Salmonella is a common cause of human foodborne illness, which is frequently associated with consumption of contaminated or undercooked poultry meat. Serotype Infantis is among the most common serotypes isolated from poultry meat products globally. Isolates of serotype Infantis carrying the pESI plasmid, the most dominant strain of Infantis, have been shown to exhibit oxidizer tolerance. Therefore, 16 strains of Salmonella with and without pESI carriage were investigated for susceptibility to biocide chemical processing aids approved for use in US poultry meat processing: peracetic acid (PAA), cetylpyridinium chloride (CPC), calcium hypochlorite, and sodium hypochlorite. Strains were exposed for 15 s to simulate spray application and 90 min to simulate application in an immersion chiller. All strains tested were susceptible to all concentrations of PAA, CPC, and sodium hypochlorite when applied for 90 min. When CPC, calcium hypochlorite, and sodium hypochlorite were applied for 15 s to simulate spray time, strains responded similarly to each other. However, strains responded variably to exposure to PAA. The variation was not statistically significant and appears unrelated to pESI carriage. Results highlight the necessity of testing biocide susceptibility in the presence of organic material and in relevant in situ applications.

Campylobacter jejuni is the leading foodborne bacterial pathogen that causes human gastroenteritis worldwide linked to the consumption of undercooked broiler livers. Application of bacteriophages during poultry production has been used as an alternative approach to reduce contamination of poultry meat by Campylobacter. To make this approach effective, understanding the presence of the bacteriophage sequences in the CRISPR spacers in C. jejuni is critical as they may confer bacterial resistance to bacteriophage treatment. Therefore, in this study, we explored the distribution of the CRISPR arrays from 178 C. jejuni isolated from chicken livers between January and July 2018. Genomic DNA of C. jejuni isolates was extracted, and CRISPR type 1 sequences were amplified by PCR. Amplicons were purified and sequenced by the Sanger dideoxy sequencing method. Direct repeats (DRs) and spacers of CRISPR sequences were identified using the CRISPRFinder program. Further, spacer sequences were submitted to the CRISPRTarget to identify potential homology to bacteriophage types. Even though CRISPR-Cas is reportedly not an active system in Campylobacter, a total of 155 (87%) C. jejuni isolates were found to harbor CRISPR sequences; one type of DR was identified in all 155 isolates. The CRISPR loci lengths ranged from 97 to 431 nucleotides. The numbers of spacers ranged from one to six. A total of 371 spacer sequences were identified in the 155 isolates that could be grouped into 51 distinctive individual sequences. Further comparison of these 51 spacer sequences with those in databases showed that most spacer sequences were homologous to Campylobacter bacteriophage DA10. The results of our study provide important information relative to the development of an effective bacteriophage treatment to mitigate Campylobacter during poultry production.

With the changing regulations in poultry processing, increased pressure is placed on integrators to reduce the number of human enteropathogenic bacteria on the final carcass and/or parts. Reducing the total number of bacteria on broiler carcasses before entering the evisceration side of the processing plant is projected to reduce the number of bacteria on the carcasses after chilling. This study was designed to evaluate the efficacy of a prototype wash cabinet using low volume, fluidic nozzles in combination with high pressure (450 psi) and hot water (60 °C) to remove bacteria from pre-scald, post-scald, or post picked carcasses. Carcasses (n = 5) from each location were obtained from a commercial processing plant, placed into individual sterile sample bags, placed into an insulated container, and transported to the U.S. National Poultry Research Center Pilot Plant within 30 min of collection. Carcasses were hung in standard shackles and sampled pre-wash with pre-moistened, cellulose swabs. All carcasses were washed in the prototype wash cabinet with 60 °C water at 450 psi at a line speed of 52 birds/minute on 15.24 cm centered shackles. Post-wash breast sponge samples were collected identical to pre-wash swabs. Buffered peptone water (BPW) was added, sponges stomached and serially diluted before plating onto total aerobic count (TAC), Enterobacteriaceae (ENT) and Escherichia. coli (EC) Petrifilm® cards. All PetriFilm® cards were incubated at 37 °C for 24 ± 2 h. After incubation, bacterial counts were recorded and converted to log10 CFU/swab. Samples were processed for Campylobacter species using the Tempo® CAM protocol. Four replications were conducted on separate dates. Paired t-tests were used to compare numbers recovered from breast swabs collected before and after the wash cabinet, significance reported at p < 0.05. Pre-scald samples had significant reductions of 2.50, 2.01, and 1.73 log10 colony-forming units/carcass (CFU/carcass) for TAC, Ent, and EC Petrifilm®, respectively, and a 2.21 CFU/mL reduction of Campylobacter species using Tempo® CAM. Post-scald, there were significant reductions of 2.09, 1.23, and 0.90 CFU/carcass for TA, Ent, and EC Petrifilm®, respectively, and a 1.14 CFU/mL reduction of Campylobacter species using Tempo® CAM. Post-pick, significant reductions of 0.73, 1.53, and 0.99 CFU/carcass for TA, Ent, and EC Petrifilm®, respectively, and a 0.86 CFU/carcass reduction of Campylobacter species using Tempo® CAM were reported. These data indicate that hot water at high pressure can reduce total bacterial load on carcasses and reduce pathogenic bacteria on carcasses prior to evisceration.

Human Campylobacter infections have been associated with chicken and other poultry meat products. Environmental conditions such as temperature and season can affect Campylobacter recoverability from chicken meat products. In the presented study, we sought to investigate the relationship between ambient weather conditions and the isolation of Campylobacter from chicken flocks, as well as the subtype of these isolates. Campylobacter was isolated from the ceca of broilers collected in a commercial processing facility over 7 years, representing 452 flocks. Isolates were subjected to whole-genome sequencing and subtyping by multilocus sequence typing (MLST). Approximately 60% (269/452) of flocks sampled were positive for Campylobacter. There was no significant effect on the presence of detectable Campylobacter by month, season, temperature, or rainfall during grow-out or transportation. Sixty-eight different STs were detected; 45 C. jejuni and 23 C. coli. Diversity as measured by Shannon's diversity index was higher in the spring and fall than in mid-winter and summer. We concluded that in the warm temperate climate of the Southeastern U.S., seasonality does not affect the rate of Campylobacter isolation from broilers, but the diversity of isolates was higher in the milder spring and fall seasons.

The pH of pre-enrichment media containing feed/ingredients can become acidic during incubation due to bacterial utilization of feed carbohydrates. This decrease in pH can result in cell injury or death, negatively impacting the detection of Salmonella. Our objective was to evaluate a new triple buffered peptone (TBP) against buffered peptone water (BPW) and lactose broth (LB) for the recovery of Salmonella from feed. Liquid cultures of nalidixic acid resistant strains of Salmonella (Enteritidis, Heidelberg, Kentucky or Typhimurium) were added to the pre-enrichment media alone, to pre-enrichment media containing feed or to artificially inoculated feed stored 1 or 7 d to evaluate the effect of the medium on the recovery of Salmonella. Three replicates per treatment were conducted. After incubation at 37 °C for 24 h, the pH of the medium was measured prior to plating onto brilliant green sulfa agar plates supplemented with 200 ppm nalidixic acid (BGSNA). Plates were incubated and evaluated for presence of typical Salmonella colonies. The experiment was replicated. TBP was observed to exhibit significantly better buffering capacity than BPW or LB. Additionally, TBP was able to recover Salmonella 100% of the time compared to BPW (97.9%) and LB (61.5%). TBP shows promise to maintain neutral pH during pre-enrichment which may allow for a more accurate detection of Salmonella in feed.

Salmonella enterica is a major cause of human foodborne illness and is often attributed to poultry food sources. S. enterica serovar Infantis, specifically those carrying the pESI plasmid, has become a frequently isolated serotype from poultry meat samples at processing and has caused numerous recent human infections. In 2016, the USDA-Food Safety and Inspection Service changed the official sampling method for raw poultry products from BPW to using neutralizing BPW (nBPW) as the rinsing agent in order to prevent residual antimicrobial effects from acidifying and oxidizing processing aids. This change was contemporaneous to the emergence of pESI-positive ser. Infantis as a prevalent serovar in poultry, prompting some to question if nBPW could be selecting for this prevalent serovar. We performed two experiments: a comparison of ser. Infantis growth in BPW versus nBPW, and a simulation of regulatory sampling methods. We found that when inoculated into both broths, ser. Infantis initially grows slightly slower in nBPW than in BPW but little difference was seen in abundance after 6 h of growth. Additionally, the use of nBPW to simulate poultry rinse sample and overnight cold shipping to a regulatory lab did not affect the survival or subsequent growth of ser. Infantis in BPW. We concluded that the change in USDA-FSIS methodology to include nBPW in sampling procedures has likely not affected the emergence of S. ser. Infantis as a prevalent serovar in chicken and turkey meat product samples.

The similarity of the Listeria innocua genome with Listeria monocytogenes and their presence in the same niche may facilitate gene transfer between them. A better understanding of the mechanisms responsible for bacterial virulence requires an in-depth knowledge of the genetic characteristics of these bacteria. In this context, draft whole genome sequences were completed on five L. innocua isolated from milk and dairy products in Egypt. The assembled sequences were screened for antimicrobial resistance and virulence genes, plasmid replicons and multilocus sequence types (MLST); phylogenetic analysis of the sequenced isolates was also performed. The sequencing results revealed the presence of only one antimicrobial resistance gene, fosX, in the L. innocua isolates. However, the five isolates carried 13 virulence genes involved in adhesion, invasion, surface protein anchoring, peptidoglycan degradation, intracellular survival, and heat stress; all five lacked the Listeria Pathogenicity Island 1 (LIPI-1) genes. MLST assigned these five isolates into the same sequence type (ST), ST-1085; however, single nucleotide polymorphism (SNP)-based phylogenetic analysis revealed 422-1,091 SNP differences between our isolates and global lineages of L. innocua. The five isolates possessed an ATP-dependent protease (clpL) gene, which mediates heat resistance, on a rep25 type plasmids. Blast analysis of clpL-carrying plasmid contigs showed approximately 99% sequence similarity to the corresponding parts of plasmids of L. monocytogenes strains 2015TE24968 and N1-011A previously isolated from Italy and the United States, respectively. Although this plasmid has been linked to L. monocytogenes that was responsible for a serious outbreak, this is the first report of L. innocua containing clpL-carrying plasmids. Various genetic mechanisms of virulence transfer among Listeria species and other genera could raise the possibility of the evolution of virulent strains of L. innocua. Such strains could challenge processing and preservation protocols and pose health risks from dairy products. Ongoing genomic research is necessary to identify these alarming genetic changes and develop preventive and control measures.

Controlling Salmonella in poultry processing continues to be important to processors and consumers. Cetylpyridinium chloride (CPC) has proven to be effective in vitro in controlling Salmonella. This study evaluated the recovery of Salmonella after overnight storage in 4°C filter-sterilized carcass rinsate containing CPC from 0.44 to 909 ppm (μg/mL). Ten Salmonella serotypes (18 strains), of which 6 serotypes are commonly isolated from poultry products, were grown in Bacto-Tryptic Soy Broth overnight at 37°C. Serial dilutions of a CPC/propylene glycol solution were prepared in 24-well tissue culture plates containing filter-sterilized carcass rinsate. Approximately 107 cfu/mL of each Salmonella serotype was added to the appropriate wells. Inoculated plates were stored overnight at 4°C. After storage, triplicate plates of brilliant green agar with sulfapyridine (BGS) were surface inoculated with 10 μL of the contents for each well, streaked for isolation, and incubated at 37°C for 24 h. Three replications were conducted. The presence of typical colonies on BGS plates was recorded as growth and verified through biochemical and serological testing. Of the serotypes chosen, Salmonella Kentucky, Dublin, and Enteritidis were the least resistant to CPC with a median minimum inhibitory concentration (MIC) of 14.22 μg/mL (range from 3.55 to 56.88 μg/mL); S. Typhimurium demonstrated a median MIC of 114.00 μg/mL (range from 28.44 to 114.00 μg/mL). Residual CPC potentially remaining attached to a carcass or in the weep after processing could potentially alter which Salmonella serotype is recovered from a carcass rinse due to different growth patterns during regulatory testing, with a potential for more virulent strains not to be recovered.

Campylobacter spp. are a leading cause of human foodborne illness associated with chicken meat products in the United States. Chicken livers, including exudate from packaging, commonly carry Campylobacter and could be a source of illness if mishandled. Survivability of naturally occurring Campylobacter, total aerobic bacteria, and coliforms was determined under drying conditions in two consumer simulated environments: moist sponge and solid surface. Fresh chicken liver exudate was dispensed onto sponges and glass slides and allowed to dry under ambient conditions for 7 days. Bacterial concentration was measured at 0, 6, 24, 48, 72, and 168 h. Total aerobic population did not decrease by more than one log over 7 days and did not correlate to water activity or time in either simulation. Coliform concentrations increased in sponge simulations but decreased in solid surface simulations. Further, coliform concentrations were significantly higher in sponge simulations than in solid surface. Campylobacter was naturally present in exudate and survived at least to 6 h in every trial. Campylobacter was recoverable at 24 h in some sponge trials. However, Campylobacter concentration was strongly correlated to water activity. Fresh chicken liver exudate could present a risk of campylobacteriosis to consumers if mishandled even after drying.

Campylobacter jejuni causes gastroenteritis in humans, and chickens and environmental water are leading sources of infection. We tested the hypothesis that Campylobacter isolated from chicken ceca and river water in an overlapping geographic area would share genetic information.

Conventional Salmonella surveillance requires a week for isolation, confirmation, and subse-quent serotyping. We previously showed that this could be reduced by 24 h by combining the pre-enrichment and enrichment steps into a single selective pre-enrich-ment step and was tested on directly after picking. The goal of this study was 2-fold: 1) to evaluate the use of selective pre-enrichment through each step of process-ing, including postintervention when the Salmonella load is reduced, and 2) to assess any changes in serovar populations in Salmonella positive samples. Duplicate carcass drip samples, each representative of 500 broiler carcasses, were collected by catching processing water drip under moving carcass shackle lines in each of three commercial broiler slaughter plants. Samples were col-lected post-pick, post-inside-outside bird wash (IOBW), and post-chill; duplicate wing rinses were per-formed pre-and post-antimicrobial parts dip. Each proc-essing plant was sampled 6 times for a total of 180 samples collected. The number of Salmonella positives identified with selective pre-enrichment conditions (48/ 180) was similar to traditional selective enrichment cul-ture conditions (52/180), showed good concordance in recovery rate between the 2 culture methods (Fisher's exact test, P = 0.72). We also found that the incidence of Salmonella reduced dramatically after antimicrobial intervention (post-pick 66.7% vs. post chill 8.3%). When serovar populations were evaluated in Salmonella posi-tive samples using CRISPR-SeroSeq, we detected four different Salmonella serovars, Kentucky, Infantis, Schwarzengrund, and Typhimurium, and their incidence rose between post-pick and post-IOBW. The relative abundance of Infantis within individual samples increased between post-pick and post-IOBW while the relative abundance of the other 3 serovars decreased. These results suggest that a selective pre-enrichment step reduces the time required for Salmonella isolation without negatively affecting detection and serovar pro-files in culture positive samples were not altered between culture conditions used.

Campylobacter is a bacterial pathogen that causes human foodborne illnesses worldwide, and outbreaks have been associated with consumption of undercooked chicken livers. The objectives of this study were to compare two PCR assays of 250 Campylobacter isolates for identification to species, to assess antibiotic resistance of the isolates, and to analyze genetic diversity of the quinolone resistance determining regions (QRDRs) of the isolates. A double-blind design was used to identify the species of Campylobacter; 181 (72%) of the isolates were identified as Campylobacter jejuni, and 69 (28%) isolates were identified as Campylobacter coli by both PCR assays. A total of 93 (37.2%) isolates were resistant to at least one antibiotic. Among 88 C. jejuni isolates, 33 (18%) were resistant to nalidixic acid (NAL) and ciprofloxacin (CIP), 25 (14%) were resistant to tetracycline (TET), and 18 (10%) were resistant to NAL and TET. Two C. jejuni isolates were resistant to four of the tested antibiotics, and one isolate was resistant to five antibiotics. Two C. coli isolates were resistant to TET, and two were resistant to NAL, CIP, and TET. The amino acid sequences of the QRDRs for the isolates had eight point mutations and could be classified into 12 groups. Thirty-eight C. jejuni isolates resistant to NAL and CIP had a point mutation at residue 86 (substitution from threonine to isoleucine). However, six isolates without this substitution were resistant to NAL and/or CIP. Ten isolates with a point mutation at residue 86 were susceptible to NAL and CIP. This observation suggests that in addition to the substitution at residue 86 other mechanisms may confer resistance to quinolones. Further studies are needed to elucidate mechanisms for quinolone resistance in Campylobacter. The Campylobacter spp. isolated from chicken livers in this study were resistant to quinolones and other classes of antibiotics.

Food processors face serious challenges due to the ubiquity and prevalence of Listeria monocytogenes in production and processing facilities. Environmental monitoring for Listeria within the industry is important and detection of Listeria spp. is often used as an indicator for the potential presence of L. monocytogenes in the food processing environment. Environmental monitoring data from the frozen food manufacturing industry was compiled and analyzed to evaluate the adequacy of current practices in mitigating risks of L. monocytogenes in the processing environment and to determine if there are trends that could be used to further refine industry practices. A method to collect anonymous data for analysis to build a strong aggregate data set from multiple facilities was used. Information included general descriptions of each facility and specific information about individual environmental monitoring test results from environmental zones 2-4, or nonfood contact surfaces. Information collected from facilities included the size of the facilities and how environmental monitoring samples were collected and analyzed. For each individual sample, information collected included the area or equipment sampled and the result of each sample. Descriptors were provided to allow for grouping of similar results. The data represents environmental monitoring samples that were collected between six months to a year prior to the beginning of this research. Twenty-seven facilities provided 42,799 environmental monitoring observations. Environmental zones 3 and 4 had a higher estimated marginal probability of Listeria positive results compared to zone two for routine environmental monitoring samples (p < 0.05). Pre-lethality and post-lethality production areas had a higher probability of Listeria positive results compared to lethality areas for all environmental samples (p < 0.05). Cold storage locations, i.e., coolers and freezers had a significantly higher probability for a Listeria positive result than noncold storage areas (p < 0.01). Applying statistical modeling and comparing estimated marginal probabilities of aggregate Listeria prevalence data from processing operations across the industry can provide information for guidance on focusing sanitation within processing operations to reduce risk related to L. monocytogenes. These data help to identify common themes in the industry and determine potential areas on which to focus when testing for prevalence of Listeria spp. within the food processing environment. This study serves as a useful baseline for continued improvement in environmental monitoring across all frozen food manufacturing facilities.

In chickens, early life exposure to environmental microbes has long-lasting impacts on gastrointestinal (GI) microbiome development and host health and growth, via mechanisms that remain uncharacterized. In this study, we demonstrated that administrating a fecal microbiome transplant (FMT) from adults to day-of-hatch chicks results in significantly higher body mass of birds and decreased residual feed intake (RFI), implying enhanced feed efficiency, at 6 weeks of age. To assess the potential mechanisms through which FMT affects adult bird phenotype, we combined 16 S rRNA gene amplification, metagenomic, and comparative genomic approaches to survey the composition and predicted activities of the resident microbiome of various GI tract segments. Early life FMT exposure had a long-lasting significant effect on the microbial community composition and function of the ceca but not on other GI segments. Within the ceca of 6-week-old FMT birds, hydrogenotrophic microbial lineages and genes were most differentially enriched. The results suggest that thermodynamic regulation in the cecum, in this case via hydrogenotrophic methanogenic and sulfur-cycling lineages, potentially serving as hydrogen sinks, may enhance fermentative efficiency and dietary energy harvest capacity. Our study provides a specific mechanism of action through which early-life microbiome transplants modulate market-relevant phenotypes in poultry and, thereby, may represent a significant advance toward microbiome-focused sustainable agriculture.

Foodborne campylobacteriosis has been traced to undercooked chicken liver dishes; thus, it is important to use the best available culture methods when testing for the presence of Campylobacter. We compared two Campylobacter enrichment broths-Bolton formulation and Neogen formulation-in combination with three selective plating media-Campy-Cefex, Campy-Line and RF Campylobacter agars-for detection of Campylobacter from fresh retail chicken livers. In each of three experiments, nine replicate tubs of chicken livers were sampled by drawing exudate and a pooled rinse of five whole liver lobes. Results are reported as number positive and compared by Fisher's exact test. In experiment 1, no combination of enrichment and plating media significantly outperformed another for detection of Campylobacter (P > 0.05); all tubs were found to include Campylobacter in both exudate and liver rinse. In experiment 2, serial dilutions of samples were plated before and after enrichment. Exudate was found to be significantly more likely than rinse to support detection of Campylobacter by direct plating (P < 0.05); most exudate samples included at least 10 CFU Campylobacter per mL. Enrichment improved detection from rinse, but not exudate; all enrichment and plating combinations resulted >= 1,000 CFU/mL from most enriched samples. In experiment 3, samples were diluted before enrichment to determine effect of enrichment on ever lower numbers of Campylobacter. Enrichment did not improve recovery of Campylobacter from exudate or undiluted rinse (P > 0.05). However, when rinse samples were diluted to lower Campylobacter numbers, enrichment improved detection (P < 0.05). Overall, all media combinations tested were equivalent for detection of Campylobacter from chicken livers; sensitivity for detection seemed to be increased by using liver exudate compared with a pooled rinse of liver lobes.

Semicarbazide (SEM) is routinely employed as an indicator for the use of nitrofurazone, a banned antimicrobial. The validity of SEM as a nitrofurazone marker has been scrutinized because of other possible sources of the compound. Nonetheless, a U.S. trade partner rejected skin-on chicken thighs because of SEM detection and suspected nitrofurazone use. Because nitrofurazone has been banned in U.S. broiler production since 2003, we hypothesized that incidental de novo SEM formation occurs during broiler processing. To assess this possibility, raw leg quarters were collected from 23 conunercial broiler processing plants across the United States and shipped frozen to our laboratory, where liquid chromatography-mass spectrometry was used to quantitatively assess for SEM. Leg quarter samples were collected at four points along the processing line: hot rehang (transfer from the kill line to the evisceration line), prechill (before the chilling process), postchill (immediately following chilling), and at the point of pack. Thigh meat with skin attached was removed from 535 leg quarters and analyzed in triplicate for SEM concentrations. The concentrations ranged from 0 to 2.67 ppb, with 462 (86.4%) of 535 samples below the regulatory decision level of 0.5 ppb of SEM. The 73 samples over the 0.5-ppb limit came from 21 plants; 53 (72.6%) of positive samples were in meat collected after chilling (postchill or point of pack). The difference in both prevalence and concentration of SEM detected before and after chilling was highly significant (P < 0.0001). These data support our hypothesis that SEM detection in raw broiler meat is related to de novo creation of the chemical during processing.

Antimicrobial resistance (AR) spread is a worldwide health challenge, stemming in large part, from the ability of microbes to share their genetic material through horizontal gene transfer (HGT). Overuse and misuse of antibiotics in clinical settings and in food production have been linked to this increased prevalence and spread of AR. Consequently, public health and consumer concerns have resulted in a remarkable recent reduction in antibiotics used for food animal production. This is driven by the assumption that removing this selective pressure will favor the recovery of antibiotic susceptible taxa and will limit AR sharing through HGT, allowing the currently available antibiotic arsenal to be effective for a longer period. In this study we used broiler chicks raised antibiotic-free and Salmonella enterica serovar Heidelberg (SH), as a model food pathogen, to test this hypothesis. Our results show that neonatal broiler chicks challenged with an antibiotic susceptible SH strain and raised without antibiotics carried susceptible and multidrug resistance SH strains 14 days after challenge. SH infection perturbed the microbiota of broiler chicks and gavaged chicks acquired antibiotic resistant SH at a higher rate. We determined that the acquisition of a plasmid from commensal Escherichia coli population conferred multidrug resistance phenotype to SH recipients and carriage of this plasmid increased the fitness of SH under acidic selection pressure. These results suggest that HGT of AR shaped the evolution of SH and that antibiotic use reduction alone is insufficient to limit antibiotic resistance transfer from commensal bacteria to Salmonella . Importance The reported increase in antibiotic resistant bacteria in humans have resulted in a major shift away from antibiotics use in food animal production. This has been driven by the assumption that removing antibiotics will select for antibiotic susceptible bacterial taxa, and this in turn will allow the currently available antibiotic arsenal to be more effective. This shift in practice has highlighted new questions that need to be answered to assess the effectiveness of antibiotic removal in reducing the spread of antibiotic resistance bacteria. This research demonstrates that antibiotic susceptible Salmonella Heidelberg strains can acquire multidrug resistance from commensal bacteria present in the gut of neonatal broiler chicks, even in the absence of antibiotic selection. We demonstrate that exposure to acidic pH drove the horizontal transfer of antimicrobial resistance plasmids and suggests that simply removing antibiotics from food-animal production might not be sufficient to limit the spread of antimicrobial resistance.

Studies were conducted to investigate the recovery of Campylobacter from feed. The impact of feed moisture, water activity, pH, number of background microflora and the use of different antibiotic supplements in Campylobacter enrichment broth (CEB) on Campylobacter recovery were evaluated in five studies. Broiler starter feed was inoculated with 10(4) -10(5) cfu of Campylobacter/g and stored at 24 degrees C and 43% RH. Enrichment culture was conducted on the day of inoculation or 24 h post inoculation and every 48 h of storage thereafter for 14 d. Feed moisture, water activity, pH and level of background microflora were not correlated with Campylobacter recovery. The incubation of feed in CEB with no antibiotic supplement resulted in the number of background microflora increasing to 10(9) cfu/g and the pH of the media decreasing to pH 4-5 impacting recovery. Addition of certain antimicrobial supplements to CEB reduced background microflora growth and maintained a near neutral pH. Campylobacter was recovered up to 10 days post inoculation when using CEB containing antibiotic supplements compared to 1 day in CEB. These findings suggest that Campylobacter can be recovered from feed and the type of antimicrobial supplement utilized influences recovery by controlling extraneous microbial growth which occurs during enrichment.