Article Text


Transmission of MRSA CC398 strains between pig farms related by trade of animals
  1. C. Espinosa-Gongora, DV1,
  2. E. M. Broens, DVM, Ph3,
  3. A. Moodley, MSc, Ph1,
  4. J. P. Nielsen2 and
  5. L. Guardabassi, DVM, Ph1
  1. Department of Veterinary Disease Biology, University of Copenhagen, Frederiksberg C, Denmark
  2. Department of Large Animal Sciences, University of Copenhagen, Frederiksberg C, Denmark
  3. Quantitative Veterinary Epidemiology Group, Wageningen Institute of Animal Sciences, Wageningen University, The Netherlands
  1. E-mail for correspondence ceg{at}

Meticillin-resistant Staphylococcus aureus (MRSA) clonal complex (CC) 398 is a genetic lineage associated with livestock, especially pigs. The authors investigated the role of pig trade in the transmission of MRSA CC398 between farms using pulsed-field gel electrophoresis (PFGE), a highly discriminatory method for strain typing. PFGE analysis of 58 MRSA isolates from a retrospective study in the Netherlands and a prospective study in Denmark provided molecular evidence that the strains present in five of the eight recipient farms were indistinguishable from those occurring in the corresponding supplying farm. The molecular typing data confirm the findings of a previous risk-analysis study indicating that trading of colonised pigs is a vehicle for transmission of MRSA CC398.

Statistics from

THE livestock-associated clone of meticillin-resistant Staphylococcus aureus (MRSA) clonal complex (CC) 398 is widespread in pig farms and represents a potential hazard to farm workers and other people exposed to pigs (van Loo and others 2007, van Cleef and others 2010). The transmission routes of MRSA CC398 in the pig production system are largely unknown. A recent Dutch study by Broens and others (2011) showed that 79 per cent of the farms with a MRSA-positive supplier were MRSA positive, while 23 per cent of herds with a MRSA-negative supplier were MRSA positive. This indicates that purchasing pigs from a positive farm may be an important risk factor for acquisition of MRSA although other transmission routes may exist. The current typing approach used for subtyping of MRSA CC398 include spa typing alone or in combination with Staphylococcal cassette chromosome mec (SCCmec) typing (van Duijkeren and others 2008, Broens and others 2011). However, this approach is not sufficiently discriminatory for studying transmission between farms as only a single locus is sequenced. The objective of this study was to study whether the MRSA CC398 strains found in pig production farms were the same as those in pig suppliers. For this purpose, MRSA isolates from farms linked by commercial trade of pigs were typed using pulsed field gel electrophoresis (PFGE), a highly discriminatory method allowing differentiation between farm-specific lineages of MRSA CC398 (Espinosa-Gongora and others 2011). The present study hypothesis was that PFGE types found in the recipient farms would also be present in the corresponding supplying farms.

Material and methods

Fifty-seven isolates of MRSA CC398 were included in the study, which was structured in two parts: a retrospective study in the Netherlands and a prospective study in Denmark.

In the retrospective study, the authors analysed 21 isolates from a previous epidemiological study in the Netherlands (Broens and others 2011). These isolates originated from 14 farms organised in six chains: Chain F (six isolates), Chain G (two isolates), Chain P (six isolates), Chain I (two isolates), Chain Q (three isolates) and Chain R (two isolates). In the original study, a chain was defined as a group of farms related to each other by trade of animals. Each chain contained farms at two or three levels of the production pyramid, including breeders, which supply gilts to farrowers; farrowers, which supply pigs to finishers and finishers, which produce pigs for slaughter. Some farms were categorised into two of the production levels, for example, breeder/farrower. Farms were included in the study by Broens and others (2011) if they had a maximum of two pig suppliers which had not changed within a year. The authors analysed a total of 21 isolates comparing the PFGE profiles of MRSA isolated from farms situated at the bottom of the production pyramid (finishers) with those from farms situated at the upper levels (farrowers and breeders).

The prospective study was a follow-up to a previous study in Denmark in 2009, where farm-specific PFGE profiles of MRSA CC398 were observed in six farrower/finisher farms (Espinosa-Gongora and others 2011). Two out of the six farms accepted to participate in the present study. Each of the farms (Farm 1 and Farm 2) received gilts from a single pig supplier during the last four years. In August 2011, the authors visited the farms to take individual samples from newly purchased gilts on the transport truck. The authors compared the PFGE profiles of MRSA isolated from gilts to those of MRSA isolated from animals within the farms, including strains isolated in the present study and representative strains of the farm-specific PFGE profiles observed in 2009. Only one isolate per pig was included in the study. The authors analysed a total of 41 MRSA isolates organised in two chains: Chain 1 (six isolates from Farm 1 and 12 isolates from gilts at the transport truck) and Chain 2 (four isolates from Farm 2 and 19 isolates from gilts at the transport truck).

All isolates originated from nasal swabs, were isolated by selective enrichment (Mueller Hinton broth 6.5 per cent NaCl) and media (MRSA Brilliance Agar), were identified as MRSA CC398 by a multiplex PCR for sau1-hsdS1 and mecA (Stegger and others 2011) and typed by a CC398-specific PFGE protocol using Cfr9I (Bosch and others 2010). The GelCompar II software (Applied Maths) was used for the PFGE cluster analysis (unweighted pair group method with arithmetic mean (UPGMA) based on the Dice similarity coefficient, with optimisation and position tolerance set 0.1 per cent and 1.0 per cent, respectively). Epidemiological relationship between isolates was assessed using the Tenover criteria (Tenover and others 1995). In addition, spa typing was done in the Danish isolates and re-done in the Dutch isolates to confirm the results shown by Broens and others (2011), according to Harmsen and others (2003).


Five of the eight recipient farms under study harboured MRSA CC398 strains that were indistinguishable from those occurring in their supplying farms (Table 1).

Table 1

Occurrence of indistinguishable MRSA CC398 strains in farms related by trade of animals in the Netherlands (retrospective study) and in Denmark (prospective study). The retrospective study compared the PFGE profiles of 21 isolates obtained from six chains of farms related by pig trade. The prospective study compared the PFGE profiles of 41 isolates from two farms and newly purchased gilts sampled at the transport truck before entering the farms

Retrospective study

Within chains F, P, Q and R, MRSA CC398 isolates from farms situated at the bottom of the production pyramid (finishers) were not distinguishable from the isolate(s) from at least one farm situated at the upper levels (farrowers and breeders) (Fig 1). All isolates within Chains F and R displayed the same PFGE profiles. In Chain P, two different (at least three genetic changes) PFGE profiles were present at both production levels. Isolates from recipient and supplying farms within Chains G and I displayed genetically unrelated PFGE profiles. The PFGE profile displayed by one of the isolates from the breeder/farrower in Chain Q did not cluster with the other two isolates from that chain but was identical to the isolates from Chain R. The four spa types identified in the study by Broens and others (2011) were confirmed to be t011, t108, t943 and t2503. The association between spa types and PFGE profiles was not consistent since the six isolates from Chain F had indistinguishable PFGE profiles but belonged to three different spa types (Fig 1).

Fig 1

Genetic diversity of 21 MRSA CC398 isolates from six pig farms in The Netherlands (NL) and 36 MRSA CC398 isolates from two pig farms in Denmark (DK) at the end of the production chain (finisher or farrower/finisher) and their corresponding gilt or pig suppliers (breeder or farrower respectively, or breeder/farrower if the farm sells both gilts and pigs). From left to right, the information in the columns indicate the PFGE profile (I-XI), identification of the production Chain (F, G, I, P, Q, R, 1, and 2), farm type (breeder, farrower, breeder/farrower, finisher or farrower/finisher), spa type and the number of isolates displaying the same PFGE profile

Prospective study

Isolates from the study in 2009 and the present study in Farm 1 clustered together (PFGE profile IV in Fig 1). At this chain, MRSA isolates from the purchased gilts (profiles VIII and XI) were not genetically related to the farm isolates (at least three genetic changes) according to the Tenover criteria (Tenover and others 1995). On the contrary, all typeable isolates from Chain 2 displayed indistinguishable or closely-related PFGE pattern (profiles IX and X in Fig 1). Four gilt isolates from Chain 1 and one farm isolate from Chain 2 were not typeable by PFGE. All Danish isolates except two from Chain 1 (t1928 and t011) were spa type t034.


The isolation of indistinguishable strains from farms related by animal trade (Chains F, P, Q and R in the Dutch study and Chain 2 in the Danish study) suggests that the strains found in the recipient farms originated from the supplying farms. These molecular typing data support the findings of the risk-analysis study by Broens and others (2011), which illustrated a large risk associated with purchasing pigs from MRSA-positive farms. On the contrary, three out of the 10 MRSA-positive herds purchasing pigs from MRSA-positive herds (Chains G and I in the Dutch study and Chain 1 in the Danish study) were not epidemiologically linked on the basis of PFGE typing of MRSA CC398 isolates. Furthermore, indistinguishable strains were isolated from farms not related by pig trade (Chains F, G and P; Chains G and I; and Chains Q and R). These findings suggest that the role played by alternative transmission routes may be even higher than that foreseen by Broens and others (2011), who observed that 23 per cent of MRSA-positive farms had a supplier with MRSA-negative status and 46 per cent of farms without supplier were MRSA positive. However, the results obtained in the retrospective study should be interpreted carefully since only one isolate per farm was available for typing and therefore the study was unable to detect the presence of multiple unrelated strains within a farm. Although another study in which higher numbers of isolates representative of six Danish farms were typed using the same PFGE protocol has shown a high degree of homogeneity among isolates originating from the same farm (Espinosa-Gongora and others 2011), it is possible that a second MRSA strain occurred in the recipient and/or the supplying farm. It is also possible that MRSA-positive farms harbouring a strain that was not found in the MRSA-positive supplier (Chains G and I) had acquired the strain from another supplier in the past.

The Cfr9I PFGE protocol described by Bosch and others (2010) was confirmed to be a useful tool for studying the epidemiology of MRSA CC398. PFGE has been indicated as a better genotyping tool than spa typing for epidemiologic assessment due to its higher discriminatory power based on the numerical index of discrimination of both techniques (Argudín and others 2010). The present study confirms the higher discriminatory power of PFGE (11 genotypes detected) over spa typing (five genotypes detected). This is only apparently in contrast with our observation that identical PFGE profile may occasionally be seen among MRSA CC398 isolates displaying different spa types (Chain F, Fig 1). It should be noted that the three spa types (t108, t943 and t2503) that were found associated with the same PFGE profile in Chain F (Fig 1) only differ by a single deletion or duplication of repeat r25 in the region of the spa gene. Sporadic lack of association between PFGE profiles and spa types has previously been observed in MRSA CC398 isolated from dairy mastitis (Fessler and others 2010).

In conclusion, the present study supports the conclusions by Broens and others 2011 by providing molecular evidence that MRSA CC398 can be transmitted by pig trade. This notion has important implications for the development of intervention strategies for control and prevention of MRSA CC398 in pig farming. However, the study also suggests that analysis of alternative transmission routes is needed to identify other risk factors and ultimately to design effective measures to control transmission of MRSA CC398 between farms.


This study was co-financed by the EU-HEALTH project PILGRIM (223050) of the Seventh Framework Programme (FP7), the Research School for Animal Production and Health (RAPH) and the Faculty of Health and Medical Sciences, University of Copenhagen.


View Abstract


  • E. M. Broens is also at the Centre for Infectious Disease Control Netherlands, National Institute for Public Health and the Environment, Bilthoven, The Netherlands

  • Provenance not commissioned; externally peer reviewed

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.