|Home | About | Journals | Submit | Contact Us | Français|
Forty-four percent of Swedish chicken meat fillets were contaminated with extended-spectrum or transferable AmpC beta-lactamase-producing Escherichia coli strains. Isolates from Swedish chicken meat and broilers were closely related to isolates from chicken meat imported into Sweden; these results indicate a common source of the contamination.
The increasing trend of resistance to broad-spectrum cephalosporins in Enterobacteriaceae in human and animal settings is of great concern, due to the fact that cephalosporins are deemed critically important in human health care (1–4). The emergence of this type of resistance is mainly linked to the spread of genes encoding extended-spectrum beta-lactamases (ESBLs) and/or transferable AmpC beta-lactamases (pAmpC) (5). Poultry, primarily broilers, have been suggested as a source for these types of genes and/or the resistant bacteria (6–8). In Sweden, broilers are rarely treated with antibiotics and no usage of cephalosporins was registered in 2010 (9). It was therefore surprising when 34% of the Swedish broilers were identified as carrying ESBL- or pAmpC-producing E. coli in 2010 (9).
A potential swift route for transmission of ESBL- and pAmpC-producing Escherichia coli from broilers to humans can be through contaminated chicken meat and meat products. As a response to the current situation in broiler production in Sweden, this study investigated whether domestically produced chicken meat contained ESBL- and/or pAmpC-producing E. coli and characterized potential isolates phenotypically and genotypically. In addition, the relatedness of pAmpC-producing E. coli isolated from Swedish chicken meat to isolates from Swedish broilers in a previous study was also determined in order to establish whether the contamination originated from broilers. This part of the study was limited to blaCMY-2 E. coli only, since earlier studies have described this gene variant to be almost universal in the Swedish broiler population (9, 10). Furthermore, these isolates were compared with isolates from imported chicken meat by using pulsed-field gel electrophoresis (PFGE), as identical PFGE patterns can indicate a potential common source.
Over a 10-week period from September to November 2010, 100 frozen chicken meat fillets intended for sale on the Swedish market were collected in collaboration with the Swedish Poultry Meat Association. The samples selected each represented a unique batch, and the number of samples was distributed between the largest processing facilities in proportion to the volume of broilers slaughtered. A 25-g portion of each sample was placed in 225 ml buffered peptone water and subjected to stomaching. One hundred milliliters of homogenate with addition of 1 mg/liter cefotaxime (Sigma-Aldrich) was incubated overnight at 37°C. After enrichment, ESBL/pAmpC-producing E. coli was isolated using CHROMagarESBL and CHROMagarOrientation with 1 mg/liter cefotaxime (EMM Life Science AB, Hägersten, Sweden). Identification of E. coli was performed using API20E (bioMérieux Sweden, Askim, Sweden), and ESBL/pAmpC phenotypes were verified using EtestESBL and EtestAmpC (bioMérieux Sweden). One E. coli isolate per sample was collected.
Antimicrobial susceptibility of the isolates was assessed using VetMIC GN-mo plates, applying epidemiological cutoff values by EUCAST, and tested by EtestImipenem. ESBL/pAmpC genes were detected by PCR (11–13). The gene variants were determined by sequencing using BigDye v1.1 and the following primers: blaCTX-M-1, 5′-CAGAATAAGGAATCCCATGGTT-3′ and 5′-GGCGATAAACAAAAACGGAAT-3′; blaCMY-2, 5′-AAATCGTTATGCT(G/C)CGCTCT-3′ (14) and 5′-CATGGGATTTTCCTTGCTGT-3′; and blaTEM (5′-TTCTTGAAGACGAAAGGGC-3′ and 5′-ACGCTCAGTGGAACGAAAAC-3′) (15). Isolates were genotyped by multilocus sequence typing (MLST) according to http://mlst.ucc.ie/.
Transferability of cephalosporin resistance from the broiler isolates was tested by conjugation to E. coli HMS174.
Transformation was performed on isolates unable to transfer ESBL/pAmpC phenotypes by conjugation and transconjugants PCR positive for multiple plasmid replicon types. Plasmid DNA was transformed using ElectroMax DH10B (Gibco Invitrogen). Transconjugants and transformants were subjected to PCR-based plasmid replicon typing (16) and PCRs for ESBL/pAmpC genes (11–13).
Additional isolates from previous studies on Swedish broilers (S. Börjesson, C. Jernberg, A. Brolund, P. Edquist, M. Finn, A. Landén, B. Olsson Liljequist, K. Tegmark-Wisell, B. Bengtsson, and S. Englund, submitted for publication) and imported chicken meat (17) were included in the present study based on identical plasmid replicon type and MLSTs and antibiotic resistance patterns. These isolates were compared with isolates from Swedish meat by PFGE (18) and analyzing the patterns obtained in BioNumerics v6.6 (Applied Maths, Ghent, Belgium).
In total, 44 of the 100 Swedish chicken fillets analyzed carried E. coli strains producing ESBL or pAmpC. Four of these isolates carried blaCTX-M-1, 38 blaCMY-2, and 2 blaCMY-2 plus blaTEM-1. None of the isolates showed decreased susceptibility to imipenem.
The four blaCTX-M-1 isolates belonged to three different sequence types (STs) and were resistant to sulfamethoxazole and tetracycline (Table 1). blaCTX-M-1 was identified on plasmids belonging to replicon type incI. The blaCTX-M incI1plasmid also carried sulfamethoxazole and tetracycline resistance phenotypes.
Twelve of the blaCMY-2 isolates belonged to MLST ST38, and the remaining isolates were associated with 13 different STs (Table 1). Twenty-three (out of 40) blaCMY-2 isolates were sensitive to all additional antibiotics tested, and six were defined as multiresistant (i.e., resistant to more than three antimicrobial classes) (Table 1). Thirty-six of the isolates carried blaCMY-2 on an incK plasmid, and three isolates carried it on an incI plasmid. In the remaining isolate, both the corresponding transconjugant and transformant were positive for incK, IncF, and incFIB (Table 1). None of the plasmids carrying blaCMY-2 transferred any additional resistance phenotypes, and neither of the two isolates carrying blaTEM-1 transferred blaTEM-1 together with blaCMY-2.
A total of 18 isolates from Swedish meat, four isolates from Swedish broiler cecum, four isolates from Danish meat, and one isolate from Finnish meat were compared using PFGE (Fig. 1). Isolates belonging to the same STs were found to be related or closely related (Fig. 1). Three isolates from Swedish broiler cecum, Swedish meat, and Danish meat were found to share identical PFGE patterns.
The occurrence of ESBL/pAmpC-producing E. coli on Swedish meat is most likely due to fecal contamination, because isolates with identical or closely related PFGE patterns were found in both broiler cecum and meat. A previous study has also shown that contamination of meat by fecal bacteria is a common event during slaughter (19).
The Swedish situation is similar to that described in Denmark, where 44% of the chicken meat was reported to be contaminated (20). However, the prevalence is still lower than, for example, that in The Netherlands, where up to 94% of the meat may contain ESBL-producing E. coli (7, 8). The high percentage of chicken meat contaminated is worrying, because studies have suggested that isolates, the plasmid, and/or the genes can be transferred from broilers to humans via the meat (6–8, 21, 22). In situ experiments have demonstrated that ESBL-producing E. coli strains of poultry origin can establish themselves and that the genes can be transferred in the human commensal bowel microbiota even without antimicrobial pressure (23). It is therefore possible that such a high occurrence on chicken meat increases the risk of contribution to the human ESBL/pAmpC load.
The majority of the E. coli isolates producing ESBL or pAmpC found in the present study carried blaCMY-2 on an incK plasmid, reflecting the situation in Swedish broilers (Börjesson et al., submitted). The blaCMY-2 gene has generally been identified on other replicon types in humans and animals, but blaCMY-2 on incK is increasing in poultry, and recent studies have identified this combination in human clinical settings (1, 24–28; Börjesson et al., submitted). In addition, the combination of incI1 and blaCTX-M-1 identified in this study has frequently been identified in both human clinical and broiler settings (29, 30).
It was found that blaCMY-2 E. coli isolates from chicken meat imported from Denmark and Finland were identical or closely related to isolates from Sweden (Fig. 1). It has been suggested that the occurrence of ESBL- and pAmpC-producing E. coli strains in Swedish broilers is due to vertical transmission from imported breeding stock (10). Therefore, the likely connection between the Swedish, Danish, and Finnish isolates is via vertical transmission through the broiler production pyramid. Swedish and Danish broilers share the same parent birds, as Swedish parent birds are exported to Denmark (31). In addition, Sweden and Finland import grandparents from the same companies (Thomas Carlson, Aviagen-Swechick, personal communication, 2012). A large proportion of the Swedish meat isolates tested were also resistant to ciprofloxacin, nalidixic acid, sulfamethoxazole, and tetracycline, the same resistance pattern seen in isolates from broiler cecum. This is noteworthy because these types of antibiotics are not used in Swedish production (9, 32). This may indicate that these resistant E. coli isolates were not selected through antibiotic pressure, but introduced into the flocks.
A large proportion of the samples from Swedish chicken meat were contaminated with ESBL- or pAmpC-producing E. coli, primarily carrying the blaCMY-2 gene. A selection of isolates from meat were identical or closely related to those identified in slaughtered Swedish broilers, indicating that the occurrence on meat is due to fecal contamination at slaughter. The results offer support for the theory that occurrence of ESBL- and pAmpC-producing E. coli in the Swedish broiler population is due to introduction and spread from imported breeding stock.
This work was supported by The Swedish Civil Contingencies Agency.
Published ahead of print 25 January 2013