Background

For the beef industry to be able to respond appropriately to the public health risks posed by zoonotic agents, it’s necessary to first define the nature of the risk. For the specific case of hide contamination with MDR Salmonella, this definition should include identification of the specific classes of slaughter cattle most likely to carry MDR Salmonella and to determine whether this risk is equally distributed by season and by geographic region. 

The stated objectives for this work were: 

1. Compare the prevalence of Salmonella hide contamination on cull dairy cattle, cull beef cattle and fed cattle.
2. Determine whether the PNW differs from the Eastern / Central / Southern regions of the US in the frequency of hide contamination with Salmonella by comparing WSU results with collaborators in Nebraska and Texas. 
3. Compare the prevalence of hide contamination with Salmonella before and after pre-slaughter washes with hyper-chlorinated water.

Methodology

Laboratory procedures
Samples were cultured for Salmonella. Hide swabs and diluent were agitated by laboratory stomacher for 60 seconds and 300 microliters of diluent was removed and spread plated on XLT-4 and MAC-ACS plates to detect and enumerate Salmonella in highly contaminated hide swabs. An additional 1 ml of the diluent was added to Rappaport-Vassiliadis (RV) and Tetrathionate (TT) selective enrichment broth tubes and incubated (42oC, overnight). Selective enrichment broths were each streaked onto separate XLT-4 selective differential agar plates and incubated (35oC, overnight). Colonies with typical Salmonella morphology were confirmed by biochemical and immunological tests (LIA, TSI, Urea tubes and serogrouping agglutination). Serotypes are being determined by the NVSL laboratory at Ames IA. 

All isolates were subjected to Kirby-Bauer agar diffusion antimicrobial susceptibility testing in strict accordance with the CLSI (NCCLS) guidelines. Antimicrobial disks used contained ampicillin, amoxicillin/clavulanic acid, ceftazidime, ceftazidime/clavulanic acid, chloramphenicol, gentamicin, kanamycin, streptomycin, sulphonamide, sulfa/trimethoprim, tetracycline, and nalidixic acid. 

Interlaboratory comparison
Sixty hide swabs were obtained from the perineal hide of cull dairy cattle for use in an inter-laboratory comparison involving all the laboratories in the study at each quarterly sampling. The hide swabs were diluted by the addition of 30 ml BPW then agitated by laboratory stomacher for 30 seconds. Three 10 ml aliquots were then removed and placed in sterile 15 ml. screw top tubes with identical coded labels and express shipped for Salmonella assay to the three laboratories involved in the study (including our own). An insulated, non-chilled FedEx parcel was prepared for each laboratory to ship these 60 coded aliquots; parcels were shipped within 72 hours of sampling and so were received at each of the three laboratories within 96 hours of sampling. Since all FedEx parcels are handled at a single national processing point, this strategy ensured that each parcel received very similar handling. These samples were assayed at each laboratory according to that laboratory’s routine bacteriologic culture protocol for isolation and identification of Salmonella from hides, and the results from each coded specimen tube and all isolates were forwarded to WSU for analysis of comparative assay results. All isolates resulting from the inter-laboratory comparison were confirmed as Salmonella. Where multiple isolates were obtained from single specimens, all were serogrouped to determine if diverse isolates resulted from the multiple cultures.

Findings

Objective 1: Compare the prevalence of MDR Salmonella hide contamination on cull dairy cattle, cull beef cattle and fed cattle. 

Inter-laboratory culture sensitivity comparison
The first consideration in this objective was a comparison of the sensitivity of the culture methods in place in the three collaborating laboratories using quarterly samples sub-divided to each laboratory. The first quarterly sample set was obtained from a fed cattle packing plant in Washington State. The second quarterly sample set was obtained by the MARC collaborating group from a packing plant in the state of California. The third quarterly sample set was obtained at a cull cattle packing plant in Washington State. The numbers of samples detected containing Salmonella by each laboratory at each sampling period are presented in Table 1. The numbers of positive specimens detected by the three laboratories did not differ statistically (Chi-square= 3.033 with 2 degrees of freedom, P = 0.220). However, there was a clear tendency for more sensitive Salmonella detection by the MARC laboratory, which uses a sophisticated IMS enrichment protocol not used by the other laboratories. The results obtained by the WSU and Texas laboratories were similar, not surprising in light of the very similar conventional selective enrichment, selective differential plating methods they use. 

The concordance of the three laboratories in Salmonella detection from individual specimens is presented in Table 2. Despite the differing methodologies and likely sensitivity differences, concordances from individual sample sets were quite high, ranging from 81 – 100% and the overall concordances over the study to date ranged from 90 – 96%. 

The second inter-laboratory comparison set provided a useful look into the diversity of Salmonella that were present in the single hide swab specimens taken from a heavily contaminated group of cattle. Two specimens tested Salmonella negative by all laboratories. Two, 4 and 55 specimens, respectively, were found to be Salmonella positive by 1, 2 or all 3 labs. Because the labs used more than one selective enrichment broth per specimen, the 65 specimens generated a total of 255 Salmonella isolates. The Texas lab isolates were not available for serogroup and resistance typing comparison, leaving 200 isolates available for this study. The serogroups and resistance types of these isolates are presented in Table 3. 

Twenty-eight specimens resulted in only a single Salmonella type (based on serogroup and resistance type. However, 23 specimens resulted in two Salmonella types and 12 specimens resulted in three different Salmonella types. Therefore, in these heavily contaminated specimens, multiple diverse Salmonella types were revealed by analysis of multiple colonies by two laboratories. 

Washington State hide contamination study
In the three seasonal sampling visits at the two Washington State packing plants sampled in this project, Salmonella was not frequently isolated (Table 4). There was no significant difference in rate of detected hide contamination between beef and dairy breed culls (p=0.55, Table 4). There was a highly significant difference in the rate of detected hide contamination between cull cattle and fed cattle (21/538 vs. 0/361, p<0.0001, Table 4).  

An additional Spring sampling visit was scheduled at the Washington State fed cattle packing plant after a remodeling project allowed access to kill floor hide-on carcasses. With this improved access, we were able to conduct brisket and belly swabbing using the MARC sampling protocol. However, these specimens also were uniformly negative for Salmonella contamination. 

Direct plating was used to enumerate highly contaminated specimens. However, none of the 878 specimens was positive by direct plating, estimated to detect contamination at approximately 50 cfu per swab or higher. This result is generally in agreement with the low prevalence in indicating the relative lack of Salmonella contamination in the Washington State cattle population.  

The prevalence of detected Salmonella hide contamination at the cull cattle plant declined seasonally from Fall through Winter to Spring sampling periods (Chi square = 15.324, 2 degrees of freedom, P < 0.001, Table 5, Chi square for trend = 14.43, 1 degree of freedom, p<0.001.) 

All Salmonella isolates from this study were tested for antimicrobial resistance. However, the low prevalence of detected Salmonella was not sufficient to draw conclusions about the occurrence of multiple drug resistance on cattle hide contaminants in our region. All 15 Salmonella isolates from the Fall sampling belonged to a single serotype (Newport) and all were multi drug-resistant. These isolates all belonged to a strain of Salmonella Newport that is currently epidemic and causing clinical disease among cattle populations in our region of the US, although the cattle sampled here were healthy. Similarly, the 5 Salmonella isolates from the Winter sampling also all shared a common serogroup (E1) and resistance pattern (susceptible to all drugs tested). With this low rate of contamination, it is possible that these contaminated cattle each resulted from contact with a single infected shedding animal or an environment contaminated by a previous single infected animal, either in transit or lairage. 

Objective 2: Determine whether the PNW differs from the Eastern / Central / Southern regions of the US in the frequency of hide contamination with Salmonella by comparing WSU results with those collaborators in Nebraska and Texas. 

The principal data to evaluate this aim will be forthcoming with the final reports from our collaborators in comparison with this report. However, the inter-laboratory methods comparison performed as part of this project provides strong preliminary data in support of the existence of a marked regional difference in hide Salmonella contamination with lower contamination rates in the Washington state plants compared to the other study locations. None of the three laboratories detected any Salmonella contamination in the first quarterly sample set, obtained from the fed cattle plant in Washington State. All of the three laboratories detected high prevalence of Salmonella contamination in the second quarterly sample set, obtained from a plant in California. All of the three laboratories detected a low Salmonella prevalence in the third quarterly sample set, obtained at a cull cow packing plant in Washington State. Therefore, all three laboratories were concordant in identification of low prevalence of hide contamination at two samplings in Washington State. Analysis of the complete results of hide contamination prevalence at the three study sites will increase the power of this comparison. 

Objective 3: Compare the prevalence of hide contamination with MDR Salmonella before and after pre-slaughter washes with hyper-chlorinated water. 

Salmonella was more frequently detected in hide swabs from pre-sanitized cattle than from cattle before sanitizing washes (Chi square = 4.494, 1 degree of freedom, P = 0.034, Table 6).

Implications

This project resulted in three significant findings: 1: hide contamination with Salmonella was relatively uncommon in Washington State packing plants compared to other regions, 2) hide contamination with Salmonella was more common on culled (older) dairy and beef breed cattle than on feedlot fed (younger) cattle, and 3) washing cattle prior to slaughter with chlorinated water did not reduce Salmonella contamination. Of these findings, the most interesting is the lower frequency of Salmonella contamination in Washington state study sites compared to other regions. Further research to understand the basis for this difference could lead to new approaches to reducing the potential of Salmonella contamination on beef products.