The increase of methicillin-resistant Staphylococcus aureus(MRSA) and the presence of an unusual sequence type ST49 in slaughter pigs in Switzerland

Increased prevalence and molecular typing

MRSA (n = 31) were detected in samples from 5 of 9 slaughterhouses (A, B, C, E, G) in 2009 and in 7 of 9 slaughterhouses (A-D, G-I) in 2010 (Table 1). MRSA was mostly found in pigs raised in cantons where the pig population is the highest (Figure 1), that is, Thurgau (n = 11), Lucerne (n = 6), St. Gallen (n = 5), Bern (n = 4), and Aargau (n = 2). Single isolates were found in pigs from Appenzell Innen Rhodes, Jura, and Zurich (Table 1 and Table 2). In 2009, the prevalence of MRSA in Swiss slaughter pigs was 2.0% (95% CI 0.9-3.9) with eight out of 405 pig nasal samples being positive and increased significantly (p-value < 0.05) to 5.9% (95% CI 3.8-8.7) in 2010 with 23 out of 392 nasal swabs containing MRSA. MRSA was detected every month in 2010, except in September, with up to five positive samples per month, whereas MRSA was only sporadically detected over the year in 2009.

	2009					2010
Abattoir	Total no. of samples	No. of MRSA positive samples	Percentage of MRSA positive samples	Origina	Genotype (n)	Total no. of samples	No. of MRSA positive samples	Percentage of MRSA positive samples	Origina	Genotype (n)
A	84	2	2.4	TG1, LU2	ST398-t011-V (2)	68	11	16.2	LU4, LU5, SG3, SG5, TG5, TG8, TG9, BE3, BE4 ZH1, JU1	ST398-t034-V (11)
B	90	1	1.1	SG1	ST398-t011-V	101	2	2.0	AG2, SG6	ST49-t208-V (2)
C	63	1	1.6	BE1	ST398-t034-V	65	4	6.2	TG6 BE2, TG7, LU6	ST49-t208-V ST398-t034-V (3)
D	62	0	0	-	-	58	2	3.4	AG1, TG10	ST398-t034-V (2)
E	39	1	2.6	LU1	ST398-t1451-V	37	0	0	-	-
F	28	0	0	-	-	22	0	0	-	-
G	19	3	15.8	TG3, TG2, SG2	ST1-t2279-IVc ST49-t208-V (2)	15	2	13.3	TG4, SG4	ST49-t208-V (2)
H	10	0	0	-	-	14	1	7.1	LU3	ST398-t034-V
I	10	0	0	-	-	12	1	8.3	AI1	ST398-t011-V
A-I (all)	405	8	2.0			392	23	5.9

a) Geographical origin of the farms: letters indicate the Cantons (AI, Appenzell Innen Rhodes; AG, Aargau; BE, Bern; JU, Jura; LU, Luzern; SG, St Gallen; TG, Thurgau; ZH, Zurich). and numbers indicate individual farms.

With the exception of one isolate (SCCmecIVc), most of the MRSA contained the SCCmecV element (Table 2). The eight MRSA isolated in 2009 belonged to five genotypes: ST398-t011-V (n = 3), ST398-t034-V (n = 1), ST398-t1451-V (n = 1), ST49-t208-V (n = 2), and ST1-t2279-IVc (n = 1); each isolate displayed a different resistance profile (Table 2). In 2010, only three clonal lineages were identified with ST398-t034-V (n = 17) and ST49-t208-V (n = 5) being the most common, whereas ST398-t011-V was only detected once (Table 1 and Table 2).

MRSA isolated from a same slaughterhouse in 2009 and 2010 displayed a different genetic profile and were isolated at different days, except for some MRSA ST398-t034 and ST49-t208 isolates. They were also detected several times from samples from the same slaughterhouses at different days [slaughterhouses A, C, D (ST398-t034)] as well as during the same sampling day in 2010 [slaughterhouses A (ST398-t034) and G (ST49-t208)]. All samples originated from pigs raised in different husbandry (Table 1 and Table 2).

Resistance profiles

Isolates belonging to the most commonly detected genotype, ST398-t034-V, shared the same resistance profile; however, one isolate was susceptible to streptomycin and did not contain the streptomycin adenyltransferase gene str (Table 2). Otherwise, ST398-t034-V isolates showed resistance to ß-lactams specified by mecA and blaZ, tetracycline [tet(K), tet(M)], macrolide-lincosamide-streptogramin B (MLS_B) antibiotics [erm(A)], spectinomycin [ant(9)-Ia], streptomycin [str], trimethoprim [dfr(G)], and tiamulin. The tiamulin resistance could not be attributed to either of the five known pleuromutilin-streptogramin A resistance genes vga(A), vga(A)v, vga(B), vga(C), vga(D) [24], or to the cfr rRNAmethylase transferase gene which confers cross resistance to phenicols, lincosamides, oxazolidones, pleuromutilins, and streptgramin A antibiotics [25]. Mutations in the 23 rRNA and rplC gene were not investigated [26].

Different resistance profiles were observed among the ST49-t208-V isolates, which was the second most prevalent lineage. The difference was mainly due to the presence, or absence, of either of the MLS_B resistance genes erm(A) or erm(C) (Table 2). In general, ST49 isolates displayed resistance to ß-lactams [mecA], tetracycline [tet(K), tet(M)], tiamulin [vga(A)v], and sulphonamides (these mechanisms were not characterised). Notably, ST49-t208 isolates were susceptible to trimethoprim and resistant to sulphonamides, whereas ST398-t034 isolates were resistant to trimethoprim and susceptible to sulphonamides.

Different antibiotic resistance profiles were also found among the less frequent isolates, ST398-t011 and ST398-t1451. They all displayed resistance to ß-lactams [mecA] and tetracycline [tet(K), tet(M)]; however, resistance to MLS_B antibiotics [erm(C)], tiamulin [vga(A)v], streptomycin [str], and spectinomycin varied from strain to strain (Table 2). The ST1-t2279-IVc isolate was only detected once and showed resistance to ß-lactams [mecA, blaZ], tetracycline [tet(M)], MLS_B antibiotics [erm(A)], and spectinomycin [ant(9)-Ia]; additionally, a resistance to ciprofloxacin was detected (Table 2). The resistance mechanism to fluoroquinolones which may be due to mutations in the topoisomerase genes [27] was not determined. In contrast, MRSA isolates from 2009 and 2010 were susceptible to vancomycin, gentamicin, kanamycin, chloramphenicol, quinupristin/dalfopristin, fusidic acid, mupirocin, rifampicin, and linezolid; importantly, none of the isolates carried the Panton-Valentine Leukocidin toxin.

Molecular typing and epidemiology

This study was the first to reveal the presence of MRSA ST49 in pigs. MRSA ST49 was already present in 2009, but its prevalence increased in 2010. The presence of the new clonal lineage ST49-t208, which is unique to MRSA from pigs in Switzerland, suggests that the selection occurred within the Swiss pig population. This idea is supported by the fact that MRSA ST49 was detected in seven different pigs from seven different farms slaughtered in three different slaughterhouses, and that ST49 isolates displayed different resistance profiles mainly due to the acquisition of an additional erm(A) or erm(C) gene (Table 2). The spa type t208, which is associated with ST49, was previously found in seven out of 48 methicillin-susceptible S. aureus MSSA from pigs in Switzerland [20], suggesting a possible acquisition of the SCCmec element. ST49 is rarely mentioned in the literature; it was not detected among 2890 MSSA and MRSA from human infections in the 26 EU countries, the 133 S. aureus colonising healthy adults in Switzerland nor within the 572 MSSA isolates from Swiss children [28–30]. ST49 has only been associated with MSSA involved in one human infection in the United Kingdom and in three cases of skin lesions and laryngeal ulceration in wild squirrels [31, 32]. In the latter three cases, the authors mentioned close contact with persons who fed the squirrels as a possible source of infection. Further screening of humans, including people working with food-producing animals, would be necessary to determine if humans also play a role in the dissemination of MRSA ST49, as demonstrated with ST398. Human carriage of ST398 has been widely documented in other European countries, and the exchange of ST398 between humans and pigs has been reported in the Netherlands and Denmark [13, 33, 34]. Furthermore, ST398-t034 was the most predominant lineage found in Swiss pigs, and has been detected in one out of 133 veterinarians in Switzerland [22]. In our study, ST1 was only detected once, although it has been associated with MRSA infections in humans in the United Kingdom and Spain, and in horses in Austria [31, 35, 36]. ST1 represented, however, 15.6% of the MRSA-positive pig finishing holdings in Italy [37]. Cases of transmission of ST1 between humans and cows have also been reported [38].

Transmission of MRSA among pigs during animal transport and at slaughterhouse has been described [39]. In our study, no direct association of STs to a specific abattoir has been observed, except for slaughterhouse A, where ST398-t034 has been predominantly detected in the samples. Transmission which may have occurred at slaughterhouse may not be excluded for this abattoir. Otherwise, MRSA ST398 and ST49 were found in different slaughterhouses at different sampling dates and the pigs originated from different farms.

Increase of MRSA prevalence

This study is also the first to describe MRSA prevalence in slaughter pigs in Switzerland using a sampling plan which was representative of the Swiss slaughter pig population and which was evenly distributed over time. This strategy allowed us to obtain statistically significant and representative data for the evaluation of the chronological distribution of MRSA. Such representative sampling criteria were not considered in the two previous studies reporting MRSA in pigs in Switzerland, resulting in a less accurate estimation of the prevalence over the entire country [20, 22]. However, these studies indicated that the prevalence of MRSA was low among the proportion of pigs analysed before 2009. Our study also showed that the prevalence of MRSA in 2009 was low, and indicated an approximate three-fold increase within one year. The increased diversity of MRSA and the emergence of the clonal lineage ST49 unique to MRSA from Swiss pigs suggested that MRSA-positive pigs imported into Switzerland have not introduced this MRSA. However, Switzerland sporadically imports a low number of breeding pigs, from which MRSA ST398 may have been introduced. In 2010, for example, Switzerland imported pigs from Denmark (n = 10), The Netherlands (n = 4), and Germany (n = 1), where the prevalence of MRSA ST398 is high (Federal Veterinary Office, pers. communication). Like Switzerland, the predominant clone in Denmark and Germany is ST398-t034, whereas MRSA ST398-t011 is the predominant in The Netherlands [40–42].

Sampling

Representative samples were taken in accordance with the framework of a national monitoring program on antimicrobial resistance in food animals [23]. A minimum required sample size of 382 randomly selected animals was calculated with the assumptions of an infinite population size, a prevalence of 5%, a desired confidence level of 97.5% and an accuracy of 2.5%. The samples were randomly taken at the nine biggest abattoirs (A to I), where over 85% of the pigs in Switzerland were slaughtered. Only one sample was taken per animal holding. The number of samples in the sampling frame collected from each slaughterhouse was proportional to the number of pigs slaughtered at each establishment per year. One to 8 samples were taken per sampling day and abattoir. The samples collected on Mondays and Tuesdays were equally distributed over the years. Based on these data, random sampling plans were conducted, resulting in a total of 405 nasal swabs of fattening pigs in 2009 and 392 swabs from pigs in 2010.

Isolation of MRSA

Samples were taken using transport swabs (Oxoid Ltd, Basingstoke, England) from the nares of the pigs subsequent to stunning by officials of the Swiss abattoir authorities and immediately after sampling, were transported to the laboratory without cooling. After arrival, swabs were transferred into tubes containing 10 ml Mueller Hinton Broth supplemented with 6.5% NaCl and incubated aerobically at 37°C for 24 h under agitation. One ml from this pre-enrichment was inoculated into 9 ml tryptone soy broth containing 3.5 mg/L cefoxitin and 75 mg/L aztreonam, and further incubated aerobically at 37°C for 24 h. A loopful was then spread onto MRSA selective agar plates (BBL ™ CHROMagar ™ MRSA; Becton Dickinson, Franklin Lakes, NJ), which were incubated at 37°C for 24 h. Pink to mauve-colored colonies were regarded as suspicious and five presumptive colonies were cultivated onto tryptone soy agar plates containing 5% sheep blood (TSA-SB) (Oxoid Ltd, Basingstoke, England) at 37°C for 24 h. S. aureus was identified using Vitek 2 with Gram-Positive (GP) cards (BioMérieux, Mary l'Etoile, France) following manufacturer's recommendations.

Molecular typing and antibiotic resistance

The minimal inhibitory concentration (MIC) of the antibiotics was determined by broth microdilution in Müller-Hinton using the Sensititre susceptibility plate EUST (Trek Diagnostics Systems, East Grinstead, England; MCS Diagnostics BV, Swalmen, The Netherlands), except for spectinomycin, which was tested using homemade microbroth dilution plates. The MIC was interpreted according to the European Committee on Antimicrobial Susceptibility Testing (EUCAST) guidelines http://www.eucast.org using clinical resistance breakpoints for cefoxitin (> 4 mg/L), chloramphenicol (> 8 mg/L), clindamycin (> 0.5 mg/L), ciprofloxacin (> 1 mg/L), erythromycin (> 2 mg/L), fusidic acid (> 1 mg/L), gentamicin (> 1 mg/L), linezolid (> 4 mg/L), penicillin (> 0.125 mg/L), quinupristin/dalfopristin (> 2 mg/L), rifampicin (> 0.5 mg/L), tetracycline (> 2 mg/L), trimethoprim (> 4 mg/L), and vancomycin (> 2 mg/L). Otherwise, resistance breakpoints were tentatively derived from epidemiological MIC cut off values from EUCAST for kanamycin (> 8 mg/L), mupirocin (> 1 mg/L), streptomycin (> 16 mg/L), spectinomycin (> 128 mg/L), sulfamethoxazole (> 128 mg/L), and tiamulin (> 2 mg/L). Antibiotic resistance genes, including mecA, were detected using a microarray [45]. SCCmec types were determined by multiplex Polymerase Chain Reaction PCR assays [46]. Sequence types (ST) were determined using multilocus sequence typing (MLST) [31]. Spa type was determined as previously described and analysed using the Ridom StaphType software (Ridom StaphType, Ridom GmbH, Würzburg, Germany) [47]. The presence of the Panton-Valentine Leukocidin operon lukS-lukF was determined by PCR as described previously [48].

Statistical Analysis

Sampling calculations were performed using the Win Episcope 2.0 software http://www.clive.ed.ac.uk/winepiscope. The difference in the prevalence between the years was analysed by Chi-square test. The level of significance was 5%. All statistics were performed using the NCSS 2007 statistical software (Hintze J. NCSS (2008) Version 07.1.8. Kaysville, UT, USA).