Carbonic dioxide for eutanasia

7/30/2019 Carbonic dioxide for eutanasia

1/25

Carbon dioxide for euthanasia: concerns

regarding pain and distress, with special

reference to mice and rats

K M Conlee1, M L Stephens1, A N Rowan1 and L A King2

1The Humane Society of the United States, Animal Research Issues, 2100 L Street NW, Washington, DC20037, USA; 2Linacre College, Oxford University, St Cross Road, Oxford OX1 3PS, UK

Summary

Carbon dioxide (CO2) is the most commonly used agent for euthanasia of laboratory rodents,used on an estimated tens of millions of laboratory rodents per year worldwide, yet there is agrowing body of evidence indicating that exposure to CO2 causes more than momentary painand distress in these and other animals. We reviewed the available literature on the use ofCO2 for euthanasia (as well as anaesthesia) and also informally canvassed laboratory animalpersonnel for their opinions regarding this topic. Our review addresses key issues such asCO2 flow rate and final concentration, presence of oxygen, and prefilled chambers (theanimal is added to the chamber once a predetermined concentration and flow rate have beenreached) versus gradual induction (the animal is put into an empty chamber and the gasagent(s) is gradually introduced at a fixed rate). Internationally, animal research standardsspecify that any procedure that would cause pain or distress in humans should be assumed to

do so in non-human animals as well (Public Health Service 1986, US Department ofAgriculture 1997, Organization for Economic Cooperation and Development 2000).European Union guidelines, however, specify a certain threshold of pain or distress, such asskilled insertion of a hypodermic needle, as the starting point at which regulation of the useof animals in experimental or other scientific procedures begins (Biotechnology RegulatoryAtlas n.d.). There is clear evidence in the human literature that CO2 exposure is painful anddistressful, while the non-human literature is equivocal. However, the fact that a number ofstudies do conclude that CO2 causes pain and distress in animals indicates a need for carefulreconsideration of its use. Finally, this review offers recommendations for alternatives to theuse of CO2 as a euthanasia agent.

Keywords Carbon dioxide; euthanasia; pain; distress; anaesthesia; welfare; rodents

Carbon dioxide (CO2) is commonly used foreuthanasia and anaesthesia of laboratoryrodents, largely because of its ease of use,relative safety, and low cost, as well as itscapacity to euthanize large numbers ofanimals in a short time span (Ambrose et al.2000). In large institutions and those with

significant rodent-breeding programmes,large numbers of rodents are ofteneuthanized in a short time (Kline et al. 1963)

and an appropriate gas agent is often the bestway to approach such a challenge. However,CO2 is not physiologically inert and thepublished evidence on whether or not CO2administration causes pain or distress inanimals raises questions about its routineuse. This paper reviews this published

evidence and includes information on theeffects of CO2 in both humans and non-humans. Methodological details are includedwhen possible in order to provide a clear

Accepted 3 December 2004 r Laboratory Animals Ltd. Laboratory Animals (2005) 39, 137161

REVIEW ARTICLE

Correspondence: K M Conlee. Email: [email protected]


2/25

picture of how studies were conducted.Alternatives to the use of CO2 as a sole agentfor euthanasia are also suggested.

Animal welfare should be the main factortaken into consideration when choosing anappropriate method of euthanasia (AmericanVeterinary Medical Association [AVMA]2001). The least stressful procedure shouldbe utilized whenever possible the termeuthanasia being derived from a Greekterm meaning good death.

The current standards for euthanasia inthe US are set out in the 2000 Report of theAVMA Panel on Euthanasia. This report,written by the AVMA (2001), indicates that agood death is one that occurs without pain

and distress and the technique shouldminimize any stress and anxiety experiencedby the animal prior to unconsciousness.Similar statements are indicated in variousinternational guidelines and recommenda-tions regarding euthanasia (e.g. CanadianCouncil on Animal Care [CCAC] 1993,Close et al. 1996, 1997, ANZCCART2001).

Existing guidelines offer no clear,quantitative guidance on what techniques

might be considered without pain anddistress nor what would qualify as rapidunconsciousness seconds, tens of seconds,minutes? Nonetheless, the literature ondecapitation gives some sense of what maynot be acceptable in regards to length oftime. The 1986 and 1993 AVMA reports oneuthanasia cautioned against the use ofdecapitation because a single study of ratsfound that the decapitated head continued toproduce electroencephalogram (EEG) tracesfor an average of 13.6 s after decapitation(5.629.5 s) (see references in AVMA 2001).Therefore, due to lack of information, suchtimes specified in guidelines can start toguide us. The published literature indicatesthat CO2, at least under some circum-stances, causes pain and distress and doesso for longer than 10 s (there are conflictingreports regarding duration, but 10 s is theminimal amount of time reported in the

literature reviewed). Despite the evidencein the published literature that CO2 causesdistress, the authors have received anec-dotal reports from laboratory animal

veterinarians that CO2 appears humanewhen euthanasia is done properly by well-trained personnel.

Existing policies relevant to this issue

The Organization for Economic Co-operation and Development (OECD)guidance document on humane endpointsfollows the principle that if something isknown to cause pain, distress and sufferingin humans, it should be assumed to causethe same in animals (OECD 2000). Thisprinciple is also included in animal researchregulations and guidelines in the UnitedStates, including those promulgated by the

United States Department of Agriculture(1997) and the Public Health Service (1993),as well as those in other countries (e.g.CCAC 1993). As mentioned, EuropeanUnion guidelines follow this principle butspecify a certain threshold of pain or distressfor regulating animal use, for example usingskilled insertion of a hypodermic needleas guidance (Biotechnology RegulatoryAtlas n.d.).

Physiological effects/actions of CO2

CO2 causes a range of neurochemical,respiratory and vascular responses in bothhumans and non-humans (Woodbury &Karler 1960). Low to moderate concen-trations of CO2 (ranging from 5% to 35%)cause changes in respiration rate (Thomas &Spyer 2000), heart rate and blood pressure(Dripps & Comroe 1947, Kety & Schmidt1947, Smith & Harrap 1997), as well as HPAaxis activity (Barbaccia et al. 1996). Highconcentrations of CO2 initially cause similarresponses and may induce hyperventilationbefore respiratory and cardiac depression andsubsequent failure (Martoft et al. 2003). Theaccumulation of CO2 also causes acidi-fication of nasal mucosa (Anton et al. 1992,AVMA 2001). There is evidence that non-myelinated nerve endings that sensechemicals (Thurauf et al. 1991) and CO2-

sensitive olfactory receptors are present inthe nasal mucosa of mammals, includingrats (Coates 2001) and humans (Alvaro et al.1993). The existence of these nerve endings

Laboratory Animals (2005) 39

138 K M Conlee et al.


3/25

and olfactory receptors indicate the ability tosense any pain that may be associated withCO2.

Relevant human dataThere is evidence from human studies thatinhalation of CO2 at various concentrationscan cause pain and/or distress. For example,Danneman et al. (1997) asked 20 adulthumans to take a full breath of CO2 atdifferent concentrations ranging from 50%to 100% and to score each concentrationaccording to the level of discomfort. Resultsindicated that higher concentrations of CO2

were perceived to be increasingly noxious.Dannemans subjects used the followingterms in reference to every concentration ofCO2 tested: burning, tingling or prickling,and unpleasant (taste or odour); these termswere used more frequently at higherconcentrations. Many described 100%CO2 as piercing, stabbing, painful or causingthe eyes to burn or water, and 18 outof 20 subjects indicated that they wereunable to take a full breath of this

concentration.Dripps and Comroe (1947) found that

710% CO2 in oxygen caused increases inpulse rate, respiratory rate, and bloodpressure in male humans. Symptoms uponinhalation of CO2 included headache,dizziness and dyspnoea. Additionaldescriptions by the subjects includedirritation of the nose, palpitation, faintness,generally uncomfortable, muscle tremorand substernal pain. McArdle (1959)examined the effects of 30% CO2:70% O2 onthe heart rate of humans and found noserious disturbance to cardiac function.However, CO2 caused hyperventilation,severe acidosis, a significant rise in arterialpressure, and was associated with an overallsubstantial degree of stress.

The ability of CO2 to induce pain inhumans is underscored by the use of CO2 asa pain stimulus in humans. For example,

Anton et al. (1992) examined pain thresholdsinduced by CO2 in the nasal mucosa inhumans. A linear relationship was foundbetween CO2 concentration and pain

sensation. Average individual tolerancethresholds ranged from 40% to 55% CO2.

Animal studies regarding CO2 as

a euthanasia agent: generalinformation

CO2 euthanasia occurs via administration ofthe inhalant gas in a sealed container, withthe purpose of inducing unconsciousnessand death (Britt 1986). One source of CO2 isdry ice, but a pressurized cylinder of CO2 isnow viewed by a number of internationalanimal research oversight authorities as theonly acceptable source (United Kingdom

1997). CO2 may be administered in the homecage or in a specialized compartment andmay be used to kill individuals or smallgroups of animals.

Research regarding CO2 use in differentspecies has examined several parameterssuch as blood pressure (Smith & Harrap1997), heart rate (Coenen et al. 1995, Smith& Harrap 1997, Leach et al. 2002a,b),behaviour (e.g. Smith & Harrap 1997, Leachet al. 2002a,b), times to anaesthesia and

death (Blackshaw et al. 1988, Dannemanet al. 1997, Kohler et al. 1999), EEG activity(Thurauf et al. 1991), histology (Britt 1986,Danneman et al. 1997) and blood pH (Hewettet al. 1993).

Procedural factors such as the CO2concentration, flow rate and the presence ofoxygen have also been included in the assess-ment of CO2. There are wide variations inthe methods used, measurements takenand the resulting recommendationsthroughout the published literature; thesefactors limit the comparability of results,although some trends can be discerned.Discussions of CO2 euthanasia with variouspeople working in laboratory animalmedicine and care (e.g. veterinarians,vivarium directors, technicians) reveal thatthere are conflicting CO2 practices andrecommendations within the animalresearch community. As one example, some

institutions require that the euthanasiachamber be prefilled with CO2, while othersprohibit the use of prefilled chambersbecause they appear to cause animal distress.


Reassessing carbon dioxide use for euthanasia 139


4/25

Similar discrepancies in practice have alsobeen noted in regards to concentration, flowrate and presence of oxygen.

Pain and acute stress studies utilizing CO2

CO2 has been used as a pain and/or stressstimulus in animals. In some cases, theconcentrations used correlate withconcentrations used for CO2 euthanasia.Thurauf et al. (1991) exposed rats to variousconcentrations of CO2 (090%) andmeasured evoked potentials in EEGrecordings in order to determine the origin ofnegative mucosal potential (NMP negativepotential recorded from the respiratory

mucosa following painful stimulation withCO2). Local anaesthetics eliminated NMPsand EEG cortical responses, signifying thatthe pain response had ceased. When no localanaesthetic was administered, the result wasincreased NMPs, indicating an increasednociceptive response.

Barbaccia et al. (1996) used CO2 to elicit astress response in rats in order to examinethe effects of acute stress on brain steroidconcentrations and GABAA receptor

function. A combination of 35% CO2 and65% O2 inhaled from gas cylinders for oneminute caused a sufficient stress responsefor the study. It was concluded that exposureto CO2 is correlated with an increase invarious brain neuroactive steroid concen-trations.

In sum, it is noteworthy that CO2 is usedto induce pain and stress in animals, and theconcentrations used in these studies are thesame or similar to those used for anaesthesiaand euthanasia.

Effects of CO2 administration

CO2 effects have been examined at variousconcentrations. A comparison of resultsillustrates concerns regarding pain anddistress at concentrations ranging from 25%to 100%. The results of a number of studiesare discussed here and specifics are provided

in Tables 1 and 2 for mice and rats.Danneman et al. (1997) euthanized rats

with various concentrations of CO2(50100%) combined with oxygen via either

prefilled or gradual induction (GI) chambersand found a number of adverse effects.Measures included time to recumbency,time to anaesthesia (shallow breathing orlack of response to toe pinch) and anyclinical adverse effects. Adverse reactions,prior to death, included seizures, convulsivechewing, nasal haemorrhage, sero-sanguinous nasal discharge and excessivesalivation. The frequency and severity of theadverse reactions were inversely related toCO2 concentration. It appears that thesereactions occurred prior to or at onset ofanaesthesia, but the publication does notmake a clear distinction. Whether adversereactions occur prior to or after unconscious-

ness is an important consideration andshould be made clear in publications. Basedon Dannemans results, it appears that lowerconcentrations (at least when considering50% CO2 and higher) cause increasedincidence of adverse effects.

Ambrose et al. (2000) found that aconcentration of 60% CO2, supplied from agas cylinder, caused many adverse effects inmice, whereas 30% was not as problematic.Simply introducing CO2 into the chamber

caused behaviours such as increasedlocomotion, rearing, defaecation andurination. These behaviours are consideredby Ambrose and others to be indicative ofdistress, but further investigation is neededfor verification. Overall, the authorsconcluded that 60% CO2 caused an undueamount of stress (Ambrose et al. 2000).

Concentration, induction method andpresence of oxygen with CO2 were examinedin rats by Coenen et al. (1995). Gas cylinderswere used and gas was humidified in allconditions. The authors reported four phasesin each case of CO2 use: normal behaviour(phase I); continuous abnormal activity,excitation and agitation at a higher rate thannormal (phase II); sagging of hindlegs andloss of body control (phase III); disappearanceof muscle tone and head sinking (phase IV).Phase II was seen more frequently when100% CO2 was used. The same adverse signs

were also observed at lower concentrations(without O2), but less frequently than in100% CO2. Animals experienced asphyxia(as evidenced by gasping with mouth open




5/25



Table1S

pecificsofmicestudiesthatexaminedcarbondioxideexposure

Re

ference

Pre

fille

d(PF)

orgra

dua

l

induct

ion

(GI)

Gasm

ixt

ure/fina

l

concentr

ation

Est

imated

timeto

unconsc

ious-

ness

(s)

Est

imated

timeto

death(s)

Be

hav

iour

be

fore

unconsc

iousness

Physio

logy/

neuro

log

ica

l

measures

Comments

Am

broseet

al

.

(2000)

GI

CO2:O2/5

0%:50%

within

110s

93.4

250.5

m

Locomotion,

rearing,

de

faecationan

d

urination

Postmortem:m

alveo

lar

haemorr

hag

ic

conso

lidationre

lative

toCO2

alone

GI

CO2:O2/5

0%:50%

within

304

0s

65.0

177.8

m

Locomotion,

rearing,

de

faecationan

d

urination

(sign

ifi-

cantly

higherthan

thetwoother

groups)

Postmortem:m

alveo

lar

haemorr

hag

ic

conso

lidationre

lative

toCO2

alone

60%:20%

pro

duces

fa

sterunconsc

ious-

ne

ss;Authors:

distressbe

hav

iour

may

be

dueto

ra

pidityo

fm

inCO2

co

ncentration

Am

broseet

al

.

(2000)

GI

CO2

alon

e/50%

within

110s

93.4

235.6

m

Locomotion,

rearing,

de

faecationan

d

urination

Postmortem:m

alveo

lar

haemorr

hag

ic

conso

lidationre

lative

toCO2

+O2

Britt(1986

)

GI

CO2:O2/

70.8%

:5.5

%

55.0

242.0

Force

dbreathing

Hyperventilation

Some

lung

haemorr

hage

Urination,

de

faecation

Iwarsson&

Re

hbinder

(1993)

GI

CO2:O2/8

0%:20%

13.0

130.0

Mildhyperpnoeaan

d

shivering,

mildstress

response

Lungs:oe

demaan

d

haemorr

hage

Blackshaw

et

al.

(1988)

PF

CO2

alon

e/97%

12.8

48.0

Nonerecorde

d(on

ly

movementwas

measure

dan

don

ly

for10s)

NS

Used

dry

icemay

be

difficu

ltto

ma

intainconsistent

co

ncentration

Britt(1986

)

PF

CO2

alon

e/100%

14.0

38.0

Force

dbreathing,

hyperventilation;

sha

king;urination,

de

faecation

Highconcentration=

more

lung

haemorr

hag

ing

be

fore

death

(continue

d)


6/25



Table1(continued)

Re

ference

Pre

fille

d(PF)

orgra

dua

l

induct

ion

(GI)

Gasm

ixt

ure/fina

l

concentr

ation

Est

imated

timeto

unconsc

ious-

ness

(s)

Est

imated

timeto

death(s)

Be

hav

iour

be

fore

unconsc

iousness

Physio

logy/

neuro

log

ica

l

measures

Comments

Iwarrson&

Re

hbinder

(1993)

PF

CO2

alon

e/100%

8.0

35.0

Mo

deratehyperpnoea,

bo

dyst

iffnessan

d

excitementbe

fore

losso

frig

hting

reflex,

mildstress

response

Lungs:m

ildtomo

der-

atelungemp

hysema

Leacheta

l.

(2002a)

PF

CO2

alon

e/28%

or

36%o

r53%

NA

NA

Shorterw

ithdrawa

l

an

ddwe

llingtimes

compare

dto

ha

lothane,

iso

flurane

an

den

flurane

(there

foremore

avers

ive

)

NS

Thisstu

dy

didnot

invo

lveeuthanasia

Leacheta

l.

(2002b)

PF

CO2

/28%

or36%

or

53%

NA

NA

Shortestw

ithdrawa

l

an

ddwe

llingtimes

in

comparisontoother

mousegroups

(there

foremost

avers

ive

);a

llconcentra-

tionsequa

llyavers

ive

NS

Thisstu

dy

didnot

invo

lveeuthanasia

PF

CO2:argo

n/

20%:1

3%

or

20%:7

.9%

or

20%:4

.5%

NA

NA

Shorterw

ithdrawa

lan

d

dwe

llingtimes

incom-

parisontoargona

lone

;

allconcentrations

equa

llyavers

ive

NS

Thisstu

dy

didnot

invo

lveeuthanasia

PF

CO2:argo

n/

30%:1

4%

or

30%:1

0.1

%

or

30%:5

.3%

NA

NA

Shorterw

ithdrawa

lan

d

dwe

llingtimes

incom-

parisontoargona

lone

;

allconcentrations

equa

llyavers

ive

NS

Thisstu

dy

didnot

invo

lveeuthanasia

Thistableincludestheresultsofstudiesofmiceonly,

forgroupsthatwereexposedtoCO2.

Therefore,

ifastudycomparedCO2

withotheragents,

informationofotheragentswasnotinclu

dedinordertokeepthe

tableasshortaspossible.

Timetounconsciousnesswas

determinedconservativelyastheonsetof

posturalcollapseandlossofmuscletone(

Woodburyet

al.

1958,

Coenenet

al.

1995);thepresentauthors

calculatedthesetimeswhentheywerenotspecifieddir

ectlybytheauthorsoftheoriginalpapers.

NS=notspecified


7/25



Table2S

pecificsofratstudiesthatexaminedCO

2

exposure

Re

ference

Pre

fille

d(PF)

orgra

dua

l

induct

ion

(GI)

Gasm

ixtu

re/fina

l

concentra

tion

Est

imated

timeto

unconsc

ious-

ness

(s)

Estim

ated

time

to

death(s)

Be

hav

iour

be

fore

unconsc

iousness

Physio

logy/neuro

log

ica

l

measures

Comments

Britt(1986

)

GI

CO2:O2

/61

.5%:

7.8

8%

70.0

395.0

Force

dbreathing

Hyperventilation

Urination,

de

faecation

Some

lung

haemorr

hage

Coenenet

al.

(1995)

GI

CO2

alone

/NS

65.0

530.0

Force

dbreathing;

unspeci

fie

dsigns

ofag

itationan

d

excitation

EEGstages:norma

l;

lowere

damp

litude;

slow

wavesan

d

largesp

ikes

Slow

waveEEG

commence

dat

onseto

fpostura

l

collapse

Heartrate:norma

l;

rap

iddecrease;

cessation

GI

CO2

alone

/NS

40

395

Force

dbreathing;

authorreporte

d

signso

fag

itation

an

dexcitation

(notdescri

be

d)

Sameasa

bove

Slow

waveEEG

commence

d

be

forepostura

l

collapse

GI

CO2:O2

/80

%:20%

50.0

618.0

Nosignso

fasp

hyx

ia

or

force

dbreathing

Sameasa

bove

Slow

waveEEG

commence

d

afterpostura

l

collapse

Danneman

et

al.(1

997)

GI

CO2

alone

/60%

NS

27.3

NS

Lungs:marke

dperivascu

lar

haemorr

hage;

perivascu

laroe

dema;

intra-a

lveo

lar

haemorr

hage

Haemorr

hage

from

nose;

excess

ivesa

livation;

nasa

ldischarge

GI

CO2

alone

/60%

un-

tilanae

sthetize

d

then100%

NS

15.1

NS

Haemorr

hage

from

nose;

excess

ivesa

livation;

serosangu

inousnasa

l

discharge

(continue

d)


8/25



Danneman

et

al.(1

997)

GI

CO2

alone

/70%

NS

16.4

NS

Lungs:marke

dperivascu

lar

haemorr

hage;

perivascu

laroe

dema

an

dintra-a

lveo

lar

haemorr

hage;

Serosangu

inousnasa

l

discharge;nasa

l

haemorr

hage

GI

CO2

alone

/70%

un-

tilanae

sthetize

d

then100%

NS

16.1

NS

Serosangu

inousnasa

l

discharge;nasa

l

haemorr

hage

GI

CO2

alone

/80%

NS

13.5

NS

Lungs:marke

dperivascu

lar

oe

dema;perivascu

lar

an

dintra-a

lveo

lar

haemorr

hage

Serosangu

inousnasa

l

discharge;nasa

l

haemorr

hage

GI

CO2

alone

/80%

un-

tilanae

sthetize

d

then100%

NS

15.4

NS

Serosangu

inous

nasa

ldischarge

GI

CO2

alone

/100%

NS

6.0

NS

Lungs:perivascu

lar

oe

demaan

dintra-

alveo

lar

haemorr

hage

Hack

barth

et

al.

(2000)

GI

CO2

alone

/55%

455

0s

(losso

f

movement)

85s

(musc

le

relaxation

)

NS

Tachypnoea

followe

d

bystationaryp

hase

an

dsu

bsequentre-

laxationo

fmusc

les

Nochanges

inACTHan

d

cort

icosterone

leve

ls

Convu

lsions

occurre

d

between

loss

ofmovement

an

dmusc

le

relaxation

GI

CO2

+ace

proma-

zine/55

%

CO2

Sameasa

bove

NS

Sameasa

bove

HigherACTHan

d

cort

icosterone

leve

ls

thanCO2

alonegroup

Sameasa

bove

GI

CO2

+pen

tobarb

i-

tone/55

%

CO2

Sameasa

bove

NS

Sameasa

bove

Sameasa

bove

Sameasa

bove

Table2(continued)

Re

ference

Pre

fille

d(PF)

orgra

dua

l

induct

ion

(GI)

Gasm

ixtu

re/fina

l

concentra

tion

Est

imated

timeto

unconsc

ious-

ness

(s)

Estim

ated

time

to

death(s)

Be

hav

iour

be

fore

unconsc

iousness

Physio

logy/neuro

log

ica

l

measures

C

omments


9/25



Hewettet

al

.

(1993)

GI

CO2

alone

/100%

1031

09.0

(loss

ofrig

hting

)

1391

44.4

(losso

fpe

da

l

reflex

)

374.

0

NS

CO2

lea

dstok

inpH,

PaO2,

an

d%

Sat.o

fHb;

m

inPaCO2

GI

CO2:O2:N2

/

75%:20

%:5%

97.9

(losso

f

rig

hting

)

140.8

(losso

f

pe

da

lre

flex

)

NS

NS

NS

Sameasa

bove

Hornett&

Haynes

(1984)

GI

CO2

alone

/NS

160foreach

group

NS

Appre

hension;

laboure

d

breathing

NS

Iwarrson&

Re

hbinder

(1993)

GI

CO2:O2

80

%:20%

27.0

196.

0

Milduneasiness

an

dlaboure

d

breathing,

mild

stressresponse

Lungs:marke

dhyperaem

ia,

oe

dema,

an

d

haemorr

hages

Sm

ith&

Harrap

(1997)

GI

CO2

alone

/75%

97.2

541.

2

Imme

diategasp

ingor

laboure

dbreathing

Rap

iddrop

inbloo

d

pressurea

fter4m

in

followe

dbye

levationto

norma

lleve

lsfor2m

in

priortoco

llapse

Blackshaw

et

al.(1

988)

PF

CO2

alone

/97%

7.8

135.

0

Nonerecorde

d(on

ly

movementwas

measure

dan

don

ly

for10s)

NS

U

seddry

icemay

be

difficu

ltto

ma

intain

consistent

concentration

PF

CO2

alone

/97%

11.8

78.0

Sameasa

bove

NS

Sameasa

bove

Britt(1986

)

PF

CO2

alone

/100%

18.0

171.

0

Hyperventilation

Some

lung

haemorr

hage

Sha

king

Urination,

de

faecation

Coenenet

al.

(1995)

PF

CO2

alone

/100%

30

408

Signso

fasp

hyx

ia;

authorreporte

dsign

s

ofag

itationan

d

excitation

(not

descri

be

d)

EEGstages:norma

l;

lowere

damp

litude;

slow

wavesan

d

largesp

ikes

Heartrate:norma

l;

rap

iddecrease;

cessation

Slow

waveEEG

commence

d

be

forepostura

l

collapse

(continue

d)


10/25



Danneman

et

al.(1

997)

PF

CO2

alone/60%

NS

19.0

NS

Lungs:marke

dperivascu

lar

an

da

lveo

lar

haemor-

rhage;perivascu

lar

oe

dema

Haemorr

hage

from

nose;

excess

ivesa

livation;

serosangu

inousnasa

l

discharge

PF

CO2

alone

/60%

un-

tilanae

sthetize

d

then100%

NS

19.0

NS

Haemorr

hage

from

nose;

excess

ivesa

livation;

serosangu

inousnasa

l

discharge

PF

CO2

alone/70%

NS

15.3

NS

Serosangu

inousnasa

l

discharge;

haemorr

hage

from

nose

PF

CO2

alone

/70%

un-

tilanae

sthetize

d

then100%

NS

8.8

NS

Serosangu

inousnasa

l

discharge;

haemorr

hage

from

nose

PF

CO2

alone/80%

NS

11.1

NS

Lungs:marke

dperivascu

lar

haemorr

hage;

perivascu

laroe

dema

an

dintra-a

lveo

lar

haemorr

hage,

Serosangu

inousnasa

l

discharge

PF

CO2

alone/80%/un-

tilanae

sthetize

d

then100%

NS

8.7

NS

Serosangu

inousnasa

l

discharge

PF

CO2

alone/100%

NS

4.6

NS

Lungs:perivascu

lar

oe

dema;perivascu

lar

an

dintra-a

lveo

lar

haemorr

hage

Table2(continued)

Re

ference

Pre

fille

d(PF)

orgra

dua

l

induct

ion

(GI)

Gasm

ixtu

re/fina

l

concentra

tion

Est

imated

timeto

unconsc

ious-

ness

(s)

Estim

ated

time

to

deat

h(s)

Be

hav

iour

be

fore

unconsc

iousness

Physio

logy/neuro

log

ica

l

measures

Comments


11/25



Hewettet

al.

(1993)

PF

CO2

alone

/100%

19.0

(losso

f

rig

hting

)

27.8

(losso

f

pe

da

lre

flex

)

109.2

NS

CO2

lea

dstom

PaCO2;

k

pHan

dpercent

saturationo

fHb

Reporte

dno

indicationo

f

distress,

but

be

hav

iours

otherthan

locomotionnot

measure

d

Iwarsson&

Re

hbinder

(1993)

PF

CO2

alone

/100%

13.0

116.0

Mo

derateuneasiness,

tachypnoea,

urination,

de

faecation,

mildmo

derate

stressresponse

Lungs:emp

hysema,

oe

demaan

d

extravasationo

f

bloo

d

PF

CO2

alone

/100%

13.0

116.0

Mo

derateuneasiness,

tachypnoea,

urination,

de

faecation,

mild-mo

derate

stressresponse

Lungs:emp

hysema,

oe

demaan

d

extravasationo

f

bloo

d

Leacheta

l.

(2002a)

PF

CO2

alone

/25.5

%

or34.9

%

or

50.8

%

NA

NA

Shorterw

ithdrawa

l

an

ddwe

llingtimes

incomparison

withha

lothane,

iso

fluranean

d

en

flurane

(there

fore

moreavers

ive

);

sign

ificantincrease

inwash

ing

NS

Thisstu

dy

didnot

invo

lveeutha-

nasia

PF

CO2

alone

/25.5

%

or34.9

%

or

50.8

%

NA

NA

Shortestw

ithdrawa

l

an

ddwe

llingtimes

incomparisonw

ith

otherratgroups

(there

foremost

avers

ive

):me

dium

an

dhigh

concentrations

avers

ive

Thisstu

dy

did

notinvo

lve

euthanasia

(continue

d)


12/25



Leacheta

l.

(2002a)

PF

CO2:argon/

25.5

%:5%

or

31.7

%:5%

or

59.2

%:5%

NA

NA

Shorterw

ithdrawa

lan

d

dwe

llingtimes

in

comparisonw

ith

argona

lone;me

dium

an

dhigh

concentrations

mostavers

ive

NS

Thisstu

dy

did

notinvo

lve

euthanasia

PF

CO2:argon/

19.1

%:7%

or

32.1

%:7%

or

54.2

%:7%

NA

NA

Sameasa

bove

NS

Thisstu

dy

did

notinvo

lve

euthanasia

Sm

ith&

Harrap

(1997)

PF

CO2

alone/75%

26.4

311.

4

Imme

diatecirc

lingand

urination;gasp

ingo

r

laboure

dbreathing;

Rap

idk

inpu

lserate

an

dbloo

dpressure

followe

dbyslow

drop

inbloo

d

pressure

V

isua

lsignso

f

deathprece

de

vascu

larco

llapse

by1m

in,

there-

foremustma

in-

taingas

flow

for

atleast1m

in

aftersignso

f

death

Thistableinc

ludestheresultsofstudiesofratsonly,

forg

roupsthatwereexposedtoCO2.

Therefore,

ifastudycomparedCO2

withotheragents,informationofotheragentswasnotincludedinordertokeepthe

tableasshortaspossible.

Timetounconsciousnesswas

determinedconservativelyastheonsetof

posturalcollapseandlossofmuscletone(Woodburyet

al.

1958,

Coenenet

al.

1995);thepresentauthors

calculatedth

esetimeswhentheywerenotspecifieddirectlybytheauthorsoftheoriginalpapers.

NS

notspecified

Table2(continued)

Re

ference

Pre

fille

d(PF)

orgra

dua

l

induct

ion

(GI)

Gasm

ixtu

re/fina

l

concentra

tion

Est

imated

timeto

unconsc

ious-

ness

(s)

Estim

ated

time

to

deat

h(s)

Be

hav

iour

be

fore

unconsc

iousness

Physio

logy/neuro

log

ica

l

measures

Comments


13/25

and head turned up and backward) in someconditions and those in the 100% CO2group showed most evidence of thisbehaviour. These results support the notionthat higher concentrations of CO2 lead toincreased adverse reactions; this is incontrast to the conclusion by Dannemanet al. (1997) that lower concentrations ofCO2 lead to increased incidence of adversereactions.

Leach et al. (2002a,b) studied aversion tovarious gaseous euthanasia agents, includingCO2, in rats and mice at low, medium andhigh concentrations. Measures of avoidanceof, or preference for, different environmentsare well-established tools in the study of

farm animal welfare. The animals choice ofenvironment is considered to reflect itspriorities to access or avoid specific stimuli.In one study (Leach et al. 2002a), the agentswere CO2 (25%, 35%, 50%) and argon (93%,95%, 98%), as well as two combinations ofCO2 and argon. A second study (Leach et al.2002b) compared halothane, isoflurane,enflurane and CO2 at low, medium andhigh concentrations. Aversion measuresincluded time taken to initially withdraw

from the test chamber, time spent in testchamber and frequency of entries and exitsto and from the test chamber. Additionalbehavioural measurements indicative ofaversion included washing, rearing, sniffingand excreting.

Overall, CO2, alone or in combinationwith argon, caused the highest degree ofaversion in both rats and mice in bothstudies, with higher concentrations leadingto significantly shorter withdrawal times(i.e. 0.6 s at highest concentration). Inregards to the additional behaviouralmeasurements examined, increase infrequency of washing was the only cleartrend, possibly the result of nasal irritation.The mice and rats differed in regards toresponse threshold, with mice having alower threshold to CO2. Leach et al.concluded that CO2, either alone or incombination with argon, cannot be used

humanely at any concentration and istherefore unacceptable as a euthanasia agentfor laboratory rodents, particularly whenmore humane methods exist.

The findings of the studies discussedabove demonstrate that CO2 causes adverseeffects that lead to stress, and perhapsdistress, at various concentrations rangingfrom 25% to 100%.

A number of studies have examined thehistological effects of CO2 and some resultsraise further concerns regarding pain and/ordistress. Some physiological evidence of theeffects of CO2 suggests that increasingphysical trauma occurs when the extent ofexposure to the inhalant gas increases.Terminal CO2 inhalation has been shown tolead to alveolar extravasation (bleeding intothe tissue) and pulmonary oedema(Danneman et al. 1997), haemorrhaging,

lung oedema and emphysema (Iwarsson &Rehbinder 1993), and degeneration ofmyocardial tissue and other organs(Iwarsson & Rehbinder 1993). Britt (1986)reported increased pulmonary haemor-rhaging in higher concentrations of CO2.However Danneman et al. (1997)found that the severity of pulmonaryoedema and haemorrhage were inverselycorrelated with the concentration ofinhaled CO2; this may be the result of

increased time of exposure. Fawell et al.(1972) found oedema of perivascularconnective tissue in the lungs of all ratssubjected to CO2 euthanasia (concentrationand flow rate were not indicated). Anincreased incidence of extravasation wasconsidered to be related to the trauma ofasphyxiation.

Only one recent study (Ambrose et al.2000) has assessed the extent of oedema andalveolar consolidation (inflammatoryinduration of lung tissue) from samplestaken immediately at the point ofunconsciousness instead of following death.This study compared the use of 30% CO2with 30% CO2 plus 20% O2 in two strains ofmice and it demonstrated greater alveolarconsolidation in the CO2 O2 condition.This consolidation may have occurred justprior to loss of consciousness and led theauthors to propose that the animal may not

only demonstrate hyperventilation withexposure to CO2O2, but may alsoexperience a state similar to drowning whilethe animal is still conscious (Ambrose et al.




14/25

2000). Danneman et al. (1997) also noteddamage to the lungs and/or respiratory tractvia observations of serious serosanguinousnasal discharge during exposure to 50% CO2,but it is unclear whether this occurred whenanimals were still conscious. This evidencedoes raise the question of whether, withoutvisible behavioural effects, the animal mayexperience a highly distressing state whilestill conscious, as oedema and haemor-rhaging develop.

Prefilled (PF) chamber versus gradualinduction (GI)/times to anaesthesia anddeath

The manner by which animals are exposedto CO2, PF chamber versus GI, has beenexamined fairly extensively. The results ofthese studies do not clearly indicate whichmethod is preferred in regards to animalwelfare, but these studies do demonstratethat both methods raise concerns in regardsto animal welfare. Smith and Harrap (1997)compared GI (from 0% to 80% CO2: 3% O2)versus PF (75% CO2: 3% O2) methods inrats. PF caused an immediate fall in blood

pressure while GI caused increased bloodpressure for the first 4 min and then a rapiddecline. Additionally, time to unconscious-ness was 30 s in PF subjects and 99 s in GIsubjects; time to death was 5.4 min in PFsubjects and 9 min in GI subjects. Althoughhead-raising, urination, defaecation andgasping or laboured breathing were reported(behaviours were predetermined by theauthors to be indicators of stress, pain,distress, anxiety and fear), the authorsreported that CO2 caused few overt signs ofdistress in either group.

Hewett et al. (1993) examined thedifferences between PF and GI methods inrats at a final concentration of 100%.Measurements included blood gas quanti-tation, and times to ataxia, immobility, lossof righting ability, loss of pedal reflex andanaesthesia (two consecutive negative pedalreflexes). The shortest amount of time to

loss of pedal reflex was 28 s, which occurredin the subjects exposed to the PF chamber.Loss of pedal reflex in the GI study occurredat 140 s, or six times longer than the PF

method. Overall, the authors reported thatthe animals responses were not indicative ofdistress by either method. However, thestudy did not clearly indicate that any speci-fic behaviours were looked for and thereforeit is impossible to determine whether theanimals were observed for anything otherthan ataxia and loss of pedal reflex.

Britt (1986) presented information on thehumaneness of CO2 use based on a compari-son of PF versus GI methods using rats andmice. Evaluations were based on types ofbehaviour, time spent in each activity,number of times activity was changed perunit of time, changes in breathing patternswith time and position in cabinet at loss of

consciousness. Britt noted that time tocollapse was shorter with PF, but PF causedmore signs of distress. Abnormal behaviours(referred to by the author as signs of distress)included: shaking (frequent), moving inreverse, tail thrashing (uncommon) andincrease in frequency of urination anddefaecation. Urination and defaecation mayresult from the physiological effects of CO2on the autonomic nervous system andconclusions regarding their use as indicators

of stress should be cautious; however, theydid arise in association with increasedmicturation, which may indicate distur-bance or distress (Britt 1986). Behaviouralresponses varied between species andindividuals. The author concluded thatneither method is stress-free; however, hefavoured GI.

Again, there is evidence from these studiesthat CO2 is associated with pain anddistress, regardless of induction method.Overall, two studies concluded that neithermethod is problematic in regards to pain ordistress (Hewett et al. 1993, Smith & Harrap1997), while one study (Britt 1986) indicatedthat prefilling leads to increased adversereactions. One could conclude, therefore,that GI is preferred due to absence of adversereactions, despite the fact that it may takelonger to euthanize the animals.

Presence of supplemental oxygen

The use of oxygen supplementation in orderto minimize pain and distress during CO2




15/25

use has also been a point of debate. Existingstudies, again, provide conflicting results,depending on the effects examined.

Iwarsson and Rehbinder (1993) examinedCO2:O2 mixtures for euthanasia of rats, miceand guineapigs. Animals were exposed toCO2 from a gas cylinder at either 100% CO2or 80% CO2: 20% O2 via GI. The presence ofoxygen doubled the amount of time tounconsciousness in rats and mice but,according to the authors, appeared todecrease distress. Stress response wasdetermined by examining variousbehaviours, such as breathing, urination anddefaecation. Postmortem examinations oflung tissue revealed that both methods

generated adverse physiological reactions. Itis not clear, however, if these changes inlung tissue occurred before or afterunconsciousness.

As previously mentioned, Coenen et al.(1995) reported that there were four phases inthe response to CO2 inhalation in rats. Thephase of continuous abnormal activity(excitation and agitation) was completelyabsent in the presence of oxygen. While100% CO2 produced the shortest time to

death, the authors concluded that thepresence of added oxygen produced a longertime to death but with a reduction of adverseeffects. Therefore, adding oxygen wasconsidered to be the preferred method.However, the difference in concentrations ofCO2, and not the use of supplementaloxygen, could be the cause of these results.

The results from Ambrose et al. (2000),however, appear to conflict with those ofCoenen et al. (1995). Ambrose et al.compared 30% chamber volume per minuteof CO2 with and without supplementaloxygen. There were no behaviouraldifferences observed between the twoconditions and presence of oxygen was foundto increase alveolar consolidation. Theauthors concluded that [a]s haemorrhage islikely to be stressful to the mice by inducinga feeling of drowning, any alveolarconsolidation above basal level indicates

potential poor welfare and that addition of20% chamber volume per minute of O2 isnot recommended as a refinement to CO2euthanasia. The reason for the conflict with

Coenen et al. (1995) results is unclear, butneither author described the behaviours thatwere assessed in order to determineaversion; differences in behavioural datacollection and analysis could explaindifferences between results of this as well asother studies.

Time to and measurements ofunconsciousness

It is important to determine the point atwhich unconsciousness occurs in order toclearly identify whether adverse effects occurprior to or following unconsciousness.Experienced veterinary clinicians have

indicated that anaesthesia sets in so quicklywhen CO2 is used that there is not sufficienttime for the animals to experience significantpain and distress. However, a review of theliterature finds that times to unconsciousnessvary greatly when using recommendedmethods. As discussed, many authors usedifferent terminology, such as: anaesthesia,collapse, recumbency, unconsciousness andimmobility. The following are observations ofrats exposed to the AVMA-recommended

condition of prefilling the chamber with 70%CO2: Danneman et al. (1997) recorded thatanaesthesia (onset of slow, shallow breathingand loss of response to toe pinch withhaemostats) occurred at an average of4.01min; Mischler et al. (1994) recordedanaesthesia (appeared comatose, exhibitedmuscle flaccidity and were unresponsive todeep pressure applied to the tail and/orhindpaw) within a maximum of 10 s; andSmith and Harrap (1997) reported loss ofrecumbency at an average of 26.6s. Whatcould account for this variability? Somepossibilities are variations in gasconcentrations, sample sizes, methods andequipment used, condition of the animals, andhow unconsciousness was actually measured.

Although many papers have described theonset of ataxia as symptomatic of the onsetof the effects of CO2 (Hewett et al. 1993,Coenen et al. 1995, Danneman et al. 1997),

it is not clear whether ataxia itself isdistressing. It is possible that the inability tobehave adaptively or complete purposefullocomotion might be unpleasant. Further




16/25

research is needed to assess whether anataxic state is aversive.

Overall, as the inhaled CO2 concentrationincreases, both time to unconsciousness(Britt 1986, Danneman et al. 1997, Ambroseet al. 2000, Leach et al. 2002a,b) and time todeath (Britt 1986, Ambrose et al. 2000)decrease, while the addition of oxygen to theinhalant gases increases time to death(Coenen et al. 1995). Additional details areprovided in Figure 1 and Tables 1 and 2. Itwould appear that if time to unconscious-ness were the only criteria influencingdistress, higher concentrations of CO2would be beneficial. However, before sucha decision is made, it is important to deter-

mine what the intensity of such distressmight be, while also taking into consider-ation reaction of humans to high concen-trations of CO2.

Summary

The literature demonstrates that CO2 ispainful and/or distressful in humans atconcentrations ranging from 7% to 100%,and that there is sufficient evidence that

CO2 likely causes pain and distress inanimals. Table 3 summarizes the incidence

of adverse reactions in the literaturereviewed here. Adverse reactions toCO2 in species other than those discussedin this review, including cats (Simonsenet al. 1981), dogs (Eisele et al. 1967) andguineapigs (Iwarsson & Rehbinder 1993),

can be found in the scientific literature. Insummary, the human and animal literature



0

20

40

60

80

100

120

140

160

0 20 40 60 80 100 120

Concentration (%)

Timetounconsciousness(seconds)

Gradual induction Prefilled

Figure 1 Effects of concentration on times tounconsciousness: gradual induction and prefilledchamber

Table 3 Comparison of publications reviewed for this paper: those that reported adverse effects and thosethat reported no adverse effects in response to carbon dioxide

Author (year) No adverse effects reported Adverse effects reported

Ambrose et al. (2000) OBarbaccia et al. (1996) OBlackshaw et al. (1988) O

Britt (1986) OCoenen et al. (1995) ODanneman et al. (1997) OFawell et al. (1972) OHackbarth et al. (2000) OHewett et al. (1993) OHornett & Haynes (1984) OHoenderken (1983) OIwarsson & Rehbinder (1993) OJongman et al. (2000) OLeach et al. (2002a) OLeach et al. (2002b) ORaj & Gregory (1994) ORaj & Gregory (1995) ORaj & Gregory (1996) ORobb et al. (2000) OSmith & Harrap (1997) OThurauf et al. (1991) O


17/25

signals that the use of CO2 may beproblematic.

Rodent studies that compare the use

of CO2 alone with other euthanasiamethods

A fair amount of the published data onCO2 stems from studies that compared CO2with other euthanasia methods. CO2concentration, induction method andpresence of oxygen remain potentiallyconfounding factors. The studies discussedhere conclude that CO2 is a preferableeuthanasia agent for rodents in comparison

to ether, chloroform, nitrogen, and that theuse of anaesthesia with pentobarbital orsedation prior to induction with CO2 has noparticular benefits.

Blackshaw et al. (1988) compared CO2,ether and chloroform in rats, mice andchickens. Dry ice was used to generate97% CO2 in a prefilled chamber. Time tocollapse and death were shortest withCO2 for all three species in comparisonwith ether and chloroform. Adverse

behavioural signs to CO2 were not observed,except for wing flapping in chickens.Arguing that the shortest time to death ispreferable when choosing a euthanasia agent(while also taking behaviour prior to uncon-sciousness into consideration), the authorsconcluded that CO2 was the preferred agentfor rats and mice, but not chickens. Onepotential problem with this study is that thebehaviours observed focused on the animalsmovement and did not include otherbehaviours that may be indicative of stress(e.g. urination and defaecation) that wereobserved in other CO2 studies.

Hornett and Haynes (1984) compared CO2and nitrogen in rats and concluded that CO2was the preferred euthanasia method. CO2from a gas cylinder was gradually inducedand behaviours measured included perceivedsigns of fear (behaviours were not specified)and oxygen deprivation, as well as laboured

breathing, loss of co-ordination and balance,collapse (correlated with unconsciousness)and respiratory arrest. Overall, the rats hadan extreme response to nitrogen including

panic, attempts to escape and convulsions,while CO2 caused increased nasal move-ment and a stage of laboured breathing. Theauthors concluded that CO2 at 19.5%chamber volume per minute achieved aquiet delivery into unconsciousness.

Finally, Hackbarth et al. (2000) examinedwhether sedation (acepromazine in choppedmeat) or anaesthesia (injection of sodiumpentobarbitone) prior to the use of CO2would decrease the stress-induced effects ofCO2 euthanasia. Rats were subjected to oneof four conditions followed by CO2: meatwith acepromazine, meat without acepro-mazine, injection with pentobarbitone orinjection with saline. Behavioural obser-

vations were recorded and adrenocortico-trophic hormone (ACTH), glucose andcorticosterone measurements were taken atvarious times after CO2 induction began. Nogroup exhibited behavioural signs of pain ordistress. Glucose and corticosterone levelsdid not differ between the groups. ACTHwas higher in the subjects given injections incomparison with those given oral treatments(assumed to be a result of handling stress).Given the relative absence of effects of

sedation and anaesthesia, the authorsconcluded that yeuthanasia with CO2[without prior sedation or anesthetization] isin concordance with animal welfare as it israpid and does not cause distress in theanimal and therefore can be recommendedas humane .

These studies do not show evidence ofdistress as some of the previously discussedstudies have. Again, this may be a result ofthe behaviours examined and differencesbetween studies in interpreting behaviourand other measures. Differences in behav-iour are important to recognize because eachindividual, as well as each species, has aparticular coping strategy and pain threshold(Leach et al. 2002b).

Literature on stunning for slaughter

The use of CO2 for stunning prior to

slaughter has been examined in a number ofstudies with different species. Overall, manystudies found that CO2 is aversive, asdemonstrated by avoidance by the animals.




18/25

Raj and Gregory (1995, 1996) compared theuse of argon (90%) and CO2 (30% or 90%) forstunning in pigs and found high concentrat-ions of CO2 to be highly aversive. Aversionwas assessed from the pigs reluctance toenter or remain in the gaseous atmospheresfor a reward of apples. When 90% CO2 wasin the box, the pigs immediately withdrewtheir heads, repeatedly attempted to feed,but withdrew their heads when they beganto hyperventilate. The following day, threeout of six pigs hesitated to enter the box. Itwas concluded that high concentrations(90%) of CO2 were aversive to the majorityof pigs (88%) and that they would notattempt to get the apples, even after 24 h of

fasting. By contrast, pigs did not demonstrateaversion to the presence of argon and themajority did not demonstrate aversion to30% CO2 in air (Raj & Gregory 1995). Sincethe design of the study allowed for escapeand did not force the animals to be exposedto the gas, distress was not examined.However, the study does demonstrate thatwhen given a choice, pigs will avoid highconcentrations of CO2. Comparable resultswere found in a similar experiment on

turkeys (Raj & Gregory 1994).A second study (Raj & Gregory 1996)

examined the severity of respiratory distresswhen pigs were placed in a well of variousconcentrations of argon with or withoutoxygen and/or CO2 in air. Respiratorydistress was scored and it was found thatargon with 2% oxygen induced minimalrespiratory distress, while the combinationof 30% CO2 and argon with residual oxygeninduced moderate distress, and allconcentrations of CO2 in air induced severerespiratory distress in the pigs (Raj &Gregory 1996).

Hoenderken (1983) compared CO2 andelectrical stunning methods in pigs and alsofound CO2 to be aversive. Behaviours(specific behaviours were not indicated) andEEG were recorded while CO2 was used;specific concentrations and inductionmethods were not clearly indicated in the

publication. It was found that pigs showsigns of excitation after 12 s of exposure toCO2 and this excitation lasts 26 s on average(isoelectric EEG was recorded at 56 s). The

author concluded that there is a long periodof excitement during CO2 exposure.Jongman et al. (2000), however, found shockwith an electric prod to be more aversivethan 90% CO2 based on the pigs reluctanceand time to enter the treatment areafollowing previous exposure to the stimuli.Behaviour other than time to enter treat-ment area was not measured in this study.

Finally, four slaughter methods(exsanguination without stunning, CO2stunning followed by gill cutting, percussivestunning and spiking) were compared insalmon; behaviour and visual evokedresponse to a flash of light (lack ofresponse good indicator of brain failure)

were examined (Robb et al. 2000). CO2caused the fish to shake their heads and tailsvigorously for 2 min, which resumed againafter gill cutting. Loss of visual evokedresponses also took much longer withexsanguination and CO2. Overall, theauthors recommended the use of stunning orspiking over exsanguination or CO2.

These studies, overall, demonstrate thatCO2 stunning for slaughter causes adverseeffects and is avoided when the animals are

provided with the opportunity to do so. Insome instances, as described by Jongmanet al. (2000), the animals will choose CO2over other methods; however, this does notindicate that CO2 is the preferred of allavailable methods.

Pain and distress: considerations

Pain is considered to be the experience of anunpleasant sensory and emotional state thatarises from central perception of nociceptionmechanisms in response to actual or poten-tial tissue damage (International Associationfor the Study of Pain 1979). Distress isprimarily considered to encompass bothphysiological and psychological states, andmay be observed through physiological andbehavioural responses which indicate thatthe animal is experiencing stimuli asnoxious and from which the animal is

motivated to avoid or withdraw (Dawkins1990, Rolls 1999).

There are a range of measures ofphysiological function and behaviour that




19/25

may be used to determine whether ananimal is in a state of distress. These includeheart rate, blood pressure, circulating freecortisol, measures of behavioural aversion oravoidance and behavioural correlates of painor distress (Appleby & Hughes 1997, Masonet al. 2001). A range of authors havemeasured a variety of behavioural andphysiological responses as indicators of painand distress in studies of CO2, such asbehavioural signs of asphyxia (Coenen et al.1995), hyperventilation (e.g. Britt 1986,Kohler et al. 1999), escape behaviours (Britt1986, Leach et al. 2002a,b), tail thrashing(Britt 1986), disturbance behaviours such aswashing (Britt 1986) and physiological signs

of nociception (Anton et al. 1992). Theinterpretation of such behavioural measurescan be difficult if the association betweenthe behaviour and an underlying state hasnot been validated, or where behaviours arepoorly described, or where quantitativemeasures are not used and the reliability ofobservational recordings confirmed (seeMartin and Bateson (1993) for a discussion).Further complications arise as some authorsmay not have been consistent in measuring

particular behaviours, or may have drawndifferent conclusions regarding similarbehavioural phenomena (e.g. Britt 1986,Blackshaw et al. 1988, Hewett et al. 1993,Danneman et al. 1997).

As seen in Tables 1 and 2, the currentevidence regarding CO2 inhalation and painand distress is based on studies using adiverse range of exposure methods, gasconcentrations and mixtures, as well asspecies and genetic strains. The types ofvariables measured are inconsistent, andoften physiological and behavioural datahave not been collected from the sameexperiment. For behavioural data to bemeaningful, rigorous ethological techniquesare required, including organized samplingmethods, quantifiable measurement andgood quality objective description, allowingfor future interpretation and replication(Martin & Bateson 1993). Behaviours

interpreted as indicating underlyingphysiological changes should be validatedwhere possible. The increasing sophisti-cation of techniques from animal welfare

science, such as measures of aversion, arevaluable tools for understanding theanimals priorities and choices in light ofexposure to environmental stimuli. The useof control animals (where CO2 is notpresent) would also be useful to indicate theextent to which the behaviours andphysiological responses observed in previousstudies arise from the novelty, unfamiliarityor mixing with unfamiliar conspecifics.Despite some of the problems with the datain the above studies, it can be concluded thatthe information available is more thansufficient to raise concerns about the painand distress associated with CO2 euthanasia.

Existing guidelines for euthanasia

Various organizations and agencies such asthe AVMA (2001), CCAC (1993), theEuropean Commission (1996, 1997), theAustralia and New Zealand Council for theCare of Animals in Research and Teaching(ANZCCART 2001), the UK Home Office(1997b) and the Universities Federation forAnimal Welfare (UFAW 1986) provideguidelines on euthanasia. These guidelines

provide conflicting recommendations inregards to CO2 euthanasia, specifically theoptimal concentrations, rates of inductionand sources of CO2.

In July 2002, the Office of LaboratoryAnimal Welfare (OLAW) of the US NationalInstitutes of Health published a guidancenotice that high concentrations of CO2 maybe distressful to some species and thereforeprefilling the chamber is only recommendedwhen such use has not been shown to cause

distress (OLAW 2002). This recommend-ation conflicts with the AVMA guidelinesthat suggest using a prefilled chamber,illustrating the continuing confusionregarding CO2 use.

Discussion of CO2 use

The evidence that CO2 causes pain anddistress in both humans and animals can be

summarized as follows.

(1) CO2 has been used in animals as apainful stimulus (Thurauf et al. 1991),




20/25

at a concentration that is typically usedfor euthanasia, and in studies to elicitan acute stress response (Barbacciaet al. 1996); this supports the claimthat CO2 causes pain and distress inrodents.

(2) Although high concentrations of CO2have been found to cause more pain anddistress in comparison with low con-centrations, pain and distress have alsobeen reported at low concentrations,such as 20% CO2 in pigs (Raj & Gregory1996) and 25% in rats and mice (Leachet al. 2002a,b). Reported findings thatthere is increased noxiousness withincreased concentrations of CO2 are of

concern because some literature andmost available euthanasia guidelinesrecommend the use of higher concen-trations of CO2 due to its shorter timeto recumbency, anaesthesia andeuthanasia.

(3) According to published findings andanecdotal reports, GI and prefilled eu-thanasia chambers both lead to con-cerns about pain and distress.

(4) The histological effects of CO2 on lungs,

heart and other organs, including oede-ma and haemorrhage, raise concernsabout animal distress.

Policies of the OECD and individualcountries indicate that if a procedure causespain and distress in humans, it must beconsidered to cause pain and distress inanimals (US Public Health Service Policyalso indicates in the absence of evidence tothe contrary). The evidence that CO2 elicitsa painful response in humans (Dripps &Comroe 1947, McArdle 1959, Anton et al.1992, Hummel et al. 1994, Danneman et al.1997) is very strong. Therefore, these policiesrequire that CO2 use should be consideredpainful and distressful in animals.

Alternatives to use of CO2

Discontinuing use of CO2 alone as aeuthanasia agent is unlikely to happen soonbecause it is so widely used, but there aresome reasonable alternatives.

Pre-anaesthetic followed by CO2

The use of a pre-anaesthetic to induceunconsciousness prior to induction of CO2for euthanasia should be considered.Halothane was found to be the least aversive

by Leach et al. (2002b) when distress associ-ated with induction of halothane, isoflurane,enflurane and CO2 were compared.Halothane, isoflurane and enflurane werefound to be more aversive than air, but weresignificantly less aversive than CO2.

Use of pre-anaesthetic agents with CO2should be acceptable for the euthanasia oflarge numbers of animals not involved inresearch protocols or for the euthanasia ofsmall numbers of animals where the effect of

the anaesthetic gas is not a problem. Wherethe research protocol calls for tissuesunaffected by anaesthetics, decapitation maybe the best method for small numbers of ratsor mice. However, this method is stillquestioned because of the debate over themeaning of the EEG trace followingdecapitation and also because of the highchance for error due to inadequate operatortraining or guillotine maintenance.

Inhalational gas agents are particularlyadvantageous because they require minimalhandling of the animals and larger numbersof animals can be euthanized simult-aneously. As Hackbarth et al. (2000) report,an injectable pre-anaesthetic can cause ahandling-induced stress response.

One concern about the use of volatileanaesthetics is exposure of personnel to thegas. However, there are portable scavengingunits that would allow for safe use of both a

pre-anaesthetic and CO2. The EuropeanCommission recommendations for eutha-nasia indicate that halothane, enflurane andisoflurane are all acceptable agents withappropriate gas scavenging apparatus.

Argon

Argon, an odourless, inert gas that is non-flammable and non-explosive (AVMA 2001),is another option for euthanasia that might

be an improvement over the use of CO2.Leach et al. (2002b) found argon to be muchless aversive to rats and mice in comparisonto CO2, although all examined concen-




21/25

trations of argon (ranging from 25% to 53%)caused a degree of aversion in comparison toair. Raj and Gregory (1995) found no aversionto argon (90%) in pigs. The AVMA (2001)considers argon to be only conditionallyacceptable, but this recommendation did notcite the scientific literature. The existingliterature, however, does generallydemonstrate that argon is preferable to CO2in regards to animal welfare. The FarmAnimal Welfare Council (2003), for example,indicates that argon is less aversive in pigs incomparison to CO2. The council alsoexpresses support for additional research inorder to determine how to best use argon foreuthanasia purposes.

Lawson et al. (2003) compared the use ofCO2, nitrogen and argon for euthanasia andfound that CO2 induced apparentunconsciousness within 40 s and deathwithin approximately 3 min, and argon hadsimilar effects (unconsciousness at 55 s anddeath at 4 min). However, CO2 caused arapid and significant increase in bloodpressure (measured by radio-telemetry),while argon and nitrogen caused musclerigidity and spasms. In this case, the authors

concluded that CO2 is preferable foreuthanasia due to the rapidity of effects andlack of muscle rigidity and spasms. Finally,the authors did not rule out that CO2 couldhave caused a stress effect.

Overall, argon is as easy to use as CO2, isnon-flammable, and sinks to the bottom ofthe chamber since it is heavier than air,therefore reducing danger to thoseconducting the procedure. One drawback,however, is that argon is more expensivethan CO2; however, utilizing a method thatis less painful and distressful to the animalsshould take priority over cost. Moreover, thecosts of both argon and CO2 are trivialcompared with the total financialinvestment in an animal when all real costsare taken into account (e.g. technician andinvestigator time, costs of the animal andthe animals care, etc.).

Decapitation

Decapitation is a possible alternative to theuse of CO2 for the euthanasia of (relatively)

small numbers of animals, but this remainssomewhat controversial. Mikeska andKlemm (1975) found that decapitationcauses an EEG trace for an average durationof 13.6 s; this study led to much of thecontroversy surrounding decapitation andprevious recommendations by the AVMA(1986, 1993) that decapitation should not bedone without justification.

Various authors have since challengedMikeska and Klemms (1975) findings. Forexample, Allred and Berntson (1986)presented a counter-argument, citingreferences that demonstrated an activatedEEG pattern after hypoxia or damage to thelower brain stem as evidence that the

Mikeska and Klemm data did not necessarilyprovide evidence of consciousness anddistress. Vanderwolf et al. (1988) conductedstudies that examined the source of lowvoltage fast activity (LVFA) and found thatsuch EEG traces are not associated withnoxious stimuli or consciousness.Additionally, Derr (1991) reported that itwould take 2.7 s for a decapitated rats headto become anoxic (indicating a presumedloss of consciousness).

Finally, Holson (1992) found that, by 1986,all eight papers reporting on the EEG ofdecapitated rodent heads agree thatdecapitation triggers an immediate, slowdirect current EEG trace of 24 s durationfollowed by an LVFA trace that is usuallygone in 1013 s, a pattern that is notassociated with consciousness.

One reason that decapitation is sometimeschosen over other methods is that someagents, such as anaesthetics, can causechanges in tissue and alter brain metabolismparameters (see, e.g. Miller et al. 1988) andthe aim of the study may be to collectundisturbed tissue. In this case, however, itis important to consider that the stress ofhandling during the decapitation proceduremay also modulate metabolite levels (e.g.Faupel et al. 1972). There may, of course, beoccasions where the stress of handling isunlikely to change the parameters a scientist

wishes to measure, but these should bereadily justifiable.

In sum, decapitation produces an EEGtrace that can last for 1013 s, but the weight




22/25

of the evidence indicates that consciousnessfollowing decapitation is unlikely to persistfor more than 36 s. It seems clear thatdecapitation produces a much quicker loss ofconsciousness than the recommended CO2protocols. Therefore, from an animal welfareperspective, decapitation properly performedmay be preferable to CO2. However,decapitation is not an easy procedure andtechnicians who use a guillotine must beproperly trained and the equipment must beproperly maintained.

Recommendations and finaldiscussion

While CO2 has long been used as a methodof euthanasia, questions have arisen thatthis practice may not fulfill the criteria foran easy and gentle death, as required by bothethical concerns and policy. The primaryconcerns are that exposure to CO2 gas maybe painful; CO2 may cause onset of asphyxiawhile the animal is still conscious;physiological effects of CO2 on nasalmucosae and on the autonomic nervoussystem may be distressing; the cognitive/

perceptual/behavioural effects of CO2, suchas ataxia, may be disturbing to the animal;and the process of CO2 inhalation may behighly aversive. Some or all of these resultsmay cause pain and distress.

Information regarding concerns with theuse of CO2 for euthanasia, described in thisreview, has led organizations such as TheHumane Society of the United States torecommend changes, such as the use of aninhalant pre-anaesthetic in conjunction withCO2 for euthanasia of a large number ofsurplus rodents or a few to moderate numberof rodents being used for research.Decapitation is strongly discouraged whenlarge numbers of animals are beingeuthanized because of the chances foroperator error. For euthanasia of relativelyfew rodents being used for research that doesnot permit contamination of tissues, TheHumane Society of the United States

recommends decapitation by well-trainedpersonnel if there is scientific justification,and when CO2 in conjunction with a pre-anaesthetic has well-founded scientific

concerns. The use of CO2 alone is stronglydiscouraged in all instances.

Personnel play a big role in the issuesdiscussed in this paper. Proper training ofpersonnel in all techniques is essential forimproved animal welfare. Centralizedfacilities in which the veterinary care staffperforms such techniques are recommended;it is easier to ensure adequate training incentralized facilities and the expertise istypically higher among those who areperforming euthanasia.

We recognize that many clinicalveterinarians and laboratory animaltechnicians appear to be comfortable withCO2 alone as a euthanasia agent. However,

we would argue that the conflicting data inthe literature and the arguments aroundprefilled chambers versus GI, the presence orabsence of oxygen, the lack of agreement ontime to unconsciousness and the fact thatCO2 produces significant physiologicalchanges and behavioural indicators of painand distress, all demand a carefulreassessment of the use of the gas by itself asa euthanasia agent. Careful studies ofdistress and objective assessments of such

distress using behavioural and physiologicalmeasures are clearly possible, as evidencedby the progress made in assessing farmanimal welfare (see, for example, Appleby &Hughes 1997).

The limitations in the currently availabledata outlined above (such as small samplesizes, examination of only certain measureswhile excluding others, and so on) suggestthe need for a structured, multi-factorialstudy utilizing a range of dependentphysiological and behavioural measures,evaluating a representative range of modes ofCO2 and CO2gas mixture exposures. Thistype of study would reveal the effects of CO2exposure while controlling for methodo-logical and procedural differences in practicebetween institutions. Most importantly,with effective coordination, this type ofstudy could be completed without incurringfurther potential animal distress, by

gathering data from existing practices, withwelfare measures taken as add-on toprocedures that would be performed onexisting animals. This type of applied animal




23/25

welfare research has already led tosignificant increases in the understanding ofanimal distress within the farm animalarena (e.g. King 2003).

Acknowledgements We would like to thank

Victoria Hampshire and our colleagues Cheryl Rossand Gina Alvino for their contribution to this review.We would also like to thank all of the laboratoryanimal personnel who shared their expertise on thisissue with us.

References

Allred JB, Berntson GG (1986) Is euthanasia of rats bydecapitation humane? Journal of Nutrition 116,185961

Alvaro RE, de Almeida V, Kwiatkowski K, Cates D,

Kryger M, Rigatto H (1993) A developmental studyof the doseresponse curve of the respiratorysensory reflex. American Review of RespiratoryDisease 148, 1013 17

Ambrose N, Wadham J, Morton D (2000) In: Refine-ment of Euthanasia. Progress in the Reduction,

Refinement and Replacement of Animal Experi-

mentation (Balls M, van Zeller AM, Halder ME,eds). Amsterdam: Elsevier

Anton F, Euchner I, Handwerker HO (1992) Psycho-physical examination of pain induced by definedCO2 pulses applied to the nasal mucosa. Pain 49,

5360ANZCCART (Australian and New Zealand Council

for the Care of Animals in Research and Teaching)(2001) In: Euthanasia of Animals Used for Scien-tific Purposes. 2nd edn. (JS Reilly, ed). Australia:Glen Osmond

Appleby MC, Hughes BO (1997) Animal Welfare.Wallingford: CAB International

AVMA (American Veterinary Medicine Association)(2001) 2000 Report of the AVMA panel on eu-thanasia. Journal of the American VeterinaryMedical Association 218, 66996

Barbaccia ML, Roscetti G, Trabucchi M, et al. (1996)Time-dependent changes in rat brain neuroactivesteroid concentrations and GABAA receptor func-tion after acute stress. Neuroendocrinology63,16672

Blackshaw JK, Fenwick DC, Beattie AW, Allan DJ(1988) The behaviour of chickens, mice and ratsduring euthanasia with chloroform, carbon dioxideand ether. Laboratory Animals 22, 6775

Britt DP (1986) The humaneness of carbon dioxide asan agent of euthanasia for laboratory rodents. In:Euthanasia of Unwanted, Injured or Diseased

Animals for Educational or Scientific Purposes.

London: Universities Federation for AnimalWelfare

Canadian Council on Animal Care (1993) Guide tothe Care and Use of Experimental Animals. Vol 1.

2nd edn. (Olfert ED, Cross BM, McWilliam AA,eds). Ontario, Canada: CCAC

Close B, Banister K, Baumans V, et al. (1996)Recommendations for euthanasia of experimentalanimals: Part 1. DGXI of the European Commis-sion. Laboratory Animals 30, 293316

Close B, Banister K, Baumans V, et al. (1997)Recommendations for euthanasia of experimentalanimals: Part 2. DGXT of the European Commis-sion. Laboratory Animals 31, 132

Coates EL (2001) Olfactory CO2 chemoreceptors.Respiration Physiology129, 21929

Coenen AM, Drinkenburg WHIM, Hoenderken R, vanLuijtelaar ELJM (1995) Carbon dioxide euthanasiain rats: oxygen supplementation minimizes signsof agitation and asphyxia. Laboratory Animals 29,2628

Danneman PJ, Stein S, Walshaw SO (1997) Humaneand practical implications of using carbon dioxide

mixed with oxygen for anaesthesia or euthanasia ofrats. Laboratory Animal Science 47, 37685

Dawkins MS (1990) From an animals point of view:motivation, fitness and animal welfare. Beha-vioural and Brain Sciences 13, 19

Derr RF (1991) Pain perception in decapitated ratbrain. Life Sciences 49, 1399402

Dripps RD, Comroe JH (1947) Respiratory andcirculatory response of normal man to inhalationof 7.6 and 10.4 per cent CO2 with a comparison ofthe maximal ventilation produced by severe mus-cular exercise, inhalation of CO2 and maximal

voluntary hyperventilation. American Journal ofPhysiology149, 4351

Eisele JH, Eger EI, Muallem M (1967) Narcoticproperties of carbon dioxide in the dog. Anesthe-siology28, 85665

European Commission (1996) Working Party Report:Recommendations for euthanasia of experimentalanimals: Part 1. Laboratory Animals 30, 293316

European Commission (1997) Recommendations foreuthanasia of experimental animals: Part 2. DGXTof the European Commission. Laboratory Animals31, 132

Biotechnology Regulatory Atlas (homepage on theInternet) (cited 5 August 2004). Available from:http://plus.i-bio.gov.uk/ibioatlas/textc2_1.html

Farm Animal Welfare Council (2003) Report on theWelfare of Farmed Animals at Slaughter or Killing,

Part 1: Red Meat Animals (monograph on theinternet) London, England: Defra Publications(cited 2 September 2004). Available from: http://www.fawc.org.uk/reports/slaughter/report.pdf

Faupel RP, Seitz HJ, Tarnowski W, Thiemann V, WeissC (1972) The problem of tissue sampling fromexperimental animals with respect to freezingtechnique, anoxia, stress and narcosis. A new

method for sampling rat liver tissue and thephysiological values of glycolytic intermediatesand related compounds. Archives of Biochemistryand Biophysics 148, 50922




24/25

Fawell JK, Thomson C, Cooke L (1972) Respiratoryartefact produced by carbon dioxide and pentobar-bitone sodium euthanasia in rats. LaboratoryAnimals 6, 3216

Hackbarth H, Kuppers N, Bohnet W (2000) Euthanasiaof rats with carbon dioxide animal welfare

aspects. Laboratory Animals34

, 916Hewett TA, Kovacs MS, Artwohl JE, Bennett BT(1993) A comparison of euthanasia methods in rats,using carbon dioxide in prefilled and fixed flow ratefilled chambers. Laboratory Animal Science 43,57982

Hornett TD, Haynes AR (1984) Comparison of carbondioxide/air mixture and nitrogen/air mixture forthe euthanasia of rodents. Design of a systemfor inhalation euthanasia. Animal Technology35,939

Hoenderken R (1983) Electrical and carbon dioxidestunning of pigs for slaughter. In: Stunning Ani-

mals for Slaughter (Eikelenboom G, ed). Boston:Martinus Nijhoff Publishers, 5963

Holson RR (1992) Euthanasia by decapitation: evi-dence that this technique produces prompt, pain-less unconsciousness in laboratory rodents.Neurotoxicology and Teratology14, 2537

Hummel T, Gruber M, Pauli E, Kobal G (1994)Chemo-somatosensory event-related potentials inresponse to repetitive painful chemical stimulationof the nasal mucosa. Electroencephalography andClinical Neurophysiology92, 42632

International Association for the Study of Pain (1979)

Pain terms: a list with definitions and notes onusage. Pain 6, 249

Iwarsson K, Rehbinder C (1993) A study of differenteuthanasia techniques in guinea pigs, rats, andmice. Animal response and postmortem findings.Scandinavian Journal of Laboratory Animal

Science 20, 191205Jongman EC, Barnett JL, Hemsworth PH (2000) The

aversiveness of carbon dioxide stunning in pigs anda comparison of the CO2 stunner crate vs. the V-restrainer. Applied Animal Behaviour Science 67,6776

Kety SS, Schmidt CF (1947) The effects of alteredarterial tensions of carbon dioxide and oxygen oncerebral blood flow and cerebral oxygen consump-tion of normal young men. Journal of ClinicalInvestigation 27, 48492

King LA (2003) Behavioral evaluation of the psycho-logical welfare and environmental requirements ofagricultural research animals: theory, measure-ment, ethics and practical implications. ILARJournal 44, 21121

Kline BE, Peckham V, Heist HE (1963) Some aids inthe handling of large numbers of mice. LaboratoryAnimal Care 13, 8490

Kohler I, Meier R, Busato A, Neiger-Aeschbacher G(1999) Is carbon dioxide (CO2) a useful short actinganaesthetic for small laboratory animals? Labora-tory Animals 33, 15561

Lawson DM, Sharp JL, Azar TA (2003) A comparisonof carbon dioxide, argon and nitrogen for euthana-sia of mice. Contemporary Topics in LaboratoryAnimal Science 42, 82

Leach M, Bowell VA, Allan TF, Morton DB (2002a)Degrees of aversion shown by rats and mice to

different concentrations of inhalational anaes-thetics. Veterinary Record 150, 80815Leach M, Bowell VA, Allan TF, Morton DB (2002b)

Aversion to gaseous euthanasia agents in rats andmice. Journal of Comparative Medicine 52, 24957

Martoft L, Stdkilde-Jrgensen H, Forslid A, PedersenHD, Jrgensen PF (2003) CO2 induced acuterespiratory acidosis and brain tissue intracellularpH: a 31P NMR study in swine. LaboratoryAnimals 37, 2418

Mason G, Cooper J, Clarebrough C (2001) The welfareof fur-farmed mink. Nature 410, 356

Martin P, Bateson P (1993) Measuring Behaviour: An

Introductory Guide. 2nd edn. Cambridge: Cam-bridge University Press

McArdle L (1959) Electrocardiographic studies duringthe inhalation of 30 per cent carbon dioxide inman. British Journal of Anaethesiology31, 14251

Mikeska JA, Klemm WA (1975) EEG evaluation ofhumaneness of asphyxia and decapitation eutha-nasia of the laboratory rat. Laboratory AnimalScience 25, 1759

Miller LP, Mayer S, Braun LD, Geiger P, OldendorfWH (1988) The effect of pretreatment with pheno-barbital on the extent of [C] incorporation from [U-

C] glucose into various rat brain glycolytic inter-mediates: relevance to regulation at hexokinaseand phosphofructokinase. Neurochemical Re-search 13, 37782

Mischler SA, Alexander M, Battles AH, Raucci JA,Nalwalk JW, Hough LB (1994) Prolonged antinoci-ception following carbon dioxide anesthesia in thelaboratory rat. Brain Research 640, 3227

OECD (2000) Guidance Document on the Recogni-tion, Assessment, and Use of Clinical Signs as

Humane Endpoints for Experimental Animals

Used in Safety Evaluation. Paris, France: Organi-sation for Economic Co-operation and Develop-ment

Office of Laboratory Animal Welfare (2002) PHSPolicy on Humane Care and Use of Laboratory

Animals Clarification Regarding Use of Carbon

Dioxide for Euthanasia of Small Laboratory Ani-

mals (cited 19 March 2004). Available from: http://grants.nih.gov/grants/guide/notice-files/NOT-OD-02-062.html

Public Health Service (1993) PHS Policy on HumaneCare and Use of Laboratory Animals. Washington,DC: US Department of Health and Human Services

Raj ABM, Gregory NG (1994) An evaluation of

humane gas stunning methods for turkeys. Veter-inary Record 135, 2223

Raj ABM, Gregory NG (1995) Welfare implications ofthe gas stunning of pigs 1. Determination of



7/30/2019 Ca

Carbonic dioxide for eutanasia

Documents

Transcript of Carbonic dioxide for eutanasia