UNITED
STATES
ENVIRONMENTAL
PROTECTION
AGENCY
WASHINGTON,
D.
C.
20460
OFFICE
OF
PREVENTION,
PESTICIDES
AND
TOXIC
SUBSTANCES
October
28,
2003
MEMORANDUM
SUBJECT:
Review
of
Atrazine
Cancer
Epidemiology
DP
Barcode
D295200,
Chemical
#
080803
FROM:
Jerome
Blondell,
Ph.
D.,
Health
Statistician
Chemistry
and
Exposure
Branch
Health
Effects
Division
(
7509C)

Vicki
Dellarco,
Ph.
D.,
Senior
Science
Advisor
Health
Effects
Division
(
7509C)

THRU:
Francis
B.
Suhre,
Chief
Chemistry
and
Exposure
Branch
Health
Effects
Division
(
7509C)

Margaret
Stasikowski,
Director
Health
Effects
Division
(
7509C)

TO:
Eric
Olson,
Ph.
D.,
PM
Team
Reviewer
Michael
Goodis,
Product
Manager
Special
Review
and
Reregistration
Division
(
7508C)

This
review
summarizes
scientific
studies
related
to
atrazine
and
cancer
epidemiology
including
a
consideration
of
animal
mode
of
action
related
to
selected
cancers.
The
review
starts
with
an
examination
of
animal
mode
of
action
issues
related
to
prostate
and
ovarian
cancers
and
non­
Hodgkin's
lymphoma.
This
section
is
followed
by
a
review
of
EPA
findings
concerning
prostate
cancer
in
a
manufacturing
plant,
the
Agricultural
Health
Study,
and
two
California
studies.
The
following
section
addresses
the
evidence
concerning
non­
Hodgkin's
lymphoma
and
then,
other
cancers.
The
review
concludes
with
a
summary
of
three
published
reviews
of
atrazine
and
cancer
epidemiology
published
1996­
1999,
followed
by
a
listing
of
those
studies
published
since
1999
and
considering
their
impact
on
the
weight­
of­
evidence.
In
order
to
be
comprehensive
all
known
epidemiologic
studies
of
atrazine
and
triazines
have
been
included.
The
U.
S.
Environmental
Protection
Agency
(
EPA)
requested
a
review
of
atrazine
and
2
prostate
cancer
by
the
Scientific
Advisory
Panel
(
SAP)
made
up
of
outside
experts.
A
report
of
the
SAP
findings
(
meeting
minutes)
from
the
July
17,
2003
meeting
can
be
found
at
http://
www.
epa.
gov/
scipoly/
sap/.
This
review
considers
the
SAP
findings
and,
where
appropriate,
revises
the
Agency's
findings
related
to
atrazine
and
prostate
cancer.
This
review
also
explains
why
studies
of
other
cancers
where
not
presented
to
the
Panel
at
the
July
2003
meeting.
To
assist
the
reader,
an
outline
of
this
review
is
provided
below:

I.
Examination
of
prostate
and
ovarian
cancers
and
NHL
in
the
context
of
the
animal
mode
of
action
II.
Atrazine
and
prostate
cancer
epidemiology
A.
Prostate
Cancer
­
Manufacturing
plant
study
B.
Prostate
Cancer
­
Agricultural
Health
Study
in
Iowa
and
North
Carolina
C.
Prostate
Cancer
­
Pesticide
use
data
and
cancer
incidence
in
California
counties
D.
Prostate
Cancer
­
Overall
conclusion
III.
Atrazine
and
epidemiology
of
Non­
Hodgkin's
Lymphoma
IV.
Atrazine
and
epidemiology
related
to
other
cancers
V.
Results
of
other
reviews
of
cancer
and
atrazine
through
1999
VI.
Studies
published
since
1999
or
not
included
in
the
three
published
reviews
VII.
EPA
Conclusion:
Atrazine
exposure
and
NHL
and
other
cancers
I.
Examination
of
prostate
and
ovarian
cancers
and
NHL
in
the
context
of
the
animal
mode
of
action
As
discussed
in
the
January
31,
2003
Interim
Reregistration
Eligibility
Decision,
atrazine's
mode
of
action
(
i.
e.,
a
decrease
LH
surge,
failed
ovulation
and
estrous
cycle
disruption)
for
induction
of
mammary
gland
tumors
(
the
only
tumor
observed
in
animal
bioassays)
in
SD
female
rats
is
not
considered
relevant
to
humans.
Because
the
epidemiology
literature
on
atrazine
(
triazines)
report
that
atrazine
exposure
may
be
associated
with
an
increased
incidence
of
prostate
and
ovarian
cancers
and
NHL,
the
available
chronic
toxicity
animal
bioassays
were
closely
examined
for
any
indication
of
either
prostate,
ovarian
cancer
or
NHL,
as
well
as
the
etiology
of
these
tumor
types.

The
prostate
glands
in
male
laboratory
animals
do
not
appear
to
be
a
target
of
atrazine
toxicity.
Subchronic
and
chronic
rodent
and
dog
studies
and
the
multigeneration
rat
studies
conducted
for
atrazine
and
its
major
metabolites
have
not
demonstrated
treatment
related
tumors
or
prostatitis
and
only
shown
inconsistent
changes
in
prostates
weights.
The
SD
and
Fisher
rats
and
CD­
1
mice
are
poor
models
for
evaluating
prostate
cancer.
Atrazine
has
been
shown
to
result
in
prostatic
inflammation
of
the
adult
rat
offspring
(
Stoker
et
al.,
1999)
but
this
is
not
due
to
direct
treatment
of
the
offspring,
but
rather
treatment
of
the
dams.
In
this
case,
atrazine
leads
to
a
decrease
in
early
lactational
exposure
to
prolactin
(
via
treatment
of
the
mothers).
Alterations
in
neonatal
prolactin
regulation
lead
to
hyperprolactemia,
which
in
turns
lead
to
the
prostatic
inflammation
found
in
the
adult
offspring.
This
work
supports
the
neuroendocrine
mode
of
action
3
for
atrazine
rather
than
a
mode
of
action
that
would
explain
prostate
cancer
in
adult
male
workers.
Although
the
etiology
of
prostate
cancer
is
not
clearly
understood,
the
hormonal
changes
caused
by
atrazine
would
be
in
the
opposite
direction
(
i.
e.,
decreased
prolactin)
of
what
would
be
expected
for
the
development
of
prostate
cancer.
Furthermore,
it
has
not
been
clearly
established
that
prostatitis
increases
the
risk
to
prostate
cancer.
Therefore,
the
animal
data
do
not
support
a
mechanism
for
atrazine
contributing
to
the
onset
or
promotion
of
prostate
cancer
in
humans.

Ovarian
tumors
in
most
laboratory
animals
are
a
rare
occurrence
(
Damjanov,
1989).
A
dose­
related
increase
in
ovarian
tumor
incidence
is
not
seen
in
any
study
using
atrazine,
simazine
or
propazine,
and
in
all
studies,
the
incidences
of
ovarian
tumors
are
(
as
would
be
expected)
very
low.
Although
the
causes
of
ovarian
cancer
are
not
definitively
known,
the
key
events
in
the
mode
of
action
established
for
atrazine,
i.
e.,
decreased
serum
LH
levels
and
a
decreased
number
of
ovulations
over
a
lifetime,
are
the
opposite
of
the
events
hypothesized
to
be
associated
with
ovarian
carcinogenesis.
Some
hypotheses
have
been
advanced
for
ovarian
carcinogenesis
with
the
predominant
hypotheses
being
the
"
incessant
ovulation"
hypothesis
and
the
"
gonadotropin"
hypothesis
(
e.
g.,
Fathalla,
1971).
This
hypothesis
suggests
that
damage
to
the
ovarian
epithelium,
resulting
from
frequent
ovulations,
leads
to
increased
risk
of
cancer
as
this
leads
to
increased
epithelial
cell
proliferation.
It
should
be
noted
that
epidemiology
studies
show
that
factors
which
decrease
the
number
of
lifetime
ovulations
­
such
as
pregnancy,
breast­
feeding
and
oral
contraceptive
use
­
reduce
ovarian
cancer
risk
(
Berchuck
and
Carney,
1997).
Therefore,
animal
data
do
not
support
a
biologically
plausible
mechanism
for
atrazine
contributing
to
ovarian
cancer
in
humans.

There
was
not
an
increase
in
any
dose
group
for
a
lymphoma
of
any
type,
including
non­
Hodgkin's
lymphomas
(
NHL)
in
atrazine
treated
SD
or
F344
rats.
The
simazine
chronic
bioassay
using
both
sexes
of
SD
rat
also
failed
to
see
any
lymphomas
in
any
animal
in
any
dose
group
(
McCormick,
1988).
Likewise,
no
animal
in
either
of
the
two
groups
examined
(
control
and
high
dose
tested
­
1000
ppm)
in
the
propazine
chronic
bioassay,
had
a
lymphoma
of
any
type
(
Jessup,
1980).
The
alterations
in
reproductive
hormones
that
atrazine
exposure
is
associated
with
have
not
been
linked
to
an
increased
risk
of
NHL.
NHLs
are
a
broad
group
of
neoplasias
which
originate
from
lymphoid
tissues
such
as
B­
cells,
T­
cells
and
histiocytes,
though
the
vast
majority
are
B­
cell
in
origin.
The
etiology
of
NHLs
are
unclear.
Generally
speaking,
increased
risk
of
developing
NHL
appears
to
be
associated
with
conditions
or
xenobiotic
exposures
that
result
in
immune
dysfunction
(
Scherr
and
Mueller,
1996).
An
association
between
NHL
and
reproductive
hormones
such
as
LH,
FSH,
estrogens
and
prolactin,
does
not
appear
to
be
present.
A
mechanistic
role
for
atrazine
contributing
to
NHL
has
not
been
identified
in
laboratory
studies.

In
summary,
multiple
animal
bioassays
do
not
reveal
an
increased
incidence
of
tumors
at
any
endocrine
site
other
than
mammary
gland
in
female
SD
rats.
Other
endocrine
tumors
that
have
been
raised
in
epidemiological
studies
can
not
be
biologically
tied
to
atrazine's
mode
of
action
(
i.
e.,
decrease
prolactin,
decrease
LH
and
suppression
of
ovulation).
Thus,
at
this
time,
based
on
the
available
animal
cancer
and
mode
of
action
data
and
epidemiological
studies,
there
is
no
tumor
endpoint
on
which
to
base
a
cancer
risk
assessment
for
atrazine.
EPA
has
considered
4
other
possible
modes
of
action
(
e.
g.,
stimulation
of
aromatase
activity)
and
finds
that
there
are
inadequate
data
to
support
these
hypotheses.
EPA's
Office
of
Research
and
Development
National
Health
and
Environmental
Effects
Laboratory
is
currently
conducting
a
detailed
investigation
of
the
effects
of
atrazine
and
related
chlorotriazines
on
aromatase
activity
and
steroidogenesis
in
an
effort
to
determine
whether
or
not
these
compounds
can
change
estrone,
estradiol
and
testostosterone
synthesis
in
response
to
atrazine
treatment
in
rats.
As
additional
data
become
available,
EPA
will
review
them.

Berchuck
A,
Carney
M.
Human
ovarian
cancer
of
the
surface
epithelium.
Biochem
Pharmacol.
1997
Sep
1;
54(
5):
541_
4.

Bosland
MC.
Male
reproductive
system.
In
Carcinogenesis.
Eds.,
MP
Waalkes,
JM
Ward.
New
Your:
Raven
Press,
Ltd.,
1994,
pp.
339­
402.

Damjanov
I.,
Ovarian
tumours
in
laboratory
and
domestic
animals.
Curr
Top
Pathol.
1989;
78:
1_
10.

Fathalla
MF.
Incessant
ovulation__
a
factor
in
ovarian
neoplasia?
Lancet.
1971
Jul
17;
2(
7716):
163.

Jessup,
D.
1980.
Two
year
oral
chronic
toxicity
study
in
rats.
IRDC
study
no.
382­
007.
MRID
00041408;
Acc.
No.
219502401.

McCormick,
C.
C.
and
Arthur,
A.
T.
Simazine­
Technical:
104­
Week
Oral
Chronic
Toxicity
and
Carcinogenicity
Study
in
Rats.
1988.
Study
Number:
2­
0011­
09.
MRID
number:
406144­
05.
Pharmaceuticals
Division,
Ciba­
Geigy
Corp.,
Summit,
NJ.

Scherr
PA
and
Mueller
NE,
Non­
Hodgkin's
lymphomas,
in
Eds
D.
Schottenfeld
and
JF
Fraumeni
Jr..
Cancer
Epidemiology
and
Prevention
New
York,
Oxford
University
Press,
1996,
pgs.
920­
945.

Stoker,
T.
E.,
Robbinette,
C.
L.,
Cooper,
R.
L.
2000.
Maternal
exposure
to
atrazine
during
lactation
suppresses
suckling_
induced
prolactin
release
and
results
in
prostatitis
in
the
adult
offspring.
Toxicol
Sci.
1999
Nov;
52(
1):
68_
79.

II.
Atrazine
and
prostate
cancer
epidemiology
The
following
review
addresses
epidemiology
related
to
prostate
cancer,
non­
Hodgkin's
5
lymphoma,
other
cancers
and
includes
a
brief
summary
of
earlier
reviews
and
the
most
recently
submitted
studies
related
to
atrazine
and
triazines.

A.
Prostate
Cancer
­
Manufacturing
plant
study
An
epidemiology
study
was
conducted
of
workers
at
the
Syngenta
St.
Gabriel
plant
where
atrazine
is
manufactured.
That
study
reported
a
statistically
significant
increase
in
the
incidence
of
prostate
cancer
among
plant
workers.
The
Agency,
upon
review
of
this
study,
requested
additional
information
on
the
exposure
profile
of
the
employees
diagnosed
with
prostate
cancer
and
this
information
was
provided
and
reviewed.
To
further
analyze
the
question
of
exposure
a
nested
case­
control
study
was
proposed
by
Syngenta
and
conducted
for
them
by
Health
Practice
Exponent
Inc.
(
Hessel
et
al.
2003).
Preliminary
results
of
this
study,
a
review
by
the
Agency,
comments
from
four
external
peer
reviewers,
a
Syngenta­
sponsored
expert
panel
review,
and
comments
by
the
Natural
Resources
Defense
Council
were
provided
to
the
EPA's
Scientific
Advisory
Panel
in
July
of
2003.
EPA's
view
of
the
study
was
that
the
increase
in
prostate
cancer
observed
in
the
St.
Gabriel
workers
was
probably
due
to
the
increase
in
PSA
screening
for
these
workers.

The
Panel
was
requested
to
comment
on
the
Agency's
conclusion
regarding
prostate
cancer
and
particularly
the
results
from
this
study.
The
specific
Agency
conclusion
that
EPA
asked
the
SAP
to
comment
on
was:
"
Due
to
the
lack
of
a
detailed
exposure
analysis
based
on
job
history
and
the
limited
statistical
power
due
to
the
small
sample
size,
atrazine
could
not
be
ruled
out
as
a
potential
cause
but
a
role
for
atrazine
seems
unlikely."

The
Panel's
analysis
of
the
St.
Gabriel
study
differed
to
a
degree
from
the
Agency's
conclusion.
The
SAP
did
conclude
that
"
the
increase
in
Prostate
Specific
Antigen
(
PSA)
screening
at
the
St.
Gabriel
plant
likely
led
to
an
increase
in
the
detection
of
cases
of
prostate
cancer."
Further,
the
Panel
noted
that
"[
s]
ubstantive
and
persuasive
arguments
have
been
made
to
support
the
EPA's
conclusion
that
PSA
screening
could
explain
the
observed
increase
in
prostate
cancer
incidence
in
the
workers."
Nonetheless,
the
Panel
did
not
believe
there
was
sufficient
evidence
to
conclude
that
it
was
"
unlikely"
that
atrazine
had
a
role
in
the
increased
prostate
cancer
cases
"
given
the
severe
limitations
of
the
St.
Gabriel
study,
particularly
those
pertaining
to
small
sample
size,
questionable
exposure
assessment
and
lack
of
an
appropriate
comparison
group."
According
to
the
SAP,
PSA
screening
may
be
only
a
"
partial
explanation"
for
the
increase
in
prostate
cancer
and
that
"
atrazine
cannot
be
ruled
out
as
a
potential
cause."

The
Agency
agrees
with
the
SAP's
analysis
and
has
rewritten
its
conclusion
as
follows:

The
increase
in
prostate
cancer
incidence
at
the
St.
Gabriel
plant
in
Louisiana
is
consistent
6
with
the
intensive
PSA
screening.
This
is
because
prostate
cancer
was
found
primarily
in
active
employees
who
received
intensive
prostate
specific
antigen
(
PSA)
screening,
there
was
no
increase
in
advanced
tumors
or
mortality,
and
proximity
to
atrazine
manufacturing
did
not
appear
to
be
correlated
with
risk.
No
evidence
was
identified
that
permit
a
determination
that
some
of
the
increase
was
likely
due
to
exposure
to
atrazine
although
atrazine
exposure
cannot
be
ruled
out
at
this
time.
However,
the
study
was
insufficiently
large
and
suffered
from
other
limitations
that
prevent
a
determination
that
all
of
the
increase
in
prostate
cancer
was
probably
due
to
the
intensive
screening
program.

One
of
EPA's
external
reviewers
agreed
with
this
finding
regarding
the
role
of
PSA
screening.
Dr.
Edward
Giovannucci
of
the
Harvard
School
of
Public
Health
stated,
"
Thus,
the
increased
excess
of
prostate
cancer
observed
in
the
Novartis
study
is
compatible
with
increases
expected
in
a
population
that
is
receiving
intensive
PSA
screening."
Another
reviewer,
Dr.
Aaron
Blair
of
the
National
Cancer
Institute,
though
not
in
full
agreement
with
Dr.
Giovannucci,
agreed
that
there
was
evidence
to
"
suggest
that
PSA
screening
may
well
explain
the
excess
incidence
of
prostate
cancer
in
this
cohort."

The
Scientific
Advisory
Panel
suggested
that
the
Agency
consider
additional
analysis
of
the
St.
Gabriel
cohort.
However,
the
resulting
sample
size
would
still
limit
the
opportunity
to
draw
further
conclusions.
The
Agency
questions
whether
additional
analysis
is
warranted
for
other
potential
risk
factors
(
such
as
smoking,
diet
and
previous
work
history,
and
nonoccupational
or
pre­
employment
exposure
to
triazine
herbicides).
Because
of
the
way
the
study
was
designed
this
information
is
not
available
to
investigators
and
it
may
not
be
feasible
to
obtain
such
information.
The
same
applies
to
the
suggested
analysis
of
family
history,
history
of
prostate
disease,
(
e.
g.,
benign
prostatic
hypertrophy
and
prostatitis),
and
additional
biologic
samples
that
allow
DNA
extraction.
The
St.
Gabriel
study
was
a
study
based
on
available
records
and
it
might
be
difficult
or
impossible
for
investigators
to
obtain
permission
from
all
or
most
subjects
or
nextof
kin
to
get
the
additional
information
outlined
by
the
SAP.
The
SAP
repeatedly
acknowledges
"
The
St.
Gabriel
cohort
study
suffered
from
several
limitations
that
could
lead
to
negative
findings
in
epidemiologic
studies
of
similar
design,
particularly
with
regard
to
the
very
small
sample
size"
which
can
greatly
hinder
the
statistical
power
to
detect
an
effect.

In
October,
2003,
Syngenta
provided
a
completed
report
on
the
nested
case­
control
study
(
cases
and
controls
selected
from
within
the
cohort)
by
Health
Practice
Exponent
Inc.
that
examined
in
more
detail
the
exposure
of
12
of
17
prostate
cancer
cases
and
examined
the
effect
of
screening
on
prostate
cancer
incidence
(
Hessel
et
al.
2003).
EPA
has
not
yet
reviewed
this
study
in
depth,
but
a
preliminary
reading
did
not
find
any
evidence
that
prostate
cancers
could
be
attributed
to
atrazine
exposure.
Statistical
analysis
suggest
that
PSA
screening
would
explain
some
or
all
of
the
elevated
rates
of
prostate
cancer.
The
study
authors
concluded
"
There
is
no
evidence
for
an
association
between
atrazine
exposure
and
prostate
cancer
among
the
workers
at
the
Syngenta
plant
in
St.
Gabriel.
The
increased
incidence
of
prostate
cancer
observed
in
the
previous
study
could
be
explained
by
the
PSA
screening
program."
7
B.
Prostate
Cancer
­
Agricultural
Health
Study
in
Iowa
and
North
Carolina
Tied
into
the
assessment
of
atrazine
and
prostate
cancer
is
the
recently
published
study
Alavanja
et
al.
(
2003).
This
large
prospective
cohort
study
of
55,332
male
pesticide
applicators,
known
as
the
Agricultural
Health
Study,
reported
on
the
risk
of
prostate
cancer
and
computed
odds
ratios
for
individual
pesticides
within
the
cohort.
For
atrazine
the
reported
odds
ratio
(
ratio
of
odds
in
favor
of
disease
among
exposed
to
the
odds
of
disease
among
the
unexposed,
an
odds
ratio
of
1.0
implies
no
increased
risk
from
exposure)
was
0.94
for
ever/
never
use
reported
by
questionnaire
with
a
95%
confidence
interval
of
0.78
to
1.14.
The
Agricultural
Health
Study
has
a
number
of
advantages
over
other
epidemiologic
studies
of
pesticides.
It
is
the
largest
study
of
its
kind,
determines
exposure
prior
to
disease
(
thus,
eliminating
recall
bias),
analyzes
a
wide
variety
of
potential
and
known
confounders
including
other
pesticide
exposures,
and
has
greater
statistical
power
to
detect
small
effects.

The
Scientific
Advisory
Panel
expressed
concern
that
the
use
of
ever/
never
use
atrazine
was
"
likely
an
inappropriate
exposure
metric"
and
that
other
factors
such
as
measures
of
continuous
or
intermittent
use
should
be
considered.
The
Panel
appeared
to
place
little
weight
on
the
dose­
response
analysis
based
on
cumulative
exposure
which
combined
duration,
frequency,
and
intensity
into
one
metric
that
did
not
show
any
association
between
atrazine
and
prostate
cancer.
The
Agency
agrees
that
other
exposure
metrics
might
be
considered
but
disagrees
that
ever/
never
use
and
cumulative
exposure
is
an
inappropriate
measure.
Available
pesticide
usage
data
suggest
that
the
pattern
of
use
of
atrazine
as
a
preharvest
herbicide
limits
variability
in
duration,
frequency,
and
intensity
of
use
and
the
dose­
response
analysis
is
a
sufficient
measure
to
account
for
this
source
of
variation.

The
Scientific
Advisory
Panel
expressed
concern
that
this
study
had
a
short
follow­
up
period
"
with
exposure
information
collected
at
the
start
of
follow­
up"
and
incorrectly
stated
this
was
less
than
five
years.
Because
study
subjects
were
queried
about
past
as
well
as
present
use
of
atrazine,
the
follow­
up
period
was
much
longer
than
five
years.
The
National
Cancer
Institute
is
planning
to
redo
the
prostate
cancer
study
with
a
much
larger
cohort
next
year
when
the
sample
size
will
be
approximately
twice
as
large.
However,
given
the
relatively
tight
confidence
interval
on
the
current
estimate
and
the
lack
of
any
evidence
of
dose­
response,
the
Agency
does
not
expect
the
new
study
to
produce
results
different
from
those
already
reported.
Nevertheless,
the
Agency
will
revisit
and
revise
these
conclusions
if
the
updated
prostate
cancer
study
produces
different
results
suggesting
a
risk
from
exposure
to
atrazine.

C.
Prostate
Cancer
­
Pesticide
use
data
and
cancer
incidence
in
California
counties
Two
studies
were
conducted
in
California
which
has
maintained
a
population­
based
cancer
registry
since
1988
and
a
state­
wide
pesticide
use
reporting
system.
Mills
(
1998)
obtained
1993
8
pesticide
usage
data
for
six
pesticides
with
a
suspicion
of
carcinogenicity
based
on
other
toxicologic
and
epidemiologic
studies.
These
data
were
compared
using
regression
analysis
with
county
age­
and
race­
adjusted
cancer
incidence
rates
(
1988­
92).
A
borderline
statistically
significant
correlation
was
found
between
atrazine
usage
and
prostate
cancer
in
black
males,
but
not
among
Hispanic,
White,
or
Asian
males.
This
study
is
subject
to
aggregation
bias
because
the
exposure
of
individuals
in
the
county
was
not
measured.
EPA
considers
such
studies
useful
for
guiding
future
studies,
but
not
for
reaching
conclusions
about
causation.

A
second
study
by
Mills
and
Yang
(
2003)
examined
the
effect
of
simazine
rather
than
atrazine
on
prostate
cancer
among
members
of
the
United
Farm
Workers
of
America.
The
study
found
a
borderline
significant
association
between
high
simazine
use
and
prostate
cancer.
Like
the
earlier
Mills
(
1998)
study,
this
study
suffers
from
aggregation
bias
and
a
crude
measure
of
exposure
(
total
poundage
of
active
ingredient
by
crop
and
county
for
a
given
time
period)
which
may
not
reflect
exposure
among
farmworkers,
90%
of
whom
are
not
actively
involved
in
applying
or
handling
pesticides
(
1999­
2002
data
from
presentation
on
National
Agricultural
Workers
Survey
by
S.
Gabbard,
J.
Nakamoto,
and
D.
Carroll,
September
24,
2003,
funded
by
the
Department
of
Labor).
The
use
of
total
poundage
of
active
ingredient
by
county
was
not
normalized
by
number
of
workers
and
is
especially
problematic
because
it
correlates
with
the
size
of
the
crop
and
acreage
in
the
county
and
many
other
factors
which
might
have
little
to
do
with
exposure
of
farmworkers.
For
example,
a
county
with
double
the
poundage
would
be
counted
as
having
double
exposure,
even
if
it
also
had
double
the
number
of
farmworkers
and
their
exposure
was
the
same
in
both
counties.

D.
Prostate
Cancer
­
Overall
conclusion
Studies
of
manufacturing
and
farming
populations
do
not
support
a
finding
that
atrazine
is
a
likely
cause
of
prostate
cancer.
The
Scientific
Advisory
Panel
stated
that
neither
the
Syngenta
St.
Gabriel
Plant
study
or
the
Agricultural
Health
Study
were
"
sufficient
for
EPA
to
conclude
that
there
is
no
causal
association
between
atrazine
exposure
and
prostate
cancer."
However,
the
Agency
does
not
find
any
results
among
these
studies
that
would
lead
us
to
conclude
that
potential
cancer
risk
is
likely
from
exposure
to
atrazine.

III.
Atrazine
and
epidemiology
of
non­
Hodgkin's
lymphoma
The
National
Cancer
Institute
has
performed
a
number
of
studies
of
non­
Hodgkin's
lymphoma
(
NHL)
in
farming
populations.
The
key
findings
related
to
atrazine
are
noted
below.

A
study
by
Schroeder
et
al.
2001
examined
two
subtypes
of
NHL
for
association
with
pesticide
exposure
based
on
earlier
data
from
Iowa
and
Minnesota.
The
negative
NHL
subtype
was
significantly
associated
with
5
pesticides.
For
atrazine,
the
odds
ratio
was
borderline
significant
(
Odds
ratio
=
1.7
with
95%
confidence
interval
1.0­
2.8).
The
Schroeder
study
9
cautiously
stated
"
In
conclusion,
we
found
weak
to
relatively
strong
associations
between
many
agricultural
exposures
and
t(
14;
18)­
positive,
but
not
t(
14;
18)­
negative
NHL."
and
"
Causal
relationships
.
.
.
are
plausible,
but
associations
should
be
confirmed
in
a
larger
study."

The
study
by
Schroeder
et
al.
(
2001)
contrasted
with
an
earlier
study
by
Zahm
et
al.
(
1993)
conducted
in
these
same
two
states
plus
Kansas
and
Nebraska
which
concluded,
"
In
our
judgment,
these
data
provide
little
evidence
that
atrazine
is
associated
with
NHL
among
white
men."

Added
to
these
two
conflicting
studies
is
a
study
by
DeRoos
et
al.
(
2003)
published
electronically
in
September
2003.
EPA
has
not
had
time
to
review
this
study
in
depth,
but
did
not
find
evidence
sufficient
to
implicate
atrazine
as
a
likely
cause
of
NHL.
This
study
stated
that
"
Reported
use
of
several
individual
pesticides
was
associated
with
increased
NHL
incidence"
but
that
"
limitations
of
our
data
hinder
the
inferences
we
can
make
regarding
specific
pesticides".
The
hierarchical
regression
odds
ratio
(
odds
ratio
adjusted
for
the
effect
of
exposure
to
other
pesticides)
for
atrazine
was
1.5
(
95%
CI
=
1.0
to
2.2)
which,
like
the
study
by
Schroeder
et
al
(
2001),
has
borderline
significance.
Studies
with
borderline
significance
increase
the
likelihood
that
chance
or
some
confounder
may
be
an
explanation
for
the
observed
findings.
The
authors
caution
that
"
some
of
the
positive
results
could
be
due
to
chance"
though
adjustment
for
the
influence
of
other
pesticides
in
the
analysis
makes
this
somewhat
less
likely.

Given
the
conflicting
results
and
the
extreme
caution
exhibited
by
the
National
Cancer
Institute
(
NCI)
in
making
its
conclusion
regarding
specific
pesticides,
the
EPA
has
concluded
that
evidence
is
not
sufficient
to
implicate
atrazine
as
a
likely
cause
of
non­
Hodgkin's
lymphoma.
Nevertheless,
EPA
will
consider
these
studies
in
conjunction
with
other
evidence
and
may
request
additional
external
review.
An
exhaustive
and
thorough
analysis
of
non­
Hodgkin's
lymphoma
and
pesticide
use
is
planned
by
NCI
for
2004­
5.
This
analysis
will
include
consideration
of
NHL
subtypes
and
many
other
factors
as
well.
Absent
compelling
information
in
the
interim,
EPA
has
determined
that
a
thorough
review
of
atrazine
and
NHL
should
be
conducted
when
the
NCI
data
are
available.

IV.
Atrazine
and
epidemiology
related
to
other
cancers
Other
cancers
besides
prostate
and
non­
Hodgkin's
lymphoma
were
found
to
have
an
elevated,
though
not
statistically
significant,
increase
in
risk
at
the
St.
Gabriel
plant
(
Delzell
et
al.
2001).
Other
studies
have
suggested
an
increased
risk
for
ovarian,
breast,
and
other
cancers.
However,
these
studies
are
at
best
preliminary
and
should
not
serve
as
a
basis
for
implicating
atrazine
as
a
human
carcinogen
due
to
their
methodological
limitations
and
the
absence
of
replication
in
other
populations.
The
National
Cancer
Institute
is
planning
a
review
of
atrazine
and
all
types
of
cancers
in
2004.
Given
the
much
larger
sample
size
and
strengths
of
this
Agricultural
Health
Study,
the
Agency
has
decided
further
review
of
other
cancers
and
atrazine
should
take
place
when
results
from
this
planned
analysis
are
available.
However,
if
other
studies
10
or
additional
compelling
evidence
becomes
available
in
the
interim,
the
Agency
will
expeditiously
review
the
new
evidence
and
the
potential
risk
resulting
from
exposure
to
atrazine.
A
summary
of
evidence
from
previous
reviews
of
atrazine
and
cancer
epidemiology,
and
studies
published
subsequently,
are
reviewed
in
the
following
section
V.
Results
of
other
reviews
of
cancer
and
atrazine
through
1999
Three
reviews
of
triazines
including
atrazine
and
cancer
epidemiology
have
been
reported
in
the
1996­
1999
time
period:

IARC.
1999.
IARC
Monographs
on
the
evaluation
of
carcinogenic
risks
to
humans.
Volume
73:
Some
chemicals
that
cause
tumours
of
the
kidney
or
urinary
bladder
in
rodents
and
some
other
substances.
World
Health
Organization,
Lyon,
France.

Neuberger
JS.
1996.
Atrazine
and/
or
triazine
herbicides
exposure
and
cancer:
an
epidemiologic
review.
Journal
of
Agromedicine
3(
2):
9­
30.

Sathiakumar
N,
Delzell
E.
1997.
A
review
of
epidemiologic
studies
of
triazine
herbicides
and
cancer.
Critical
Reviews
in
Toxicology
27:
599­
613.

These
three
reviews
identified
the
10
case­
control
studies
(
see
numbered
references
at
the
end
of
this
review,
1­
10)
and
two
published
cohort
studies
of
workers
exposed
to
triazines
at
manufacturing
plants
(
11­
12).

Neuberger
(
1996)
concluded
"
based
on
the
data
to
date
.
.
.
there
is
no
convincing
evidence
of
a
causal
association
between
atrazine
and/
or
triazine(
s)
and
colon
cancer,
soft
tissue
sarcoma,
Hodgkin's
disease,
multiple
myeloma,
or
leukemia.
.
.
.
There
is
a
suggestion
of
a
possible
association
between
atrazine
and/
or
triazine(
s)
with
ovarian
cancer
and
non­
Hodgkin's
lymphoma.
However,
the
ovarian
cancer
study
needs
to
be
replicated
and
the
NHL
studies
fall
short
of
providing
conclusive
evidence
of
risk
because
the
results
could
be
due
to
chance,
bias,
or
confounding."

Sathiakumar
and
Delzell
(
1997)
concluded
"
The
available
epidemiologic
studies,
singly
and
collectively,
do
not
provide
any
consistent,
convincing
evidence
of
a
causal
relationship
between
exposure
to
triazine
herbicides
and
cancer
in
humans."

The
International
Agency
for
Research
on
Cancer
(
1999)
summarized
the
human
carcinogenicity
data
as
follows:

A
combined
analysis
of
results
of
two
cohort
studies
of
agricultural
chemical
11
production
workers
in
the
United
States
showed
decreased
mortality
from
cancers
at
all
sites
combined
among
the
subset
of
workers
who
had
had
definite
or
probable
exposures
to
triazine.
Site­
specific
analyses
in
this
subset
of
workers
yielded
no
significant
findings;
a
non­
significant
increase
in
the
number
of
deaths
from
non­
Hodgkin's
lymphoma
was
seen,
but
was
based
on
very
few
observed
cases.

A
pooled
analysis
of
the
results
of
three
population­
based
case­
control
studies
of
men
in
Kansas,
eastern
Nebraska
and
Iowa­
Minnesota,
United
States,
in
which
the
risk
for
non­
Hodgkin's
lymphoma
in
relation
to
exposure
to
atrazine
and
other
herbicides
on
farms
was
evaluated,
showed
a
significant
association;
however,
the
association
was
weaker
when
adjustment
was
made
for
reported
use
of
phenoxyacetic
acid
herbicides
or
organophosphate
insecticides.
A
sub­
analysis
of
results
for
farmers
in
Nebraska,
the
State
in
which
the
most
detailed
information
on
atrazine
use
was
available,
showed
no
excess
risk
for
non­
Hodgkin's
lymphoma
among
farmers
who
had
used
atrazine
for
at
least
15
years,
after
adjustment
for
use
of
other
pesticides.
In
a
case­
control
study
of
non­
Hodgkin's
lymphoma
among
women
in
eastern
Nebraska,
a
slight,
nonsignificant
increase
in
risk
was
seen.
In
all
these
studies,
the
farmers
tended
to
have
an
increased
risk
for
non­
Hodgkin's
lymphoma,
but
the
excess
could
not
be
attributed
to
atrazine.

Less
information
was
available
to
evaluate
the
assoication
between
exposure
to
atrazine
and
other
cancers
of
the
lymphatic
and
haematopoietic
tissues.
One
study
of
Hodgkin
disease
in
Kansas,
one
study
of
leukaemia
in
Iowa­
Minnesota
and
one
study
of
multiple
myeloma
from
Iowa
gave
no
indication
of
excess
risk
among
persons
handling
triazine
herbicides.

In
a
population­
based
study
in
Italy,
definite
exposure
to
triazines
was
associated
with
a
two­
to
threefold
increase
of
borderline
significance
in
the
risk
for
ovarian
cancer.
The
study
was
small,
and
potential
confounding
by
exposure
to
other
herbicides
was
not
controlled
in
the
analysis.

Based
on
these
findings
the
IARC
concluded
"
There
is
inadequate
evidence
in
humans
for
the
carcinogenicity
of
atrazine."
In
a
critique
of
IARC
monographs,
Huff
(
2002)
questioned
the
decision
on
atrazine
and
animal
carcinogenicity
and
based
this
partly
on
the
Delzell
et
al.
(
2001)
report
showing
increased
risk
for
prostate
cancer
and
above
expected
levels
for
certain
cancers
including
buccal
cavity
(
3
observed,
2.1
expected),
esophagus
(
2
observed,
0.7
expected),
stomach
(
2
observed,
0.9
expected),
bladder
(
3
observed,
1.6
expected),
thyroid
(
2
observed,
0.6
expected)
and
leukemia/
lymphomas
(
7
observed,
4.5
expected).
These
data
are
based
on
Table
7
of
the
Syngenta
report
by
Delzell
et
al.
(
2001).
This
table
shows
9
different
estimates
of
risk,
not
counting
prostate
(
already
discussed)
and
certain
grouped
categories.
Therefore,
6
of
9
categories
exhibited
an
excess,
though
statistically
insignificant
risk.
Chance
alone
is
a
possible
explanation
for
such
findings.
In
addition,
bias
and
confounding
could
produce
such
results.
Therefore,
these
elevated,
nonsignficant
incidence
ratios
must
be
considered
preliminary
findings.
Until
these
findings
are
replicated
in
other
studies
that
address
the
serious
methodological
12
limitations
(
especially
the
low
statistical
power)
of
the
present
study,
they
should
be
regarded
as
spurious
or
suggestive
at
best.
Therefore,
the
EPA
disagrees
with
Huff
(
2002)
that
the
epidemiology
studies
provide
support
for
a
revision
of
the
IARC
classification
for
atrazine.

Independent
of
these
three
reviews,
EPA
has
performed
internal
reviews
of
all
of
the
above
studies
which
had
statistically
significant
findings
relevant
to
atrazine
or
triazines
including
additional
updates
to
the
manufacturing
plant
studies
submitted
to
EPA
but
not
published.
With
the
exception
of
the
possible
association
with
ovarian
cancer,
which
EPA
reviewers
stated
needed
to
be
confirmed
in
other
populations,
the
Agency
did
not
find
convincing
evidence
of
an
association
between
triazines
or
atrazine
and
cancer.

Huff
J.
2002.
IARC
monographs,
industry
influence,
and
upgrading,
downgrading,
and
under_
grading
chemicals:
a
personal
point
of
view.
International
Agency
for
Research
on
Cancer.
Int
J
Occup
Environ
Health.
8(
3):
249­
70.

VI.
Studies
published
since
1999
or
not
included
in
the
three
published
reviews
The
following
six
studies
were
reviewed
by
EPA
and
submitted
to
the
Scientific
Advisory
Panel
for
review
in
July
2003.

Alavanja
MCR,
Samanic
C,
Dosemeci
M,
et
al.
2003.
Use
of
agricultural
pesticides
and
prostate
cancer
risk
in
the
Agricultural
Health
Study
cohort.
Am
J
Epidemiol
157:
800­
814.

Delzell
E,
et
al.
2001.
Cancer
Incidence
Among
Workers
in
Triazine­
related
Operations
at
the
Novartis
St.
Gabriel
Plant"
Oct.
12,
2001.
MRID#
451521­
01
and
455184­
01,
Chemical
#
080803.
[
Technical
Report
170
pp.]

MacLennan
PA,
Delzell
E,
Sathiakumar
N,
et
al.
2002.
Cancer
incidence
among
triazine
herbicide
manufacturing
workers.
J
Occup
Environ
Med.
44:
1048­
1058.

MacLennan
PA,
Delzell
E,
Sathiakumar
N,
et
al.
2003.
Mortality
among
triazine
herbicide
manufacturing
workers.
J
Toxicol
Environ
Health
A
66(
6):
501­
517.

Mills
PK.
1998.
Correlation
analysis
of
pesticide
use
data
and
cancer
incidence
rates
in
California
counties.
Arch
Environ
Health.
53:
410­
3.

Mills
PK,
Yang
R.
Prostate
cancer
risk
in
California
farm
workers.
2003.
J
Occup
Environ
Med.
45:
249­
258.

Results
from
the
Alavanja
(
2003),
Delzell
(
2001),
Mills
(
1998),
and
Mills
and
Yang
(
2003)
have
already
been
discussed
above.
The
two
reports
by
MacLennan
et
al.
are
updates
to
the
two
earlier
reports
by
Sathiakumar
(
1992,
1995).
Most
of
the
results
in
these
two
studies
are
covered
in
much
more
detail
by
Delzell
et
al.
(
2001)
which
has
already
been
discussed
above
in
13
the
sections
on
prostate
cancer
and
other
cancers.
The
mortality
study
(
MacLennan
et
al.
2003)
did
find
a
borderline
significant
result
for
non­
Hodgkin's
lymphoma
based
on
4
observed
deaths
versus
1.1
deaths
expected.
The
authors
noted,
however,
that
"
one
of
the
decedents
whose
death
certificate
included
a
diagnosis
of
NHL
had
medical
records,
including
a
biopsy
report
that
indicated
a
diagnosis
of
poorly
differentiated
nasopharyngeal
cancer.
This
case
was
not
removed
from
our
analysis.
To
have
done
so
would
have
introduced
a
bias
because
there
is
no
satisfactory
procedure
for
removing
similarly
misclassified
cases
from
the
numerator
of
general
population
mortality
rates
used
to
calculate
the
expected
number
of
deaths.
Our
data
were
not
of
adequate
statistical
precision
to
demonstrate
trends
in
NHL
rates
or
SMRs
by
years
worked
and
years
since
hire."
This
acknowledgment
of
bias
based
on
a
misclassified
case
means
that
borderline
statistically
significant
finding
would
no
longer
be
significant
if
the
case
were
excluded.
As
stated
above,
this
evidence
is
not
sufficient
to
support
a
finding
that
atrazine
is
a
likely
cause
of
non­
Hodgkin's
lymphoma.

The
following
six
studies,
including
two
published
since
July
2003
meeting
were
not
submitted
to
the
Scientific
Advisory
Panel.

De
Roos
AJ,
Zahm
SH,
Cantor
KP,
Weisenburger
DD,
Holmes
FF,
Burmeister
LF,
Blair
A.
2003.
Integrative
assessment
of
multiple
pesticides
as
risk
factors
for
non­
Hodgkin's
lymphoma
among
men.
Occup
Environ
Med
60:
e11.(
http://
www.
occenvmed.
com/
cgi/
content/
full/
60/
9/
e11)

Hessel
PA,
Kalmes
R,
Smith
TJ,
Lau
E,
Mink
P,
Mandel
J.
2003.
A
Nested
Case­
Control
Study
of
Prostate
Cancer
and
Atrazine
Exposure.
Final
Report,
October
3,
2003
performed
by
Health
Practice
Exponent,
Inc.
and
submitted
by
Syngenta
Crop
Protection,
Inc.

Hopenhayn­
Rich
C,
Stump
ML,
Browning
SR.
2002.
Regional
assessment
of
atrazine
exposure
and
incidence
of
breast
and
ovarian
cancers
in
Kentucky.
Arch
Environ
Contam
Toxicol
42:
127­
136.

Kettles
MA,
Browning
SR,
Prince
TS,
Horstman
SW.
1997.
Triazine
herbicide
exposure
and
breast
cancer
incidence:
an
ecologic
study
of
Kentucky
counties.
Environmental
Health
Perspectives
105:
1222­
1227.

Schroeder
JC,
Olshan
AF,
Baric
R,
et
al.
2001.
Agricultural
risk
factors
for
t(
14;
18)
subtypes
of
non­
Hodgkin's
lymphoma.
Epidemiology
12:
701­
709.

Van
Leeuwen
JA,
Waltner­
Toews
D,
Abernathy
T,
et
al.
1999.
Associations
between
stomach
cancer
incidence
and
drinking
water
contamination
with
atrazine
and
nitrate
in
Ontario
(
Canada)
agroecosystems,
1987­
1991.
Int
J
Epidemiol.
28:
836­
40.

DeRoos
et
al.
(
2003),
Hessel
et
al.
(
2003),
Schroeder
et
al.
(
2001)
have
already
been
discussed
above.
Hessel
et
al.
(
2003)
is
discussed
at
the
end
of
the
section
on
"
Prostate
Cancer
­
Manufacturing
Plant
Study".
DeRoos
et
al.
(
2003)
and
Schroeder
et
al.
(
2001)
are
discussed
in
14
the
section
on
non­
Hodgkin's
lymphoma.

The
other
three
studies
by
Hopenhayn­
Rich
et
al.
(
2002),
Kettles
et
al.
(
1997),
and
Van
Leeuwen
et
al.
(
1999)
are
ecological
studies
where
the
unit
of
analysis
are
populations
or
groups
of
people
rather
than
individuals.
An
earlier
study
by
Kettles
et
al
(
1997)
suggested
an
association
between
triazine
exposure
and
breast
cancer
in
Kentucky.
However,
a
later
follow­
up
study
by
Hopenhayn­
Rich
et
al.
(
2002)
did
not
support
this
finding
for
Kentucky
and
instead
found
results
suggesting
a
protective
effect
for
atrazine
for
ovarian
cancer
and
no
effect
on
breast
cancer.
The
Hopenhayn­
Rich
et
al.
study
was
based
on
5
year,
age­
adjusted
cancer
rates
which
are
likely
to
be
more
stable
than
the
two
year
rates
used
by
Kettles
et
al.
A
study
by
Van
Leeuwen
et
al.
(
1999)
found
a
positive
association
between
atrazine
water
contamination
levels
and
stomach
cancer
among
40
ecodistricts
in
Ontario,
Canada,
and
a
negative
association
with
colon
cancer
suggesting
a
protective
effect
for
atrazine.
The
authors
"
noted
that
so­
called
`
ecologic
studies',
the
type
of
analysis
conducted
in
this
research,
have
a
number
of
weaknesses,
including
ecologic
fallacy
and
multiple
collinearity."
Stomach
cancer
has
been
declining
for
many
years
and
is
likely
associated
with
a
number
of
dietary
and
lifestyle
factors,
not
controlled
for
in
the
Van
Leewen
et
al.
(
1999)
study
(
Cancer
Rates
and
Risks,
4th
edition,
National
Cancer
Institute
1996).
All
of
these
studies
are
subject
to
aggregation
bias
because
the
actual
exposures
of
individuals
in
the
county/
district
or
how
long
they
resided
there
is
not
known.
As
noted
in
standard
epidemiology
texts,
ecologic
studies
"
can
suggest
avenues
of
research
that
may
be
promising
.
.
.
In
and
of
themselves,
however,
they
do
not
demonstrate
that
a
causal
association
exists"
(
Gordis
L.
Epidemiology.
W.
B.
Saunders
Company,
Philadelphia,
1996).
The
authors
themselves
warn
"
conclusions
concerning
causality
cannot
be
drawn"
(
Kettles
et
al.
1997).
An
ecologic
study
in
Kentucky
of
similar
design
to
those
above,
found
a
whole
host
of
factors
that
vary
across
an
urban­
rural
gradient
(
Blondell
JM.
Urban­
rural
factors
affecting
cancer
mortality
in
Kentucky,
1950­
1969.
Cancer
Detection
and
Prevention
11:
209­
223,
1988).
Persons
living
in
rural
areas
differ
not
only
in
terms
of
pesticide
exposures,
but
also
diet,
parity,
physical
activity,
exposure
to
viruses,
and
other
lifestyle
factors.
Appropriate
controls
are
critical
when
studying
the
relationship
between
pesticide
exposure
and
cancer.

VII.
EPA
Conclusion:
Atrazine
exposure
and
NHL
and
other
cancers
The
Agency
does
not
find
any
results
among
the
available
studies
that
would
lead
us
to
conclude
that
potential
cancer
risk
is
likely
from
exposure
to
atrazine.
EPA
plans
to
revisit
this
conclusion
upon
receipt
of
new
studies,
especially
those
from
NCI's
Agricultural
Health
Study
on
atrazine
and
all
cancers,
prostate
cancer,
and
non­
Hodgkin's
lymphoma,
all
of
which
are
planned
for
completion
in
the
next
1­
2
years.

Numbered
references
1.
Brown
LM,
Blair
A,
Gibson
R,
et
al.
1990.
Pesticide
exposures
and
other
agricultural
risk
15
factors
for
leukemia
among
men
in
Iowa
and
Minnesota.
Cancer
Research
50:
6585­
6591.

2.
Brown
LM,
Burmeister
LF
Everett
GD,
et
al.
1993.
Pesticide
exposures
and
multiple
myeloma
in
Iowa
men.
Cancer
Causes
Control
4:
153­
156.

3.
Burmeister
LF.
1990.
Cancer
in
Iowa
farmers:
recent
results.
Am
J
Ind
Med
18:
295­
301.

4.
Cantor
KP,
Blair
A,
Everett
G,
et
al.
1992.
Pesticides
and
other
agricultural
risk
factors
for
non­
Hodgkin's
lymphoma
among
men
in
Iowa
and
Minnesota.
Cancer
Research
52:
2447­
2455.

5.
Donna
A,
Crosignani
P,
Robutti
F,
et
al.
1989.
Triazine
herbicides
and
ovarian
epithelial
neoplasms.
Scan
J
Work
Environ
Health
15:
47­
53.

6.
Hoar
SK,
Blair
A,
Holmes
FF,
et
al.
1986.
Agricultural
herbicide
use
and
risk
of
lymphoma
and
soft­
tissue
sarcoma.
JAMA.
256:
1141­
7.

7.
Hoar
SK,
Blair
A,
Holmes
FF,
et
al.
1985.
Herbicides
and
colon
cancer.
Lancet.
1(
8440):
1277­
8.

8.
Hoar­
Zahm
S,
Weisenburger
DD,
Babbitt
PA,
et
al.
1988.
Case­
control
study
of
non­
Hodgkin's
lymphoma
and
agricultural
factors
in
eastern
Nebraska
(
abstract).
Am
J
Epidemiol
128:
901.

9.
Zahm
SH,
Wisenburger
DD,
Cantor
KP,
et
al.
1993a.
Role
of
the
herbicide
atrazine
in
the
development
of
non­
Hodgkin's
lymphoma.
Scan
J
Work
Environ
Health
19:
108­
114.

10.
Zahm
SH,
Weisenburger
DD,
Saal
RC,
et
al.
1993b.
The
role
of
agricultural
pesticide
use
in
the
development
of
non­
Hodgkin's
lymphoma
in
women.
Arch
Environ
Health.
48(
5):
353­
358.

11.
Sathiakumar
N,
Delzell
E,
Austin
H,
Cole
P.
1992.
A
follow­
up
study
of
agricultural
chemical
production
workers.
American
Journal
of
Industrial
Medicine
21:
320­
330.

12.
Sathiakumar
N,
Delzell
E,
Cole
P.
1995.
Mortality
among
workers
at
two
triazine
manufacturing
plants.
American
Journal
of
Industrial
Medicine
29:
143­
151.

cc:
atrazine
file
(
080803)
Catherine
Eiden
(
7509C)
