Abstract
We estimated the proportion of retail beef and pork products containing bla CMY-mediated third-generation cephalosporin resistance in commensal Escherichia coli and Salmonella enterica. Samples were obtained from 50 grocery stores located in two U.S. states. From each store, 20 fresh meat products were purchased, including 7 packages of ground beef, 3 packages of beef steak, 6 packages of pork chops, and 4 packages of pork ribs. The resulting 1000 packages of fresh meat product were individually screened for the presence of E. coli or Salmonella harboring bla CMY. Over 8% of all retail meat packages contained E. coli with bla CMY, whereas 4% contained Salmonella and only 0.5% contained Salmonella with bla CMY. Retail pork products more frequently yielded E. coli with bla CMY than did beef products (12.2% vs. 4.0%; p = 0.06). Salmonella were also recovered more frequently from pork than beef (5.8% vs. 2.4%; p < 0.01). In addition, all five Salmonella isolates with bla CMY were recovered from pork products. Our data suggest that enteric bacteria carrying bla CMY are frequently present in fresh retail meat products. However, we did not quantify the number of resistant bacteria present in these products, which makes the public health implications of this result unclear.
Introduction
Over a decade since being initially reported in the United States (Fey et al., 2000), bla CMY has become a prominent third-generation cephalosporin resistance gene among Salmonella enterica clinical isolates of both human and animal origin (Arlet et al., 2006; Frye and Fedorka-Cray, 2007; CDC, 2010; USDA, 2010). In addition, Escherichia coli have been recognized as a potential bla CMY reservoir for enteric pathogens (Neidhardt and Curtiss, 1996; Winokur et al., 2001). To better understand the potential for human exposure to resistant commensal bacteria and pathogens via fresh meat products, we investigated the occurrence of bla CMY-mediated third-generation cephalosporin resistance in Salmonella and E. coli recovered from retail beef and pork products.
Materials and Methods
Sampling
Fresh retail meat products were purchased from grocery stores located in Ohio and North Carolina. In each state, 5 retail grocery stores representing different corporate chains were identified in each of 5 metropolitan areas for a total of 50 grocery stores. At each store, 20 fresh retail meat products were purchased during a single store visit. Of these, 10 fresh beef and 10 pork products were purchased, including 7 packages of ground beef, 3 packages of beef steak, 6 packages of pork chops, and 4 packages of pork ribs. This distribution of products was intended to approximately represent their relative consumption as reported by the USDA Economic Research Service (Davis and Lin, 2005a, 2005b).
Bacterial culture
Samples from each package of beef steak, pork chops, and pork ribs were pre-enriched by aseptically collecting a 450 g aliquot into a 4 L plastic reclosable bag with 500 mL of buffered peptone water (BPW; Becton Dickinson, Sparks, MD) incubated overnight at 37°C. Ground beef packages were pre-enriched by placing a 10 g aliquot into 90 mL of BPW incubated overnight at 37°C.
Beef steak, pork chop, and pork rib samples were screened for the presence of E. coli containing bla CMY by inoculating 1 mL of pre-enriched BPW into 9 mL of nutrient broth (Becton Dickinson) with cefoxitin 4 μg/mL (Sigma-Aldrich, St. Louis, MO) incubated overnight at 37°C. Ground beef packages were screened for E. coli with bla CMY by inoculating three 10 g aliquots from each package directly into 90 mL Nutrient broth with cefoxitin 4 μg/mL incubated overnight at 37°C. The following day, the Nutrient broth was inoculated with a sterile cotton swab to MacConkey agar (Becton Dickinson) containing ceftriaxone 8 μg/mL (Sigma-Aldrich) and streaked for isolation. After overnight incubation at 37°C, a single, presumptive E. coli colony from each plate with growth was selected for confirmation using the indole test and polymerase chain reaction (PCR).
For the Salmonella culture of all meat types, 100 μL of pre-enriched BPW was inoculated into 10 mL of Rappaport-Vassiliadis R10 broth (Becton Dickinson) on day 2. After overnight incubation at 42°C, Rappaport-Vassiliadis R10 broth was inoculated to and streaked for isolated onto xylose lysine tergitol-4 agar (Becton Dickinson). From the xylose lysine tergitol-4 agar, an individual characteristic black colony was streaked to MacConkey agar. The resulting isolated colorless, lactose-negative colonies were used for S. enterica biochemical confirmation including triple sugar iron and urea, as well as latex agglutination and PCR.
The presence of the ampC β-lactamase bla CMY in E. coli and Salmonella isolates expressing the expected phenotype was confirmed by PCR (Heider et al., 2009). Estimation and comparison of the proportion of samples from which E. coli with bla CMY, Salmonella, or Salmonella with bla CMY were recovered was accomplished using logistic regression with generalized estimating equations to account for the clustering of samples within stores.
Results and Discussion
Of the 1000 packages of retail beef and pork products purchased, E. coli harboring bla CMY were recovered from 81 (8.1%), Salmonella from 40 (4.0%), and Salmonella carrying bla CMY from 5 (0.5%). Pork products more frequently (p = 0.06) yielded E. coli with bla CMY (12.2%) than did beef products (4.0%). Salmonella spp. were also recovered more frequently (p < 0.01) from pork (5.8%) than from beef products (2.4%). In addition, the 5 Salmonella isolates containing bla CMY were all recovered from pork products (Table 1). There was no difference in the frequency of recovery of E. coli with bla CMY, Salmonella, or Salmonella with bla CMY between ground beef and beef steak, or between pork chops and pork ribs (Table 1). The frequency of recovery of E. coli containing bla CMY ranged from 4% to 15% in individual cities, with both the highest and lowest prevalence cities located in Ohio. Salmonella spp. were recovered from 13% of the pork packages in one city in NC, and 10% of the pork packages in another NC city. Four of the five Salmonella containing bla CMY came from two NC stores in the same city. However, isolates from different stores in the same city belonged to different serogroups.
Our data suggest that enteric bacteria carrying bla CMY are frequently present in retail beef and pork products. Others have reported the presence of multiresistant E. coli (LeJeune and Christie, 2004) and Salmonella spp. (White et al., 2001; Chen et al., 2004) present in retail meat products, including Salmonella carrying bla CMY-2. Salmonella present in retail meat poses an immediate health threat to consumers as it is an important enteric pathogen resulting in ∼500 deaths per year in the United States (Mead et al., 1999), most of which arise from foodborne exposures. E. coli may serve as a primary pathogen (Jakobsen et al., 2010), or as a potential reservoir of resistance genes for other pathogens (Winokur et al., 2001). While our study was not designed to determine the source of the E. coli and Salmonella present in the meat products, it is likely that fecal contamination of the carcass during processing introduced the animal's own enteric flora onto the surface of the meat products. Ceftiofur is a veterinary antimicrobial drug that provides selection pressure resulting in the transient detection of bla CMY in the fecal flora of treated animals (Singer et al., 2008; Alali et al., 2009). Veterinary ceftiofur use has also been associated with the dissemination of bla CMY in livestock populations (Tragesser et al., 2006; Lowrance et al., 2007). However, we do not know if the retail meat products included in this study originated from livestock populations exposed to ceftiofur. It is also important to note that we did not attempt to quantify the E. coli present in the meat, making the public health implications of this result unclear.
Footnotes
Disclosure Statement
No competing financial interests exist.