PII: S0003-4878(00)00081-8
Ann. occup. Hyg., Vol. 45, No. 6, pp. 457–465, 2001
 2001 British Occupational Hygiene Society
Published by Elsevier Science Ltd. All rights reserved
Printed in Great Britain.
0003-4878/01/$20.00
Exposure Assessment to Airborne Endotoxin, Dust,
Ammonia, Hydrogen Sulfide and Carbon Dioxide in
Open Style Swine Houses
C. W. CHANG†*, H. CHUNG†‡, C. F. HUANG§ and H. J. J. SU§
†Institute of Occupational Safety and Health, Council of Labor Affairs, Executive Yuan, Taipei,
Taiwan, ROC; ‡Present address: 3M Taiwan Ltd, 13th floor, Lotus Building, #136, Sec. 3, Jen-Ai Rd,
Taipei, 106, Taiwan, ROC; §Department of Environmental and Occupational Health, National Cheng
Kung University, Tainan, Taiwan, ROC
Information is limited for the exposure levels of airborne hazardous substances in swine feed
buildings that are not completely enclosed. Open-style breeding, growing and finishing swine
houses in six farms in subtropical Taiwan were studied for the airborne concentrations of
endotoxin, dust, ammonia, hydrogen sulfide and carbon dioxide. The air in the farrowing
and nursery stalls as partially enclosed was also simultaneously evaluated. Three selected
gases and airborne dusts were quantified respectively by using Drager diffusion tubes and a
filter-weighing method. Endotoxin was analyzed by the Limulusamoebocyte lysate assay.
Average concentration of airborne total endotoxin among piggeries was between 36.8 and
298 EU/m3, while that for respirable endotoxin was 14.1–129 EU/m3. Mean concentration of
total dust was between 0.15 and 0.34 mg/m3, with average level of respirable dust of 0.14
mg/m3. The respective concentrations of NH3, CO2 and H2S were less than 5 ppm, 600–895
ppm and less than 0.2 ppm. Airborne concentrations of total dust and endotoxin in the nursery house were higher than in the other types of swine houses. The finishing house presented
the highest exposure risk to NH3, CO2 and H2S. Employees working in the finishing stalls
were also exposed to the highest airborne levels of respirable endotoxin and dust. On the
other hand, the air of the breeding units was the least contaminated in terms of airborne
endotoxin, dust, NH3, CO2 and H2S. The airborne concentrations of substances measured in
the present study were all lower than most of published studies conducted in mainly enclosed
swine buildings. Distinct characteristics, including maintaining swine houses in an open status
and frequent spraying water inside the stalls, significantly reduce accumulation of gases and
airborne particulates.  2001 British Occupational Hygiene Society. Published by Elsevier
Science Ltd. All rights reserved
Keywords: endotoxin; pigs; swine; ventilation
INTRODUCTION
The confinement buildings, which are enclosed structures with high density of livestock, have been used
in the swine production industry in recent decades.
Because of the building design, the feed materials,
swine, and their wastes are concentrated indoors,
resulting in high levels of airborne hazardous particulates and toxic gases. Development of respiratory dis-
Received 21 July 2000; in final form 9 October 2000.
*Author to whom correspondence should be addressed. Tel.:
+886-2-87701567;
fax:
+886-2-25451048;
e-mail:
[email protected]
orders has been reported associated with working in
swine confinement buildings (Crook et al., 1991;
Donham and Gustafson, 1982; Olson and Bark,
1996). Endotoxin, swine dust and toxic gases that are
identified in the air of piggeries are considered as the
important causes (Donham and Popendorf, 1985;
Heederik et al., 1991; Reynolds et al., 1996; Zejda et
al., 1994).
Studies on airborne endotoxin, a lipopolysaccharide (LPS) component of cell walls of gram-negative
bacteria, have shown significant relation to decrement
in lung function of swine workers (Donham et al.,
1989; Heederik et al., 1991; Vogelzang et al., 1998).
Exposure to airborne swine dust above 2.5 mg/m3 is
457
458
C. W. Chang et al.
also reported to be associated with symptoms of respiratory disease (Donham et al., 1989) and decrease
in FEV1 (forced expiratory volume at 1 s) (Donham
et al., 1995; Reynolds et al., 1996). In addition to
airborne particulates, toxic gases may be produced
during degradation of fecal matter, animal respiration,
and building operations. Levels of carbon dioxide,
ammonia and hydrogen sulfide in swine confinement
units have been reported in excess of threshold limit
values for occupational exposure (Donham et al.,
1977). A number of acute deaths during swine farming operation have been associated with exposure to
high levels of hydrogen sulfide (Donham and Gustafson, 1982). Some studies also demonstrated that the
magnitude of the decrease in baseline (Preller et al.,
1995) and cross-shift (Donham et al., 1995) lung
function of swine workers is associated with increasing ammonia concentration in swine confinement
units.
Many studies conducted in the enclosed swine
buildings demonstrate the exposure levels of hazardous substances and their associations with observed
respiratory disorders, but little information is available for the open style swine houses that are modified
to adjust for the subtropical climate to achieve optimal pig growth and maximal production. Although
pigs in open style swine houses are managed similarly
to those in enclosed buildings, it was hypothesized
that the difference in building structure, interacting
with other characteristics of swine farms such as the
cleaning practice and pig density, might affect the
levels of airborne hazardous matters in stalls. Consequently, the concentrations of airborne hazardous substances and the possibility of developing adverse
health effects on workers employed in the open style
piggeries might be different from those reported from
swine confinement buildings. Thus, the open style
swine houses, located in subtropical Taiwan, were
selected as the subjects of the present study. The airborne endotoxin, dust, ammonia, hydrogen sulfide
and carbon dioxide in different types of piggeries
were assessed. The results were compared to the published data that mostly obtained from the enclosed
swine buildings.
Fig. 1. View of a breeding stall with individually separated pigs
in an open building.
nursery piggeries where piglets are raised to a weight
of approximately 20 kg. Then, pigs are fattened in the
growing houses until reaching 50 kg approximately,
and moved again to the finishing buildings where they
are kept fattened to 100–110 kg before they are sent
for slaughter. An example of growing and finishing
swine houses that are commonly observed in Taiwan
is shown in Fig. 2.
Characteristically of a subtropical climate, the
ambient temperature in summer is normally above
30°C, especially in southern Taiwan where most
swine farms are located. Thus, swine houses are built
and operated in a very open style, typified by the
breeding, growing and finishing piggeries shown in
Figs 1 and 2. In order to protect newborn and growing
piglets from cold weather occasionally occurs in winter, two types of buildings are constructed for the farrowing and nursery stalls. One is the open structure
similar to Fig. 2, but with plastic curtains on sides
of the buildings and adjustable for weather condition.
Another type is partially enclosed with many windows on the sides (Fig. 3).
MATERIAL AND METHOD
The environments
In Taiwan, there are five types of distinct piggeries
involved in pig production industry, i.e., breeding,
farrowing, nursery, growing and finishing houses.
The breeding house is the place primarily designed
for pregnant pigs. As shown in Fig. 1, each pregnant
sow lives in an individual pen until approximately
one week prior to delivery of babies. They are then
transferred to the farrowing buildings where piglets
are born and lactated for four weeks. Sows then return
to the breeding buildings and prepare for the next
pregnancy, while weaned piglets are moved to the
Fig. 2. Illustration of a growing piggery as an example of open
style swine houses commonly found in the growing and finishing stalls in Taiwan.
Endotoxin, dust, and gases in open style swine houses
459
above. The cassette was then clipped onto collars of
workers with pumps operated at 1.7 l./min.
Depending on the working hours of each participating
person on the sampling day, the sampling periods for
respirable dust ranged from 6 to 8 h. All of collected
filters were subsequently re-conditioned in the thermostatic cabinet set for 19°C and 50% relative
humidity for 24 h, and reweighed in the laboratory
to determine the weight concentrations of total and
respirable airborne dust.
Fig. 3. View of an example of partially enclosed style swine
houses built for farrowing and nursery purposes.
Sampling strategy and procedure
In order to simultaneously investigate and compare
levels of airborne contaminants among five major
types of swine buildings, 10 large-scale farms located
at southern Taiwan were randomly selected. Each
raised more than 10 000 pigs annually and consisted
of five distinct types of piggeries. Walkthrough of
selected farms and preliminary survey on levels of
airborne dust, ammonia, hydrogen sulfide and carbon
dioxide were conducted in September and repeated in
December 1994. The information was used to determine a feasible sampling protocol for the contaminants measured.
A two-day comprehensive sampling survey was
then performed in April–May 1995 at each type of
swine house on six recruited swine farms. In total 30
swine houses were investigated. Area sampling was
performed for total endotoxin, total dust, ammonia,
hydrogen sulfide and carbon dioxide. These samples
were simultaneously collected at 1.5 m above the
floor of central walkway in the examined piggeries.
The samples were also collected at the same height
in the administration office as well as the perimeter
area of swine farms that were at least 600 m from the
nearest piggery. Personal sampling was also conducted for determination of airborne levels of respirable
endotoxin and dust.
Dust sampling and analysis.
Polycarbonate
membrane filters (0.4 µm pore size, 37 mm diameter,
Nuclepore) were positioned in a thermostatic cabinet
set for 19°C and 50% relative humidity for 24 h. Filters were then weighted, housed into 3-piece polystyrene cassettes (37 mm diameter, SKC) with the
support pads (cellulose backing pad, Nuclepore),
sealed and ready for use. In measurement, air was
then drawn through the cassette at a flow rate of 1.7
l./min for 6 h. Three samples were simultaneously
collected at 1.5 m height above the floor of central
walkway in each selected house with one field blank.
Respirable dust was obtained from the breathing zone
of swine workers by attaching 10-mm Gilian nylon
cyclone onto the dust collection device as described
Endotoxin sampling and analysis.
Glass fiber
backing pads (AP40, Millipore) were autoclaved at
270°C for 30 min. Polycarbonate membrane filters
(0.4 µm pore size, 37 mm diameter, Millipore) supported by sterile glass fiber pads were subsequently
housed in 3-piece polystyrene cassettes (SKC). One
endotoxin sample per sampling day was collected at
1.7 l./min for 6 h at each measured site with one field
blank. After sampling, a desiccant tube was plugged
into the bottom inlet of each cassette, which was then
sealed, transported to the laboratory and stored at
4°C. For analysis, exposed and blank filters were
placed into glass bottles containing 5 ml TAP Buffer
(0.05 M potassium phosphate, 0.01% triethylamine,
pH 7.5), and ultrasonically shaken for 1 h at 20°C.
Serial dilutions were then prepared with Tap Buffer.
With addition of Limulus amoebocyte lysate KineticQCL reagent (BioWhittaker, Walkersville, Maryland,
USA) the optical density of diluted samples was
recorded at 405 nm. The level of endotoxin was then
determined by the kinetic Limulus assay with resistant-parallel-line estimation (KLARE) (Walters et al.,
1994; Milton et al., 1990), expressed as endotoxin
unit (EU).
Respirable endotoxin sampling was carried out in
the breathing zone of swine workers, in a similar way
to collecting respirable dust. The sterilization of support pads and treatment of collected samples for respirable endotoxin were identical to that for the total
endotoxin as mentioned above.
Gases sampling and analysis. Concentrations of
NH3, H2S and CO2 were determined with Drager colorimetric detection tubes (Zuskin et al., 1992). Readings were recorded at 4, 8 and 24 h at each measured
site after employees started working, approximately
at 7:30–8:00 AM. The reading values were presented
as the average concentration of the tested gases in the
respective period of measurement.
Characteristics of examined piggeries and environmental factors. Data on the characteristics of the
stalls where actual measurements were taken were
obtained through visual inspection and interviews
with swine workers. These consisted of structure of
piggery, swine density, feed material and frequency
as well as cleaning practice and frequency. Environmental factors including temperature, relative
humidity and wind velocity were recorded at 1.5 m
above the floor in the central walkways as well as at
460
C. W. Chang et al.
two opening ends of swine houses using direct-reading instruments. Three readings for each environmental factor were daily taken at four occasions, i.e., at
7:30–8:30 AM, 10:30–11:30 AM, 1:30–2:30 PM and
4:00–5:00 PM.
Statistical analysis
SAS software was used to test the normality of
original and transformed data. For normally distributed data, analysis of variance (ANOVA) followed
by Duncan’s multiple range test, Student–Newman–
Keuls (SNK) test and Scheffe test was performed to
detect significant difference between means of contaminants among various measurement sites. Student
t test or nonparametric analysis was used to test the
difference of exposure levels that were taken in stalls
and outside area.
RESULTS
Endotoxin
The airborne endotoxin levels are presented in
Table 1. Comparison of the amounts of total airborne
endotoxin shows that the nursery, growing and finishing buildings had substantially higher endotoxin
levels than the farrowing and breeding buildings. The
highest average level of 298 EU/m3 was present in
the nursery units while the breeding piggeries had the
lowest mean of 36.8 EU/m3. The growing and finishing units had similar concentrations of total airborne endotoxin, 145 and 136 EU/m3, respectively.
Twenty four samples taken from the administrative
offices and perimeters of swine farms, indicated as
surrounding in Table 1, showed an average total
endotoxin of 8.9 EU/m3. This concentration was significantly lower than those measured in piggeries.
Table 1 also indicates that swine employees working in the finishing piggeries had the highest average
exposure level of respirable endotoxin in their breathing zone areas (129 EU/m3), followed by workers in
farrowing, growing, nursery and breeding units.
There was approximate tenfold difference in respirable endotoxin level between persons working in the
finishing piggeries and in the breeding houses. However, no statistically significant difference was
detected.
Dust
As shown in Table 2, the means of total airborne
dust in various swine houses were between 0.15 and
0.34 mg/m3 with individual values ranging from 0.03
to 1.11 mg/m3. Results of statistical analysis showed
that dust concentration in the nursery stalls was significantly higher than in the other piggeries, and total
dust concentrations in swine houses were all significantly higher than the surrounding area. Table 2 also
demonstrates that workers in the finishing units were
exposed to the highest level of respirable dust of 0.24
mg/m3, but no statistical difference was found
between stalls for respirable dust levels.
Gases
Table 3 illustrates that mean concentrations of
NH3, CO2 and H2S in the piggeries were less than 5
ppm, between 600 and 895 ppm, and less than 0.2
ppm, respectively. Concentrations of these three gases
inside the swine houses were all significantly higher
than those measured in the surrounding area except
for H2S in the breeding units. When compared to
other types of stalls, the finishing houses continuously
contained the highest levels of all three measured
gases over the 24-h sampling period, and the difference reached statistical significance (P⬍0.05). The
levels of the three gases, on the other hand, were the
lowest in general in the air of the breeding units.
Table 3 also shows that, NH3 concentrations read
at 4 hours after starting working at approximately
7:30–8:00 AM were higher than those read at 8 h,
indicating that NH3 levels in the morning were higher
Table 1. Concentrations of total and respirable airborne endotoxin in various swine buildingsa
Subject measured
Total endotoxin
Mean±SD (EU/m3)
Range (EU/m3)
C.V. (%)b
N
Respirable endotoxin
Mean±SD (EU/m3)
Range (EU/m3)
C.V. (%)
N
a
Breeding
Farrowing
Nursery
Growing
Finishing
Overall
Surroundingc
36.8±18.1
15.8–73.2
49
12
82.1±85.0
14.4–277
103
12
298±249d
32.0–818
83
12
145±81.4d
40.1–298
56
12
136±105d
30.8–418
77
12
140±132
14.4–818
95
60
8.9±7.2e
3.2–32.9
81
24
14.1±16.0
3.4–56.6
113
10
48.6±166
3.5–837
341
25
20.9±31.9
1.6–155
153
22
21.8±45.6
0.02–217
209
21
129±396
5.6–1643
308
17
47.0±190
0.02–1643
404
95
–
–
–
–
Area sampling was performed for total endotoxin while respirable endotoxin was determined by personal sampling.
C.V.: coefficient of variation.
Surrounding area included administrative offices and farm perimeters.
d
P⬍0.05, compared to each of other types of stalls with lower mean of endotoxin concentration.
e
P⬍0.05, compared to each type of piggeries.
b
c
Endotoxin, dust, and gases in open style swine houses
461
Table 2. Concentrations of total and respirable airborne dust in various swine buildingsa
Subject measured
Total dust
Mean±SD (mg/m3)
Range (mg/m3)
C.V. (%)b
N
Respirable dust
Mean±SD (mg/m3)
Range (mg/m3)
C.V.(%)
N
Breeding
Farrowing
Nursery
Growing
Finishing
Overall
Surroundingc
0.15±0.04
0.09–0.23
26
18
0.23±0.12
0.08–0.47
52
18
0.34±0.13d
0.16–0.54
39
18
0.28±0.28
0.03–1.11
97
18
0.21±0.07
0.09–0.34
34
18
0.24±0.15
0.03–1.11
64
90
0.12±0.05e
0.06–0.31
46
24
0.12±0.13
0.01–0.49
111
12
0.08±0.05
0.01–0.22
60
13
0.13±0.15
0.02–0.57
121
11
0.15±0.18
NDf–0.64
123
12
0.24±0.46
0.03–1.45
102
9
0.14±0.22
ND–1.45
158
57
–
–
–
–
a
Area sampling was performed for total dust while respirable dust was determined by personal sampling.
C.V. coefficient of variation.
c
Surrounding area included administrative offices and farm perimeters.
d
P⬍0.05, compared to each of other types of stalls with lower mean of dust concentration.
e
P⬍0.05, compared to each type of piggeries.
f
ND: non-detected.
b
Table 3. Mean concentrations of NH3, H2S and CO2 measured from six swine farms and surrounding area
Gas
NH3
Reading
time, ha
4
8
24
H2S
4
8
24
CO2
4
8
24
Mean conc.±SD, ppm (N)
Breeding
Farrowing
Nursing
Growing
Finishing
Overall
Surroundingb
NDc
(12)
ND
(11)
0.33±0.40
(6)
ND
(12)
ND
(12)
0.02±0.04
(6)
711±156
(12)
633±78
(12)
600±50
(6)
0.83±1.85
(12)
0.63±1.24
(12)
1.18±1.30
(6)
0.08±0.20
(12)
0.11±0.24
(11)
0.04±0.09
(5)
791±87
(12)
621±57
(12)
700±40
(6)
2.21±3.66
(12)
2.43±3.58
(12)
2.32±3.17
(6)
0.16±0.37
(12)
0.16±0.26
(12)
0.13±0.16
(6)
849±201
(12)
690±142
(12)
623±27
(5)
2.54±2.70
(12)
1.85±2.01
(12)
3.05±2.72
(6)
0.03±0.11
(12)
0.06±0.11
(12)
0.03±0.08
(6)
765±79
(12)
619±54
(12)
699±49
(6)
3.93±3.11d
(12)
3.38±2.74d
(12)
4.15±3.91d
(6)
0.18±0.33d
(12)
0.18±0.26d
(12)
0.13±0.12d
(6)
895±116d
(12)
720±57d
(12)
704±73d
(5)
1.90±2.60
(60)
1.66±2.28
(59)
2.21±2.63
(30)
0.09±0.24
(60)
0.10±0.20
(59)
0.07±0.11
(29)
802±136
(60)
657±85
(60)
665±50
(28)
ND
(17)
ND
(16)
ND
(6)
ND
(17)
ND
(16)
ND
(6)
595±99e
(17)
518±41e
(16)
552±27e
(12)
a
The readings of gas detection tubes were taken at 4, 8 and 24 h after swine workers beginning working at approximately
7:30–8:00 AM.
b
Surrounding area included administrative offices and farm perimeters.
c
ND: non-detected.
d
P⬍0.05, compared to each of other types of stalls with lower mean of measured gas concentration.
e
P⬍0.05, compared to each type of piggeries.
than the averages of the whole workday. However,
the highest concentrations were found at 7:30–8:00
AM of the next day (i.e., 24-h readings). This trend
was typically shown in the farrowing, growing and
finishing stalls. As for CO2, the readings taken at
11:30 AM–12:00 PM (i.e. 4 h after starting working)
were the highest in all types of piggeries. No consistent change in CO2 concentration was demonstrated
for 8 and 24 h readings.
Characteristics of examined piggeries and environmental factors. Table 4 shows that on average the
finishing stall was the largest with a mean surface
area of 936 m2. Other stalls were averaged between
530 and 586 m2. Swine density was estimated to be
0.32, 0.85 and 0.78 hog/m2 in the breeding, growing
and finishing houses, and 1.23 and 1.63 piglet/m2 in
the farrowing and nursing stalls, respectively. All of
the investigated breeding, growing and finishing
houses were open style, with no curtains or curtains
scrolled up on sides of building, and respectively, 50,
20 and 17% of breeding, growing and finishing
houses were also mechanically ventilated. On the
other hand, 84% of farrowing and 67% of nursery
houses were partially enclosed, which were recognized as enclosed buildings with open windows or
16
84
67
100
33
67
0
1.2±0.4
1±0
100
0
50
50
83
0
0
1±0
1±0
0.61±0.48
30.7±2.1
61.4±8.4
0.10±0.00
1.23±0.14
0.32±0.16
–
0.80±0.44
30.4±2.1
61.8±8.8
530±166
534±190
Farrowing
0.60±0.38
30.6±2.3
62.6±8.7
67
0
0
1±0.1
1.3±0.7
34
67
67
83
b
0.74±0.34
30.6±2.5
61.9±8.4
50
0
17
0.9±0.2
25±53.1
100
0
20
17
0.85±0.24
–
–
1.63±0.73
586±543
a
Growing
547±193
Nursery
Not applicable.
Open structure with no curtains or curtains scrolled up.
c
Including the enclosed building with open windows and the open structure with curtains partially scrolled down.
d
Percentage in occurrence of specified characteristic.
e
Total not equal to 100%, some observations not classifiable.
a
Wind velocity (m/s)
Temperature (°C)
Relative humidity (%)
Surface area of stalls, m2
Pig density, #/m2
Hog
Piglet
Stall style
Openb
Partially enclosedc
Mechanically ventilated (%)d
Slatted pen floor (%)d
Cleaning methode (%)d
Manual, watering
Manual, sweeping
Automated
Cleaning frequency (per day)
Cleaning interval (day)
Breeding
Table 4. Characteristics, practices and environmental factors in the swine houses studied
0.74±0.39
30.5±2.2
62.5±8.3
50
0
33
1.2±0.5
30.1±59.5
100
0
17
17
0.78±0.25
–
936±419
Finishing
0.7±0.4
30.5±2.2
62.0±8.5
57
16
10
1.1±0.3
11.9±35
70
30
45
53
0.41±0.39
0.57±0.79
681±362
Overall
462
C. W. Chang et al.
Endotoxin, dust, and gases in open style swine houses
open structures with curtains partially scrolled down.
Mechanical ventilation was used in 67% of farrowing
and nursery buildings as well. The pen flooring was
slatted in all of farrowing and 83% of nursery piggeries, in contrast to only 17% of growing and finishing stalls. The manure pit was not observed under
the slatted floors of the investigated stalls.
Investigated stalls were mainly cleaned by water
spraying except for the farrowing units, where dry
sweeping was conducted. The cleaning in the breeding, farrowing and nursery stalls was performed routinely on daily basis. For growing and finishing piggeries, the average intervals of cleaning practice were
25 and 30.1 days, respectively. It was observed in all
piggeries that swine were fed 1–2 times per day with
dry formula (powder-type materials found in 80% of
studied stalls).
The averages of wind velocity, temperature and
relative humidity among swine houses were measured
to be 0.6–0.8 m/s, 30.4–30.7°C, and 61.4–62.6%,
respectively.
DISCUSSION
Exposure levels in piggeries
Present results show the air of nursery stalls was
contaminated with the significantly highest levels of
total endotoxin as well as total dust. Attwood et al.
also demonstrated that the nursery and farrowing
buildings had substantially higher dust levels and
endotoxin concentrations than the fattening buildings
(Attwood et al., 1987). They considered the reduced
ventilation for heat conservation, greater activity of
young pigs and higher pig density were the reasonable explanations. Although they did not analyze the
data specifically for the nursery stalls, their explanations seem also adequate to justify the findings of
the present study except for the reduced ventilation
for heat conservation. In the present study, 67% of
nursery houses were mechanically ventilated along
with natural air exchange by open windows. Similar
levels of temperature and wind velocity within all of
investigated stalls further imply significant reduced
ventilation was unlikely occurred in the investigated
nursery piggeries. Under such condition, great
activity of young swine and high pig density are probably the major reasons for the existence of high airborne contaminants in the nursery stalls of the
present study.
The growing houses contained the second highest
levels of both total endotoxin and total dust. Growing
pigs in this type of piggeries were not as active as
young piglets in nursery stalls. However, the growing
stalls were still densely stocked. Unlike the nursery
stalls with daily cleaning, the growing piggeries were
cleaned at an average interval of 25 days. Probably
because of high pig density and infrequent cleaning
practice, significant high concentration of airborne
culturable gram-negative bacteria was identified in
463
the air of this type of stalls (Chang et al., 2001).
Although the biological activity of endotoxin is not
dependent on bacterial viability (Jacobs, 1989),
increasing concentration of culturable gram-negative
bacteria found in the growing stalls is likely to lead
to elevation of airborne endotoxin level. This explanation also supports the high endotoxin found in the
finishing houses (Table 1), where crowded adult pigs
were stocked in the pens that were also found
infrequently cleaned (interval of 30.1 days) and significantly contaminated by gram-negative bacteria
(Chang et al., 2001).
Based on the findings on the nursery, growing and
finishing stalls, it is expected that decreases in pig
density, pig activity and gram-negative bacteria level
would reduce the amount of airborne endotoxin and
dust in the stalls. This finding is also supported by our
other study, which found that the breeding buildings,
housing the lowest density of inactive pregnant pigs
and being cleaned on daily basis, had relatively low
concentration of airborne gram-negative bacteria
(Chang et al., 2001) and the lowest levels for both
endotoxin and dust, as shown in Tables 1 and 2.
While the nursery houses contained the highest airborne concentrations of total dust and total endotoxin,
the workers in this type of stall were not exposed to
the highest level of either respirable dust or respirable
endotoxin. Instead, this was found for those working
in the finishing piggeries. This inconsistency might
be due to the variation of respirable portion in total
endotoxin and dust among different stalls. The difference in the length of time that workers stayed in the
stalls might also contribute to this finding. Among all
of the investigated stalls, the surface area of the finishing units was the largest (936±419 m2) and almost
twice that of nursery houses (547±193 m2). It was
observed that workers in the finishing units needed to
stay longer than those in the nursery houses in order
to finish their work such as checking, vaccinating,
cleaning or feeding pigs. Since endotoxin and dust
concentrations were significantly higher in piggeries
than in the surrounding area of swine farms, the average amount of endotoxin or dust measured by personal sampling would be increased as if workers stay
longer in the contaminated piggeries than in the surrounding area.
NH3 is mainly produced by microbial activity on
breakdown of urea in animal urine (Essen and Donham, 1999), while CO2 results from animal respiration (Olson and Bark, 1996). Concentrations of NH3
and CO2 in the finishing units were detected as significantly highest among piggeries. This finding was
properly justified by high concentration of airborne
microbes (Chang et al., 2001), presence of swine
excrements on non-slatted floor and high density of
grown pigs observed in the finishing units of the
present study. On the other hand, lower concentrations of both NH3 and CO2 detected at the end of
workday than at noon are likely to be related to water
464
C. W. Chang et al.
cleaning practices and mist supply activities performed mainly in the afternoon, as these gases are watersoluble and consequently quickly removed from the
air. Pigs were usually fed again before workers left
the farms around 4:30 PM. During the night, pig
excrements were continuously produced, accumulated
and decomposed by microbes, possibly created more
NH3. As neither cleaning practice nor significant disturbance by animal activities likely occurred overnight in piggeries, NH3 concentration tended to be
continuously elevated. This might explain the findings of the highest NH3 concentrations at 7:30–8:00
AM of the next day. However, CO2 concentrations
were not consistently increased overnight, probably
because of decreased respiration rate of pigs during
sleeping. Relatively low concentrations of H2S in the
present study limit the possibility of observing consistent concentration fluctuations over time.
Comparison with the published reports
Previous reports demonstrated the total endotoxin
in the air of swine confinement buildings ranges
between 0.006 and 31.25 µg/m3 (Attwood et al.,
1986, 1987; Clark et al., 1983; Crook et al., 1991;
Donham et al., 1986, 1989; Haglind and Rylander,
1987; Heederik et al., 1991; Mackiewicz, 1998), or
averaged as 11443 EU/m3 (Zejda et al., 1994), 176
and 203 EU/m3 (Reynolds et al., 1996). Meanwhile,
most studies conducted in the Netherland, Sweden
and America reported mean endotoxin levels ranging
from 100 to 200 ng/m3 (Attwood et al., 1986, 1987;
Clark et al., 1983; Donham et al., 1986, 1989; Heederik et al., 1991; Vogelzang et al., 1997). In our
study, the RSE (reference standand endotoxin,
U.S.P.C., Inc. Rockville, MD. USA) used was from
Lot EC-7 (Lot G). The CSE (control standard endotoxin, Associates of CAPE CODE, Inc. Woods Hole,
MA, USA) was Escherichia coli O113:H10. Based on
the two standards used, we achieved a conversion factor of 0.0157 Eu/pg. By applying this conversion factor to our results, the means of total endotoxin concentration obtained in the present study were varied
from 2.36 to 19.07 ng/m3. These levels are relatively
lower than most studies referenced above, and are
only similar to those reported by Crook et al. (1991)
and Reynolds et al. (1996). The average levels of respirable endotoxin in the breathing zone of swine
workers, between 0.9 and 8.26 ng/m3 after conversion, were also lower than those of Donham et al.
(1989) and of Vogelzang et al. (1998), but comparable to the findings reported by Reynolds et al. (1996).
As for dust, total airborne dust levels determined
in the present study, averaged between 0.15 and 0.34
mg/m3, were approximately tenfold smaller than
those of other published studies averaged from 1.32
to 8.8 mg/m3 (Attwood et al., 1986, 1987; Clark et
al., 1983; Cormier et al., 1990; Donham and Gustafson, 1982; Donham et al., 1986, 1989; Haglind and
Rylander, 1987; Heederik et al., 1991; Holness et al.,
1987; Mackiewicz, 1998; Zejda et al., 1994; Zuskin
et al., 1992). Means of respirable airborne dust in piggeries have been reported between 0.13 and 2.5
mg/m3 (Donham and Gustafson, 1982; Donham et al.,
1986, 1989; Haglind and Rylander, 1987; Holness et
al., 1987; Reynolds et al., 1996; Zejda et al., 1994;
Zuskin et al., 1992). Our data of 0.08–0.24 mg/m3
are only comparable to the results of Zejda et al..
The concentrations of NH3 and CO2 in our findings
were also well below most of the published data, with
average levels respectively as 5.15–34 ppm (Attwood
et al., 1987; Donham and Popendorf, 1985; Donham
et al., 1989; Haglind and Rylander, 1987; Heederik
et al., 1991; Reynolds et al., 1996; Zejda et al., 1994)
and 1640–2632 ppm (Attwood et al., 1987; Donham
and Popendorf, 1985; Donham et al., 1989; Zejda et
al., 1994). H2S levels shown in the present study were
also lower than 1.4 ppm reported by Donham and
Popendorf (1985), and less than 5–10 ppm determined in totally enclosed swine buildings during wintertime (Donham et al., 1977).
In addition to as an indicator of the potential health
effect, CO2 level serves as a measure of general air
quality as it reflects the ventilation status of a building. A level of CO2 above 1500 ppm is likely to be
associated with a higher risk of occupational respiratory diseases since it usually reflects elevated airborne concentrations of other dangerous substances
in insufficiently ventilated workplaces (Essen and
Donham, 1999). Low level of CO2 in the present
study illustrated appropriate performance of ventilation in the investigated swine houses. Buildings
maintained as open structures with additional mechanical ventilation enhance efficient dilution of gases
and aerosols indoors, resulting in the relatively low
levels of gases, endotoxin and dust observed in the
present study. In addition, frequent water spraying
inside the stalls also help reducing the accumulation
of toxic gases and hazardous particulates.
CONCLUSION
Our present data showed that the finishing units
were the most contaminated stalls with the highest
levels of ammonia, hydrogen sulfide and carbon dioxide in the air. Workers in this type of stalls were also
exposed to the highest airborne levels of respirable
endotoxin and dust. Our previous study that surveyed
on the same farms also identified the finishing houses
as the highest contaminated stalls with total bacteria
and gram-negative bacteria (Chang et al., 2001). In
addition to the finishing houses, the present study also
revealed that the nursery houses were contaminated
with relatively high endotoxin and dust in the air. The
workers in the finishing and nursery stalls were probably exposed to relatively higher levels of airborne
hazards than working in other piggeries. On the other
hand, the air of the breeding units was the least contaminated in terms of endotoxin, dust, ammonia,
Endotoxin, dust, and gases in open style swine houses
hydrogen sulfide and carbon dioxide, for which low
pig density, inactive characteristic of pregnant sows,
daily water cleaning practice and open feature of
building style are reasonable explanations.
The concentrations of airborne contaminants measured in the present study were all lower than most of
the published studies that were conducted mainly in
enclosed swine buildings. The maintenance of open
style swine houses minimizes the accumulation of
gases and airborne particulates indoors. Spraying
water routinely on swine to reduce body temperature
also enhances the reduction of airborne contaminants,
but this may provide high moisture for microbial
growth and multiplication, as observed in the finishing and growing houses (Chang et al., 2001). Factors such as the flooring feature of pens, stocking density, characteristic of pig activity and cleaning
practices also account for the diverse levels of airborne contaminants found in different piggeries of the
present study. Further studies are warranted to investigate the effects of the aforementioned factors on
interventions for reducing airborne contaminant levels and preventing workers from developing adverse
health effects.
Acknowledgements—The authors are indebted to the swine farmers for their participation in this study and to Mr ShouwLiang Cheng and Mr Shen Chen for technical assistance. This
study was funded by the Institute of Occupational Safety and
Health, Council of Labor Affairs, Executive Yuan of Taiwan,
Republic of China, IOSH84-H308.
REFERENCES
Attwood P, Versloot P, Heederik D, de Wit R, Boleij JSM.
Assessment of dust and endotoxin levels in the working
environment of Dutch pig farmers: a preliminary study. Ann
Occup Hyg 1986;30:201–8.
Attwood P, Brouwer R, Ruigewaard P, Versloot P, de Wit R,
Heederik D, Boleij JSM. A study of the relationship between
airborne contaminants and environmental factors in Dutch
swine confinement buildings. Am Ind Hyg Assoc J
1987;48:745–51.
Chang CW, Chung H, Huang CF, Su HJJ. Exposure of workers
to airborne microorganisms in open-style swine houses.
Appl. Envir. Microbiol. 2001;67(1):155–61.
Clark S, Rylander R, Larsson L. Airborne bacteria, endotoxin
and fungi in dust in poultry and swine confinement buildings.
Am Ind Hyg Assoc J 1983;44:537–41.
Cormier Y, Tremblay G, Meriaux A, Brochu G, Lavoie J. Airborne microbial contents in two types of swine confinement
buildings in Quebec. Am Ind Hyg Assoc J 1990;51:304–9.
Crook B, Robertson JF, Travers Glass SA, Botheroyd EM,
Lacey J, Topping MD. Airborne dust, ammonia, microorganisms, and antigens in pig confinement houses and the
respiratory health of exposed farm workers. Am Ind Hyg
Assoc J 1991;52:271–9.
Donham KJ, Gustafson KE. Human occupational hazards from
swine confinement. Ann Am Conf Gov Ind Hyg
1982;2:137–42.
465
Donham KJ, Popendorf W. Ambient levels of selected gases
inside swine confinement buildings. Am Ind Hyg Assoc J
1985;46:658–61.
Donham KJ, Rubino M, Thedell TD, Kammermeyer J. Potential health hazards to agricultural workers in swine confinement buildings. J Occup Med 1977;19:383–7.
Donham KJ, Popendorf W, Palmgren U, Larsson L. Characterization of dusts collected from swine confinement buildings. Am J Ind Med 1986;10:294–7.
Donham K, Haglind P, Peterson Y, Rylander R, Belin L.
Environmental and health studies of farm workers in Swedish swine confinement buildings. Br J Ind Med
1989;46:31–7.
Donham KJ, Reynolds SJ, Whitten P, Merchant JA, Burmeister
LF, Popendorf WJ. Respiratory dysfunction in swine production facility workers: dose–response relationships of
environmental exposures and pulmonary function. Am J Ind
Med 1995;27:405–18.
Essen SV, Donham K. Illness and injury in animal confinement
workers. Occup Med 1999;14:337–50.
Haglind P, Rylander R. Occupational exposure and lung function measurements among workers in swine confinement
buildings. J Occup Med 1987;29:904–7.
Heederik D, Brouwer R, Biersteker K, Boleij JSM. Relationship of airborne endotoxin and bacteria levels in pig farms
with the lung function and respiratory symptoms of farmers.
Int Arch Occup Envir Health 1991;62:595–601.
Holness DL, O’Blenis EL, Sas-Kortsak A, Pilger C, Nethercott
JR. Respiratory effects and dust exposure in hog confinement
farming. Am J Ind Med 1987;11:571–80.
Jacobs RR. Airborne endotoxins: an association with occupational lung disease. Appl Ind Hyg 1989;4:50–6.
Mackiewicz B. Study on exposure of pig farm workers to
bioaerosols, immunologic reactivity and health effects. Ann
Agric Envir Med 1998;5:169–75.
Milton DK, Gere RJ, Feldman HA, Greaves IA. Endotoxin
measurement: aerosol sampling and application of a new
limulus method. Am Ind Hyg Assoc J 1990;51:331–7.
Olson DK, Bark SM. Health hazards affecting the animal confinement farm worker. Am Assoc Occup Health Nur J
1996;44:198–204.
Preller L, Heederik D, Boleij JSM, Vogelzang PFJ, Tielen
MJM. Lung function and chronic respiratory symptoms of
pig farmers: focus on exposure to endotoxins and ammonia
and use of disinfectants. Occup Envir Med 1995;52:654–60.
Reynolds SJ, Donham KJ, Whitten P, Merchant JA, Burmeister
LF, Popendorf WJ. Longitudinal evaluation of dose–
response relationships for environmental exposures and pulmonary function in swine production workers. Am J Ind Med
1996;29:33–40.
Vogelzang PFJ, van der Gulden JWJ, Preller L, Tielen MJM,
Schayck CP, Folgering H. Bronchial hyperresponsiveness
and exposure in pig farmers. Int Arch Occup Envir Health
1997;70:327–33.
Vogelzang PFJ, van der Gulden JWJ, Folgering H, Kolk JJ,
Heederik D, Preller L, Tielen MJM, Schayck CP. Endotoxin
exposure as a major determinant of lung function decline in
pig farmers. Am J Resp Crit Care Med 1998;157:15–8.
Walters M, Milton DK, Larsson L, Ferd T. Airborne environmental endotoxin:a cross-validation of sampling and analysis
techniques. Appl Envir Microbiol 1994;60:996–1005.
Zejda JE, Barber E, Dosman JA, Olenchock SA, McDuffie HH,
Rhodes C, Hurst T. Respiratory health status in swine producers relateds to endotoxin exposure in the presence of low
dust levels. J Occup Med 1994;36:49–56.
Zuskin E, Zagar Z, Schachter EN, Mustajbegovic J, Kern J.
Respiratory symptoms and ventilatory capacity in swine confinement workers. Br J Ind Med 1992;49:435–40.