Available via license: CC BY-NC-ND 4.0
Content may be subject to copyright.
©2007 Poultry Science Association, Inc.
Identifying Process Variables for a Low
Atmospheric Pressure Stunning-Killing System
J. L. Purswell,
1
* J. P. Thaxton,† and S. L. Branton*
*USDA-ARS Poultry Research Unit, and †Poultry Science Department,
Mississippi State University, Mississippi State 39762
Primary Audience: Poultry Processors, Researchers
SUMMARY
Current systems for preslaughter gas stunning and killing of broilers use process gases such
as CO
2
,N
2
, Ar, or a mixture of these gases with air or O
2
. These systems, known as controlled-
atmosphere stunning-killing systems, work by displacing O
2
, ultimately to induce hypoxia in the
bird, leading to unconsciousness and death. In this study, mechanical removal of O
2
by rapidly
reducing air pressure was investigated as an alternative to controlled-atmosphere stunning-killing
systems. Low atmospheric pressure systems could offer advantages in worker safety and operational
gas cost because they operate solely with atmospheric air. This study comprised 2 experiments,
one to define the initial range of effective pressures, and the second to determine a recommended
process pressure. In experiment 1, 48 female broilers, aged 63 d, were subjected to 6 different
pressure treatments, ranging from 70.9 to 17.8 kPa. In experiment 2, 56 male broilers, aged 60 d,
were subjected to 7 different pressure treatments, ranging from 35.3 to 17.8 kPa. Birds were
individually placed in an airtight vessel and exposed to a pressure treatment for 2 min after the
final pressure was attained. Results from experiment 1 showed that the effective range of pressure
was between 29.5 and 17.8 kPa, with only 25% of the birds exposed to 29.5 kPa surviving and
none of the birds exposed to 17.8 kPa surviving. Experiment 2 used a finer resolution of pressure
increments, and the estimated pressure level lethal for 99.99% of the birds was determined to be
19.4 kPa.
Key words: broiler, gas stunning, slaughter
2007 J. Appl. Poult. Res. 16:509–513
doi:10.3382/japr.2007-00026
DESCRIPTION OF PROBLEM
Previous research focusing on the develop-
ment and optimization of gas stunning-killing
systems for broilers has used process gases such
as CO
2
,N
2
, Ar, or a mixture of these gases with
air or O
2
to incapacitate hens and broilers prior to
shackling and exsanguination [1, 2, 3]. However,
birds can rapidly recover from exposure to these
gas mixtures; therefore, a stun-to-kill process
has been recommended [2, 4, 5]. Unconscious-
1
Corresponding author: jpurswell@msa-msstate.ars.usda.gov
ness was reported in hens when the O
2
concen-
tration was reduced below 5% O
2
by volume
[6], and less than 2% O
2
by volume is recom-
mended for anoxic stun-to-kill processes [7, 8].
However, in gas mixtures containing CO
2
, un-
consciousness may be induced when higher con-
centrations of O
2
are present with an extended
exposure time [9, 10, 11].
One limitation of gas stunning systems is in
achieving uniform concentrations of gases in the
JAPR: Research Report510
atmosphere surrounding the birds as a conse-
quence of inadequate mixing, which may lead
to pockets of air between the birds, reducing the
effectiveness of the process and prolonging the
process time [7]. Mechanically removing air to
reduce the atmospheric pressure, thereby reduc-
ing the partial pressure of oxygen (P
O
2
), may be
an alternative to controlled-atmosphere stun-
ning-killing systems, because pressure is exerted
uniformly in a vessel and does not depend on
the type of process gas used. The use of reduced
atmospheric pressure as a means of slaughter has
been approved for use in farmed game species
(quail, partridge, and pheasant) in Europe [12].
However, little information exists on the re-
sponses of broilers to reduced atmospheric pres-
sure. Thus, the objectives of this study were to
1) determine the operating range of atmospheric
pressures, and 2) to determine the optimum pres-
sure level required to reliably stun and subse-
quently kill broilers by inducing irreversible ces-
sation of respiratory ventilation movements.
MATERIALS AND METHODS
Test System
The test system was composed of an 83.3-
L cylindrical vessel [13] connected directly to
a rotary vane vacuum pump [14] with a flow
rate of 16.9 m
3
/h; the vessel was equipped with
a translucent acrylic lid for observation. A PC-
based data acquisition and control system [15]
was used to monitor tank pressure and control
pump operation, and tank pressure was mea-
sured with a strain gauge-based pressure trans-
ducer [16]. Inlet and exhaust airflow was routed
by 2 manually actuated ball valves. During the
tests, the inlet valve was closed to isolate the
tank from the external atmosphere, and airflow
was directed through the pump via the second
valve. At the conclusion of the test, the tank was
returned to atmospheric pressure through the in-
let valve. The experiments in this study were
approved by the animal care and use committee
at the USDA-ARS Mississippi State location.
Experimental Design
The upper pressure level (70.9 kPa) was se-
lected based on the allowable range for human
habitats in long-term space exploration [17], and
the lowest level (17.8 kPa) was selected based
on O
2
partial pressures shown to induce uncon-
sciousness in hens [6]. Birds in both experiments
were exposed to the respective treatments for 2
min after final pressure had been attained, and
in each of the 2 experiments, 8 replications per
treatment were used.
In experiment 1, 48 Ross ×Ross 708 [18]
female broilers, aged 63 d, were individually
subjected to the following target pressure levels:
70.9, 60.8, 50.7, 40.5, 29.5, and 17.8 kPa. Pres-
sure levels used in experiment 2 were taken from
the lower range of pressures in experiment 1
that resulted in loss of posture (LOP). With
these pressures from experiment 1, 56 Ross ×
Ross 708 male broilers, aged 60 d, were individ-
ually subjected to target pressure levels of 35.3,
32.1, 29.5, 26.6, 23.6, 20.7, and 17.8 kPa.
Loss of posture, resulting from the inability
to maintain a sitting position or neck tension,
has been noted to occur at the onset of uncon-
sciousness [19, 20]. The occurrence of LOP and
cessation of respiratory ventilation movements
were recorded in experiment 1 as primary re-
sponses of interest from which to determine the
range of operating pressures; a bird was consid-
ered dead when respiratory ventilation move-
ments had ceased. Movement of the keel bone
was used as the major indicator of respiratory
ventilation movement. Elapsed times to LOP
and cessation of respiratory ventilation move-
ments were also recorded in experiment 2 and
were determined with a stopwatch.
Statistical Analysis
Loss of posture and survival were coded as
binary data (occurrence =1; no occurrence =0).
Binary data were compared by using logistic
regression [21, 22] with PROC GENMOD [20].
Time data were compared by using ANOVA
with PROC MIXED [23]. Dose-response rela-
tionships can be described by a sigmoid-shaped
curve [24] of the form (equation [1]):
y=a
1+⎛
⎜
⎝
x
x
0
⎞
⎟
⎠
b
, [1]
where yis the probability of survival, ais the
asymptotic maximum probability, xis the slope
of transition (1/kPa), x
0
is the midpoint of transi-
PURSWELL ET AL.: LOW ATMOSPHERIC PRESSURE STUNNING-KILLING 511
Table 1. Effect of atmospheric pressure set point on
incidence of loss of posture (LOP) and mortality
(experiment 1)
Incidence Incidence of
Pressure of LOP mortality
(kPa) (%) (%)
17.8 100 100
29.5 100 70
40.5 0 0
50.7 0 0
60.8 0 0
70.9 0 0
tion (kPa), and bis the inflection point for the
asymptotic maximum/minimum.
Survival data were fitted to this equation,
and nonlinear regression analysis was used to
determine the dose-response relationship be-
tween atmospheric pressure and bird responses
[25]. Statistical significance was established at
P≤0.05.
RESULTS AND DISCUSSION
Experiment 1
Data from experiment 1 are shown in Table
1. The higher pressure treatments (70.9, 60.8,
50.7, and 40.5 kPa) elicited no responses of in-
terest from broilers in this experiment and were
subsequently excluded from further experi-
ments. Loss of posture was observed in all birds
exposed to the remaining 2 treatments (29.5 and
17.8 kPa). Of those birds, 75% exposed to 29.5
kPa and 100% exposed to 17.8 kPa did not sur-
vive the treatment. These data showed complete
separation, and no further statistical analysis was
warranted. The pressure treatments used in ex-
periment 2 were subsequently selected to include
this range.
Experiment 2
The goal of experiment 2 was to determine
the optimal operating pressure for a low-pressure
system. As observed in experiment 1, the propor-
tion of birds surviving and exhibiting LOP in-
creased with increasing atmospheric pressure.
Time to LOP ranged from 34.1 to 50.5 s, and
means are presented in Table 2. The times to
LOP for the lowest 4 pressures (≤26.6 kPa) are
very similar, ranging from 34.1 to 34.9 s. Time
to death increased with increasing atmospheric
Figure 1. Pressure data for the pumping and holding
phases. Evacuation rates were identical for all
treatments, and only the time to pressure differed.
Holding times were likewise identical (120 s) for all
treatments (experiment 2).
pressure and ranged from 79.1 to 142.8 s (Table
2). Time to cessation of respiratory ventilation
movement was not different between pressures
≤23.6 kPa; however, these 3 set points differed
from the 26.6- and 29.5-kPa set points (P≤0.01).
Survival data from experiment 2 were fitted
to equation [1], and the resulting coefficients
were recorded: a=93.3672 (P<0.0001), b=
−21.1147 (P=0.01), and x
0
=29.8918 kPa (P
<0.0001), with an SE of 4.2. By using this
equation, we could determine the pressure re-
sulting in the desired proportion of mortality;
for 99.99% mortality, a maximum pressure of
19.4 kPa should be used.
Time to LOP in experiment 2 for birds ex-
posed to pressures ≤26.6 kPa fell within a narrow
range, showing little correlation with atmo-
spheric pressure. This resulted from the nature
of the pump-down cycle of the test system. With
a constant flow rate and vessel volume, evacua-
tion of the system was highly repeatable, and
the vessel reached the same pressure, with little
variation in time. Pressure measurements were
monitored and recorded throughout the pumping
and holding phases (Figure 1). Taking the aver-
age time to LOP for the lowest 4 pressures (34.5
s) and then determining the pressure at that stage
during the pumping phase, an estimate of the
pressure at which LOP occurs could be deter-
mined. For all birds (n =32) in these 4 treat-
ments, the average pressure after 34.5 s of evacu-
ation was 21.1 kPa, with an SE of 0.8.
JAPR: Research Report512
Table 2. Effect of atmospheric pressure set point on incidence of loss of posture (LOP), time to LOP, incidence
of mortality, and time to cessation of respiratory ventilation movement (experiment 2)
1,2
Incidence Time to Incidence of Time to
Pressure of LOP LOP mortality death
(kPa) (%) (s) (%) (s)
17.8 100 34.5 ±0.7
c
100.0 79.1 ±1.6
b
20.7 100 37.9 ±1.0
c
100.0 85.5 ±1.5
b
23.6 100 34.1 ±1.3
c
100.0 83.4 ±3.8
b
26.6 100 34.6 ±1.6
c
100.0 128.4 ±8.3
a
29.5 100 38.1 ±2.3
bc
62.5 142.8 ±8.7
a
32.1 75 50.5 ±5.4
a
12.5 —
3
35.3 75 46.7 ±4.3
ab
12.5 —
3
a–c
Means within a column with no common superscript differ (P<0.05).
1
Birds were exposed to low-pressure conditions for 2 min after the final pressure was attained.
2
Table values represent mean ±SEM.
3
Time to death was excluded for these pressure settings because of the low incidence of mortality associated with them.
Raj and Gregory [7] recommended an O
2
concentration of 2% by volume under nominal
atmospheric pressure, equating to approximately
2.0 kPa P
O
2
, to minimize problems that might
arise from uneven gas distribution. Results of
this study showed that for the case of reduced
atmospheric pressure, reducing P
O
2
to approxi-
mately 3.7 kPa at the lowest pressure of 17.8
kPa was sufficient to stun and kill broilers. The
pressure required for 99.99% mortality (19.4
kPa) would result in a P
O
2
of 4.0 kPa, which is
similar to that of other anoxic systems at nominal
atmospheric pressure with good gas distribution.
Mean time to death at pressures ≤23.6 kPa was
82.7 s and was similar to those reported for the
Ar, CO
2
, and N
2
processes [9, 10, 26, 27].
The American Veterinary Medical Associa-
tion lists decompression as an unacceptable
means of euthanasia for animals, with the pri-
mary concern that pain and distress could occur
from gases trapped within the body [28]. How-
ever, concerns regarding air trapped in the body
cavity are not applicable to birds. The anatomy
of the avian respiratory apparatus differs from
mammals in many respects. Lungs are fixed and
do not expand and contract; the lungs are
CONCLUSIONS AND APPLICATIONS
1. Controlled-atmospheric pressure reduction appears to be an effective method for humanely
stunning and killing chickens.
2. The mean time to LOP was 34.5 s for pressures ≤26.6 kPa.
3. The mean time to death was 82.7 s for pressures ≤23.6 kPa and is similar to times reported for
other controlled-atmosphere processes.
attached to air sacs and ramifications of these
air sacs extend into many bones. Air sacs com-
pletely fill all the vacant space in the thoracic
cavity and much of the abdominal cavity as well.
Birds have only a rudimentary diaphragm, which
does not extend across the interface of the tho-
racic and abdominal cavities. Thus, in poultry,
trapped air pockets in the body cavity are impos-
sible because of the organization of the extensive
respiratory apparatus. Fedde [29] provides an
excellent discussion of the avian respiratory
system.
A low-pressure system may provide eco-
nomic and safety advantages over other gas stun-
ning processes. The process uses atmospheric
air, eliminating the need to purchase process
gases or generate them on site. Safety concerns
about worker exposure to process gases are elim-
inated, and any leaks would bring atmospheric
air into the system, rather than discharging it.
Further studies should address the influence of
evacuation rate and exposure time to refine the
process, determine the effects of this process on
physiological responses and on carcass or meat
quality, and evaluate the welfare aspects of using
this process.
PURSWELL ET AL.: LOW ATMOSPHERIC PRESSURE STUNNING-KILLING 513
4. The estimated operating pressure for a low-pressure stunning-killing system was 19.4 kPa.
REFERENCES AND NOTES
1. Hoen, T., and J. Lankhaar. 1999. Controlled atmosphere
stunning of poultry. Poult. Sci. 78:287–289.
2. Raj, M., and A. Tserveni-Gousi. 2000. Stunning methods for
poultry. World’s Poult. Sci. J. 56:291–304.
3. Gregory, N. G. 2005. Recent concerns about stunning and
slaughter. Meat Sci. 70:481–491.
4. Raj, A. B. M., N. G. Gregory, and S. B. Wotton. 1990. Effect
of carbon dioxide stunning on somatosensory evoked potentials in
hens. Res. Vet. Sci. 49:355–359.
5. Raj, M. 1998. Welfare during stunning and slaughter of
poultry. Poult. Sci. 77:1815–1819.
6. Woolley, S. C., and M. J. Gentle. 1988. Physiological and
behavioral responses of the domestic hen to hypoxia. Res. Vet. Sci.
45:377–382.
7. Raj, A. B. M., and N. G. Gregory. 1990. Investigation into
the batch stunning/killing of chickens using carbon dioxide or argon-
induced hypoxia. Res. Vet. Sci. 49:364–366.
8. Raj, A. B. M., N. G. Gregory, and S. B. Wootton. 1991.
Changes in the somatosensory evoked potentials and spontaneous
electroencephalogram of hens during stunning in argon-induced an-
oxia. Br. Vet. J. 147:322–330.
9. Gerritzen, M. A., E. Lambooij, S. J. W. Hillebrand, J. A. C.
Lankhaar, and C. Pieterse. 2000. Behavioral responses of broilers to
different gaseous atmospheres. Poult. Sci. 79:928–933.
10. Gerritzen, M. A., B. Lambooij, H. Reimert, A. Stegeman,
and B. Spruijt. 2004. On-farm euthanasia of broiler chickens: Effects
of different gas mixtures on behavior and brain activity. Poult. Sci.
83:1294–1301.
11. Lambooij, E., M. A. Gerritzen, B. Engel, S. J. W. Hillebrand,
J. Lankhaar, and C. Pieterse. 1999. Behavioural responses during
exposure of broiler chickens to different gas mixtures. Appl. Anim.
Behav. Sci. 62:255–265.
12. European Commission, Directorate General, Health and Con-
sumer Protection. 2003. Council Directive 93/119/EC. European
Commission, Brussels, Belgium.
13. LVC1820FP75-VIA, Laco Technologies, Salt Lake City, UT.
14. DS302, Varian Inc., Palo Alto, CA.
15. USB-1208FS, Measurement Computing Corp., Norton, MA.
16. PX-603, Omega Engineering Inc., Stamford, CT.
17. Lange, K. E., and C. H. Lin. 1998. Advanced Life Support
Program: Requirements Definition and Design Considerations. JSC-
38571. NASA Johnson Space Center, Houston, TX.
18. Aviagen Inc., Huntsville, AL.
19. Raj, A. B. M., and N. G. Gregory. 1990. Effect of rate
of induction of carbon dioxide anesthesia on the time of onset of
unconsciousness and convulsions. Res. Vet. Sci. 49:360–363.
20. Raj, A. B. M., S. B. Wotton, J. L. McKinstry, S. J. W.
Hillebrand, and C. Pieterse. 1998. Changes in the somatosensory
evoked potentials and spontaneous electroencephalogram or broiler
chickens during exposure to gas mixtures. Br. Poult. Sci. 39:686–695.
21. McCullagh, P., and J. A. Nelder. 1989. Generalized Linear
Models. 2nd ed. Chapman and Hall, New York, NY.
22. Hosmer, D. W., and S. Lemeshow. 1989. Applied Logistic
Regression. John Wiley and Sons, New York, NY.
23. SAS Institute. 2004. SAS User’s Guide. Statistics. Version
9.1 Edition. SAS Inst. Inc., Cary, NC.
24. Hudson, R. J., and T. K. Henthorn. 2006. Basic principles
of clinical pharmacology. Pages 263–264 in Clinical Anesthesia. 5th
ed. P. G. Barash, B. F. Cullen, and R. K. Stoelting, ed. Lippincott
Williams and Wilkins, Philadelphia, PA.
25. SigmaPlot, Version 8, Systat Software Inc., San Jose, CA.
26. Raj, A. B. M., S. B. Wotton, and N. G. Gregory. 1992.
Changes in the somatosensory evoked potentials and spontaneous
electroencephalogram of hens during stunning with a carbon dioxide
and argon mixture. Br. Vet. J. 148:147–156.
27. Poole, G. H., and D. L. Fletcher. 1995. A comparison of
argon, carbon dioxide, and nitrogen in a broiler killing system. Poult.
Sci. 74:1218–1223.
28. American Veterinary Medical Association. 2000. Report of
the AVMA Panel on Euthanasia. 2000. J. Am. Vet. Med. Assoc.
218:669–696.
29. Fedde, M. J. 1986. Respiration. Pages 191–220 in Avian
Physiology. 4th ed. P. D. Sturkie, ed. Springer-Verlag, New York,
NY.