Occurrence of methicillin-resistant Staphylococci in surgically treated dogs and the environment in a Swedish animal hospital


  • A. Bergström,

    1. University Animal Hospital, Uppsala, Sweden
    Search for more papers by this author
  • C. Gustafsson,

    1. Department of Biomedical Sciences and Veterinary Public Health, Division of Bacteriology and Food Safety, Swedish University of Agricultural Sciences, Uppsala, Sweden
    Search for more papers by this author
  • M. Leander,

    1. Department of Biomedical Sciences and Veterinary Public Health, Division of Bacteriology and Food Safety, Swedish University of Agricultural Sciences, Uppsala, Sweden
    Search for more papers by this author
  • M. Fredriksson,

    1. Department of Biomedical Sciences and Veterinary Public Health, Division of Bacteriology and Food Safety, Swedish University of Agricultural Sciences, Uppsala, Sweden
    Search for more papers by this author
  • U. Grönlund,

    1. Department of Animal Health and Antimicrobial Strategies, National Veterinary Institute (SVA), Uppsala, Sweden
    Search for more papers by this author
  • G. Trowald-Wigh

    1. Department of Biomedical Sciences and Veterinary Public Health, Division of Bacteriology and Food Safety, Swedish University of Agricultural Sciences, Uppsala, Sweden
    Search for more papers by this author



To investigate whether hospitalised dogs treated surgically may become culture positive for methicillin-resistant Staphylococcus pseudintermedius or methicillin-resistant Staphylococcus aureus.


Surgically treated dogs (n=45) were sampled for methicillin-resistant Staphylococcus pseudintermedius or methicillin-resistant Staphylococcus aureus on admission, before and after surgery and at the time of removal of surgical stitches. The hospital environment (n=57), including healthy dogs in the veterinary hospital environment (n=34), were sampled for methicillin-resistant Staphylococcus pseudintermedius or methicillin-resistant Staphylococcus aureus. Genetic variations among methicillin-resistant Staphylococcus pseudintermedius or methicillin-resistant Staphylococcus aureus isolates were identified through detection of restriction fragment polymorphisms.


No dogs developed a wound infection due to methicillin-resistant Staphylococcus pseudintermedius or methicillin-resistant Staphylococcus aureus. However, there was a significant increase in the number of dogs carrying methicillin-resistant Staphylococcus pseudintermedius after hospitalisation compared to admission (P<0·001). No methicillin-resistant Staphylococcus aureus was isolated from dogs, but was present in the environment. Methicillin-resistant Staphylococcus pseudintermedius isolates were recovered from environmental surfaces and hospitalised animals, but not from healthy dogs. Methicillin-resistant Staphylococcus pseudintermedius isolates representing nine different restriction endonuclease digestion patterns were found, with two of these occurring in both the environment and on dogs.

Clinical Significance

Dogs may contract methicillin-resistant Staphylococcus pseudintermedius in association with surgery and hospitalisation. Resistant bacteria may be transmitted between dogs, staff and the environment. Dogs colonised with methicillin-resistant Staphylococcus pseudintermedius may be a source for hospital- and community-acquired infections.


Staphylococcus pseudintermedius and Staphylococcus aureus commonly colonise the skin, mucous membranes, urogenital tract and occasionally alimentary tract in dogs and humans (Harvey and Noble 1998, Kluytmans and Wertheim 2005). Coagulase-positive staphylococci are the most common cause of surgical site infections (SSIs) in dogs (Vasseur and others 1988) as well as in humans (Dohmen 2008).

Methicillin-resistant S. pseudintermedius (MRSP) and S. aureus (MRSA) are microorganisms known to cause complicated infections, because of difficulties in finding effective antimicrobial treatment. In dogs, MRSP is a significant pathogen (Weese and van Duijkeren 2010, van Duijkeren and others 2011) and has occasionally been found in humans (Chuang and others 2010, Stegmann and others 2010). Hospitalisation and antibiotic treatment are risk factors for MRSP carriage in dogs (Nienhoff and others 2011). Infection with MRSA is associated with previous antimicrobial treatment, number of days in a veterinary hospital, surgical implants and contact with hospitalised humans (Soares Magalhaes and others 2010). SSIs caused by MRSA have been reported in dogs (Tomlin and others 1999, Rich and Roberts 2006, Grinberg and others 2008). A possible source of methicillin-resistant staphylococci is environmental surfaces in small animal veterinary clinics (Heller and others 2009, Ishihara and others 2010).

The main purpose of this study was to investigate whether hospitalised dogs who undergo surgery may become culture positive for MRSP or MRSA. The occurrence of both MRSP and MRSA within the animal hospital environment and MRSP and MRSA colonisation in healthy dogs (personnel dogs and colony dogs) at the animal hospital were investigated during the same period.

Materials and Methods

Animal welfare

This study was approved by Uppsala Ethical Committee, Uppsala, Sweden. The Board of Agriculture approved the use of client owned dogs in this study. Dog owners provided written consent before dogs were enrolled in the study. Dogs were sampled using methods with limited invasiveness using Copan swabs (25125 Brescia, Italy) in Amies media.

Sampling of surgery patients

To be enrolled in the study, dogs were required to meet several criteria: (1) be admitted for surgery, (2) the type of surgery would require overnight stay at hospital and (3) the surgery would generate sutures requiring removal 10 to 12 days postoperatively to allow for bacteriological sampling. Dogs admitted to the University Animal Hospital for surgical treatment and hospitalisation were sampled during summer and autumn 2009. Dogs treated with odontologic surgery and dogs presenting with infected surgical sites were excluded. There were no further classifications of the surgeries into clean, clean-contaminated and contaminated. Information regarding antimicrobial therapy was collected from the medical records including anamnesis. Antimicrobial treatment was not an exclusion criterion for the study.

Perioperative antimicrobial therapy was used in surgeries with prolonged surgical time (more than 90 minutes), if implants were used, or if the animal had known or perceived decreased immunocompetence (e.g. prior immunosuppressive therapy, Cushings disease). The antimicrobial treatment was started before incision, after induction of anaesthesia, but was not continued after surgery.

Samples for determining if a dog was colonised with MRSP and MRSA were taken at admission to the hospital. The samples were taken from the sites recommended by the National Board of Health and Welfare for sampling humans for MRSA, (Socialstyrelsen 2010) with some modifications. With one swab gingival surface of the mouth, the left nostril and perianal region (inserted 0·5 cm into the anus) were sampled, and labelled as carrier sample 1 (CS1).

After anaesthetic induction, the surgical area was shaved. In soft tissue surgeries, the initial cleaning was with soap and water; if the surgery was orthopaedic, the soap was exchanged for 4% chlorhexidine-gluconate solution (Hibiskrub 40 mg/mL, Regent Medical, Manchester, UK). The washing took at least 4 minutes. Thereafter, a surgical wash with 5% chlorhexidine gluconate and ethanol was used. The surgical area was allowed to dry before the first surgical sample (S1) was collected. Then, the dog was transported to the operating theatre for a final sterile surgical wash with 5% chlorhexidine and ethanol before being draped. When the surgery was completed, a second sample [called surgery sample 2 (S2)] was collected from the surgical site by swabbing along the edges of the wound and around the sutures before the drapes were removed. On the day of discharge, a swab was taken from the surgical site along the edges of the wound and denoted surgery sample 3 (S3).

At 10 to 12 days postoperatively, the dog returned for removal of the stitches. The surgical site was again swabbed before stitch removal [surgery sample 4 (S4)] and a second carrier sample (CS2) was taken to determine whether the dog had become a carrier since the first sample was taken. Six owners who did not plan to return to the animal hospital for stitch removal received a sampling kit, instructions, requests and a prepared envelope so that the referral clinic or owner removing the stitches could sample the dog.

Samples taken at the University Animal Hospital were cultured within 24 hours from sampling and samples from outside the hospital were cultured within 36 hours from sampling.

If dogs were confirmed positive for MRSP or MRSA, new samples were taken to evaluate if the colonisation of the methicillin-resistant bacteria remained (CS3). The CS3 samples were taken two to five months postoperatively, and in one dog, a fourth sample was taken six months postoperatively (CS4) until the animal was found to be negative for MRSP culture. A summary of the different sampling steps is presented in Fig 1.

Figure 1.

Flow chart of the bacterial sampling of surgically treated dogs in the study

Sampling of healthy dogs in the animal hospital environment

Fourteen pet dogs of different breeds and ages, owned by the animal hospital personnel, and 20 beagle dogs belonging to the research dog colony at the Department of Clinical Sciences were included in the study. All dogs were housed close to the animal hospital ward and were walked in the same environment, therefore considered useful indicators of bacterial contamination from the environment. These dogs were sampled once in the same manner as surgically treated dogs (CS).

Surgery patients and healthy dogs

Healthy dogs and hospitalised dogs in this study shared the same environment, and were walked in the same corridor but lived in different rooms. The beagle dogs were mainly cared for by dedicated staff and the healthy pet dogs by their owners who comprised part of the hospital staff. The hospitalised dogs were nursed by staff and the dogs lived in separate cages but shared a room with other patients.

Sampling of animal hospital environment

Fifty-seven environmental samples were taken for MRSP/MRSA culture during the same time frame as the rest of this study (Table 2). The environment was divided into animal and human contact surfaces, depending on the most likely sources of potential contamination. Animal contact areas sampled were in areas that the dogs came into contact with during their hospital stay. Human contact surfaces were areas that the staff frequently touched. Sterile cloths (Sodibox, Névez, France), premoistened with a buffered peptone solution with a 10% additive of a neutralising substance (lecitine, Tween 80, L-histidine and sodiumthiosulphate) to reduce the effect of disinfectants, were rubbed thoroughly on the sampling site. On sampling, sterile gloves were worn to prevent the samples from becoming contaminated by study personnel.

Bacterial isolation and phenotypic identification

Swab samples from dogs were placed in tubes containing 10 mL Tryptone Soy Broth (TSB) (Oxoid Ltd, Basingstoke, UK) to which 4% NaCl, 1% mannitol, 16 µg/mL phenol Red, 100 μL/mL (1 µg/mL) cefoxitin (Sigma-Aldrich, Stockholm, Sweden) and 100 μL/mL (50 µg/mL) aztreonam (Bergman-Labora, Danderyd, Sweden) were added. The tubes were incubated for 48 hours at 37°C.

Environmental sampling cloths were placed in individual stomacher bags containing 100 mL Mueller Hinton Broth (Becton, Dickinson & Company, Le Pont de Claix, France) with an additive of 6·5% sodium chloride, and blended for 1 minute in a laboratory paddle blender (Stomacher®-80 Biomaster lab system, Seward Ltd., Worthing, UK) (Nilsson and others 2009). Each sample was incubated in the stomacher bag for 24 hours at 37°C, 1 mL of the sample was transferred to a tube containing 9 mL TSB and incubated as described above for swab cultures.

Following incubation in TSB, all samples were processed in the same manner. From the TSB culture, 10 μL was inoculated on bovine blood agar (Oxoid Ltd, Basingstoke, UK) and 10 μL on mannitol salt agar (MAST) with lithium chloride (LiCl) (MAST DM 160, MastGroup Ltd, Merseyside, UK). The plates were incubated for 24 hours and further for 24 hours at 37°C and then examined for colonies morphologically consistent with S. aureus and S. pseudintermedius (Bannerman and Peacock 2007). Five colonies were selected from each sample and recultured onto bovine blood agar for 24 hours at 37°C. A tube coagulase test (Håtunaholm, National Veterinary Institute, Uppsala, Sweden) was performed and coagulase-positive strains were selected for phenotypic species identification using the DNase-test, aerobic acid production from maltose in bromcresol-purle agar (Difco Purple agar base, Becton, Dickins & Company, Le Point de Claix, France; Maltos+monohydrat, Merck KGaA, Darmstadt, Germany), and trehalose broth (LabLemco broth with trehalose number: 321180, National Veterinary Institute, Uppsala, Sweden) (Devriese and others 2005, Bannerman and Peacock 2007).

Antimicrobial susceptibility test

Antimicrobial susceptibility testing was performed by broth microdilution with microdilution panels (VetMIC™ MRS-detection plate 395119, National Veterinary Institute, Uppsala, Sweden). Staphylococcus pseudintermedius and S. aureus were tested for oxacillin resistance. For S. aureus the resistant breakpoint of >2 mg/L for oxacillin was used according to the standards of the Clinical and Laboratory Standards Institute guidelines, document M100-S19 (CLSI 2007) and for S. pseudintermedius the resistance breakpoint of >0·5 mg/L for oxacillin was used as approved by the CLSI subcommittee on Veterinary Antimicrobial Susceptibility Testing (Bemis and others 2009, Schissler and others 2009). Resistant isolates underwent further molecular identification with polymerase chain reaction (PCR).

DNA isolation and PCR amplification

Bacterial DNA was extracted from overnight cultures (Capurro and others 2009) and the isolated DNA was preserved at −20°C until use. A multiplex PCR, targeting two species-specific fragments of the nuc gene encoding thermonuclease, was used to distinguish S. aureus from S. pseudintermedius (Baron and others 2004). A conserved region of the Staphylococcus genus-specific 16S rRNA gene was included in the assay as an internal positive control. Detection of the mecA gene was determined by PCR, as described previously (Strommenger and others 2003). After amplification, the size of the PCR products was determined by gel electrophoresis.

Pulse-field gel electrophoresis and spa-typing

Genomic DNA from all MRSP and MRSA isolates was digested with SmaI (New England Biolabs GmbH, Frankfurt am Main, Germany), then separated by pulse-field gel electrophoresis (PFGE) using the standardised HARMONY protocol (Murchan and others 2003) and the pulse switch times recommended elsewhere (Perreten and others 2010). Each gel included SmaI-digested genomic DNA from S. aureus reference standard NCTC 8325 which provided an internal control for the procedure. The restriction fragments from S. aureus NCTC 8325 together with a low lamda ladder and the universal size marker Salmonella serotype Braenderup (Hunter and others 2005) were also used on every gel as they provided a consistent pattern by which samples from different isolates separated on different gels could be compared. Gels were analysed by visual interpretation of the banding patterns according to the criteria of Tenover and others (1995), including DNA fragments between approximately 36 and 650 kb (Tenover and others 1995).

MRSA strains were characterised further through the detection of polymorphisms of tandem repeats within the protein A gene, i.e. spa-typing (Frenay and others 1994). PCR amplification and subsequent analysis of products were done as described elsewhere (Harmsen and others 2003).

Statistical analysis

Results are presented as mean ±standard deviation using JMP 8 (SAS Institute, Cary, NC, USA). McNemar's chi-square test was used for repeated measurement analysis.


Clinical results

Forty-five surgically treated dogs (aged 0·5 months to 12·5 years, mean 5.9 ±3.5 years) were included in the study. Thirty dogs underwent elective surgeries and 15 dogs underwent emergency surgeries (with pyometra being the most common diagnosis). All dogs in this study were negative for MRSA.

At admission (CS1), one dog (case 4) was MRSP positive (2·2%). Surgery samples S1 and S2 were negative for MRSP. At discharge one dog (case 4) was positive for MRSP in the surgical wound (S3) (Table 1). At the time of stitch removal, 39 dogs were sampled at the University Animal Hospital and 6 dogs were sampled by their owners. Five dogs (12·8%) were MRSP positive in carrier sample CS2 and one dog (case 6) (2·6%) was positive for MRSP at the surgical site (S4) (Table 1). This dog had received antimicrobial therapy because of SSI, caused by Staphylococcus schleiferi.

Table 1. Presentation of the MRSP-positive dogs including age (years), surgical procedure, HL (number of days), occurrence of AMT, positive samples and PFGE pattern
Dog NoAgeSurgeryHLAMTPositive samplePFGE pattern
  1. HL Hospitalisation length, AMT Antimicrobial therapy, PFGE Pulse-field gel electrophoresis, CS Carrier sample 1 to 3,S Surgery sample 1 to 4

  2. aAmoxicillin with clavulan treatment for 10 days, started 2 months before surgery
  3. bTibial plateau levelling osteotomy
  4. cFirst-generation cephalosporins
  5. dTibiotarsal arthrodesis
  6. eAmoxicillin 10 days until the day before surgery
  7. fAmoxicillin started seven days postoperatively
19Mammary tumour12 months preopaCS2D
35TT arthrod1PeriopcCS2G
42TPLO1PeriopbCS1, CS2, CS3, S3E
511Mammary tumour1PreopeCS2H
69Mammary tumour1PostopfS4, CS3A&I

Altogether, six dogs were MRSP positive in one or several samples and these dogs were sampled again two to five months after surgery (CS3). Case 4 was still positive for MRSP at CS3 (Table 1), but six months postoperatively the dog was MRSP negative (CS4). The dog had no signs of complications or infection at the surgical wound during the study time. Case 6 was positive for CS3 two months after surgery, but was MRSP negative three months postoperatively. There was a significant increase in the number of MRSP colonised dogs after hospitalisation, at the time of stitch removal (5/45) compared to the time of admission (1/45) (P<0·001).

All MRSP-positive dogs were treated with antimicrobial therapy, three of them only had perioperative treatment, one was treated only preoperatively, one was treated two months before surgery and one dog was treated postoperatively (Table 1). Of the 39 MRSP-negative dogs, 20 dogs were treated with antimicrobial therapy, 11 of which were treated perioperatively only, 4 dogs were treated perioperatively and postoperatively. One dog was treated from the third day after surgery because of an SSI with P. aeruginosa, and this was the only MRSP-negative dog developing a SSI. One dog was treated preoperatively and perioperatively with antimicrobials. Three dogs were treated before, during and after surgery (urinary tract infection, abdominal abscess and bite wound with P. aeruginosa infection).

Healthy dogs and the animal hospital environment

The healthy pet dogs and research colony dogs were all negative for MRSP and MRSA.

Thirteen of the 57 environmental samples (22·8%) were positive for S. pseudintermedius and 22 (38·6%) were positive for S. aureus. Ten isolates (17·5%) were confirmed as MRSP and three isolates were confirmed as MRSA (5·3%). MRSP as well as MRSA isolates were retrieved from both animal and human contact areas. A list of the MRSP- and MRSA-positive sample sites is presented in Table 2.

Table 2. Environmental sampling sites positive for MRSA and MRSP and PFGE pattern of the isolates
Environmental sampling sitesAnimal/human contact surfaceMRSAMRSPPFGE pattern
  1. a

    MRSA Methicillin-resistant S Staphylococcus aureus, MRSP Methicillin-resistant

  2. b

    S. pseudintermedius, PFGE Pulse-field gel electrophoresis

  3. c

    1 m2

  4. d

    The whole carpet 1×1/2 m

  5. e

    4×1 m2 including chair and table legs

  6. f

    Light switches, door handles, lamp handels, taps

  7. g

    Dogs lying areas 3×1 m2

  8. h

    Human contact areas in corridor connecting different wards

  9. i


  10. j

    Cage floor, lower part of grid

Entrance floorcAnimal +A
Waiting room, carpetdAnimal +A
Waiting room, flooreAnimal +A
Intensive care§Human +C
Intensive care, floorgAnimal+ J
On call theatre§Human+ J
Orthopaedic theatre§Human+ J
ScaleAnimal +A
Isolation ward§Human +B
CorridorhHuman +C
Dermatology roomfHuman +B
KeyboardiHuman +A
Anaesthesia recovery roomjAnimal +D

PFGE and spa-typing

The MRSA isolates displayed identical PFGE patterns, named J (Table 2) and belonged to spa-type t002. The MRSP isolates could be divided into nine different PFGE patterns, named A-I (Tables 1 and 2). Within each group the pattern was indistinguishable.

Among the environmental samples positive for MRSP (n=10), there were four different PFGE patterns (Table 2). The MRSP-positive isolates collected from dogs had seven different PFGE patterns (Fig 2). MRSP isolated at different times from case 4 had indistinguishable PFGE patterns. The S4 isolate from case 6 was indistinguishable when compared to the environmental isolates from five different locations in the entrance and waiting room areas and keyboards (Fig 2). However, CS3 from the same dog two months later displayed a different PFGE pattern. The PFGE pattern retrieved from case 1 (pattern D) was consistent with the environmental samples from the recovery room (Tables 1 and 2).

Figure 2.

PULSE-field gel electrophoresis (PFGE) pattern of methicillin-resistant Staphylococcus pseudintermedius (MRSP) strains digested with SmaΙ: Lane 1: low lamda ladder; Lane 2: Salmonella serotype Braenderup; Lane 3: NCTC 8325; Lane 4: entrance floor; Lane 5: waiting room, carpet; Lane 6: waiting room, floor; Lane 7: scale; Lane 8: Keyboard; Lane 9: dog no 6, S4; Lane 10: NCTC 8325; Lane 11: dog no 4, CS1; Lane 12: dog no 4, CS2; Lane 13: dog no 4, CS3; Lane 14: dog no 4, S3; Lane 15: NCTC 8325; Lane 16: Salmonella serotype Braenderup; Lane 17: low lamda ladder. Selected lanes were copied from original gel images and reordered to summarise the results for clarity of presentation, but were not otherwise manipulated


Dogs may be carriers for at least six months after MRSP/MRSA acquisition and therefore constitute a potential cause for spread in the hospital and in the community (Laarhoven and others 2011). The present study was done to help identify potential sources of MRSP/MRSA exposure to dogs. Surgery and hospitalisation have been reported as important factors leading to MRSP carriage of dogs (Nienhoff and others 2011). The results of this study are consistent with that study and revealed a significant increase in the number of MRSP-positive dogs following admission to the animal hospital for surgery. While MRSA may also pose a significant concern for SSI in human patients undergoing surgery or other medical procedures (Coello and others 1997, Harinstein and others 2011), this did not appear to be a significant concern in the present study regarding MRSP. MRSP was cultured from the surgical sites of only two dogs (4%, cases 4 and 6). The wound of case 6 was clinically normal without evidence of bacterial infection/inflammation other than a positive culture. The SSI, caused by S. schleiferi, was healing in case 6 at the time of stitch removal when MRSP was cultured from the wound. The carrier sampling procedure might be a weakness in this study as samples were pooled from three different sites, which might reduce sensitivity. Many sampling regimens are represented, such as carrier sample from the nose and perineum, respectively (Laarhoven and others 2011), pooled samples from nose and pharynx with one swab and one swab from perineum (Nienhoff and others 2011), and swabs from different locations pooled in the same broth (Loeffler and others 2010).

The source of most MRSP isolates obtained from hospitalised dogs in this study is unknown. Only two of the MRSP isolates in the dogs (cases 1 and 6) shared PFGE patterns with isolates from the hospital environment (Fig 2). This finding was surprising; the hospital environment including floors, cages and equipment is considered to be an important source for the spread of both MRSP and MRSA in animals (Portner and Johnson 2010). MRSA of human origin pose a risk for hospitalised animals (Seguin and others 1999). Results of environmental sampling in the present study found widespread distribution of MRSP/MRSA contamination around the hospital, most notably in areas associated with human contact. MRSA environmental isolates are consistent with a human origin; all isolates shared the same spa-type t002 common for humans in Sweden (Smittskyddinstitutet 2009), yet no MRSA isolates were obtained from patient samples. If the hospital was the source of MRSP in the dogs testing positive following surgery, then sampling of additional hospital areas is needed to determine the source of infection. The finding that each animal yielded a MRSP strain with a different restriction fragment pattern suggests that the dogs were exposed to different strains, from different locations within or outside the hospital. The low percentage of animals that developed postoperative SSI (4%) suggests that good biocontainment procedures were in place, yet continued improvement in hand hygiene routines may contribute to a reduction in the infection rate.

Antimicrobial therapy can alter the composition of normal flora (Looft and others 2012). Alterations in microflora may reduce competition for nutrients and allow methicillin-resistant bacteria to colonise; Nienhoff and others (2011) showed an association between antimicrobial treatment and MRSP carriage. In the present study, all dogs testing positive for MRSP had been treated with antimicrobials. Although healthy dogs included in this study shared the same general environment as surgical patients, none of them were MRSP positive. Therefore, antimicrobial treatment should be considered as one potential factor contributing to MRSP isolation from patients, but other factors, including surgical stress, anaesthesia and change in environment, may also contribute to the isolation of MRSP following surgery.

In conclusion, in this study the number of dogs who became MRSP positive significantly increased in dogs after surgery and hospitalisation as compared to that before entrance to the hospital. Although both MRSP and MRSA were identified in the hospital environment, few bacterial isolates from the dogs matched those within the hospital environment, suggesting that MRSP exposure most commonly occurred following the hospital stay, or originated from unknown sites within the hospital or community. Animals colonised with MRSP in connection with hospital stay may harbour methicillin-resistant bacteria for months, increasing the risk of introducing MRSP into veterinary environments. The MRSP colonisation may also cause clinical problems in the dogs, and may contribute to community-acquired infections affecting other animals and introducing potential health problems in humans.


The authors wish to thank both Mikaela Eldh for help with the sampling of the dogs and Lise-Lotte Fernström for performing the bacterial analysis, respectively. Also, the authors extend sincere thanks to Professor Richard Zuerner for assistance with the manuscript. This study was funded by Thure F and Karin Forsberg's Foundation.

Conflict of interest

None of the authors of this article has a financial or personal relationship with other people or organisations that could inappropriately influence or bias the content of the paper.