Article Text

Download PDFPDF

Occurrence of Salmonella, Campylobacter, Clostridium and Enterobacteriaceae in raw meat-based diets for dogs
  1. Josefin Hellgren1,
  2. Lovisa Staaf Hästö1,
  3. Camilla Wikström2,
  4. Lise-Lotte Fernström1 and
  5. Ingrid Hansson1
  1. 1 Department of Biomedical Sciences and Veterinary Public Health, Swedish University of Agricultural Sciences, Uppsala, Sweden
  2. 2 Department of Microbiology, National Veterinary Institute, Uppsala, Sweden
  1. E-mail for correspondence; ingrid.hansson{at}


The practice of feeding raw meat-based diets (RMBD) to dogs has increased in popularity in recent years. However, RMBD are based on offal that has not undergone any type of treatment to reduce the microbial content, so there is a risk of potential pathogenic microorganisms being present. Frozen samples from 60 RMBD packs produced by 10 different manufacturers were analysed for their content of bacteria belonging to the family Enterobacteriaceae, for Clostridium perfringens and for the presence of Salmonella and Campylobacter. Enterobacteriaceae were detected in all 60 samples and in 31 samples exceeded a level of 5000 bacteria/g, which is the threshold for satisfactory microbial hygiene according to EU regulations. In two samples, the amount of C. perfringens exceeded 5000 bacteria/g, which is the maximum level of anaerobic bacteria permitted by Swedish national guidelines. Salmonella species were found in four (7 per cent) and Campylobacter species in three (5 per cent) samples. These results show that it is critical to maintain good hygiene when storing, handling and feeding RMBD, in order to limit the potential health risks to animals and humans, especially young and immunocompromised individuals.

  • bacteriology
  • dogs
  • nutrition
  • zoonoses
View Full Text

Statistics from


The practice of feeding raw meat-based diets (RMBD) to dogs has increased in popularity in recent years. RMBD, which are also called Bones and Raw Food or Biologically Appropriate Raw Food (BARF) or Raw Animal Products (RAP),1 2 do not undergo any form of heat treatment before chilling or freezing, and can be homemade or commercially produced. The main ingredient in RMBD is offal, consisting primarily of uncooked meat, edible bones and organs from ruminants, pigs, poultry, horses, game and fish. There are different opinions about the advantages and disadvantages of RMBD, with some stating that it is a healthier and more natural alternative to other commercial feeds for dogs.2 Proponents of feeding RMBD to pets claim that raw diets featuring fresh, natural ingredients are unequivocally the best nutritional choice. In a survey, owners feeding RMBD to their pets reported important health benefits such as improvements in the immune system, skin and coat, a reduction in dental diseases and a lower incidence of food allergies.2 3

One of the risks associated with RMBD is the microbiological load, since the meat and offal they contain has not undergone any type of treatment to reduce or eliminate the presence of possible pathogenic microorganisms.1 4 Most proponents of RMBDs consider that the bacteria present in the feed do not pose a disease risk for dogs, since their gastrointestinal tract is adapted to raw meat. However, several studies have found that pathogenic bacteria isolated from RMBD are excreted in faeces of dogs with gastrointestinal symptoms to a greater extent than in faeces from clinically healthy individuals.5–8 Colonisation with enteropathogens primarily represents a health hazard for individuals with an impaired immune system, such as dogs kept in stressful environments or housed in animal hospitals or kennels where many individuals are gathered within a limited area, with high infection pressure. Bacteria present in RMBD could also pose a risk of infection to humans during handling and storage of the feed, the dog’s feed bowls, possible contamination of kitchen equipment, and also the dog’s tongue, and faeces.1 9 10Salmonella isolates have been recovered from RMBD and from canine faeces samples.11 12 Dog chews made from animal products, such as dried beef and pig ears, have also been linked to diseases in humans.13 14 The risk is particularly high for people with a compromised immune system, including children, the elderly and immunocompromised individuals.15

Bacterial contamination is more often linked with RMBD than with heat-treated feed,9 10 16 17 and feeding dogs with RMBD has been proven to be a risk factor in the dog excreting certain zoonotic pathogens in the faeces.7 9 18–21 From both a veterinary and public health perspective, it is of great interest to investigate the presence of bacteria in RMBD. Potentially pathogenic bacteria such as Salmonella species, Campylobacter species, Listeria species, Escherichia coli and Clostridium species have been detected in previous studies.4 10 11 15–18 22 Another study has investigated commercial RMBD on the Swedish market for the presence of resistance to extended-spectrum beta-lactamase cephalosporins (ESBL-producing) E. coli.23

The purpose of the present study was to investigate the presence of certain bacteria in RMBD consisting of meat and offal from ruminants and poultry and intended for dogs. Specific objectives were to quantify the level of (i) Enterobacteriaceae and (ii) Clostridium perfringens, markers for faecal contamination and hygienic quality; to document the presence of two species of zoonotic bacteria, (iii) Salmonella and (iv) thermotolerant Campylobacter and (v) to evaluate the presence and amount of bacteria relative to the feed safety standards set by the EU (Commission Regulation (EU) No. 142/2011) and by the Swedish Board of Agriculture (SJVFS 2006:81, updated 2011:40).

Materials and methods


During March to September 2017, samples were taken from 60 packs of frozen RMBD purchased from various stores within a 200 km radius of the laboratory. The inclusion criteria were that (i) the RMBD had not undergone any treatment, such as drying or heat treatment before freezing; (ii) the ingredients in the samples were meat and/or offal from either ruminants or poultry and (iii) the feed was intended for dogs. The RMBD samples contained at least one of the following ingredients: uncooked meat, edible bones and/or organs from beef, chicken, lamb, turkey, pig, duck, reindeer or salmon. Some of the products also contained other ingredients, such as vegetables, fibre and minerals. According to the labels, the products originated from Sweden, Norway, Finland, Germany and the UK. Raw animal products that could only be purchased in very large volumes were excluded. The 60 samples included in the study were stored in a freezer at −24°C until the start of the analysis and thawed in a refrigerator +8°C for 43–48 hours before analysis.

Analysis of Enterobacteriaceae

The samples were analysed for Enterobacteriaceae according to Nordic Commitee on Food Analysis (NMKL)144, 3. Ed., 2005. In brief, 25 g of sample were transferred to nine times the volume (approximately 225 ml) of buffered peptone water (BPW) and homogenised in a stomacher (easyMIX Lab Blender, AES-Chemunex, Weber Scientific, Hamilton, New Jersey, USA) for one minute. A 10-fold serial dilution in 0.1 per cent (v/v) peptone water (Oxoid, Basingstoke, UK) was prepared and 1.0 ml from each dilution was mixed carefully with 10–15 ml of violet red bile glucose agar (VRBG) (Becton, Dickinson and Company, Sparks, Maryland, USA) in a petri dish (9 cm diameter), with a final overlay of an additional 5 ml VRGG. After agar solidification, the plates were incubated at 37°C for 24±3 hours. Dark purple colonies 1–2 mm in diameter and surrounded by a purple halo were included in presumptive counts. Bacterial counts were performed on plates with 15–150 colonies. Five colonies preliminarily identified as Enterobacteriaceae were cultured on blood agar and incubated at 37°C for 24±3 hours. The identity of the colonies was verified by oxidase testing. Oxidase-negative colonies were identified to species level using matrix-assisted laser desorption/ionisation-time of flight mass spectrometry (MALDI-TOF MS) (Bruker Daltonics, Bremen, Germany). The number of Enterobacteriaceae was expressed as log colony-forming units (CFU) per gram.

­C. perfringens­

The content of C. perfringens was analysed according to NMKL 56, 5 Ed., 2015. In brief, 25 g of sample were transferred to nine times the volume (approximately 225 ml) of BPW and homogenised in a stomacher for one minute. A 10-fold serial dilution in 0.1 per cent (v/v) peptone water (Oxoid) was prepared and 1.0 ml from each dilution was poured onto a basal layer of tryptose sulphite cycloserine (TSC) agar (Oxoid) on a petri dish measuring 9 cm in diameter. An additional 10 ml TSC was poured as an overlay. After agar solidification, the plates were incubated at 37°C for 24±3 hours. Presumptive C. perfringens colonies appeared as large (2–4 mm diameter) black colonies within the depth of the agar. Bacterial counts were performed on plates with 10–100 colonies. Five presumptive colonies were cultured on blood agar and incubated at 37°C for 24±3 hours. All agar plates were incubated in an anaerobic atmosphere generated by the use of AnaeroGen (Oxoid) at 37°C for 24±3 hours. Colonies with haemolysis were suspected to be C. perfringens and their identity was confirmed by MALDI-TOF MS. The number of C. perfringens was expressed as log CFU/g.

Salmonella species

The presence of Salmonella species was analysed according to NMKL 187, 2nd Ed., 2016. In brief, 25 g of RMBD sample were homogenised and preincubated in 225 ml BPW at 37°C for 18±2 hours. After incubation, three drops of the culture with enrichment broth and RMBD sample with a total volume of 100 µl were added to selective modified semi-solid Rappaport-Vassiliadis (MSRV) with 10 mg/l novobiocin (Oxoid) and incubated at 41.5°C for 24±3 hours. If no suspected Salmonella were detected, the sample was incubated for a further 24±3 hours. Putative Salmonella colonies were subcultured on Brilliant Green agar (BG) (Oxoid) and xylose lysine deoxycholate agar (XLD) with 5 per cent novobiocin (Oxoid), according to NMKL standards. Up to five suspected Salmonella colonies on BG and XLD were recultured on purple lactose agar, incubated at 37°C for 24±3 hours, and further analysed by MALDI-TOF MS. Colonies identified as Salmonella by MALDI-TOF MS were sent to the Swedish reference laboratory for Salmonella at the National Veterinary Institute (SVA) for confirmation and species identification. The results are expressed as Salmonella detected or not detected in 25 g RMBD.

Campylobacter species

The content of thermotolerant Campylobacter was analysed according to ISO 10272 part 1 (2017), with some modifications. In brief, 25 g of RMBD were placed in a plastic stomacher bag together with 90 ml Preston broth, homogenised for one minute, and incubated at 41.5°C±1°C for 24±3 hours. After incubation, the enriched cultures were spread on modified charcoal-cefoperazone deoxycholate agar (mCCDA) and the plates were incubated at 41.5°C for 44±4 hours. All enrichment broth and agar plates were incubated in a microaerophilic atmosphere generated by use of CampyGen (Oxoid). Identification of Campylobacter species was based on typical morphological aspects, white to grey colonies with a metallic sheen and phase-contrast microscopy observations of a corkscrew movement, according to ISO 10272: part 1 (2017). Suspected Campylobacter colonies were cultured on blood agar and incubated at 41.5°C±1°C for 44±4 hours and their identity was confirmed by MALDI-TOF MS.

Species identification

Bacteria considered interesting and relevant to the study were further analysed by MALDI-TOF MS. In brief, single colonies were picked from fresh agar plates and smeared/spotted on the MALDI-TOF MS target plate, followed by addition of 1 µl α-cyano-4-hydroxycinnamic acid matrix solution (bioMérieux, France). After drying, the mass spectral fingerprint was generated with the MALDI-TOF MS instrument and the spectra obtained were compared against a reference database spectrum. Genus and species identification was then performed using a Bruker Maldi Biotyper system.

Feed safety requirements

Within the EU, Annex XIII, ‘Petfood and certain other derived products’, of Commission Regulation (EU) No. 142/2011 on the health rules regarding animal by-products and associated products not intended for human consumption contains general and specific requirements. For example, at all stages of production, processing and distribution within the businesses under their control, feed business operators must ensure that foods or feeds satisfy all requirements on food/feed law which are relevant to their activities and verify that such requirements are met. Enterobacteriaceae numbers may not exceed 5000 bacteria/g and a maximum of two out of five samples may exceed the limit of 10 bacteria per gram. Regarding Salmonella species, there is zero tolerance. Thus, if Salmonella species is detected in the sample, the feed is not permitted for use in animal feed. If the maximum value is exceeded, the feed must not be offered for sale. In the specific requirements for petfood, including dog chews, the regulation states that random samples must be taken from raw petfood during production and/or during storage (before dispatch) to verify compliance with the following standards:

Salmonella: absence in 25 g, n=5, c=0, m=0, M=0.

Enterobacteriaceae: n=5, c=2, m=10, M=5000 in 1 g


n=number of samples to be tested;

m=threshold value for number of bacteria (the result is considered satisfactory if the number of bacteria in all samples does not exceed m);

M=maximum value for the number of bacteria (the result is considered unsatisfactory if the number of bacteria in one or more samples is ≥M);

c=number of samples in which the bacterial count is between m and M (the sample shall still be considered acceptable if the bacterial count in other replicate samples is ≤m).

In Sweden, the number of replicate samples to be analysed is based on the company’s risk assessment of potential hazards. The Swedish Board of Agriculture recommends that companies analyse feed and environmental samples for Salmonella species at least twice a year. If Salmonella is detected, the source of the infection must be identified and eliminated. In the Swedish Board of Agriculture regulations and general advice on feed, all values stated except those for Salmonella are recommended guideline values for the risk assessment carried out by the feed company (table 1). Feed samples exceeding these values are not necessarily unsafe for animals, but there is a suspicion that they may be unsafe. An assessment can be made in each case regarding the risk associated with using the feed. In addition, the company must consider whether special measures are required to assure the quality of continued production.

Table 1

Guideline maximum values of bacteria for fresh petfood and feed containing offal

In the Swedish Board of Agriculture guidelines governing animal feed materials of animal origin (SJVFS 2011:40), coliforms are used in the risk assessment, with a maximum limit for coliform bacteria of 5×104 CFU/g. In EU Regulation No. 142/2011, the specific requirements refer to the Enterobacteriaceae, rather than coliforms. Coliform bacteria (coliforms) are Gram-negative, non-spore-forming, rod-shaped bacteria that are also facultative anaerobics, lactose-positive (indicated by formation of gas) and oxidase-negative when grown for 24–48 hours at 37°C. All coliform bacteria belong to the family Enterobacteriaceae.



Bacteria belonging to family Enterobacteriaceae were detected in all 60 samples (figure 1). The number of bacteria varied from log 1.6 CFU/g to log 6.4 CFU/g, with a mean of log 3.9 CFU/g. There was a great variation between the 10 different manufacturers of the samples and, in some cases, between different products from the same manufacturer (figure 1). In all, 31 (52 per cent) of the 60 samples exceeded the level of 3.7 log CFU/g for Enterobacteriaceae and were thus of unsatisfactory microbial hygiene according to EU regulations. The Swedish guideline of maximum 4.7 log CFU/g was exceeded by 13 (22 per cent) samples, including all samples from manufacturers 1 and 8 (figure 1). The bacteria identified as belonging to the Enterobacteriaceae were: Hafnia species, E coli, Enterobacter species, Serratia species, Citrobacter species, Klebsiella species, Buttiauxella species, Pantoea species, Raultella species and Providencia species. All except Serratia species and Providencia species are known coliforms.

Figure 1

Incidence of bacteria belonging to the family Enterobacteriaceae (log CFU/g) in 60 samples from 10 different manufacturers of raw meat-based diets for dogs. CFU, colony-forming unit.

­C.  perfringens­

Of the 60 RMBD samples analysed in the study, 18 (30 per cent) exceeded the detection limit of log 1 CFU/g for C. perfringens (figure 2). Ten of these 18 samples contained offal from poultry as the main protein source, while the other eight contained protein from ruminants. In two of the samples, from manufacturers 1 and 8, the amount of C. perfringens exceeded the maximum value of anaerobic bacteria permitted by Swedish guidelines. One of those samples contained offal from ruminants, dried algae and herbs, while the other contained offal from ruminants, potato fibre, vegetables, calcium, minerals and vitamins.

Figure 2

Incidence of Clostridium perfringens (log CFU/g) in samples of raw meat-based diets from different manufacturers that exceeded the detection limit of 1 log CFU/g. CFU, colony-forming unit.


Salmonella species were found in 4 (7 per cent) of the 60 samples. Two of these samples were produced by manufacturer number 8. The isolated serotypes were identified as Salmonella Rissen and monophasic Salmonella Typhimurium 4,5:i:-. The RMBD with Salmonella Rissen contained offal from cattle such as rumen, meat, cartilage, heart, liver and fat, plus dried algae and herbs, while the RWMD with monophasic Salmonella Typhimurium contained 100 per cent offal from cattle such as rumen, meat, cartilage, heart and liver. In the third sample with Salmonella, which was produced by manufacturer 1 and consisted of offal from cattle and vegetables, the species isolated was identified as Salmonella Leeuwarden. The fourth sample with Salmonella consisted of 100 per cent offal from turkeys, produced by manufacturer 2, and the species present was identified as Salmonella Typhimurium.


Thermotolerant Campylobacter species was isolated by qualitative analysis in three samples, from three different manufacturers (5, 6 and 10). Campylobacter coli was found in two samples, one of which consisted of offal from pigs and cattle, and the other of offal from turkeys. Campylobacter jejuni was detected in one sample containing offal from chicken, made by manufacturer 10.



Bacteria belonging to the family Enterobacteriaceae are part of the normal intestinal flora, which is easily spread during the slaughter process. Their presence in the dogfood samples tested was expected, since no steps are taken in the production of RMBD to eliminate these bacteria. According to EU regulations, a batch of feed is classified as having unsatisfactory hygiene quality if more than two of five replicate samples contain more than 5000 bacteria per gram. In this study, 52 per cent of the samples tested exceeded this recommended limit. However, as only one sample from each batch was analysed, it is not possible to know whether the batch from which they originated was classified as having unsatisfactory hygiene quality. In 13 of the 60 samples, the number of bacteria belonging to the Enterobacteriaceae exceeded the maximum threshold for coliforms set in Swedish guidelines. Although Enterobacteriaceae were analysed here, rather than coliforms, an important finding was that most members of the Enterobacteriaceae isolated were coliforms. However, exceeding the limit does not mean that the feed is harmful for animals, but implies a suspicion of harm (SJVFS 2011: 40, Annex 17). The isolated strains belonging to the Enterobacteriaceae that were most prevalent in the samples tested are known to be non-pathogenic or opportunists, apart from to E. coli, which has some pathogenic types. Whether the strains of E. coli isolated in this study were pathogenic was not determined. In a previous Swedish study of RMBD, E. coli was isolated from all samples and antimicrobial resistance to ESBL-producing E. coli was isolated from 23 per cent of samples.23 A strong association has been found between ESBL-producing E. coli shedding and feeding raw petfood products.24 This means that close contact between dogs and humans provides the opportunity for transmission of antimicrobial-resistant bacteria belonging to the Enterobacteriaceae, posing a risk to human health.25–27

­C.  perfringens­

C. perfringens was detected in 18 samples (30 per cent) in this study. In other studies, this bacterium has been found in 20 per cent of the samples.15 According to Swedish guidelines, the maximum level for anaerobes in animal feed materials is 5×103 CFU/g (3.7 log CFU/g) (table 1). Two (3 per cent) of the samples tested in the present study had values above this limit. A small amount of C. perfringens in the RMBD samples was expected, as this bacterium is a part of the normal intestinal microflora in many animal species. It is difficult to assess whether the levels of C. perfringens were low before freezing or whether the numbers were reduced during freezing, because this species is sensitive to cold.28 The significance of C. perfringens is not fully understood, but a correlation has been found between excretion of C. perfringens in faeces and gastrointestinal symptoms.8

Salmonella species

The incidence of Salmonella (found in 7 per cent of samples) was higher than expected, as the presence of Salmonella in meat and dairy animals and feed is low in the countries of origin of the offal (Sweden, Finland, Norway, Germany and England),29 and as some of those counties have a strict Salmonella control programme.29 30 There are various transmission routes for Salmonella to RMBD, for example, the meat may have been contaminated with Salmonella originating from the intestinal tract of the various animal species from which the offal derived.29Salmonella could also have originated from the spices, herbs or vegetables used in the RMBD formulation.31 The possibility of contamination during production of the RMBD cannot be excluded, as Salmonella can pose a hazard in animal feed and can persist for a long time in the environment in slaughterhouse and feed manufacturing facilities.32–34 Thus, Salmonella may have entered the feed factory with some product in the past and then survived in the production facility. The expiry date was stated on the packaging of all products but the production date was not mentioned on all of them, making it difficult to trace the origin of the Salmonella. However, EU regulations require effective steps to be taken to ensure that the product is not exposed to contamination throughout the production chain and up to the point of sale.

In previous studies on petfood samples in Finland, Salmonella was isolated from 2 per cent of RMBD samples of Finnish origin.22 In studies in the USA and Canada, Salmonella was found in 7 per cent to 21 per cent of samples.10 16 17 35 Significantly lower prevalence in samples from northern Europe than in samples from North America can be expected, since the EU has strict regulations on feed safety requirements.36 37

Thermotolerant Campylobacter

In this study, Campylobacter species was detected in three samples. This was a slightly lower incidence than expected, since Campylobacter species is common in the normal intestinal flora in both cattle and poultry.29 The low prevalence may be related to the fact that Campylobacter species are very sensitive to freezing. According to Georgsson and others,38 the amount of Campylobacter is reduced by approximately 1 log CFU/g immediately on freezing and thereafter by a total of 0.7–2.9 log CFU/g. Therefore, it is most likely that Campylobacter was present in more RMBD samples before freezing, and that those samples in which Campylobacter was isolated contained very high levels of Campylobacter species before the freezing process, as some managed to survive the freezer. Other studies have found a significantly higher prevalence of Campylobacter species in RMBD than observed in the present study. For example, Bojanić and others18 isolated Campylobacter from 28 per cent of samples tested, although it should be noted that 76 per cent of the samples they tested were not frozen before analysis. Campylobacter can easily colonise the intestine of dogs and is frequently isolated from the faeces of healthy dogs.39 40 Dog ownership or contact with dogs constitutes a potential risk of campylobacteriosis in humans,41 with young children and infants being the most susceptible.42 43 It has been estimated that 8 per cent of cases of human campylobacteriosis may be due to contact with cats and dogs.44 Young children have particularly close contact with pets and are therefore more susceptible to faecal-oral spread. Pet-associated cases have been confirmed by molecular fingerprinting techniques, for example, for a case of neonatal sepsis in young children caused by C. jejuni spread from the family dog.45 46


The results obtained in this study show that it is highly important to handle RMBD carefully and to maintain good hygiene, due to the potential risks these feeds pose to human and animal health. The RMBD should be kept frozen until used, thawing should take place at a maximum of 10°C and the thawed product should be separated from human food and handled with separate kitchen equipment, or with the equipment properly washed after use. Bacteria in the raw juices from RMBD can splash and spread to other foods and surfaces. A great opportunity for dogs to transfer potential pathogenic and antimicrobial-resistant bacteria to humans is by ‘kissing’ people in the face immediately after they have eaten. In view of the resistance problem, dogs should not be fed RMBD while they are being treated with antimicrobials, as this could increase the risk of resistant strains being selected and multiplying. Dogs in families with infants, elderly people or immunocompromised individuals should also not be fed RMBD, as these groups are more susceptible to infections.


The authors would like to acknowledge CITES, Feed and Animal Products Unit, Department for Animal Welfare and Health, Swedish Board of Agriculture for providing information about the current guidelines.


View Abstract


  • Contributors All authors contributed substantially to interpretation of data, drafting the final manuscript and critical revision for important intellectual content. All authors have agreed this final submitted version, which fulfils all three of the ICMJE guidelines for authorship.

  • Funding This research was funded by the C August Carlsson and Elsa Paulsson Foundation at the Faculty of Veterinary Medicine and Animal Sciences, Swedish University of Agricultural Sciences.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Correction notice This article has been corrected since it was published Online First. A percentage value was previously incorrect.

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.