Journal of Food Protection, Vol. 69. No. 7, 2006, Pages 1717- / 719
copyright t International Association for Food Protection
Quantity and Distribution of Salmonella Recovered from Three
Swine Lairage Pens
A. M. OCONNOR,!* J. GAILEY,1 J. D. McKEAN,' AND H. S. HURD2
'Department of Veterinar? y Diagnostic and Production Animal Medicine, College of Veterina, Medicine, Iowa State University, Ames, Iowa 500)0:
and 'National Animal Disease Center, U.S. Department of Agriculture, Agricultural Research Service. Ames, Iowa 50010, USA
MS 05-592: Received 30 November 2005/Accepted 20 February 2006
The quantity of Salmonella recoverable from three lairage pens in a swine abattoir was determined. Using dry tour-ply
cotton gauze pads measuring 10 by 10 cm. 100 fecal slurry samples were collected from each of the three pens. Salmonella
recovery was expressed as the log CFU per milliliter of sample. Mean values were 2.5 log CFU/ml in pen A, 2.7 log CFU/
ml in pen B, and 0.89 log CFU/ml in pen C. Median values were 2.6 log CFU/ml in pen A, 2.0 log CFU/ml in pen B, and
below the detectable limit in pen C. In pen C. Salmonella was not recoverable from a high number of samples. Pen B results
suggested spatial dependency, i.e., samples close together were more similar than samples farther apart. These results indicate
that Salmonella concentrations vary within and between lairage pens. Because of the limited number of pens assessed, it was
not possible to identify factors that were associated with the observed variation in Salmonella concentrations within and
between pens. However, this variation suggests that numerous samples are required to adequately describe the concentration
of Salmonella in a lairage pen.
Several studies have documented that recovery of Sal-
monella from swine gut contents and gut-associated lymph
nodes is higher in penmates slaughtered at abattoirs than in
penmates slaughtered on the farm (5, 6). Therefore, it was
hypothesized that lairage pens expose swine to new non-
farm Salmonella serotypes. The consequence of this expo-
sure is thought to be rapid infection of swine with these
nonfarm serotypes of Salmonella and carriage of these se-
rotypes into the abattoir, accounting for the postharvest
pathogen recovery (7). Other studies have been focused on
the prevalence of Sa/mone//a-positive samples obtained
from barn floors, but rarely has the concentration of Sal-
monella been reported (1-3).
In this study, the concentration of Salmonella recov-
ered from the floor in three lairage pens from swine abat-
toirs was determined. The rationale for quantifying Sal-
inonella on the abattoir lairage floor was to determine var-
iability within and between pens. Our motivation for de-
termining whether Salmonella populations were variable
among pens was that if such variability exists then the prob-
ability of rapid swine infection may differ across pens.
However, if pen levels of Salmonella are homogeneous. dif-
ferences in Salmonella infection rates could not be attrib-
utable to differences in contamination levels among lairage
* Author for correspondence. Tel: 515-294-5012: Fax: 515-294-1072:
Present address: Utah Army National Guard. 348 East Main Street. Lehi,
UT 84043, USA.
Present address: Department of Veterinary Diagnostic and Production
Animal Medicine. College of Veterinary Medicine. Iowa State Univer-
sity. Ames, IA 50010. LISA.
pens. Although the issue of rapid infection with Salmonella
provided the motivation for this study, factors associated
with variations in Salmnonella contamination among pens
and the consequences of this variation were not addressed.
MATERIALS AND METHODS
Sample collection. Fecal slurry samples were collected from
three lairage pens at the same abattoir. At each visit, a lairage pen
was free of pigs between 7 am. and 10 am. Pigs had been in the
pen prior to that time (slatighter begins at 5 am.), and the pens
were hosed clean with water at 5 am. At each visit, a different
pen was evaluated; therefore, the sampled pens differed in size,
although all were in continuous use during the day. All the pens
were designed to hold market groups of 180 pigs. At each visit.
100 samples were collected on a sampling grid. For pens A and
B. a large-scale grid with points 3.25 m apart was laid over the
entire pen. A medium-scale grid with points 1 iii apart was also
placed in a randomly chosen location within the pens. Two small-
scale grids with points 0.5 m apart were placed in areas considered
likely to have higher concentrations of Salmonella, i.e., watering
areas and entrances. For pen C, a simple systematic grid was used
to identify the locations for sampling the floor. The rationale be-
hind changing the sampling grid was to examine how an alter-
native sample grid would describe a pen.
A dry four-ply cotton gauze pad measuring 10 by 10 cm was
placed at each intersection of the grid lines. After all 100 pads
were laid on the grid, the pads were collected (in the order laid
down) with sterilized forceps into Whirl-Pak bags marked to in-
dicate their location in the lairage pen, and the bags were trans-
ported within 2 It to the laboratory for processing.
Sample processing. Upon arrival at the laboratory, the
weight of liquid absorbed by the pad was determined, and 10 ml
O'CONNOR ET \I?
\ ol (). \?
Log CFU of Salmonella per ml of sample
Log CFU of Salmonella per ml of sample
Log CFU of Salmonella per ml of sample
FIGURE 1. Frequency distribution histograms of the log CFU of
Salmonella per milliliter of fluid collected from the three study
of phosphate-buffered saline (PBS) was added to each bag. Sam-
ples not processed were immediately stored at 4°C until process-
ing. All the samples were processed on the day of collection. After
addition of PBS, the pads were homogenized with a stomacher
(Seward Ltd., London, UK) at 260 rpm for 1 mm. For enumera-
tion of Salmonella, the most-probable-number (MPN) technique
was used with five 1:10 serial dilutions of the each sample. Bo-
rosilicate test tubes (15 by 150 mm) containing 4.5 ml of buffered
TABLE 1. Descriptive statistics of the most probable number of
Salmonella collected from three lairage pens at a commercial
Salmonella concn (log MPN/ml)"
Pen B? Pen C
° BDL, below detectable limit.
peptone water (Becton Dickinson, Sparks, Md.) were inoculated
with 0.5 ml of sample. The tubes were covered and incubated at
37°C for 24 h, and then 0.1 ml of the culture was transferred to
10 ml of Rappaport-Vassiliadis medium (Becton Dickinson) and
incubated at 42°C for 24 h. A 3-mm loopful of this culture was
then streaked onto XLT agar (Becton Dickinson) and incubated at
37°C for 24 h. Colonies on each plate were then enumerated using
the MPN technique (8), and Salmonella results are reported as the
MPN of log CFU per milliliter of pen liquid collected.
Statistical analysis. For each pen, mean, median, standard
deviation, and range of the MPN of log CFU per milliliter of
liquid was calculated. Because the formulae for the mean, median,
and standard deviation assume sample independence, we tested
this assumption by looking for spatial dependency within the pens.
PROC VARIOGRAM (version 8.2, SAS Institute, Cary,
NC.) and the spatial module in SPLUS 6 were used to compute
a directional and nondirectional empirical and theoretical vario-
gram for each pen. When there was no evidence of anisotrophy
(directionality), then the variogram was calculated using the entire
data set assuming isotrophy, i.e., no directionality in the data.
Theoretical variograms were calculated using either the exponen-
tial or the linear function of the shape of the empirical variogram.
The formulae and forms for these theoretical models were de-
scribed by Cressie (4). Evidence of strata in pens A and B were
examined by using median Polish kriging (4).
Figure 1 shows the frequency distribution histograms of
the log CFU of Salmonella per milliliter of fluid collected for
each of the three pens. Descriptive statistics for the three pens,
assuming independence, are listed in Table 1.
There was no evidence of directionality in the spatial
dependency for any of the pens. The predicted variograms
for pens A and C suggest little or no correlation between
Salmonella concentrations at various locations in the pens,
i.e., within a pen, the Salmonella concentration at one lo-
cation did not allow prediction of the concentration at an-
other location. In pen B, gamma increased distinctly until
becoming constant at 6 to 8 m, meaning that Salmonella
concentrations were correlated with locations up to 6 to 8
in and the relationship was stronger for closer loca-
tions. The parameter estimates for the pen A theoretical
variogram using a linear model were slope = 0.006 and
nugget = 0.85. For pen B, the most appropriate model was
an exponential model with range = 3.77, sill = 1.18, and
L id I'rL. \I ().?
nugget = 0.69. For pen C, a linear model was most appro-
priate, with slope = 0.02 and nugget = 1.04. There was
no evidence of strata in pens A and B as determined with
median Polish kriging (results not shown).
We had anticipated that the movement of hundreds of
pigs through barns over many days would create a high but
steady concentration of Salmonella in pens. However, in
the three lairage pens we examined the Salmonella concen-
trations were variable. This study was not designed to ex-
amine the numerous possible explanations for this variation,
which include but are not limited to differences in the Sal-
monella status of the pigs previously in the pen, the hu-
midity of the pens, use patterns, and differences in the
clearing capability, which could be related to the slopes and
drainage of the pens. Whatever the source of variation, the
consequence of this variation is that in research studies ex-
amining factors associated with Salmonella in swine, a sin-
gle or small number of pen samples is not likely a good
descriptor of the pen.
Only Salmonella recoverable from the floor by our col-
lection method, i.e., dry pads, were cultured. It is not clear
whether this technique accurately reflects the bacterial load
picked up by swine in lairage pens. Therefore, the concen-
trations of Salmonella we observed may be confounded by
the amount of water present on the lairage pen floor. Dif-
ferences in dampness were noted among pens. In particular,
pen C was drier than the other pens, which may have made
it more difficult to recover Salmonella from that pen. How-
ever, such limited sample size makes it inappropriate to
attribute the variation in Salmonella to pen dampness.
Another potential complication in our study may be
measurement error. Serial dilutions were used to quantify
the recovered Salmonella cells. As with any enumeration
method, measurement error is possible. However, in the ab-
sence of a "gold standard" for quantifying Salmonella in
environmental samples, it is not possible to make an ac-
curate determination of measurement error. Except for the
exact measurement of the number of CFU, the quantifica-
IL IlU\LLL4 RECOVERABLE FROM LAIRAGE PENS?
tion of Salmonella in pens should be compared within a
laboratory using the same methodology. Comparisons of
quantities across laboratories probably are not valid.
This project was supported by funds from the U.S. Department of
Agriculture Cooperative State Research Education and Extension Service,
National Research Initiative grant no. 2001-35212-10864. The authors also
acknowledge assistance from Dr. M. Kaiser (Department of Statistics, Col-
lege of Agriculture, Iowa State University) for his assistance with statis-
tical aspects of this study.
1. Barber, D. A., R. M. Weigel, R. E. Issaacson, P. B. Bahnson, and C.
J. Jones. 1999. Spatial and temporal patterns of the distribution of
Salmonella on swine farms in Illinois, p. 97-100. In Proceedings of
the 3rd International Symposium on the Epidemiology and Control of
Salmonella in Pork, Washington, D.C.
Beloeil, P. A., E. Eveno, P. Gerault, P. Fravalo, V. Rose, N. Rose, and
F Madec. 1999. An exploratory study about contamination of pens
of finishing pigs by ubiquitous Salmonella, p. 101-105. In Proceed-
ings of the 3rd International Symposium on the Epidemiology and
Control of Salmonella in Pork, Washington, D.C.
Carlson, A. R., and T. H. Blaha. 1999. Investigations into the infec-
tion-contamination-infection cycle of zoonotic Salmonella on swine
farms—investigations into the occurrence of Salmonella in the envi-
ronment of four selected Minnesota swine farms, p. 109-112. In Pro-
ceedings of the 3rd International Symposium on the Epidemiology
and Control of Salmonella in Pork, Washington, D.C.
Cressie, N. A. C. 1993. Statistics for spatial data. John Wiley & Sons,
Hurd, H. S., J. K. Gailey, J. D. McKean, and M. H. Rostagno. 2001.
Rapid infection in market-weight swine following exposure to a Sal-
monella typhimurium–contaminated environment. Am. J. Vet. Res. 62:
Hurd, H. S., J. D. McKean, R. W. Griffith, I. V. Wesley, and M. H.
Rostagno. 2002. Salmonella enterica infections in market swine with
and without transport and holding. AppI. Environ. Microbial. 68:
Rostagno, M., S. Hurd, J. McKean, C. Ziemer, J. Gailey, and R. Leite.
2001. Abattoir holding pens as a source of Salmonella for swine, p.
298-300. In Proceedings of the 4th International Symposium on the
Epidemiology and Control of Salmonella and Other Food-Borne Path-
ogens in Pork, Leipzig, Germany.
Wood, R. L., and R. Rose. 1992. Populations of Salmonella typhi-
murium in internal organs of experimentally infected carrier swine.
Am. J. Vet. Res. 53:653-658.