Page
1
of
8
UNITED
STATES
ENVIRONMENTAL
PROTECTION
AGENCY
WASHINGTON,
D.
C.
20460
TXR
No.
0054130
MEMORANDUM
DATE:
March
17,
2006
SUBJECT:
Human
Studies
Review
Board:
Final
Weight
of
Evidence
Comparison
of
Human
and
Animal
Toxicology
Studies
and
Endpoints
for
Hydrogen
Cyanide
Human
Health
Risk
Assessment.

DP
Barcode:
327266
PC
Code:
099801
FROM:
Ray
Kent,
Ph.
D.,
Chief
Reregistration
Branch
4
Health
Effects
Division
(
7509C)

William
Dykstra,
Toxicologist
Reregistration
Branch
4
Health
Effects
Division
(
7509C)

TO:
Tina
E.
Levine,
Ph.
D.,
Director
Health
Effects
Division
(
7509C)

This
document
presents
the
rationale
for
the
selection
of
an
acute
dietary
endpoint
and
point
of
departure
for
hydrogen
cyanide
based
on
data
in
two
human
studies
(
MRIDs
46769601
and
46769602).
OFFICE
OF
PREVENTION,
PESTICIDES,
AND
TOXIC
SUBSTANCES
Page
2
of
8
Background.
One
of
the
pesticides
currently
undergoing
reregistration
and
tolerance
reassessment
is
hydrogen
cyanide.
Although
cyanide
compounds
have
been
used
for
fumigation
of
a
variety
of
commodities
in
the
past,
the
only
remaining
use
on
food
for
cyanide
compounds
in
the
United
States
is
a
post­
harvest
fumigant
use
of
sodium
cyanide
for
citrus
grown
in
California
and
shipped
to
Arizona.
In
the
course
of
fumigation
with
sodium
cyanide,
hydrogen
cyanide
(
HCN),
a
highly
toxic
vapor
is
generated
by
acidification
of
sodium
cyanide.
Because
residues
of
HCN
may
remain
on
fumigated
citrus,
a
dietary
risk
assessment
is
required
as
part
of
the
pesticide
reregistration
and
tolerance
reassessment
process.

Dietary
risk
assessments
are
carried
out
on
all
food
use
pesticides,
and
ordinarily
both
acute
and
chronic
dietary
assessments
are
performed.
In
chronic
pesticide
dietary
assessments,
average
dietary
exposures
are
compared
to
chronic
reference
doses,
whereas
in
acute
dietary
assessments,
distributions
of
single
day
exposures
computed
from
distributions
of
residue
data
and
single
day
consumption
data
are
compared
to
acute
reference
doses.

The
HCN
risk
assessment
team
was
faced
with
a
shortage
of
toxicity
data
with
which
to
carry
out
the
required
dietary
assessments.
Ordinarily,
dietary
risk
assessments
on
pesticides
are
based
on
toxicity
and
residue
data
submitted
by
registrants,
but
there
were
no
registrant­
submitted
studies
on
cyanide
compounds
suitable
for
that
purpose.
There
is
a
chronic
Reference
Dose
(
RfD)
for
hydrogen
cyanide
in
the
Agency=
s
IRIS
database
(
http://
www.
epa.
gov/
iris/
subst/
0060.
htm).
The
chronic
RfD
is
based
on
a
two­
year
feeding
study
in
rats
(
Howard
and
Hanzal,
1955)
in
which
the
rats
were
fed
a
diet
that
had
been
fumigated
with
HCN.
The
risk
assessment
team
concluded
that
the
RfD
for
HCN
in
IRIS
was
suitable
for
assessment
of
chronic
dietary
risks.

The
HCN
team
did
not
identify
any
consensus
reference
values
for
assessment
of
acute
toxicity.
For
many
pesticides,
suitable
data
for
acute
dietary
assessment
are
lacking,
and
the
decision
is
made
to
either
not
perform
an
acute
assessment
because
the
toxicity
database
suggests
that
acute
toxicity
is
not
an
issue,
or
to
use
a
repeated
dose
point
of
departure
(
POD)
since
in
the
great
majority
of
cases,
PODs
from
repeated
dose
studies
are
lower
than
acute
study
PODs
and
therefore
protective
of
acute
toxicity.

It
was
clear
to
the
HCN
risk
assessment
team
that
acute
toxicity
of
HCN
is
very
much
an
issue,
and
that,
because
of
the
rapid
metabolism
of
cyanide
described
below,
chronic
reference
values
reflecting
day­
long
dosing
of
small
amounts
of
cyanide,
are
not
necessarily
protective
of
acute
effects
from
administration
of
the
same
total
amount
of
cyanide
administered
at
one
time.
In
the
course
of
developing
an
assessment
of
acute
dietary
risks
from
use
of
HCN,
the
risk
assessment
team
scoured
the
literature
for
appropriate
oral
toxicity
data,
considering
not
only
animal
toxicity
studies
on
cyanide
salts,
but
literature
reports
on
human
poisonings
with
cyanide
salts,
as
well
as
studies
on
cyanogenic
glycosides
­
natural
compounds
that
release
cyanide
upon
ingestion.
The
HCN
team
compared
human
data
with
animal
data
from
similar
types
of
toxicity
studies,
and
has
recommended
a
POD
based
one
human
study,
but
supported
by
a
number
of
other
studies.
The
studies
considered
in
this
analysis
are
presented
in
summary
form
and
then
the
weight
of
evidence
discussion
of
endpoint
and
POD
selection
follows.
Page
3
of
8
Chemical
Characterization.
Hydrogen
cyanide
or
hydrocyanic
acid
(
HCN)
is
a
colorless,
flammable
liquid
with
an
almond
odor.
The
boiling
point
of
HCN
is
26.5
C
and
it
has
a
very
high
vapor
pressure,
so
it
does
not
remain
very
long
in
the
liquid
state
at
atmospheric
pressure.
Hydrogen
cyanide
itself
is
not
a
registered
pesticide
in
the
U.
S.
It
is
generated
at
citrus
fumigation
sites
in
California
from
sodium
cyanide
and
sulfuric
acid.

Oral
Toxicity.
HCN
is
extremely
acutely
toxic
by
the
oral
route.
Based
on
case
history
data,
a
fatal
dose
of
HCN
in
humans
is
estimated
to
be
50­
90
mg
total
or
~
1
mg/
kg
BW
(
FDA,
1956).
The
lowest
fatal
dose
of
cyanide
(
form
not
specified)
reported
in
humans
was
estimated
as
0.56
mg
cyanide/
kg
BW
(
Gettler
and
Baine,
1938)
in
a
person
who
died
3
hours
after
being
found
in
a
coma.
The
effects
of
acute
cyanide
exposure
in
humans
are
dominated
by
central
nervous
system
and
cardiovascular
disturbances.
Cyanide
is
a
strong
inhibitor
of
cytochrome
oxidase,
the
terminal
enzyme
in
the
mitochondrial
electron
transport
chain.
Typical
signs
of
acute
cyanide
poisoning
include
tachypnea,
headache,
vertigo,
lack
of
motor
coordination,
weak
pulse,
cardiac
arrhythmias,
vomiting,
stupor,
convulsions,
and
coma.

In
contrast
to
its
high
acute
toxicity,
hydrogen
cyanide
exhibits
relatively
low
subchronic
and
chronic
toxicity
by
the
oral
route.
A
2­
year
rat
feeding
study
with
hydrogen
cyanide
fumigated
food
containing
residues
up
to
300
ppm
(
10.8
mg/
kg/
day)
showed
no
treatment­
related
effects
in
clinical
signs,
body
weight,
food
consumption,
hematology,
gross
necropsy
or
histopathological
changes
in
the
organs
examined.
Rapid
detoxification
of
hydrogen
cyanide
explains
how
the
rats
in
the
2­
year
study
could
ingest
2
to
3
times
the
fatal
acute
dose
each
day
throughout
their
lifetime
without
any
measurable
effects.
Similarly,
in
a
13­
week
repeated
dose
toxicity
study
in
which
sodium
cyanide
was
administered
in
the
drinking
water,
there
were
no
mortalities,
clinical
signs
or
histopathological
effects
in
the
brain,
thyroid,
or
other
organs
of
rats
and
mice
exposed
to
doses
up
to
12.5
and
26
mg/
kg/
day
respectively.

Metabolism.
The
unusual
pattern
of
cyanide
toxicity
is
explained
by
its
rapid
metabolism.
Hydrogen
cyanide
is
readily
converted
to
thiocyanate
in
the
liver
by
the
enzyme
rhodanese
(
Cyanide:
Thiosulfate
Sulfur
Transferase;
E.
C.
2.8.1.1).
This
metabolic
conversion,
first
demonstrated
in
1894,
is
the
major
detoxification
mechanism
for
hydrogen
cyanide.
Once
thiocyanate
is
formed,
it
is
not
converted
back
to
cyanide.
Conversion
of
cyanide
to
thiocyanate
is
enhanced
when
cyanide
poisoning
is
treated
by
intravenous
administration
of
a
sulfur
donor
(
e.
g.,
sodium
thiosulfate).
The
rapidity
of
this
conversion
was
demonstrated
in
a
case
of
humans
where
one­
twentieth
the
lethal
dose
(
0.1­
0.2
mg/
kg
BW)
of
soluble
cyanide
was
injected
intravenously
every
ten
minutes.
In
this
case,
more
than
the
equivalent
of
an
acutely
fatal
dose
was
reportedly
administered
without
any
apparent
harm
to
the
individual
(
FDA,
1956).
Based
on
the
FDA
study,
as
long
as
the
oral
intake
of
HCN
does
not
exceed
the
body=
s
detoxification
mechanisms
via
the
formation
of
thiocyanate,
HCN
can
be
ingested
in
the
form
of
dietary
residues
for
prolonged
periods
without
harm.
Page
4
of
8
Toxicity
Studies
Considered
for
Acute
Oral
Risk
Assessment
Animal
toxicity
studies
The
critical
study
(
Howard
and
Hanzal,
1955)
in
IRIS
that
serves
as
the
basis
for
the
chronic
RfD
is
described
as
follows:

In
this
2­
year
dietary
study,
rats
(
10/
sex/
group)
were
administered
food
fumigated
with
hydrogen
cyanide
(
HCN).
The
average
daily
concentrations
were
73
and
183
mg
CN/
kg
diet.
From
the
data
reported
on
food
consumption
and
body
weight,
daily
estimated
doses
were
4.3
mg
and
10.8
mg
CN/
kg
bw.
The
average
food
CN
concentrations
were
estimated
based
on
the
authors'
data
for
concentrations
at
the
beginning
and
end
of
each
food
preparation
period
and
by
assuming
a
first­
order
rate
of
loss
for
the
intervening
period.
There
were
no
treatment­
related
effects
on
growth
rate,
no
gross
signs
of
toxicity,
and
no
histopathologic
lesions
...
the
confidence
in
the
study
is
medium
because
adequate
records
of
food
consumption
and
body
weight
were
maintained,
and
animals
of
both
sexes
were
tested
at
two
doses
for
2
years
(
http://
www.
epa.
gov/
iris/
subst/
0060.
htm).

The
NOAEL
in
the
Howard
and
Hanzal
study
is
considered
to
be
the
highest
dose,
10.8
mg/
kg/
day.
In
addition
to
this
study,
the
following
information
is
available
on
the
oral
toxicity
of
cyanide
in
laboratory
animals:

$
The
ATSDR
Tox
Profile
for
cyanide
(
ATSDR,
2004
),
mentions
a
number
of
acute
lethality
studies
for
cyanide
salts.
The
acute
oral
LD50
in
the
rats,
mice
or
rabbits
from
administration
of
sodium
or
potassium
cyanide
varied
from
2
to
8
mg
of
CN­/
kg.

$
The
World
Health
Organization=
s
CICAD
document
for
hydrogen
cyanide
(
WHO,
2004)
discusses
a
90­
day
study
in
the
rat
(
Hayes,
1967)
with
potassium
cyanide
in
which
no
mortality
was
seen
at
a
dietary
dose
of
250
mg/
kg/
day
(
100
mg
CN­/
kg),
even
though
the
LD50
for
KCN
in
the
rat
is
10
mg/
kg
(
4
mg
CN­/
kg)
when
it
is
administered
by
gavage.

Human
studies
In
a
clinical
trial
of
amygdalin
(
Laetrile,
D­
mandelonitrile­
 ­
D­
glucosido­
6­
D­
glucoside),
one
hundred
seventy­
eight
cancer
patients
were
treated
with
amygdalin
plus
a
Ametabolic
therapy@
consisting
of
diet,
enzymes
and
vitamins.
Amygdalin
was
administered
at
an
intravenous
dose
of
4.5
g/
m2
of
body
surface
area/
day
(
standard
dose)
or
7
g/
m2
of
body
surface
area/
day
(
high
dose)
for
21
days.
Following
intravenous
treatment,
amygdalin
was
administered
orally
at
0.5
g
three
times
daily
at
the
standard
dose
regimen
or
0.5
g
four
times
daily
at
the
high
dose
regimen
until
the
patient
had
definite
evidence
of
progressive
malignant
disease
or
until
severe
clinical
deterioration
occurred.
Amygdalin
releases
HCN
upon
enzymatic
action
of
 ­
glucosidase
in
the
GI
tract
causing
high
blood
cyanide
level
and
produces
clinical
toxicity.
Clinical
signs
and
symptoms
consisting
of
nausea
(
30­
31%),
vomiting
(
17­
25%),
headaches
(
7­
8%),
dizziness
(
7­
Page
5
of
8
10%),
mental
obtundation
(
4­
5%)
and
dermatitis
(
2%)
were
observed
on
occasion
following
a
single
0.5
g
oral
dose
of
amygdalin.
Most
of
these
typical
symptoms
of
cyanide
toxicity
associated
with
elevated
blood
cyanide
levels
(
in
the
range
of
2
­
3
µ
g/
ml)
were
also
seen
when
two
0.5
g
oral
doses
were
taken
at
the
same
time,
or
shortly
apart,
to
compensate
for
a
missed
dose
or
when
raw
almonds
(
a
rich
source
of
 ­
glucosidase)
were
consumed
following
oral
therapy.
Therefore,
the
single
dose
of
0.5
g
of
amygdalin
was
determined
to
be
a
minimally
toxic
dose,
whereas
the
double
dose
was
generally
regarded
as
frankly
toxic.
(
Moertel
et
al.,
1982).

The
cyanide
risk
assessment
team
is
also
aware
of
another
human
study
described
in
a
Quarterly
Report
of
the
Association
of
Food
and
Drug
Officials
by
A.
J.
Lehman
(
FDA,
1956).
In
this
report
total
oral
doses
of
20
to
30
mg
sodium
cyanide
(
the
equivalent
of
11.0
to
16.5
mg
HCN
or
0.2
mg
HCN/
kg
BW)
were
administered
to
an
unknown
number
of
individuals
to
demonstrate
thiocyanate
in
saliva
without
apparent
adverse
effect.
The
risk
assessment
team
was
unable
to
obtain
additional
details
of
the
study,
and
therefore
neither
the
scientific
validity
nor
the
ethics
of
the
study
can
be
evaluated.

Weight
of
Evidence
Discussion.

The
goal
of
the
HCN
risk
assessment
team
was
to
be
able
to
assess
the
dietary
risk
of
eating
one
or
two
hydrogen
cyanide­
treated
oranges
in
one
sitting.
The
pesticide
tolerance
associated
with
the
post­
harvest
use
of
sodium
cyanide
on
citrus
is
50
ppm
(
50
mg
of
HCN
per
kg
of
citrus).
If
a
70
kg
person
were
to
eat
one
average
orange
(
including
peel)
weighing
160
g
and
containing
HCN
at
the
50
ppm
tolerance
level,
the
corresponding
dose
would
be
approximately
0.1
mg/
kg
which
is
1/
10
of
a
fatal
dose
in
humans.
The
risk
assessment
team
faced
two
tasks,
how
much
HCN
is
actually
available
for
ingestion
on
a
treated
orange,
and
what
toxicity
reference
value
is
appropriate
to
compare
with
the
estimated
exposure?

With
respect
to
an
appropriate
acute
reference
value,
it
was
clear
to
the
risk
assessment
team
that
the
IRIS
RfD
for
HCN
was
not
useful
for
assessing
acute
risks.
The
POD
on
which
the
chronic
RfD
is
based
is
a
NOAEL
in
the
rat
of
10
mg
CN­/
kg/
day
that
reflects
consumption
over
an
entire
day
of
HCN­
fumigated
diet;
whereas
the
acute
lethality
studies
in
animals
mentioned
in
the
ATSDR
(
2004)
Draft
Tox
Profile
for
cyanide
give
a
range
of
LD50s
for
sodium
and
potassium
cyanide
salts
from
2
­
8
mg
CN­/
kg.
It
is
apparent
from
a
comparison
of
the
acute
LD50
studies
and
the
chronic
study
NOAEL
of
10
mg/
kg/
day
that
cyanide
is
much
more
hazardous
when
administered
all
at
once
rather
than
throughout
the
day.

In
addition,
in
comparing
the
animal
LD50s
of
2­
8
mg
CN­/
kg
with
fatal
human
doses
of
~
1
mg/
kg,
there
is
a
suggestion
that
humans
may
be
more
sensitive
than
animals
on
a
mg/
kg
basis.
The
Cyanide
CICAD
(
WHO,
2004)
suggests
that
humans
may
be
less
able
than
rodents
to
detoxify
cyanide
because
of
a
lower
concentration
of
sulfur
donors
in
humans.
Page
6
of
8
The
HCN
team
was
not
able
to
identify
any
studies
in
laboratory
animals
suitable
for
assessment
of
acute
dietary
risks,
but
they
did
identify
the
potentially
useful
clinical
trial
on
the
cyanogenic
glycoside,
amygdalin.
In
the
clinical
study,
the
oral
dose
of
amygdalin
was
0.5
g
(
500
mg)
administered
three
times
per
day.
The
molecular
weight
of
amygdalin
is
457
and
each
molecule
of
amygdalin
contains
one
equivalent
of
HCN
(
molecular
weight
=
27).
Therefore
500
mg
of
amygdalin
has
the
potential
to
release
30
mg
of
HCN
or
0.4
mg/
kg
for
a
70
kg
person.
There
are
some
problems
with
using
the
clinical
study
of
amygdalin
for
establishing
an
acute
RfD
for
hydrogen
cyanide.
The
subjects
in
the
study
were
suffering
from
terminal
cancer
and
it
is
not
clear
from
the
study
to
what
extent
the
symptoms
of
toxicity
observed
were
attributable
to
amygdalin
administration
or
to
the
progression
of
cancer.
Oral
administration
of
amygdalin
did
not
commence
until
each
treated
patient
had
received
an
i.
v.
treatment
of
amygdalin
for
21
days,
which
confounds
the
interpretation
of
symptoms.
In
the
main
study,
the
listed
toxic
reactions
did
not
seem
to
differ
between
the
intravenous
route
and
the
oral
route,
despite
the
fact
that
only
oral
administration
of
amygdalin
resulted
in
appreciable
levels
of
blood
cyanide.

The
differences
in
blood
cyanide
concentrations
between
routes
of
exposure
was
explained
on
the
basis
that
conversion
of
amygdalin
to
cyanide
is
dependent
on
enzymes
(
s)
in
the
gut
or
present
in
foods
such
as
almonds.
Intravenous
amygdalin
is
not
exposed
to
 ­
glucosidase
in
the
liver
or
bloodstream
and
so
is
excreted
essentially
unchanged,
whereas
oral
amygdalin
is
converted
to
HCN.
There
is
information
of
an
anecdotal
nature
in
the
study
that
consumption
of
almonds
with
oral
amygdalin
enhances
toxicity,
highlighting
another
uncertainty
about
the
study
­
the
extent
and
rapidity
of
amygdalin
conversion
to
cyanide.
Other
anecdotal
information
from
the
study
indicates
that
doubling
up
of
the
500
mg
dose
clearly
results
in
more
severe
symptoms.
Taken
together,
the
data
indicate
that
500
mg
amygdalin
is
more
appropriately
considered
a
minimal
effect
level
with
respect
to
cyanide
toxicity.
Overall
attributing
those
symptoms
to
amygdalin
is
likely
to
be
conservative
Conclusion.

Despite
the
limitations
of
the
clinical
study
on
amygdalin,
OPP=
s
HCN
risk
assessment
team
concluded
that
oral
administration
of
500
mg
of
amygdalin,
equivalent
to
30
mg
of
HCN
or
approximately
0.4
mg/
kg
HCN,
provided
a
reasonable
basis
to
assess
acute
dietary
risks
from
ingestion
of
citrus
containing
residues
of
HCN.
The
potential
POD
of
0.4
mg/
kg
compares
to
the
reported
fatal
dose
in
humans
of
approximately
1
mg/
kg
(
FDA,
1956),
the
lowest
known
fatal
dose
in
humans
of
0.56
mg/
kg
(
Gettler
and
Baine,
1938)
and
the
unsubstantiated
NOAEL
for
CN­
in
humans
of
0.2
mg/
kg
(
FDA,
1956).

The
team
recommended
using
the
equivalent
NOAEL
of
0.4
mg
HCN/
kg
with
uncertainty
factors
of
10
x
for
intraspecies
variation
and
a
second
factor
of
10
x
to
account
for
a
combination
of
severity
of
effect,
steepness
of
dose­
response,
as
well
as
the
indications
from
the
clinical
trial
that
500
mg
amygdalin
per
administration
was
a
minimal
effect
level
and
not
a
NOAEL.
The
acute
RfD
is
therefore
0.42
mg/
kg
)
100
=
0.004
mg/
kg.
The
factor
is
1/
5
that
of
the
chronic
RfD
in
IRIS
based
on
chronic
dietary
administration
of
HCN­
treated
food
to
rats.
It
is
unusual
that
the
Page
7
of
8
chronic
RfD
is
exceeds
the
acute
RfD;
however,
HED
scientists
believe
that
use
of
the
acute
RfD
is
appropriate
whenever
cyanide
exposure
is
expected
to
occur
all
at
once.
Page
8
of
8
Bibliography
Agency
for
Toxic
Substances
and
Disease
Registry
(
ATSDR).
2004.
Toxicological
Profile
for
cyanide.
(
Draft
for
Public
Comment).
Atlanta,
GA:
U.
S.
Department
of
Health
and
Human
Services,
Public
Health
Service.

Food
and
Drug
Administration
(
FDA,
1956),
Association
of
Food
and
Drug
Officials:
Quarterly
Report
to
the
Editor
on
Topics
of
Current
Interest:
Hydrogen
Cyanide
and
Amygdalin;
A.
J.
Lehman,
Ph.
D.,
M.
D.,
Editor
Howard,
J.
W.
and
R.
F.
Hanzal.
(
1955).
Chronic
toxicity
to
rats
of
food
treated
with
hydrogen
cyanide.
J.
Agric.
Food
Chem.
3:
325­
329.

Moertel,
CG
et
al
(
1982).
A
Clinical
Trial
of
Amygdalin
(
Laetrile)
in
the
Treatment
of
Human
Cancer.
New
England
Journal
of
Medicine,
Vol.
306,
No.
4:
201­
206,
MRID
#
46769602,
published.

Moertel,
CG
et
al
(
1981)
A
Pharmacologic
and
Toxicological
Study
of
Amygdalin.
JAMA,
Vol.
245,
No.
6:
591­
594,
MRID
#
46769601,
published.

World
Health
Organization
(
WHO,
2004),
Hydrogen
Cyanide
and
Cyanides:
Human
Health
Aspects
(
CICADS
61,
2004),
Geneva
