a
sample
of
a
RAC.
Such
surveys
capture
the
correlations
between
the
residues.
In
the
OP
cumulative
dietary
risk
assessment
EPA
used
residue
monitoring
data
collected
by
the
United
States
Department
of
Agriculture's
Pesticide
Data
Program
(
USDA­
PDP)
supplemented
with
information
from
the
Food
and
Drug
Administration
Center
for
Food
Safety
and
Applied
Nutrition
(
FDA/
CFSAN)
monitoring
data
(
EPA,
2002c).
For
drinking
water,
data
on
concurrent
residue
levels
of
pesticides
data
were
derived
using
regional
use
data
and
the
PRZM
and
EXAMS
models.
These
types
of
data
are
expected
to
be
used
for
the
assessment
of
N­
methyl
carbamate
pesticides.

It
should
be
recognized
that
this
approach,
using
survey
data,
is
limited
to
retrospective
assessments
with
contemporary
data
applicable
to
all
commodities
involved
in
the
risk
assessment,
wherein
all
uses
of
registered
pesticides
within
the
family
have
been
included
in
the
assays.
This
approach
cannot
be
used
for
the
prospective
evaluation
of
new
pesticides
and
modification
of
use
patterns
for
existing
registered
products
for
which
EPA
considers
registration
and
tolerance
setting.
Challenges
for
the
regulator
include
situations
when
there
is
a
paucity
of
acceptable
monitoring
data,
when
the
data
are
not
contemporary
or
where
not
all
crop
uses
are
accounted
for
in
the
residue
survey
data.
THE
LIFELINE
GROUP
Page
32
of
57
Use
of
the
Model
for
Prospective
Evaluation
of
New
and
Existing
Registered
Products
Databases
exist
which
detail
the
co­
occurrence
of
multiple
pesticides
on
foods,
water
and
other
media
that
may
provide
opportunities
for
simultaneous
exposure
to
humans
during
a
single
time
interval.
The
residues
from
specific
pesticides
as
well
as
the
co­
occurrence
probability
are
in
many
instances
a
part
of
these
databases
or
can
be
derived
from
the
databases.
There
are
multiple
ways
to
assign
these
residue
values
to
the
exposure
assessment
parameters.
Some
approaches
may
yield
an
underestimation
of
the
true
exposure;
some
may
yield
an
overestimation
of
the
true
exposure.
Scenarios
which
must
be
considered
when
choosing
residue
data
for
the
analysis
input
(
for
multiple
pesticide
cumulative
exposure
assessment)
include:

1.
Registrant
makes
a
claim
of
reduced
risk.
Issue:
Is
the
risk
profile
actually
reduced
for
the
resulting
cumulative
risk
assessment
when
the
new
cumulative
residue
profile
is
constructed?

2.
A
new
product
is
coming
onto
the
market.
Issue:
What
are
the
displacement
curves
in
the
cumulative
residue
profile
for
the
new
product
and
existing
products
for
the
same
uses?

3.
Regulatory
action
is
taken
to
cancel
one
or
more
pesticides
in
a
family
of
pesticides.
Issue:
since
the
pest
pressure
remains,
how
do
the
remaining
pesticides
fill
the
void
and
what
is
the
resulting
cumulative
exposure
assessment?

4.
Some
crops
within
the
pesticide/
crop
matrices
are
imported
during
key
seasons.
Issues:
What
is
the
seasonality
of
the
residue
profiles
and
how
are
the
differences
between
pesticide
use
practices
in
the
US
versus
those
in
the
exporting
countries
taken
into
consideration?

Since
the
residue
profile
(
occurrence
and
magnitude
of
use)
for
any
one
pesticide
is
correlated
with
the
residue
profiles
of
other
co­
occurring
pesticides
registered
for
similar
uses,
one
cannot
simply
add
a
new
pesticide
profile
into
an
existing
residue
survey.
Likewise,
one
cannot
simply
extract
a
particular
pesticide
profile
from
the
residue
survey
with
an
array
of
pesticide
residues.
Addition
or
removal
of
a
pesticide
or
some
of
its
uses
from
commerce
can
result
in
changes
in
the
market
share
of
the
remaining
pesticides
in
the
group.
These
changes
may
partially
or
completely
compensate
for
the
loss
of
the
extracted
pesticide.

In
the
cumulative
assessment
for
N­
methyl
carbamate
pesticides,
each
pesticide's
exposure
profile
will
be
maintained
separately.
Because
of
this
requirement,
the
organization
of
the
residue
data
will
require
a
different
and
more
complex
structure
than
earlier
single
pesticide
assessments.
In
single
pesticide
assessments,
data
on
residues
are
entered
on
the
commodity
or
food
form
level
THE
LIFELINE
GROUP
Page
33
of
57
using
an
input
table.
In
cumulative
assessments,
the
data
will
be
entered
in
a
manner
that
maintains
a
link
between
the
specific
residue
levels
that
occurred
in
a
specific
sample
from
the
survey,
thus
preserving
the
information
on
cooccurrence

An
approach
that
achieves
this
goal
is
a
two
dimensional
array
that
lists
each
residue
of
the
pesticides
in
the
individual
samples
in
the
database.
Table
1
presents
an
example
of
how
such
data
can
be
organized
in
the
input
files
for
the
dietary
simulation.
In
this
example,
there
are
10
residue
values
for
carrots
for
each
of
six
co­
occurring
pesticides.
This
structure
will
be
repeated
for
each
of
the
commodities
in
the
residue
database.

Using
this
structure,
the
dietary
software
will
select
a
"
column
of
data"
for
a
single
sample,
such
as
column
5
in
the
above
table.
Values
in
this
column
will
be
used
to
represent
each
pesticide's
residue
in
a
food
containing
carrots,
as
the
possible
residues
in
CSFII
foods
are
calculated.
In
the
simplest
case
of
eating
a
raw
carrot,
the
dietary
exposures
from
the
six
pesticides
would
be
equal
to
the
amount
of
carrot
consumed
times
each
of
the
respective
six
residue
levels.
This
data
format
can
also
apply
to
blended
commodity
scenarios,
where
the
mean
value
of
each
pesticide
would
be
calculated
(
mean
of
each
row)
to
represent
the
pesticides'
residue
values.
This
is
in
keeping
with
the
EPA
policy
for
dealing
with
blended
commodities.

Since
the
database
of
residues
is
at
the
commodity
level,
the
raw
data
will
not
reflect
the
residues
at
the
"
Food
Form"
level.
To
define
the
residues
at
the
food
form
level,
the
user
will
have
to
modify
each
of
the
pesticides
to
reflect
how
the
residues
are
likely
to
change
with
processing.
This
modification
could
be
accomplished
by
entering
the
data
at
the
food
form
level
or
by
using
an
expanded
set
of
processing
factors
within
the
LifeLine
 
Food
Residue
Translator.
Details
on
this
program
module
are
given
in
the
Technical
Manual.
Table
1.
Example
of
the
Structure
of
Residue
Data
for
One
Commodity:
Carrots
Crop
Group
Commodity
Samples
Code
Name
Code
Name
Pesticide
1
2
3
4
5
6
7
8
9
10
1
Root
And
Tuber
Vegetables
780
Carrot
Pesticide
A
0.0001
0.0001
0.002
0.0001
0.005
0.0001
0.005
0.002
0.0001
0.005
1
Root
And
Tuber
Vegetables
780
Carrot
Pesticide
B
0.0001
0.0003
0.0001
0.0007
0.0001
0.0007
0.0001
0.0001
0.0007
0.0001
1
Root
And
Tuber
Vegetables
780
Carrot
Pesticide
C
0.0001
0.0001
0.0034
0.0001
0.0001
0.0001
0.0001
0.0034
0.0001
0.0001
1
Root
And
Tuber
Vegetables
780
Carrot
Pesticide
D
0.0001
0.002
0.0001
0.0002
0.0004
0.0002
0.0004
0.0001
0.0002
0.0004
1
Root
And
Tuber
Vegetables
780
Carrot
Pesticide
E
0.0001
0.0001
0.002
0.0001
0.0001
0.0001
0.0001
0.002
0.0001
0.0001
1
Root
And
Tuber
Vegetables
780
Carrot
Pesticide
F
0.002
0.002
0.002
0.002
0.002
0.002
0.002
0.002
0.002
0.002
THE
LIFELINE
GROUP
Page
34
of
57
Data
for
Residential
Assessment
The
residential
exposure
assessment
must
also
deal
with
the
issues
of
correlation
of
residues
for
multiple
co­
occurring
pesticides.
Residue
profiles
(
occurrence
probability
and
magnitude)
of
different
pesticides
in
an
individual's
home
are
correlated
because
of
homeowner
pest
management
practices.
If
homeowners
control
a
pest
by
using
a
single
product,
they
are
not
likely
to
then
use
a
second
competing
product.
As
a
result,
the
presence
of
residues
of
competing
products
is
expected
to
be
negatively
correlated.
In
other
scenarios
where
homeowners
could
be
anticipated
to
use
multiple
products
(
a
flea
bomb
for
a
home,
a
spray
for
pet
bedding
and
a
collar
for
the
pet)
certain
products
may
be
positively
correlated.

The
pesticide
exposure
profiles
experienced
by
an
individual
also
may
be
correlated
over
time.
If
a
pesticide
is
used
in
a
home
on
one
day,
the
residues
will
persist
over
time.
Further,
a
single
product
(
large
spray
can)
may
be
applied
multiple
times
in
a
season
to
control
pests.
Thus,
use
of
a
specific
product
in
a
home
on
one
day
may
imply
repeated
exposures
to
the
same
pesticide
over
time
rather
than
exposures
to
competing
pesticide
products.

In
the
LifeLine
 
software,
residential
exposure
assessments
are
based
on
data
on
pesticide
use
and
studies
of
the
exposure
that
will
occur
if
a
product
containing
the
pesticide
is
used.
The
probability
of
using
a
pesticide
was
modeled
by
decomposing
the
process
into
two
steps.
First,
what
is
the
probability
of
needing
to
control
a
pest,
and
then
given
that
a
person
is
controlling
a
pest,
what
is
the
probability
of
using
a
product
containing
a
specific
pesticide?

Data
are
available
on
the
first
step
of
this
process
in
the
form
of
surveys
of
"
pest
pressure".
Pest
pressure
is
defined
as
the
frequency
that
a
specific
pest
(
insect,
weed,
fungus,
etc.)
is
treated
in
a
specific
microenvironment
in
the
home
or
yard.
Data
on
pest
pressure
are
taken
from
the
1991
National
Home
and
Garden
Pesticide
Use
Survey
(
NHGPUS)
(
RTI,
1991).
Because
pest
pressure
is
a
function
of
the
climate
and
housing
stock,
data
on
the
need
to
control
pests
is
believed
to
be
somewhat
stable
over
time.
Pesticide
product
market
shares
are
much
more
variable
as
pesticides
come
on
and
off
the
market,
as
alternative
methods
are
employed
and
as
pesticide
product
prices
change.
Use
of
contemporary
market
share
data,
when
available,
allows
EPA
to
model
contemporary
pesticide
usage
in
lieu
of
new
use
surveys.

Under
this
approach,
LifeLine
 
models
the
times
that
the
homeowner
treats
a
pest
in
a
specific
microenvironment.
This
modeling
is
performed
by
selecting
a
record
from
NHGPUS
and
taking
the
number
of
treatments
to
generate
a
daily
application
frequency
that
is
applied
to
the
appropriate
seasons
when
the
pests
will
require
controls.
Each
instance
where
a
pesticide
is
applied
is
assumed
to
be
a
single
pesticide
product.
In
addition,
the
model
assumes
that
for
any
given
THE
LIFELINE
GROUP
Page
35
of
57
year,
a
pest
in
an
individual's
home
is
treated
using
products
containing
the
same
active
ingredients
(
LLG,
2004).

This
"
pest
pressure"
approach
allows
the
simulation
of
exposure
to
multiple
pest
products
containing
multiple
pesticides.
Since
each
application
is
linked
to
a
specific
product
and
pesticide,
the
residential
exposures
can
be
linked
to
the
individual
pesticides.

This
approach
only
requires
the
user
to
identify
the
market
share
of
each
of
the
N­
methyl
carbamate
pesticides
used
for
controlling
a
specific
pest.
Therefore,
the
approach
is
amenable
to
both
retrospective
analyses
and
for
the
evaluation
of
new
products
and
their
impact
on
cumulative
risk.

ii.
Modification
of
LifeLine
 
to
Separately
Calculate,
Track,
and
Save
Exposure
Information
for
Multiple
Pesticides
LifeLine
 
software
was
designed
from
its
inception
to
support
assessments
of
cumulative
risk
(
LLG,
1999,
LLG
2002).
The
design
of
the
software
uses
a
series
of
nested
loops
to
model
the
concurrent
operational
aspects
of
exposure
calculations.
As
Figure
2
indicates,
the
dietary
exposure
estimation
process
for
a
single
day
and
a
single
pesticide
uses
a
loop
(
the
Dietary
Exposure
Event
Loop)
to
cycle
through
each
food
(
and
drinking
water)
in
a
CSFII
record.
In
each
of
these
loops,
the
dietary
exposure
from
consuming
the
food
and
drinking
water
is
determined
based
on
the
amount
of
food
consumed
and
the
residues
in
the
foods.
As
indicated
in
Figure
4,
adding
a
new
loop,
The
Pesticide
Loop,
allows
the
software
to
separately
determine
and
track
the
exposures
from
each
pesticide
in
the
family
of
pesticides
in
a
cumulative
assessment.

Continuing
with
the
dietary
example,
in
the
new
Time
Step
Loop,
the
software
will
select
a
dietary
record
and
then
cycle
through
each
of
the
foods
consumed
within
the
time
step
using
the
Exposure
Event
Loop
Diet.
For
each
food
in
an
eating
event
within
the
time
step,
the
model
will
survey
the
residue
data
structure
to
determine
if
the
first
pesticide
in
the
cumulative
assessment
is
present.
If
the
residue
is
present,
the
exposure
resulting
from
that
residue
(
for
just
that
pesticide)
is
calculated
and
saved.
This
is
repeated
for
each
pesticide.
Then
the
model
returns
to
the
Exposure
Event
Loop
Diet
and
moves
to
the
next
food.
Once
all
of
the
foods
are
evaluated,
the
data
for
the
exposure
of
a
pesticide
from
each
food
are
summed
for
the
time
step
to
give
the
total
dietary
exposure
of
that
pesticide
for
that
time
step.
This
is
repeated
for
all
of
the
pesticides
within
the
family
of
pesticides.
The
result
is
a
record
of
the
total
dietary
exposure
for
each
pesticide
within
the
time
step
where
each
pesticide­
specific
exposure
value
is
correlated
to
the
others
within
that
time
step.

A
similar
approach
is
used
for
drinking
water
and
for
the
residential
exposure
assessment.
As
shown
in
Figure
4,
a
new
loop,
the
Pesticide
Loop,
is
added
to
the
non­
dietary,
non­
occupational
exposure
assessment
portion
of
LifeLine
 
.
THE
LIFELINE
GROUP
Page
36
of
57
This
loop
will
be
trigged
when
a
new
microenvironment
is
considered
and
exposures
associated
to
residues
in
that
microenvironment
are
calculated.
The
software
cycles
through
the
loop
to
separately
determine
if
each
pesticide
is
present,
and
what
exposure
would
result
if
it
were
present.
Unlike
the
dietary
portion
of
LifeLine
 
,
which
only
addressed
oral
exposure,
this
loop
will
separately
calculate
the
three
possible
routes
of
exposure
(
oral,
dermal
and
inhalation).
The
exposures
for
each
pesticide
that
occurs
by
a
given
route
from
all
of
the
microenvironments
during
a
time
step
are
then
summed
to
yield
the
route­
specific
total
residential
exposure
for
that
pesticide.
This
process
is
performed
separately
for
each
pesticide
within
the
family
of
pesticides
and
each
pesticide­
specific
exposure
value
is
thus
correlated
to
the
others
within
that
time
step.
THE
LIFELINE
GROUP
Page
37
of
57
Yes
Yes
No
Figure
4.
Flow
Chart
for
Modeling
Multiple
Pesticides
Ye
s
No
Ye
s
Determine
route
specific
doses
Start
Dietary
Non
Dietary
Save
exposure
history
No
Yes
Stop
Select
first/
next
microenv.
Is
the
first/
next
pesticide
in
this
food?

More
microenv?
Residue
data
Assign
birth
and
fixed
characteristics
Time
step
record
Another
time
step?
Time
Step
Loop
Exposure
Event
Loop
Diet
Exposure
Event
Loop
Non­

Diet
Update
exposure
More
foods?
Select
first/
next
food
Pesticide
Residue
data
Determine
oral
dose
More
pesticides?
Update
exposure
history
Ye
s
More
pesticides?

No
First/
next
pesticide
in
first/
next
microenv.?

Pesticide
No
No
No
THE
LIFELINE
GROUP
Page
38
of
57
5
d.
Matching
the
LifeLine
 
Exposure
Outputs
to
the
Definitions
of
Dose
in
the
PBPK/
PD
Model
As
discussed
in
the
earlier
sections
the
process
of
a
pesticide
moving
from
an
environment
to
the
target
organ
or
compartment
of
the
exposed
individual
is
a
multi
step
process
(
EPA,
1992;
EPA,
1997).
It
is
critical
that
the
output
of
the
exposure
model
defines
exposure
and/
or
dose
in
terms
that
are
consistent
with
the
definitions
used
by
the
PBPK/
PD
model.
The
exposure
assessment
model
must
deliver
exposure
metrics
at
the
exact
point
on
the
continuum
where
the
PBPK/
PD
risk
assessment
model
commences.
There
can
be
no
gap
(
such
as
not
accounting
for
the
factors
involved
as
the
pesticide
moves
through
the
dermis),
nor
can
such
elements
be
represented
in
both
models
yielding
duplicative
calculations
for
the
element.

Some
PBPK/
PD
models
define
doses
in
terms
of
the
dosing
protocols
used
in
toxicology
studies
on
which
dose­
effect
associations
are
modeled.
Examples
of
this
are
the
values
representing
the
airborne
level
of
the
pesticide
in
an
inhalation
chamber
in
a
toxicology
study
or
the
concentration
of
pesticide
in
the
test
animal's
diet.
These
media­
specific
concentrations
may
be
held
constant
over
the
duration
of
the
experiment
and
can
be
described
by
a
single
value
and
a
clearly
defined
duration.
This
profile
of
exposure
is
unlikely
to
represent
the
human
experience
of
exposures
to
pesticides
in
real
life,
where
the
air
concentrations
vary
from
moment
to
moment
and
dietary
exposure
changes
with
every
eating
occasion
every
day.
As
a
result,
the
dose
metrics
produced
by
the
exposure
model
for
the
PBPK/
PD
model
require
a
slightly
different
approach
from
traditional
exposure
assessment
outputs.
The
following
sections
present
the
proposed
approach
for
each
route
of
exposure
in
the
exposure
model
supporting
the
PBPK/
PD
risk
assessment
models.

i.
Oral
Exposures
Oral
exposures
result
from
three
sources
in
LifeLine
 
:
food,
drinking
water
and
incidental
hand
to
mouth
contact.
Each
food
item
has
a
different
residue
level.
The
number
of
consumed
foods,
type
of
eating
occasion
and
amount
of
each
food
item
consumed
are
defined
for
each
time
step.
Consumption
of
the
food
items
is
considered
to
be
randomly
distributed
across
the
time
step(
s)
in
which
the
eating
occasion
occurs
and
occupies
the
full
duration
of
the
time
step(
s)
associated
with
that
eating
occasion.
A
more
detailed
discussion
of
this
process
is
provided
in
the
previous
section
of
this
document.
As
a
result,
the
oral
exposure
from
diet
is
defined
as
a
mass
intake
for
the
time
step:

Di
=
 
Fij
Aj
/
S
THE
LIFELINE
GROUP
Page
39
of
57
Where
Di
is
the
exposure
of
the
ith
pesticide,
Fij
is
the
concentration
of
the
residue
of
the
ith
pesticide
in
the
jth
food,
Aj
is
the
amount
of
the
jth
food
consumed
in
the
eating
occasion
and
S
is
the
number
of
contiguous
time
steps
over
which
the
eating
occasion
occurs.

If
S
is
two
or
three,
the
paradigm
assumes
eating
occasions
occupy
10,
20
or
30
minutes,
Di
will
be
the
same
for
the
next
one
or
two
time
steps.
A
similar
approach
will
be
used
for
the
consumption
of
drinking
water
determined
to
occur
during
an
eating
event.

Oral
exposure
from
hand
to
mouth
events
will
be
modeled
based
on
the
assumption
that
the
exposure
continues
at
roughly
a
constant
rate
while
an
individual
is
in
a
microenvironment
performing
an
activity.
The
exposure
will
be
defined
as
the
amount
of
pesticide
that
is
removed
from
the
hand
by
hand
to
mouth
contact.

ii.
Dermal
Exposures
Dermal
exposure
to
pesticides
occurs
from
three
sources:
dermal
contact
with
pesticide
residues
on
contaminated
surfaces;
dermal
exposure
to
the
product
during
application;
and
from
showering
with
water
containing
pesticides.
The
exposures
for
showering
are
different
from
the
other
two
types
of
exposure
and
require
a
different
measure
of
exposure.
In
the
case
of
showering,
the
majority
of
the
body
is
in
contact
with
essentially
an
infinite
source
(
pesticide
in
water
that
is
constantly
flowing
over
the
skin.)
In
the
showering
scenario,
dermal
exposure
will
be
defined
in
terms
of
the
concentration
in
water
and
the
duration
of
contact.

In
the
case
of
dermal
exposure
to
residues
on
surfaces
and
dermal
exposure
during
pesticide
application,
a
finite
amount
of
pesticide
deposits
on
the
skin
and
remains
there
until
it
is
absorbed
or
is
removed
by
some
process.

In
the
applicator
and
post
application
scenarios,
the
amount
of
a
pesticide
on
an
individual's
skin
and
the
area
over
which
this
exposure
occurs
will
be
modeled
over
time.
In
the
case
of
the
applicator
exposures,
the
"
unit"
exposure
rate
and
the
duration
of
the
application
event
will
be
used
to
determine
the
loading
rates
and
the
area
exposed
for
each
time
step.
In
the
case
of
the
post
application
exposures,
the
dermal
transfer
rates,
clothing
and
duration
of
time
spent
in
the
microenvironment
will
be
used
to
define
the
loading
rate
for
each
time
step.
Residues
are
assumed
to
remain
on
the
skin
until
removed
by
dermal
absorption,
bathing,
hand
washing
or
hand
to
mouth
contact.
For
bathing
and
hand
washing,
a
washing
removal
efficiency
is
used.
This
approach
has
been
used
in
the
SHEDS
model
of
dermal
exposure
(
EPA,
2002c).
THE
LIFELINE
GROUP
Page
40
of
57
In
this
approach,
there
will
be
an
estimate
of
the
average
amount
of
pesticide
present
on
the
skin
and
the
area
of
skin
that
is
contaminated
for
each
time
step.
The
model
does
have
the
capacity
to
estimate
other
exposure
parameters
i.
e.
delivered
dose,
as
the
PBPK/
PD
model
requires.

iii.
Inhalation
Exposures
Inhalation
exposure
occurs
during
showering,
during
pesticide
application,
and
during
post
application
activities.
Exposures
for
all
three
inhalation
scenarios
will
be
defined
in
terms
of
the
mass
of
pesticide
inspired
in
a
time
step.
If
the
exposures
from
inhalation
scenarios
are
critical
to
an
assessment,
it
may
be
necessary
to
determine
what
fraction
of
the
inspired
exposure:
 
Reaches
the
deep
lung
and
is
absorbed:
 
Impacts
the
nasal
pharyngeal
region
and
bronchial
tubes
and
should
be
addressed
as
an
oral
exposure;
 
Is
absorbed
in
nasal
pharyngeal
region
and
bronchial
region;
or
 
Is
exhaled
unchanged.

iv.
Summary
In
summary,
the
route
specific
exposure
information
provided
to
the
PBPK/
PD
model
will
consist
of:
 
The
mass
of
each
pesticide
ingested
in
a
time
step
from
drinking
water,
from
food
and
from
incidental
hand
to
mouth
contact;
 
The
mass
of
each
pesticide
inspired
in
a
time
step;
 
The
average
mass
of
pesticide
on
the
skin
and
the
area
over
which
the
exposure
occurs
for
each
time
step;
and
 
The
concentration
of
pesticides
in
shower
water
and
the
whether
or
not
showering
occurs
in
a
time
step.

This
information
will
be
provided
for
each
pesticide,
for
each
time
step,
for
each
individual
simulated
and
concurrent
metrics
will
remain
linked.

5
e.
Modeling
Physiological
and
Genetic
Variability
in
Partitioning,
Metabolism,
and
Dose­
Effect
Relationships
in
PBPK/
PD
Models
As
discussed
above,
for
the
LifeLine
 
software
to
define
the
contact
doses
received
by
an
individual,
the
software
must
define
the
characteristics
of
the
individual
receiving
the
doses.
This
definition
of
the
individual
provides
a
framework
for
defining
or
assisting
in
the
definition
of
many
of
the
parameters
for
the
PBPK/
PD
model.
The
PBPK/
PD
cumulative
risk
assessment
model
will
deal
with
such
issues
such
as
distribution
of
the
pesticide
from
routes
of
exposure
to
various
target
organs,
partitioning
of
the
pesticide
from
blood
to
the
target
tissues,
metabolism
and
the
quantitative
dose­
effect
relationship
at
the
molecular
or
clinical
level.
Given
that
many
of
these
parameters
may
be
influenced
by
THE
LIFELINE
GROUP
Page
41
of
57
interindividual
differences
related
to
age,
gender,
ethnicity,
activity
level
or
other
characteristics
defined
in
the
LifeLine
 
exposure
assessment
model,
those
definitions
should
be
linked
to
the
exposure
data
for
each
individual
(
applied
at
the
appropriate
level
of
detail
in
the
exposure
history).
The
authors
of
the
PBPK/
PD
model
must
describe
the
way
in
which
this
information
is
employed
within
the
risk
assessment
model.

i.
The
Lifeline
 
Framework
The
current
version
of
the
Lifeline
 
software
assigns
a
number
of
characteristics
to
each
individual
in
a
simulation.
The
process
used
to
perform
this
is
described
in
detail
in
the
Lifeline
 
Technical
Manual
(
LifeLine
 
,
2004).
The
following
is
a
brief
summary
of
the
process.

The
process
begins
by
assigning
characteristics
to
an
individual
at
birth
and
modeling
how
the
characteristics
vary
over
time
(
Figure
5).
LifeLine
 
begins
by
assigning
an
individual
a
gender,
race,
and
ethnicity.
Based
in
these
fixed
characteristics,
the
software
assigns
a
body
length
to
each
individual
for
the
first
year
of
life.
Data
on
race
and
gender
specific
growth
in
height
is
then
used
to
model
height
changes
over
the
individual's
life.

The
result
is
an
estimate
of
the
individual's
height
at
each
year
of
their
life.
These
age­
specific
heights
are
then
used
to
select
the
body
weights
for
those
ages.
The
equations
used
for
modeling
are
based
on
data
taken
from
NHANES
III.
Knowing
the
age,
height
and
weight
of
the
individual
allows
the
determination
of
the
surface
area
of
the
hands
and
whole
body.

Gender,
Race,
and
Ethnicity
(
Natality
Database)

Longitudinal
model
of
height
(
NHANES
III
Data)

Model
of
weight
(
NHANES
III
Data)

Calculate
Body
Mass
Index
Calculate
resting
breathing
rate
(
Layton
1992)
Determine
total
body
surface
area
Exp.
Fact.
Handbook
Surface
areas
of
hands
Figure
5.
Process
for
Assigning
Physiological
Characteristics
to
an
Individual
THE
LIFELINE
GROUP
Page
42
of
57
In
addition,
once
the
height
and
weight
of
the
individual
are
known,
the
body
mass
index
(
BMI)
can
be
determined.
The
BMI
can,
along
with
age
and
gender,
be
used
to
predict
the
resting
breathing
rates.
These
resting
breathing
rates
can
be
used
along
with
information
on
the
level
of
activity
to
determine
the
breathing
rates
under
various
levels
of
activity.
LifeLine
 
determines
breathing
rates
for
resting
and
sleeping,
sedentary
activities
and
three
levels
of
active
behavior
(
light,
moderate
and
heavy).
The
result
of
this
process
is
a
determination
in
Lifeline
 
Version
2.1
of
the
following
physiological
criteria
of
each
individual
in
a
simulation:

 
Demographics:


Race

Ethnicity

Gender

Age
 
Physiology

Height

Weight

Body
Mass
Index
(
BMI)


Surface
area
 
Whole
body
 
Hands

Breathing
rates
 
Resting
and
sleeping
 
Sedentary
activities
 
Light
activities
 
Moderate
activities
 
Heavy
activities
ii.
Proposed
Modifications
P3M
While
the
current
version
of
LifeLine
 
provides
an
excellent
basis
for
developing
the
inputs
for
PBPK/
PD
modeling,
the
proposed
modification
of
this
framework
draws
upon
a
related
project
accomplished
by
the
LLG
scientists8.
In
that
project,
a
software
program
Physiological
Parameters
for
Physiologically
based
Pharmacokinetic
Models
(
PPPM
or
P3M)
was
created.
P3M
was
developed
to
produce
demographically
specific
and
internally
consistent
values
of
the
physiological
parameters
(
Price
et
al.
2003b).
P3M
was
created
based
on
published
studies
that
reported
correlations
between
various
physiological
8
The
described
software,
P
3
M
is
copyrighted
protected
and
distributed
via
LINEA,
Inc.
The
software
is
available
to
the
public
without
fee
via
the
LINEA
web
site:
www.
lineainc.
com.
Authors
have
granted
exclusive
license
for
its
use
in
the
LifeLine
 
model
as
proposed
in
this
document.
THE
LIFELINE
GROUP
Page
43
of
57
parameters
and
individuals'
height,
weight,
gender,
age
and
ethnicity.
Based
on
that
review
empirical
models
were
identified
that
allowed
the
prediction
of
correlated
volumes
of
many
compartments,
tissues
and
organs.
The
specific
algorithms
vary
with
the
tissue
under
consideration,
but
generally
consist
of
a
series
of
regression
models
for
various
ages
and
genders.
These
algorithms
captured
the
bulk
of
the
inter­
individual
variation.

An
example
of
the
approach
used
in
P3M
is
given
below.
In
this
example,
a
model
(
here
a
set
of
equations)
is
developed
for
predicting
kidney
volume
and
blood
flows:

Adults
(
Males
and
Females)

Total
Weight
of
Kidneys
(
g)
=
15.4
+
2.04*
BW
+
51.8*
BH
(
m)
2
(
R2=
0.64)
(
Kasiske
and
Umen
1986)

Children
Male
and
Females:

Left
Kidney
Volume
(
ml)
=
4.214
*
BW
(
kg)
0.823
(
R2=
0.97)
Right
Kidney
Volume
(
ml)
=
4.456
*
BW
(
kg)
0.795
(
R2=
0.97)
(
Dinkel
et
al.
1985)

Additional
information
on
the
model
is
contained
in
the
manuscript
"
Modeling
Inter­
individual
Variation
in
Physiological
Factors
Used
in
PBPK
Models
of
Humans"
presented
in
Appendix
A
of
this
document.

As
part
of
the
P3M
project,
the
resting
blood
flow
for
each
organ
was
also
calculated
based
on
estimated
volume
of
the
tissues
and
literature
values
of
tissue
specific
blood
flow
rates.
The
data
for
kidneys
are
as
follows:

Tissue
blood
flow:
Males
3.68
(
l
of
blood
/
min/
l
of
tissue)
Females
3.22
(
l
of
blood
/
min/
l
of
tissue)
(
Williams
and
Legett,
1989)

Thus
the
blood
flow
in
l/
m
for
a
male
child
would
be:

=
(
4.214
*
BW
(
kg)
0.823
+
4.456
*
BW
(
kg)
0.795)/
1000*
3.68
(
l
of
blood
/
min/
l
of
tissue)

The
organ
and
tissue­
specific
calculations
from
each
of
the
algorithms
can
be
summed
to
give
estimates
of
the
commonly
used
PBPK/
PD
compartments
(
well
perfused,
poorly
perfused
and
fatty
tissues)
for
each
individual.
Finally,
the
blood
THE
LIFELINE
GROUP
Page
44
of
57
flows
for
all
of
the
organs
can
be
summed
to
provide
an
estimate
of
the
total
resting
cardiac
output9.

Estimates
of
volume
and
blood
flow
are
available
for
the
following
organs,
organ
systems,
tissues,
and
compartments:

1.
Total
well
perfused
tissues
2.
Red
marrow
3.
Lungs
(
tissue
volume)
4.
Brain
5.
Kidneys
6.
Liver
7.
Pancreas
8.
Thyroid
9.
Spleen
10.
GI
organs
(
total
tissue
volume
for
stomach
and
small
and
large
intestines)
11.
Blood
12.
Plasma
13.
Blood
cells
14.
Total
poorly
perfused
tissues
15.
Dermis
16.
Epidermis
17.
Skeletal
muscle
18.
Heart
(
Tissue
volume)
19.
Tongue
20.
Total
fatty
tissues
21.
Adipose
issue
22.
Yellow
marrow
23.
Bone
tissue
Proposed
Approach
Based
upon
the
empirical
models
developed
in
the
P3M
project
and
the
characteristics
currently
assigned
to
individuals
in
the
LifeLine
 
Software,
the
volumes
and
resting
blood
flows
of
the
above
tissues
and
organs
can
be
specified
for
every
individual
at
every
age
in
their
lives.

The
P3M
project
also
developed
models
of
the
resting
breathing
rates
and
cardiac
output.
The
resting
breathing
rates
are
based
on
the
work
by
Layton
(
1993).
The
PBPK/
PD
model
being
developed
by
ORD/
EPA
for
the
cumulative
risk
assessment
uses
the
alveolar
ventilation
rate
rather
than
the
breathing
rate
(
minute
volume).
To
convert
from
the
breathing
rate
to
the
alveolar
ventilation
rate
a
correction
is
necessary
to
account
for
the
physiologic
dead
space
volume.

9
Note
the
actual
process
of
deriving
the
blood
flows
for
the
well­
perfused
tissues
and
for
the
total
cardiac
output
has
to
be
corrected
for
organs
and
organ
systems
where
the
blood
flow
occurs
in
sequence
(
GI
and
the
liver).
THE
LIFELINE
GROUP
Page
45
of
57
EPA
(
1996)
reports
that
the
alveolar
ventilation
rate
is
70%
of
the
breathing
rates.
However,
it
may
also
be
possible
to
predict
alveolar
ventilation
rates
based
on
alternative
models
(
Neder
et
al.,
2003;
Bennett
and
Zeman,
2004;
and
Harris
1978).

The
PBPK/
PD
model
being
developed
by
ORD/
EPA
for
the
cumulative
risk
assessment
is
expected
to
describe
blood
flow
using
two
types
of
data,
the
total
cardiac
output
and
the
percentage
of
the
cardiac
output
that
goes
to
each
compartment.
The
proposed
approach
will
directly
provide
the
total
cardiac
output.
The
percentage
going
to
each
compartment
can
be
estimated
based
on
the
compartment
specific
flow
rates
and
the
total
cardiac
output.

Providing
Information
Relevant
to
Interindividual
Variation
for
PBPK/
PD
Model
Elements
Parameters
such
as
partitioning,
metabolic
factors
and
dose­
effect
relationships
can
vary
across
individuals.
Some
of
this
variation
may
be
related
to
the
physiology,
gender
or
age
of
the
individuals.
In
addition,
a
great
deal
of
research
is
being
performed
on
the
genetic
contribution
to
this
inter­
individual
variation.
To
the
extent
that
the
factors
can
be
related
to
the
any
of
the
demographic
information
(
age,
gender,
or
race)
or
physiology
(
body
type)
these
factors
can
also
be
modeled.
Because
of
this
potential
opportunity,
the
proposed
LifeLine
 
outputs
to
the
PBPK/
PD
model
will
include
demographic
information
associated
to
the
calculated
exposure
values.

Modeling
Time­
Dependent
Physiological
Parameters
The
volume
of
the
various
organs,
tissue
and
compartments
do
not
vary
significantly
over
the
course
of
a
few
hours
or
days.
Thus,
the
values
for
these
parameters
are
assumed
constant
over
these
short
time
periods.
However,
blood
flow
and
the
alveolar
ventilation
rates
are
not
constant
and
vary
with
the
individuals'
activities.
Activities
that
are
more
strenuous
will
raise
both
the
cardiac
output
and
the
alveolar
ventilation
rate
and
will
change
the
fractions
of
the
cardiac
output
going
to
different
organs.
Consumption
of
food
will
change
the
fraction
of
cardiac
output
going
to
the
digestive
organs.

This
time­
dependent
and
activity­
dependent
variation
can
be
captured
by
estimating
values
for
these
parameters
for
each
time
step
of
the
exposure
history.
As
discussed
in
the
prior
section
of
this
report,
the
duration
of
the
time
steps
can
range
from
10
minutes
to
24
hours.
The
duration
of
time
over
which
these
values
can
be
modeled
(
the
exposure
history)
can
be
as
short
as
24
hours
or
as
long
as
365
days.

One
approach
for
estimating
the
impact
of
activity
on
values
for
breathing
rates
and
blood
flow
is
to
employ
estimates
of
changes
in
breathing
rates
to
estimate
the
impact
on
cardiac
output.
Layton
(
1993)
proposed
a
simple
set
of
factors
for
THE
LIFELINE
GROUP
Page
46
of
57
various
levels
of
activity
that
are
multiplied
by
the
resting
breathing
rate
to
yield
breathing
rates
associated
with
various
levels
of
exertion.
In
LifeLine
 
software,
these
factors
are
linked
to
the
various
activities
listed
in
the
activity
patterns
and
are
used
to
define
the
average
breathing
rate
for
an
individual's
macro
behaviors
over
the
course
of
a
day.
Using
this
approach,
it
is
feasible
to
estimate
the
breathing
rates
for
each
10­
minute
time
step
in
an
exposure
history.

The
breathing
rates
can
be
linked
to
alveolar
ventilation
rates
and
these
can
in
turn
be
used
to
predict
the
corresponding
cardiac
outputs.
The
cardiac
output
in
the
current
PBPK/
PD
model
is
estimated
based
on
the
alveolar
ventilation
rate
using
the
equation
CO
=
A0*
QA0
+
A1
(
QA
­
QA0)

where
A0
is
the
ratio
of
the
cardiac
output
to
alveolar
ventilation
rate
under
resting
conditions;
A1
=
the
fractional
increase
in
cardiac
output
for
an
increase
of
the
alveolar
ventilation
rate;
CO
=
Cardiac
output
for
a
given
level
of
activity;
QA
=
Alveolar
ventilation
rate
for
the
level
of
activity;
and
QA0
=
Alveolar
ventilation
rate
under
resting
conditions.

Using
this
approach,
it
will
be
possible
to
define
the
cardiac
output
and
alveolar
ventilation
rate
for
each
10
minute
time
step
of
an
individual's
exposure
history.

Finally,
since
LifeLine
 
includes
a
detailed
time­
related
profile
of
food
consumption,
the
model
can
calculate
eating­
related
temporal
changes
of
cardiac
output
and
the
changes
for
percentage
of
cardiac
output
for
each
compartment
of
the
PBPK/
PD
model.
This
can
be
presented
for
time
steps
that
include
an
eating
occasion
and
for
time
steps
that
follow
a
time
step
containing
an
eating
occasion.

In
summary,
the
current
approach
used
in
LifeLine
 
software
can
be
extended
to
address
both
the
time
dependent
and
time
independent
physiology
parameters
required
by
PBPK/
PD
models.

5
f.
Designing
the
Interface
between
LifeLine
 
Exposure
Model
and
the
PBPK/
PD
Model
of
Cumulative
Risks
Once
the
modifications
are
made
to
the
current
LifeLine
 
exposure
software,
there
remains
the
conceptually
simple
but
technically
challenging
task
of
linking
the
information
from
LifeLine
 
to
the
PBPK/
PD
model.
There
are
two
basic
approaches
to
accomplish
this.
One
is
the
integration
of
the
two
models
into
one;
THE
LIFELINE
GROUP
Page
47
of
57
the
second
is
the
creation
of
an
interface
to
deliver
a
properly
formatted
data
file
from
the
exposure
model
to
the
PBPK/
PD
model.

An
advantage
of
integrating
LifeLine
 
and
the
PBPK/
PD
model
is
that
the
resulting
merged
model
would
preserve
a
perfect
continuum
of
algorithms
beginning
with
the
various
exposure
events
and
progressing
to
the
target
tissue
dose­
effect
calculations
for
each
time
step
in
the
simulation.
There
would
be
a
seamless
interface
between
the
output
metrics
from
the
exposure
model
and
the
input
requirements
of
the
PBPK/
PD
model.
It
would
also
assure
correct
calculation
of
the
exposure
source
contributions
for
any
specific
target
tissue
dose­
effect
values.

However,
there
are
strong
advantages
to
preserving
the
independence
of
these
two
models
and
creating
a
flexible
and
responsive
interface
between
them.
This
approach
recognizes
the
fact
that
the
exposure
assessment
model
is
a
relatively
advanced
tool
that
has
evolved
over
the
years.
It
functions
on
a
fast
operating
system
which
can
process
great
numbers
of
calculations
and
store
a
giant
block
of
answers
for
convenient
viewing
options.
As
technology
has
advanced,
so
too
have
the
operational
options
and
available
functions
of
the
exposure
model.
By
comparison,
PBPK/
PD
models
such
as
the
EPA/
ORD
ERDEM
(
Exposure
Related
Dose
Estimation
Model)
model
are
not
yet
as
advanced.
This
may
also
be
the
case
for
other
PBPK/
PD
models.
The
model
architects
have
focused
on
basic
construction
elements
and
design
concepts.
Building
fast
operating
systems
are
not
yet
the
priority
issue,
and
even
the
design
elements
will
likely
change
as
experience
grows
with
these
developing
models.
Another
important
reason
to
maintain
separation
of
the
two
model
types
is
that
it
permits
both
models
to
evolve
with
their
respective
science
and
technology
and
avoids
some
of
the
resource
(
time
and
financial)
issues
intrinsic
to
the
integrated
model.

The
key
to
success
in
pursing
the
approach
of
two
separate
models
is
to
develop
a
flexible
and
reliable
interface
or
`
bridge"
between
the
exposure
model
and
the
PBPK/
PD
model
 
one
that
has
the
capacity
to
evolve
as
new
versions
of
PBPK/
PD
models
are
developed.
This
interface
is
most
logically
built
to
receive
output
data
from
the
exposure
model
and
to
modulate
it
to
mesh
with
the
input
requirements
of
the
PBPK/
PD
model.
It
must
also
take
into
consideration
the
technical
realities
of
the
PBPK/
PD
models
of
today,
and
the
likelihood
that
their
capacity
to
receive
more
data
i.
e.
the
"
delivered
data
load"
will
increase
as
the
models
evolve.

5
g.
Processing
Demands
To
design
this
interface,
it
is
helpful
to
consider
key
factors
imposed
by
the
flanking
models
and
the
user
for
whom
it
is
being
built.
The
following
figure
summarizes
the
factors
raised
earlier
in
this
document
(
Figure
6).
THE
LIFELINE
GROUP
Page
48
of
57
In
a
matter
of
minutes
or
hours
of
computational
time,
the
LifeLine
 
model
can
deliver
an
immense
block
of
exposure
profiles
that
quantify
the
source
specific/
route
specific
exposures
for
100,000
individuals
along
with
each
individual's
demographic,
physiological
and
anatomical
information,
as
described
in
this
document.
In
the
coming
years,
it
is
expected
that
the
model
will
yield
even
more
information
bits
as
the
time
steps
for
these
exposure
assessments
become
smaller
and
as
other
functions
are
incorporated.
The
model
can
express
its
answers
as
exposure
to
the
individual
or
as
absorbed
dose
by
given
routes.

The
PBPK/
PD
model
is
newly
developed
and,
as
with
most
new
models,
operates
on
a
relatively
limited
system,
which
may
have
significant
data
volume
and
operational
speed
limitations.
Yet,
the
PBPK/
PD
model
demands
great
detail
for
the
exposure
metrics
(
small
time
steps,
route
specific
exposure,
etc)
linked
to
all
details
about
the
exposed
individual.
Thus,
no
compromise
can
be
made
on
the
option
of
delivering
as
much
information
as
possible
on
the
individual
and
on
that
individual's
exposure
profile
for
the
chemical(
s)
in
question.
The
PBPK/
PD
model
is
likely
to
change
significantly
as
it
is
developed;
as
scientists
accrue
operational
experience
with
it,
and
as
it
address
different
chemicals
with
different
needs
vis­
à­
vis
the
physiological
metrics
of
importance.
This
evolution
could
be
quite
rapid
and
other
models
may
emerge
as
well.

The
user
 
EPA's
Office
of
Pesticide
Programs
 
needs
a
tool
that
expresses
the
exposure
and
risk
profiles
in
a
way
that
reflects
their
regulatory
needs.
Traditionally,
the
risk­
THE
LIFELINE
GROUP
Page
49
of
57
Figure
6.
Proposed
Linkage
of
Lifeline
to
PBPK/
PD
Model
based
regulatory
issues
have
focused
on
scenarios
dealing
with
"
highest
exposures"
and
"
sensitive
sub
populations"
and
route­
specific
risk.
Although
a
full
view
of
the
distribution
of
risk
across
populations
is
valuable
to
public
health
professionals
and
researchers,
that
view
may
have
limited
value
for
the
decisions
mandated
under
FIFRA,
FFDCA
and
FQPA.
The
focus
of
the
risk
assessment
is
driven
by
the
legislation
and
EPA/
OPP
policy.
The
filter
can
accommodate
whatever
part
of
the
distribution
of
exposure
across
the
population
that
EPA
needs
to
focus
on.
(
See
Figure
7).
LifeLine
 
Agg/
Cum
Exposure/
Risk
Assessment
Model
Interface
PBPK/
PD
Model
EPA/
OPP
Risk
assm t
for
Risk­
based
Regulatory
Decisions
Rapid
computational
system
Yields
very
large
data
block
Multiple
viewing
options
for
exposure
data
New
(
relatively
slow)

computational
system
Can
process
limited
volume
of
data
per
unit
time
Needs
exposure
source,

route
data
linked
to
demographics
and
physiological
data
for
exposed
individual
Risk
assessment
must
be
relevant
to
regulatory
focus:

highest
risk
status,

sensitive
subpopulation
etc.
Other
data
THE
LIFELINE
GROUP
Page
50
of
57
Thus,
the
interface
must
be
able
to:

 
Characterize
and
deliver
a
defined
population
that
is
sufficiently
small
to
be
run
in
the
PBPK/
PB
model
but
large
enough
to
represent
the
target
population
for
which
exposure
profiles
are
desired.


Ideally,
the
user
should
be
able
to
select
the
points
on
the
distribution
and
the
subpopulations
of
interest
to
direct
the
filtering
work
of
the
interface.
The
individuals
delivered
to
the
PBPK/
PD
models
should
reflect
the
policy
of
the
EPA/
OPP
as
to
their
regulatory
interest.
Current
regulatory
goals
are
designed
around
risk
calculations
at
the
95
to
99.9th
percentiles
of
an
exposed
population.
Definition
of
risk
at
this
extreme
tail
in
the
distribution
of
risk
values
usually
requires
that
populations
of
10,000
to
100,000
individuals
be
simulated
in
the
LifeLine
 
model.
In
addition,
separate
assessments
are
made
for
up
to
nine
different
age
and
gender
defined
populations.
Because
of
the
limited
capability
of
the
PBPK/
PD
model,
only
a
fraction
of
the
individual
exposure
histories
generated
by
LifeLine
 
could
be
transferred
the
PBPK/
PD
model
for
derivation
of
the
risk
assessments.


The
interface
must
filter
the
information
and
deliver
an
appropriate
data
density
in
the
right
format
for
the
PBPK/
PD
model.
This
is
a
technical
detail,
but
very
important.
Thus,
LifeLine
 
output
through
the
interface
should
deliver
files
that
are
easy
to
import
into
the
PBPK/
PD
model(
s),
easy
to
rearrange
and
manage.
Obscure
file
structures
would
be
detrimental
to
this
critical
requirement.
These
parameters
are
likely
to
change
as
the
PBPK/
PD
models
become
more
sophisticated
and
run
on
faster
operating
systems.
The
interface
should
respond
to
such
upgrades
by
permitting
more
information
to
flow
(
less
filtering)
at
the
user's
command.

 
Convey
the
demographic
and
physiological
parameters
that
do
not
vary
with
time
(
during
an
individual's
exposure
history)
for
the
simulated
individuals;
and
 
Convey
the
time
variant
information
on
exposure
and
activitydependent
physiological
parameters.
The
requirements
for
these
may
vary
from
one
chemical
class
to
another
as
different
mechanisms
of
action
at
different
tissue
sites
are
involved.
Ideally,
the
user
could
specify
which
parameters
are
of
interest,
allowing
the
interface
to
filter
out
unnecessary
data
elements
for
the
PBPK/
PD
model.
THE
LIFELINE
GROUP
Page
51
of
57
Figure
7.
Options
for
Selecting
Records
for
PBPK/
PD
Analysis
Finally,
the
interface
from
the
exposure
model
to
the
PBPK/
PD
model
must
be
fashioned
whereby
the
information
is
transferred
without
losing
the
interconnections
of
multiple
pesticide,
multiple
route
exposure
values
for
a
coherent
series
of
time
steps
for
each
individual
with
the
relevant
physiological
and
demographic
identifiers.
The
interface
must
faithfully
maintain
the
continuum
from
the
media
concentration
values
to
the
target
tissue
doses
without
creating
gaps
or
overlaps,
as
previously
discussed
The
proposed
approach
will
be
to
revise
LifeLine
 
to
allow
the
analyst
to
customize
the
exposure
outputs
for
the
specific
PBPK/
PD
analysis
to
be
run.
The
analyst
may
choose
the
desired
target
tissue
calculations
from
a
menu
consisting
Option:
Focus
on
Upper
X
Percentile
of
Distribution
Option:
Focus
on
Few
Points
Representing
Entire
Distribution
Option:
Focus
on
Mean
/
Median
Zone
etc
EXPOSURE
PROFILE
ACROSS
10,000
SIMULATED
INDIVIDUALS
IN
DEFINED
POPULATION
EXPOSURE
PROFILES
FOR
10,000
SIMULATED
THE
LIFELINE
GROUP
Page
52
of
57
of
the
23
tissues,
organs
and
compartments
listed
above.
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.

An
example
of
the
way
in
which
the
interface
could
work
is
presented
below.
In
this
example,
it
is
assumed
that
EPA
would
be
interested
in
the
upper
percentiles
of
the
exposure
distribution
for
its
regulatory
inquiry
and
would
thus
select
exposure
profiles
for
those
individuals
within
the
population
with
the
highest
exposures.
The
vast
bulk
of
the
LifeLine
 
data
block
would
thus
be
filtered
out
and
left
behind.

The
process
of
selecting
the
records
to
deliver
to
the
PBPK/
PD
model
requires
special
attention
and
a
transparent
prioritization
scheme
based
on
explicit
criteria
and
definition
of
each
detail
of
the
filtering
process.
Those
approaches
could
be
based
on
any
of
many
criteria.
These
criteria
deserve
careful
consideration
by
EPA
and
when
specified,
can
be
accomplished
by
the
LifeLine
 
interface.

One
example
of
a
process
for
selecting
records
would
be
as
follows:

1.
The
exposure
software
would
create
a
demographic,
physiological
and
exposure
history
for
an
individual.
2.
The
data
would
be
evaluated
against
screening
criteria
that
would
eliminate
individuals
with
a
low
potential
for
adverse
effects.
3.
If
the
record
was
found
to
exceed
the
criteria
then
it
is
included
in
the
Access
 
file
that
will
be
run
in
the
PBPK/
PD
model
of
cumulative
risk.
In
addition
the
information
would
be
saved
as
part
of
the
LifeLine
 
software
outputs
4.
If
the
record
were
not
found
to
exceed
the
screening
criteria
then
it
would
not
be
sent
on
to
the
PBPK/
PD
model
but
would
be
saved
as
part
of
the
LifeLine
 
master
output
file.
5.
The
interface
Access
 
file
is
exported
to
the
PBPK/
PD
model
and
the
high­
risk
records
would
be
run
through
that
model.
The
records
that
are
determined
to
have
exceeded
some
user­
defined
threshold
would
be
identified.
The
findings
of
which
records
exceeded
the
user­
defined
threshold
will
be
evaluated
using
the
data
set
of
all
records
saved
by
LifeLine
 
to
determine
the
factors
that
are
associated
with
these.

5
h.
Output
File
Structure
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.
The
first
contains
the
data
that
remains
constant
over
the
exposure
history.
The
second
is
a
table
of
the
time
dependent
information.
THE
LIFELINE
GROUP
Page
53
of
57
The
time
dependent
information
is
presented
as
a
set
of
values
for
each
of
the
time
steps
in
the
individual's
exposure
history.

Table
2
presents
an
example
of
a
table
of
time
invariant
information
for
a
PBPK/
PD
model
with
five
compartments.
The
information
is
divided
into
two
areas
the
demographics
of
the
individual
and
the
compartment
volumes
of
the
individual's
physiology.

Table
2.
Time
Independent
Data
Demographic
Information
Volumes
of
Selected
Compartments
of
PBPK/
PD
Model
Age
Gender
Race
Ethnicity
Comp.
1
Comp.
2
Comp.
3
Comp.
4
Comp.
5
Table
3
(
parts
a
and
b)
presents
the
time
dependent
data
for
a
cumulative
risk
model
of
three
pesticides.
Each
row
presents
data
for
a
single
time
step.
The
tables
present
the
data
for
the
first
two
and
the
last
of
the
144
ten­
minute
time
steps
in
a
24­
hour
exposure
history
for
the
individual.
Table
3a
presents
the
data
necessary
for
determining
the
route
specific
doses
of
the
three
pesticides.

Table
3a.
Time
Dependent
Data
(
Part
1)

Time
Step
Measure
of
Inhalation
Exposure
(
mg)
Measure
of
Oral
Exposure
(
mg)
Concentration
in
Shower
Water
(
mg/
l)
Measure
of
Dermal
Exposure
(
mg)
Area
of
Dermal
Exposure
(
cm2)

Begin.
Time
End.
Time
Pest.
1
Pest.
2
Pest.
3
Pest.
1
Pest.
2
Pest.
3
Pest.
1
Pest.
2
Pest.
3
Pest.
1
Pest.
2
Pest.
3
Pest.
1
Pest.
2
Pest.
3
0:
00
0:
10
0:
10
0:
20
23:
50
24:
00
Table
3b
presents
the
time
varying
physiology
of
the
individual
for
the
same
time
steps.

Table
3b.
Time
Dependent
Data
(
Part
2)

Time
Step
Fraction
of
Cardiac
Output
for
Each
Compartment
Begin.
Time
End.
Time
Cardiac
Output
Alveolar
Ventilation
Rate
Comp.
1
Comp.
2
Comp.
3
Comp.
4
Comp.
5
0:
00
0:
10
0:
10
0:
20
23:
50
24:
00
This
approach
can
also
be
used
for
linking
PBPK/
PD
models
of
a
single
pesticide
as
well
as
multiple
pesticides.
In
this
case,
data
would
be
presented
only
for
a
single
pesticide.
THE
LIFELINE
GROUP
Page
54
of
57
Obviously,
the
exact
details
of
the
output
file
will
depend
on
the
parameters
selected
for
sending
to
the
PBPK/
PD
model
(
as
discussed
above).
This
example
demonstrates
the
concept
for
delivering
such
parameters
in
a
coherent
and
linked
fashion.

The
technical
aspects
of
the
development
of
the
interface
can
be
achieved
readily
with
existing
techniques
using
the
framework
of
the
LifeLine
 
exposure
output
files.
LifeLine
 
'
s
original
configuration
anticipated
this
stage
in
the
evolution
of
risk
assessments,
and
thus
its
infrastructure
will
accommodate
flexible,
efficient
and
user­
friendly
portals
such
as
this
interface.

However,
much
attention
should
be
given
to
the
criteria
by
which
the
data
are
filtered
through
the
interface.
This
is
an
option
to
be
directed
by
EPA
 
almost
any
focus
can
be
accomplished
by
the
software.
Thereafter,
there
must
be
complete
transparency
of
the
mechanics
by
which
the
filtering
is
accomplished
by
the
interface.
The
filtering
task
of
the
interface
yields
data
representing
only
a
small
fraction
of
the
original
data
block.
The
Agency
and
all
stakeholders
should
understand
the
nature
of
the
data
delivered
to
the
PBPK/
PD
model
and
the
nature
of
the
data
left
behind
and
the
statistical
consequence
of
this
filtering.
THE
LIFELINE
GROUP
Page
55
of
57
6.
References
ACC,
2003.
Acetone
(
CAS
No.
67­
64­
1)
VCCEP
Submission,
American
Chemistry
Council
Acetone
Panel
September
10,
2003
(
The
report
can
be
down
loaded
at:
http://
www.
tera.
org/
peer/
VCCEP/
ACETONE/
ACETONEwelcome.
html).

Bennett
WD,
Zeman
KL.
2004.
Effect
of
body
size
on
breathing
pattern
and
fineparticle
deposition
in
children.
J
Appl
Physiol.
2004
Sep;
97(
3):
821­
6.

Dinkel,
E.,
Ertel,
M.,
Dittrich,
M.,
Peters,
H.,
Berres,
M.,
and
Schulte­
Wissermann,
H.
1985.
Kidney
Size
in
Childhood.
Sonographical
Growth
Charts
for
Kidney
Length
and
Volume.
Pediatr.
Radiol.
15(
1):
38­
43.

Elgethun
K,
Fenske
RA,
Yost
MG,
Palcisko
GJ.
2003
Time­
location
analysis
for
exposure
assessment
studies
of
children
using
a
novel
global
positioning
system
instrument.
Environ
Health
Perspect.:
111(
1):
115­
22.

EPA,
1992.
Guidelines
for
Exposure
Assessment.
Fed.
Reg.
57(
104).

EPA,
1997.
Exposure
Factors
Handbook.
EPA/
600/
P­
95/
002F(
a­
c),
Washington,
DC.

EPA,
2001.
General
Principles
for
Performing
Aggregate
Exposure
and
Risk
Assessment.
November
28.
U.
S.
EPA,
OPP,
Washington,
D.
C.

EPA,
2002a.
Guidance
on
Cumulative
Risk
Assessment
of
Pesticides
Chemicals
Having
a
Common
Mechanism
of
Toxicity.
January
16,
U.
S.
EPA,
OPP,
Washington,
D.
C.

EPA,
2002b
Organophosphate
Pesticides:
Revised
Cumulative
Risk
Assessment,,
U.
S.
EPA
Office
of
Pesticide
Programs
http://
www.
epa.
gov/
pesticides/
cumulative/
rra­
op/
June
10.

EPA,
2002c.
SHEDS­
Wood
Stochastic
Human
Exposure
and
Dose
Simulation
Model
Wood
Preservative
Exposure
Scenario­
Technical
Manual:
Using
SHEDS­
Wood
for
the
Assessment
of
Children's
Exposure
and
Dose
from
Treated
Wood
Preservatives
on
Play
sets
and
Residential
Decks,
U.
S.
Environmental
Protection
Agency
Office
of
Research
and
Development,
National
Exposure
Research
Laboratory.

Harris
EA,
Seelye
ER,
Whitlock
RM.
(
1978)
Revised
standards
for
normal
resting
dead­
space
volume
and
venous
admixture
in
men
and
women.
Clin
Sci
Mol
Med
55:
125­
128.
THE
LIFELINE
GROUP
Page
56
of
57
Kasiske,
B.
L.
and
Umen,
A.
J.
1986.
The
Influence
of
Age,
Sex,
Race,
and
Body
Habitus
on
Kidney
Weight
in
Humans.
Arch.
Pathol.
Lab.
Med.
110(
1):
55­
60.

Layton,
D.
W.
1993.
Metabolically
Consistent
Breathing
Rates
for
Use
in
Dose
Assessments.
Health
Phys.
64(
1):
23­
36.
LLG,
1999.
Background
Document
for
the
Session:
Review
of
an
Aggregate
Exposure
Assessment
Tool.
to
the:
FIFRA
Scientific
Advisory
Panel,
Arlington,
Virginia
September,
1999.

LLG,
2002.
Cumulative
Risk
Assessment
for
Organophosphorous
Pesticides
Using
LifeLine
 
Version
2.0,
Performed
for
the
U.
S.
EPA
Office
of
Pesticide
Programs
by
The
LifeLine
Group,
Inc.
September
20,
2002.

LLG,
2004.
The
Lifeline
 
Technical
Manual
Version
2.1.

Neder
JA,
Dal
Corso
S,
Malaguti
C,
Reis
S,
De
Fuccio
MB,
Schmidt
H,
Fuld
JP,
Nery
LE.
2003.
The
pattern
and
timing
of
breathing
during
incremental
exercise:
a
normative
study.
Eur
Respir
J.
Mar;
21(
3):
530­
8.

Price,
P,
Chaisson,
C,
Koontz
M,
Wilkes
C,
Ryan,
P.
Macintosh,
D,
Georgopoulos,
P.
(
2003a).
Construction
of
a
Comprehensive
Chemical
Exposure
Framework
Using
Person
Oriented
Modeling.
Exposure
Technical
Implementation
Panel
American
Chemistry
Council
Contract
#
1388
http://
www.
thelifelinegroup.
org/
report.
html.

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

Rappaport,
S.
and
Flynn,
M.
2003.
Two
Seminal
Contributions
of
S.
A.
Roach
to
the
Evaluation
and
Control
of
Hazardous
Substances
in
Air.
Ann.
Occup.
Hyg.,
47(
5):
343­
348.

Roach,
S.
A.
1977.
A
Most
Rational
Basis
for
Air
Sampling
Programs.
20:
65­
84.

Williams,
L.
R.
and
Leggett,
R.
W.
1989.
Reference
Values
for
Resting
Blood
Flow
to
Organs
of
Man.
Clin.
Phys.
Physiol.
Meas.
10(
3):
187­
217.
THE
LIFELINE
GROUP
Page
57
of
57
Appendix
A.
Modeling
Inter­
individual
Variation
in
Physiological
Factors
Used
in
PBPK/
PD
Models
of
Humans
(
Price
et
al.
2003b)

The
PDF
file
for
the
reference
of
this
manuscript
is
provided
separately.
