Pr oject Summar y Seasonal effects on E. coli O157:H7, multi drug-resistant Salmonella, and Listeria monocytogenes prevalence and E. coli O157:H7 and Salmonella load on hides and carcasses at cow/bull packing plants Principal Investigators: Dayna Brichta-Harhay, Terry Arthur, Mick Bosilevac, Michael Guerini, Norasak Kalchayanand, Steven Shackelford, Tommy Wheeler, and Mohammad Koohmaraie U.S. Department of Agriculture, Agricultural Research Service Study Completed May 2006 Funded by The Beef Checkoff 1
Seasonal effects on E. coli O157:H7, multi drug-resistant Salmonella, and Listeria monocytogenes prevalence and E. coli O157:H7 and Salmonella load on hides and carcasses at cow/bull packing plants Background Preventing the contamination of ground beef with pathogens such as E. coli O157:H7, that can lead to outbreaks of foodborne illness linked to consumption of hamburger, is a high priority for the meat industry as well as consumers. Meat from cow/bull processing plants makes up a significant proportion of ground beef. We have established that the prevalence of E. coli O157:H7 on the hides of fed cattle is high, that often pre-evisceration carcasses become contaminated with E. coli O157:H7. Yet little is known about the prevalence of E. coli O157:H7 on the hides and carcasses of cull cows and bulls. This experiment was designed to assess the prevalence and levels of E. coli O157:H7, Salmonella, Listeria, and aerobic plate count on the hides and carcasses of cull cows and bulls and to determine the multi drug resistance of Salmonella isolates. Salmonella have been found on the hides of fed beef cattle at higher rates than in feces. We found hide prevalence to vary from 27.7% in the winter to 91.6% and 97.7% in the summer and fall, respectively (Barkocy-Gallagher et al., 2003). Salmonella enterica serovars Typhimurium and Newport are two of the top three disease-causing Salmonella serovars. Multi drug-resistant (MDR) Salmonella, such as S. Typhimurium with the ACSSuT resistance pattern (resistant to ampicillin, chloramphenicol, streptomycin, sulfisoxazole, tetracycline), and S. Newport with the MDR-AmpC resistance pattern (resistant to at least 9 antibiotics including: amoxicillin/clavulonic acid, ampicillin, cephalothin, ceftiofur, cefoxitin, chloramphenicol, streptomycin, sulfamethoxazole, tetracycline and show decreased susceptibility to ceftriaxone) were two of the most common multi drug-resistant Salmonella phenotypes in 2002 (CDC, 2004). During 1997-2001, the number of laboratory-confirmed S. Newport infections reported to the CDC increased from 1,584 (5%) of 34,608 reported Salmonella infections to 3,152 (10%) of 31,607 (Anon., 2002). The increase in the number of S. Newport infections appears to be associated with the emergence and rapid dissemination of the multi drug-resistant strains of MDR-AmpC Newport (Anon., 2002). These strains are noted for their resistance to extended spectrum cephalosporins (ESCs) such as ceftriaxone or cephalothin. Recent studies indicate that plasmid mediated ESC resistance can be readily transferred between E. coli and Salmonella isolates (Poppe et al., 2005). Thus, the need to assess the levels of S. Typhimurium and S. Newport on hides and carcasses, and to characterize their antimicrobial resistance profiles, is evident. L. monocytogenes is commonly associated with human listeriosis. Listeriosis is a serious infection, affecting primarily pregnant women, newborns, and adults with weakened immune systems, that is caused by eating food contaminated with the bacteria and the overt form of the disease has a mortality rate greater than 25 percent. L. monocytogenes has 11 serovars, but just three (1/2a, 1/2b, and 4b) are responsible for over 98% of reported human listeriosis cases. Major outbreaks of Listeria have been caused by L. monocytogenes 4b. Listeria has been considered a problem associated with ready-to-eat (RTE) foods, but recently, the producers of RTE products have started to require their suppliers to certify raw materials are free of Listeria. Thus, there is a need to assess the prevalence of this pathogen on the hides and carcasses of animals at slaughter. The stated objectives for this work were: 1) Determine the prevalence of E. coli O157:H7, Salmonella, and Listeria and the levels of E. coli O157:H7, Salmonella and generic microbial contamination on the hides, preevisceration carcasses, and post-interventions carcasses of slaughter cows and bulls. 2
2) Determine the prevalence and multi drug resistance of Salmonella Newport and Salmonella Typhimurium isolated from hide, pre-evisceration carcasses, and postinterventions carcasses of slaughter cows and bulls. Methodology Sample collection Hide samples (1000 cm2 area), and pre-evisceration and post-intervention (hot-box) carcass samples (8000 cm2 area) were obtained by swabbing the surfaces with sterile sponges pre-wetted with 20 ml of buffered peptone water (BPW). The samples were collected from 4 cow/bull slaughter plants in 4 regions of the U.S. (north-east, mid-west, south-central and west-coast) over 4 seasons (Summer, Fall, Winter and Spring). Samples were collected over two days, 95 samples at each site per day, for a total of 9120 samples over the course of the study. Samples were transported at 2ºC in coolers and were received at USMARC within 24 hours of collection for analysis of the prevalence of E. coli O157:H7, Salmonella and Listeria. Samples were also analyzed for general aerobic plate count (APC) and for the level of contaminating E. coli 0157:H7 and Salmonella by the enumeration methods described below. Sample enrichment and pathogen isolation All samples were enriched with tryptic soy broth (TSB) at a 1:10 ratio, and incubated as per the standard USMARC method of 2 h at 25ºC and 6 h at 42ºC. Samples were held at 4ºC until further processed. E. coli O157:H7 and Salmonella were each isolated by separate specific IMS procedures. Suspect E. coli O157:H7 colonies were confirmed with latex agglutination and multiplex PCR for the presence of genes encoding O157, H7, stx1, stx2 and intimin (Hu, et. al., 1999). Suspect Salmonella were confirmed with inva PCR (Rahn et. al., 1992). Enrichment for Listeria was performed by inoculating 3 ml of Fraser broth with 0.1 ml of TSB enrichment and subsequent incubation at 35ºC for 48 h. Samples that were suspect for containing Listeria (black in color) were streaked to Listeria CHROMagar, and incubated at 37ºC for 24 h. Two suspect Listeria colonies (blue in color) and 3 suspect L. monocytogenes (blue colonies with opaque haloes) were picked and confirmed by PCR (Doumith et. al., 2005). Enumeration Hide E. coli O157:H7 were enumerated using a Spiral Plater on ntchromagar (DRG International) as per USMARC protocols. Hide Salmonella were enumerated by spiral plating onto XLDtnc medium (XLD medium with 4.6 ml/l tergitol, 15 mg/l novobiocin, and 5 mg/l cefesulodin). Presumptive Salmonella colonies (up to 10 per plate) were confirmed by latex antibody agglutination and/or PCR for the inva gene (Rahn et al., 1992). Carcass samples were enumerated for E. coli O157:H7, using IsoGrid membrane filters placed onto ntchromagar. Carcass samples were enumerated for Salmonella also using IsoGrid membranes placed onto XLDtnc plates. APC general bacterial load assessment The generic aerobic bacterial load of hide and pre-evisceration carcass samples was evaluated using a Bactometer (Biomiereux). The APC values for post-intervention carcass samples were calculated by plating dilutions on APC Petrifilm (3M) and were enumerated using a Petrifilm counter. Salmonella antimicrobial resistance profile analysis The antibiotic susceptibility profile of tetracycline-resistant Salmonella was determined using Sensititre antimicrobial susceptibility plates (CMV1AGNF) and the Sensititre automated broth micro-dilution system (TREK Diagnostic Systems, Cleveland, Ohio, USA). The plates were incubated at 37ºC for 16-18 h and the MIC for 15 antibiotics was determined using a Sensititre Auto Reader (TREK Diagnostics). To ensure that the TREK system and plates were working properly, appropriate reference control strains (ATCC strains: E. coli 25922, Enterococcus faecalis 29212, Pseudomonas aeruginosa 27853, Staphylococcus aureus 29213) were analyzed along side the test strains. 3
Findings Prevalence of E. coli O157:H7 on hides and carcasses was highly variable between plants and seasons, ranging from 5% to 80% and 2% to 60% respectively. The characteristic seasonal trend was not observed in this study, but that may have been due in part to an unusually warm winter experienced in many parts of the U.S. Prevalence of Salmonella was generally greater than 70% for hides in all plants and seasons. Prevalence on carcasses showed more variability and the overall lowest values were observed in the spring and the highest in the winter. Listeria prevalence on hides showed more regional variation, and was consistently more prevalent in the cooler regions (60% to 80% in all seasons), while prevalence was lower in the warmer regions in the summer and spring at 5% to 20%, but increased in the winter to 30% to 60%. L. monocytogenes prevalence was generally low in the summer and fall (less than 10% on hides and carcasses at all plants) but increased in the winter and spring to 5% to 40% on hides and 2% to 20% on carcasses. The L. monocytogenes isolates were tested to determine if they were serovars 1/2a, 1/2b, or 4b, the isolates noted as most frequently causing human disease. Serovar 1/2a was found to be the predominant hide serovar, followed by 1/2b and 4b. Interestingly, serovar 1/2c was not detected on hides in any plant and only sporadically on carcasses, yet it was detected on hot box carcasses in Plant A at 14.7% in the summer and 19% in the spring (bold text in Table 2), indicating a possible in-house source for this pathogen. In addition, pulsed-field gel electrophoresis analysis of 48 hot box L. monocytogenes 1/2c isolates from Plant A found that 22/23 summer isolates and 25/25 spring isolates had identical patterns, signifying they have the same source. This study entailed the collection of 9120 samples resulting in approximately 11,000 Salmonella isolates. A portion of these have been serogrouped or serotyped and tested for antimicrobial sensitivity. Midway through this study, we began screening Salmonella isolates for their resistance to tetracycline, as this is a common resistance for Salmonella and a good indicator of multi drug resistance status. Of 1110 isolates screened to date, 346 (31%) were resistant to tetracycline. Serogrouping of these isolates has shown the majority to be groups C or E/G, with nine hot box isolates in the group D category. Serotyping and antibiotic resistance screening will continue for the remaining isolates and updated results will be provided when available. Aerobic plate count (APC) data has been collected for 9120 hide, carcass and postintervention carcass samples. Enumeration data has been collected for 3040 hide samples and 1520 carcass samples, using the spiral plate count method (SPCM) and hydrophobic grid membrane filtration method (HGMF) described herein. A portion of the enumeration data for E. coli O157:H7 and Salmonella on hides and carcasses is summarized in Table 9. The enumeration data so far indicate that Salmonella levels on hides and carcasses are generally higher than those of E. coli O157:H7. Additionally, the percent of samples with pathogen load levels above the limit of detection is generally low, with 6.8% of hide and 16.2% of carcass samples for E. coli O157:H7, and 17.8% of hide and 23.5% of carcass samples for Salmonella within the detection limit. The remaining data are being analyzed and updated results will be provided when available. Implications We found levels of Salmonella on hides and carcasses to be consistently higher than E. coli O157:H7 and Listeria. The multi drug resistance status of the Salmonella isolated in this study is being assessed. Preliminary data (from 1 of the 4 seasons, Winter 2006) indicates that the percent of MDR Salmonella present on hides and carcasses varies considerably. These data demonstrate that foodborne pathogens are present on the hides and carcasses of cow/bull cattle at slaughter, but that carcass interventions in general, greatly reduce the prevalence and levels of Salmonella, MDR Salmonella and Listeria on final carcasses. 4
5
6
7
8
9
For more information contact: National Cattlemen's Beef Association A Contractor to the Beef Checkoff 9110 East Nichols Avenue Centennial, Colorado 80112-3450 (303) 694-0305 10