Title:
Determination of Salmonella Classification Levels for Selected Midwestern
Swine Herds Utilizing Abattoir-based Samples and Compare Classification
Stability over Time – NPB# 04-201
Investigator:
James D. McKean
Institution:
Iowa State University
Co-Investigator:
Annette O’Connor
Date Received:
March 21, 2006
Abstract: This study was designed to compare results of a previous NPB-supported
serologic survey of Midwestern market swine for Salmonella antibodies, and to expand
information about Salmonella prevalence in large, multi-site populations relative to
traditional production systems. For 2002 and 2004 respectively a total of 14,401 and
13,718 samples were collected from eight Iowa abattoirs. In 2002, 934 of the14,401
(6.4%), and in 2004, 639 of 13,718 (11.9%) samples were positive for Salmonella
antibodies. The average prevalence of Salmonella for low volume producers in 2002
was 6.8% (5.4-7.3), the median prevalence was 0.8% and the within producer
prevalence range was 0-59.2%. In 2004, the average Salmonella prevalence for small
volume producers was 11.8% (10.3-13.3), the median prevalence was 5.2% and the
range was 0-81.8%. These variations are not unexpected, but are not explained by the
data available. For large herds the percentage of herds classified at the various levels
was the same for both periods. The higher cumulative percentages for Level 2 and 3
classifications (29 vs. 28%) within the population relative to the smaller herds was
maintained even though the comparable smaller herds’ classification doubled (9 vs.
20%) in 2004 and the large herds sampled increased from 45 to 87. This classification
differential for both periods buttresses that these populations have different Salmonella
statuses.
Introduction:
Since 1995 the Danish swine industry has conducted an on-farm Salmonella
control program to categorize production sites as to their risks for Salmonella
contamination. This national program has undergone several changes since initiation,
but the concept of production-site risk categorization has remained a constant. Similar
programs are being implemented in other European Union countries. Danish herd
classifications for Levels I, II, and III have remained static at approximately 95%, 3.3%,
and 1.6%, respectively (Nielsen et al., 2001) for the past several years. This
classification has been used to schedule transport and to harvest herds of similar
statuses to reduce antemortem Salmonella cross contamination. It was reported that
94.4% of sero-negative Level I herds supplied animals with negative cecal cultures for
Salmonella under this controlled harvest and transport production system. No
systematic classification of USA production sites has been attempted. This paucity of
information hampers the ability to conduct farm to abattoir risk assessments, and to
determine the feasibility of on-farm control programs as a method to reduce carcass
Salmonella contamination.
In an earlier NPB/IPPA funded project we examined ~1150 swine herds
marketing through eight Iowa abattoirs in 2002 as part of a Psuedorabies (PRV)
monitoring project. This study represented the first attempt to classify Midwestern
production sites for Salmonella spp. sero-prevalence. Ignoring herd size, 90.2% of
surveyed herds were negative or Level I, 8.2% were Level II herds, and 1.6% Level III.
These results are similar to previous Danish studies (Alban et al., 2002; Mousing et al.,
1997). Studies reported by Fong and Hald (2000) suggested that herd size was
statistically associated, albeit weakly, with Salmonella sero-prevalence, but the authors
concluded the relationship was probably not biologically significant. Our earlier data
suggested that larger herds tended to have a higher sero-prevalence and more
variations between lots marketed than smaller units. In addition there appeared to be a
trend that smaller units remained within a single classification over the 12 weeks of the
initial study, indicating the potential for stability within Salmonella classifications.
The current study has been designed to compare a cohort of the
previously classified herds and a larger sample (~100 systems) of known multi-site
production systems. The large herd cohort estimated to range from over 1.5 million
head to 24,500 head marketed annually. The smaller herds were selected based on
sales of at least two lots (sub-truckload) per month.
Objectives:
1) Using the Danish salmonella herd classification scheme, classify selected
Midwestern swine herds for Salmonella status;
2) Compare the current Salmonella classification status with similar
determinations made in 2002 on the selected herds; and
3) Compare classification of ~100 largest (multi-site production) systems with
the remainder of the selected population (predominately single sites).
Materials and Methods:
Sample herds were all available production systems classified in the 2002 seroprevalence study (funded by NPB/IPPA) from eight Iowa abattoirs. All classification
results from the prior study were available within an existing database by producer
identifier for comparison purposes. All previously classified herds with sufficient volume
for analysis (minimum of 2 lots/month) were flagged for sampling. In addition the
producers with production < 20,000 as estimated from annual sales to the eight
cooperating abattoirs were flagged. Each week for a 3 month period the database of
current meat juice samples submitted for PRV analysis was queried and compared with
the two lists of flagged producers. Two samples per lot (< 8 samples/producer/week)
2
were selected, individually identified to the herd and collection date, and stored in a
commercial freezer for analysis.
After processing all identified samples were submitted to the ISU veterinary
Diagnostic Laboratory for analysis using the IDEXX Herd-chek Salmonella ELISA, the
same procedure used in 2002 for initial herd classifications. Herds were segmented
into those herds that remained in the initial (2002) classification and those that changed
classifications to a lower or higher level. In addition the within period prevalence of
positive samples was compared across periods for each herd to determine variability
within herds. All results were statistically analyzed to evaluate the changes in status
and to determine whether a difference exists between large and small herds for
prevalence variability.
Data analyses of the producer paired and unpaired subsets differentiated into
smaller (< 20,000) and larger (> 20,000) producers were completed. Estimated
production was extrapolated from the total number of animals submitted during a three
month period in 2002.After removing herds that did not have at least 12 samples in both
2002 and 2004 sampling periods, 502 smaller producers were available for analysis.
These 502 herds formed the representative cohort for comparisons across sampling
periods. In the larger production group 87 producers were identified for 2004. They
were analyzed as a cohort rather than divided by level of annualized production.
Each producer name was classified in 2002 and 2004 based on cumulative
antibody prevalence. The levels are comparable to classifications used by the Danish
system – Levels 0, 1, 2 and 3 which have been described by Mousing et al. (1997) and
Alban (2002). Level 0 is no or negligible evidence of Salmonella exposure. Level 1 was
defined as those herds with low levels (~1-25%), Level 2 with moderate (~25-50% and
Level 3 widespread (>50%) positive carcasses per producer. Comparisons of status
changes were made for each paired producer.
Results:
A total of 1,088 herds were surveyed in 2002 and 919 herds were available in
2004, with 794 identified present in both periods. In addition to the 794 producers
represented in both 2002 and 2004 data 294 smaller producers who met the sampling
algorithms in 2002 but were not part of the 2004 sample subset are not reported. The
502 producers analyzed presented the requisite minimum of two lots/month for three
month 2004 test period (> 12 samples).
For 2002 and 2004 respectively a total of 14,401 and 13,718 samples were
collected. In 2002, 934 of the14,401 (6.4%) samples were positive for Salmonella
antibodies. The average prevalence of Salmonella for the producers was 6.8% (5.47.3), the median prevalence was 0.8% and the within producer prevalence range was 059.2%. In 2004, 1,639 of 13,718 (11.9%) samples were positive for Salmonella
antibodies. In 2004, the average prevalence of Salmonella for the producers was 11.8%
(10.3-13.3), the median prevalence was 5.2% and the range was 0-81.8%.
Table 1 shows the average prevalence of Salmonella based on the estimated
annual slaughter for each producer, aggregated by herd size. In Table 2 the percentage
of herds in the Salmonella classification levels in 2002 and 2004 are presented. Table 3
demonstrates the changes in classification for producers between 2002 and 2004. For
3
example, 48% of producers showed no difference in the category received in 2004
compared to the 2002 estimate whereas the remaining 52% changed classifications
with 37.3% rising to higher levels and 14.9%. The cross tabulation in Table 4
demonstrates the break down of the changes by the Salmonella classification levels
between sampling periods for individual herds. Table 5 contains the prevalence
distribution summary of 87 high volume herds surveyed in 2004 compared to the 45
herds collected in 2002.
Discussion:
The sampling algorithms per lot changed for 2002 and 2004 because of
modifications in the PRV monitoring program. Rather than sampling 4 carcasses/lot as
in 2002 the sampling rate was 2/lot in 2004. These differences were compensated by
more sustained sampling efforts in 2004. Unfortunately these efforts did not capture all
eligible producers which resulted in the loss of 417 producers that did not meet the
minimum sampling standard of 12 samples/producer for the 12 week period. Many of
these producers would have been expected to populate the smaller herd size
classifications but because of the combined effects of their frequency of lot presentation
and lower sampling rate per lot are not fully represented.
The comparison of 2002 and 2004 herd Salmonella prevalence data was an
objective of this study. It demonstrates a higher prevalence rate for samples collected in
2004 compared to 2002. The data suggest that the prevalence of Salmonella was not
stable over long periods of time among producers and that the prevalence of
Salmonella exposure in Midwestern swine differed between the two surveys. The
results highlight that classifications assigned to producers in 2002 were not likely to be
the same as in 2004. This finding has implications for Salmonella control schemes that
may use classification to impose marketing restriction on herds based upon the
producer level prevalence.
The Danish classification scheme utilizes these within herd prevalence values to
place herds in appropriate Level designations. Even with a general (~2X) rise in
prevalence within the sampled population for 2004, the 502 herds compared by size
demonstrated a significant rise of only within herd mean prevalence for herd sizes of
501-1000, 1001-2000, and 3001-5000. A similar effect was not demonstrated in the
larger herds’ comparisons. Although 47.8% of the low volume herds did not change
Level status those herds that did change disproportionately moved to higher
classifications. An explanation of these variations within the total population, although
not surprising given the Danish experiences, is not readily apparent. The absence of
more than two observation points makes evaluation of these changes problematic.
Additional sampling under similar configurations would demonstrate whether this
increase is sustainable or represents a data aberration.
The comparison of a larger population of herds (> 20,000), which are likely to be
multiple sited production, to determine whether the higher Level 2 and 3 prevalences
observed in the 2002 data was sustainable was the third objective of this study. The
percentage of herds classified at the various levels was the same for both periods.
Cumulative percentages for Level 2 and 3 classifications (28 vs. 25%) within the
population relative to the smaller herds was maintained even though the smaller herds’
classification for Level 2 and 3 doubled (9 vs. 20%) in 2004 and the number of large
herds sampled increased from 45 to 87. This differential in Levels 2 and 3
4
classifications for both sampling periods buttresses the position that these two
populations are different in their Salmonella statuses. Questions immediately arise as to
the different facility or management practices may be impacting these populations.
Further examinations for these relationships may be useful in determining potential onfarm interventions to reduce Salmonella prevalence within herds.
Lay Interpretation:
Since 1995, the Danish swine industry has conducted an on-farm Salmonella
control program to categorize production sites as to their risks for Salmonella
contamination. This classification has been used to schedule transport and to harvest
herds of similar statuses to reduce antemortem Salmonella cross contamination. Danish
herd classifications for Levels I, II, and III have remained relatively static at
approximately 95%, 3.3%, and 1.6%, respectively (Nielsen et al., 2001). In 2002 similar
values (91%, 8%, and 1%) were obtained from a large Midwestern USA sample. In
2004 a similar sampling scheme compared 502 herds from 2002 and 2004 and found
classifications for Levels I, II, and III to be 79%, 14, and 6%, respectively. This increase
was also reflected in an approximately 2X increase in total samples testing positive for
Salmonella antibodies. Analysis of these data suggests that the prevalence of
Salmonella within production units was not temporally stable and that the prevalence of
Salmonella exposure in Midwestern swine differed between these surveys for small
volume but not large producers. The results highlight that classifications assigned to
producers in 2002 were not likely to be the same in 2004. These findings have
implications for Salmonella control schemes that use on-farm classification schemes to
impose marketing restriction on herds based upon antibody prevalence levels.
References:
D. Lo Fo Wong , T. Hald. Chapter VI. (2000) Surveillance and Control Options Based on
Epidemiology of Salmonella in Pork and Humans.
Mousing J, Jensen PT, Halgaard C, et al. (1997) Nation-wide Salmonella enterica
surveillance and control in Danish slaughter swine herds. Prev Vet Med;29:247261.
Alban L, Stege H, Dahl J. (2002) The new classification system for slaughter-pig herds
in the Danish Salmonella surveillance-and-control program.
Prev Vet
Med;53:133-146.
Nielsen, B., L. Alban, S. Helle, L. L. Sorensen, V. Migelmose, J. Bagger, J. Dahl, D. L.
Baggesen (2001): A new Salmonella surveillance and control program in Danish
pig herds and slaughter houses. Proc. 4th Int’l Symposium on the Epidemiology
and Control of Salmonella and Other Food-borne Pathogens in Pork, Leipzig,
Germany: 14-21.
5
Table 1: The distribution of the within herd prevalence to Salmonella antibodies by
annualized production estimates for 502 low volume producers in 2002 and 2004.
Herd size
Frequency
Mean (95% CLM)
Median
Range
Year 2002
< 200
2
.
0
.
201-500
56
4.5(1.8-7.3)
0
0
501-1000
159
6.1 (4.4-7.8)
0
59.2
1001-2000
168
6.8(5.2-8.4)
2.8
53.3
2001-3000
66
8.9(5.8-12.1)
3.5
54.5
3001-5000
48
4.6(2.6-6.6)
2.6
37.0
>5000
3
1.5
2.3
2.4
Year 2004
< 200
2
7.1
201-500
56
8.7(4.2-13.3)
0
68.7
501-1000
159
11.2(8.5-13.8)
4.2
69.2
1001-2000
168
11.8(9.2-14.3)
6.2
80
2001-3000
66
15.1 (10.9-19.1)
9.2
70.5
3001-5000
48
13.3 (7.4-19.7)
4.5
81.8
>5000
3
19.7
22.8
25.0
6
Table 2
The frequency distribution of categories evidence of past exposure to Salmonella in 502
low volume producers that marketed swine in Iowa in 2002 and 2004. Number of herds
(percent of row)
Year of collection
Negative
Level 1
Level 2
Level 3
# of herds
2002
2004
251
(50%)
172
(34%)
207(41%)
40(8%)
4(1%)
502
230(45%)
71(14%)
29(6%)
502
7
Table 3. The change in Salmonella level in 2004 compared to the level observed in
2002 for 502 low volume producers marketing swine in Iowa in 2002 and 2004.
Change in classification level from 2004 to 2002
-2
-1
0
1
2
3
8 (1.6%)
67(13.3%)
240(47.8%)
138 (27.5%)
42 (8.4%)
7 (1.4%)
* A producer in Level 1 in 2002 and Level 3 in 2004 would be described by a + 2 (3 -1 =
+2) level change, as would producer in Level 0 in 2002 and Level 2 in 2004 (2 - 0 =2). A
producer in Level 3 in 2002 and Level 2 in 2004 would be described as a negative one
(2 - 3=-1) level change, as would a producer in Level 1 in 2002 and Level 0 in 2004 (0 1= -1).
8
Table 4: The comparative prevalence distribution for low volume producers Salmonella
classifications in 2004 and 2002. (Red indicates same classification both periods)
Year
2002
Negative
Level 1
Level 2
Level 3
Total
2004
Negative
Level 1
117
100
(46.6%)
(39.8%)
49 (23.7%) 112
(54.1%)
6 (15.0%)
16 (40.0%)
0
2 (50%)
172
230
Level 2
27 (10.8%)
Level 3
7 (2.8%)
Total
251
31 (14.9%)
15 (7.2%)
207
11 (27.5%)
2 (50%)
71
7 (17.5%)
0
29
40
4
502
Table 5
Prevalence frequency distribution for high volume herds for 2002 and 2004
Negative Level 1
Level 2
Level 3
No
herds
High
volume 5 (11%)
27(60%)
12 (27%) 1 (2%)
45
†
population 2002
High
volume 3 (3%)
59 (68%) 22 (25%) 3 (3%)
87
†
population 2004
9
of
Appendix
The sero-prevalence cut-off points to categorize herds for past exposure to Salmonella.
Estimated
Negative
Level 1
Level 2
Level 3
Annual
Harvest
< 200
0
> 0- 50
NA
>50
201-500
0
>0-25
>25-50
>50
501-1000
0
>0-23
>23-50
>50
1001-2000
0
>0-20
>20-50
>50
2001-3000
0
>0-17
>17-50
>50
3001-5000
0
>0-17
>17-50
>50
>5000*
0
>0-17
>17-33
>33
* A herd with a prevalence >33% positive samples it was assigned to level III. Mousing
et al. (1997)
10