Clostridioides difficile is a leading cause of human antibiotic-associated diarrhoeal disease globally. Zoonotic reservoirs of infection are increasingly suspected to play a role in the emergence of this disease in the community and dogs are considered as one potential source. Here we use a canine case-control study at a referral veterinary hospital in Scotland to assess: i) the risk factors associated with carriage of C. difficile by dogs, ii) whether carriage of C. difficile is associated with clinical disease in dogs and iii) the similarity of strains isolated from dogs with local human clinical surveillance. The overall prevalence of C. difficile carriage in dogs was 18.7% (95% CI 14.8–23.2%, n = 61/327) of which 34% (n = 21/61) were toxigenic strains. We found risk factors related to prior antibiotic treatment were significantly associated with C. difficile carriage by dogs. However, the presence of toxigenic strains of C. difficile in a canine faecal sample was not associated with diarrhoeal disease in dogs. Active toxin was infrequently detected in canine faecal samples carrying toxigenic strains (2/11 samples). Both dogs in which active toxin was detected had no clinical evidence of gastrointestinal disease. Among the ten toxigenic ribotypes of C. difficile detected in dogs in this study, six of these (012, 014, 020, 026, 078, 106) were ribotypes commonly associated with human clinical disease in Scotland, while nontoxigenic isolates largely belonged to 010 and 039 ribotypes. Whilst C. difficile does not appear commonly associated with diarrhoeal disease in dogs, antibiotic treatment increases carriage of this bacteria including toxigenic strains commonly found in human clinical disease.
Citation: Albuquerque C, Pagnossin D, Landsgaard K, Simpson J, Brown D, Irvine J, et al. (2021) The duration of antibiotic treatment is associated with carriage of toxigenic and non-toxigenic strains of Clostridioides difficile in dogs. PLoS ONE 16(5): e0245949. https://doi.org/10.1371/journal.pone.0245949
Editor: Simon Clegg, University of Lincoln, UNITED KINGDOM
Received: January 8, 2021; Accepted: April 29, 2021; Published: May 12, 2021
Copyright: © 2021 Albuquerque et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the manuscript and its Supporting Information files.
Funding: Petplan Charitable Trust pump primer award (CM, GD) Reference S19-816-855 https://petplancharitabletrust.org.uk/apply-for-a-grant/ University of Glasgow Small Animal Fund (CM, GD, AR) https://www.gla.ac.uk/connect/supportus/vetfund/ BBSRC STARS scholarship (KL). Vacation scholarship, Carnegie Trust (JS) https://www.carnegie-trust.org/awardschemes/vacation-scholarships/ The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Clostridioides difficile has emerged as a leading cause of antibiotic associated diarrhoeal disease in people globally which is associated with significant morbidity, mortality and healthcare costs . In the past this disease was predominately associated with elderly patients treated with antibiotics in healthcare environments, however sentinel surveillance studies have revealed that a substantial proportion of C. difficile infections (CDI) are acquired within the community [2–5]. Whole genome sequencing has shown that only around a third of hospital cases can be linked to horizontal transmission from symptomatic patients, with the remainder caused by diverse strains of C. difficile . The source of these infections, and those arising within the community is unknown, and may include asymptomatic human carriers, zoonotic reservoirs, food and the environment [7–9].
Most research on C. difficile in animals has focused on production animals and horses with emergence of the 078 ribotype as a significant cause of enteritis in piglets and adult horses occurring around the same time as emergence of CDI in humans . The frequent isolation of this organism from the faeces of production animals including pigs, cattle and chickens and food has led to concerns that spread to humans can occur through contamination of the local environment or via the food chain [11–15]. Whole genome sequencing of identical strains of C. difficile in pig farmers and pigs on the same farm suggests that interspecies transmission is likely, although a common environmental source cannot be excluded . In contrast, less is known about the potential of companion animals including dogs, to become colonised with C. difficile, develop disease or act as a zoonotic reservoir. The frequency of dog ownership, close living relationships with people, and evidence that pet dogs can be a risk factor for human colonisation , justifies evaluation of this species as a potential reservoir host of zoonotic strains of C. difficile.
Results from published studies of C. difficile carriage by companion animals report prevalence rates in dogs from 0% to 58%, with a lower prevalence in healthy dogs [18–20] and a higher prevalence reported in hospitalised dogs  and those visiting human hospitals . Similar ribotypes have been identified in both canines and humans suggesting potential for interspecies transmission [23, 24]. A small number of studies have looked at risk factors for C. difficile carriage in dogs. These studies include risk factors which may increase individual susceptibility to colonisation such as antibiotic treatment and potential sources of infection such as diet and household factors [25–29]. Results to date are often contradictory, which may reflect differences in study design and geographic location.
Similarly, existing studies investigating associations between C. difficile and disease in dogs have found contrasting results. C. difficile in humans is largely a toxin-mediated disease with most pathogenic isolates of C. difficile producing one or both major toxins: toxin A (an enterotoxin) and toxin B (a cytotoxin) . Nontoxigenic and toxigenic strains of C. difficile have been detected by several studies in both healthy dogs and those with diarrhoea by bacterial culture and PCR testing for toxin genes [25, 31, 32]. To assess for associations between carriage of toxigenic strains and clinical disease, testing for the presence of active toxin in faecal samples from healthy dogs and those with diarrhoea is also needed . Some studies which have tested for active toxin have suggested an association between the presence of toxin in faecal samples and diarrhoeal disease in dogs [18, 33, 34]. However, others have not which may in part be explained by differences in methods used to detect active toxin in samples .
To assess the potential role of dogs as a zoonotic reservoir of C. difficile, and association with clinical diarrhoeal disease in dogs we designed a case-control study of dogs presenting to a veterinary referral practice in Scotland. Our study had the following objectives; i) to assess the risk factors associated with carriage of C. difficile by dogs, ii) to test for associations between carriage of C. difficile and diarrhoeal disease in dogs and iii) to determine if dogs carry strains of C. difficile that are frequently associated with clinical disease in humans.
Materials and methods
Ethical approval for the study was obtained from the University of Glasgow, School of Veterinary Medicine Ethics and Welfare Committee (Reference number 11a/16). Informed written consent was obtained from all participants recruited to complete a questionnaire survey and informed written consent for the use of residual clinical samples for research was given by all participants in the study. To investigate whether C. difficile carriage was associated with disease in dogs we recruited a total of 327 dogs referred from across the west of Scotland to the University of Glasgow School of Veterinary Medicine Small Animal Hospital (hereafter described as the referral hospital). A referral hospital was chosen for the study due to the large geographic catchment area, ability to include dogs with a range of potential risk factors for carriage of C. difficile and the capacity to investigate the association of C. difficile carriage with diarrhoeal disease. A fresh faecal sample collected by the owner on the day of admission to the hospital or a sample from the first stool passed within 48 hours of admission was used to evaluate C. difficile colonisation of dogs within the community . Dogs referred for treatment of either acute or chronic diarrhoea (n = 101) or for non-gastrointestinal reasons (n = 226) were recruited to the study between June 2016 and October 2019.
Assessment of risk factors for C. difficile carriage by dogs
A questionnaire designed to assess potential risk factors for the carriage of C. difficile by dogs was completed by a subset of owners (n = 200) recruited to the study (S1 File). All owners presenting to the referral hospital with their dog between June to December 2016 were invited to complete the survey unless the dog was critically ill. The survey was designed to provide information on potential risk factors for increased susceptibility for C. difficile carriage and to identify potential sources of infection. We requested information on the diet, breed, sex and age of the dog and information on the household, including co-habiting pets, elderly people or infants. We examined clinical records from the referring practice to obtain clinical information on antibiotic, antacid, immunosuppressive treatment and the number of visits and days as an inpatient at a veterinary hospital within the three months prior to admission to the referral hospital.
Detection and strain typing of C. difficile from canine faecal samples
Following collection, canine faecal samples were placed in anaerobic jars, stored at 4°C, and processed within 72 hours. One gram of each faecal sample was emulsified in 2ml of ethanol and incubated at room temperature for 30 minutes to select for resistant C. difficile spores, which were then germinated by plating of 200ml of this suspension on Brazier’s taurocholate cycloserine cefoxitin agar (TCCA) supplemented with 5% defibrinated horse blood and egg white emulsion. Each sample was cultured in duplicate and incubated for a maximum of 7 days in an anaerobic chamber (Don Whitley, UK). Colonies showing typical C. difficile colony morphology and which appeared black when subcultured onto Biomerieux chromID® C. difficile agar were selected. Clones were identified as C. difficile by amplification and analysis of the 16S (V4) ribosomal RNA sequence (S1 Table). Chromosomal DNA was prepared from each isolate using DNeasy Blood and Tissue kits, (Qiagen, Hilden, Germany), following the manufacturer’s instructions. Amplified PCR products were purified using the Qiaquick PCR purification kit, and sequenced by Source Bioscience (Livingston, UK). The bacterial species for each isolate was determined by subjecting each sequenced and trimmed PCR product to BLAST analysis using the National Centre for Biotechnology (NCBI) Nucleotide database.
All isolates were ribotyped at the Scottish Microbiology Reference Laboratory (Glasgow) using PCR ribotyping as previously described [36, 37]. Variable-length intragenic spacer regions of the rRNA complex were amplified by PCR and visualised following agarose gel electrophoresis. Ribotype patterns were assigned following analysis using BioNumerics software v7.6 (Applied Maths, Sint-Maart-ens-Latem, Belgium). Patterns were compared with libraries using the Pearson Correlation Coefficient of similarity with a 1% optimization setting. A library containing a number of examples of individual ribotype patterns allowed the correct identification of similar ribotypes such as RT014 and RT020 as evidenced by the laboratory’s performance in external quality control schemes .
Detection of toxin genes by PCR assays for tcdA and tcdB and detection of active toxin.
All isolates of C. difficile were tested for the presence of fragments of the tcdA and tcdB genes by PCR amplification with primers designed using the annotated C. difficile 630 genome (S1 Table). Successful primer binding was confirmed in silico with sequences from representative strains of C. difficile. These primers amplify the first 427 and 417bp tcdA and tcdB respectively. Amplification was shown to be specific by inclusion of DNA from the epidemic, toxin-producing strain C. difficile R20291 and from a PaLoc negative strain, 1342.
A subset of 116 faecal samples collected between 2018–19, were tested for the presence of active toxin using a cytotoxicity assay. Of these samples, 22 were from dogs presenting with diarrhoea (8 with acute diarrhoea and 14 with chronic diarrhoea) and 94 samples were from control dogs without diarrhoea. To detect the presence of toxin, aliquots of fresh faeces, were stored in anaerobic jars at 4°C and within 48 hours of collection, were emulsified in 2ml of PBS and solid material removed by centrifugation. Pilot experiments with known concentrations of toxin spiked into fresh canine faecal pellets confirmed this storage method did not alter the capacity to detect toxin activity. The supernatant was filtered through a 0.2um membrane filter and the resultant material was serially diluted in PBS and added to a prepared monolayer of Vero cells. To confirm specificity of cytotoxicity, a second set of samples, prepared in parallel, were treated with Clostridium sordellii antitoxin (NIBSC, 20 IU/ml). This antitoxin cross-reacts with C. difficile toxin and has been used to confirm the presence of C. difficile toxin activity in human samples . Treated cells were then incubated for 18–24 hours at 37°C with 5% CO2, before cells were fixed with 1% formalin for 30 minutes and stained using Giemsa stain (SIGMA-ALDRICH®, 6% diluted) for 1 hour. Cell rounding, which is associated with C. difficile toxin presence was assessed microscopically. A sample was considered positive if cell rounding was observed that was neutralized by addition of the C. sordellii antitoxin .
All statistical analyses were carried out in R version 4.0.2 (R Development Core Team, Vienna, Austria using the package lme4  for analyses. Collinearity was tested for using the variance inflation factor in the car package in ‘R’ . Prevalence was calculated using the prop.test function in ‘R’.
Assessment of risk factors for carriage of C. difficile.
C. difficile carriage in dogs (present or absent) was modelled in a binomial general linear model (GLM) with a logit link as a function of each of the following potential risk factors listed in Table 1. Risk factors with a p value of < 0.10 based on univariable analysis were included in a multivariable general linear model of C. difficile carriage (present or absent) with binomially distributed errors and a logit link. Starting from the maximum global model, stepwise backwards model selection was carried out using likelihood ratio tests.
Testing for an association between C. difficile carriage and diarrhoeal disease in dogs.
To test whether the carriage of toxigenic strains of C. difficile was associated with diarrhoea, the presence or absence of diarrhoea in each dog (n = 327) was modelled using a GLM with binomially distributed errors and a logit link, as a function of C. difficile carriage, and separately as a function of carriage of a toxigenic strain of C. difficile (based on the PCR presence of one or both toxin genes).
Comparison of strains detected in dogs with clinical surveillance for C. difficile in humans
To assess the potential for shared strains between dogs and humans, ribotypes detected from dogs in this study were compared to ribotypes recorded from human surveillance of C. difficile cases by the National Microbiology Reference laboratory in Scotland between 2015 and 2018 [41–43].
Prevalence and strain diversity of C. difficile in canine faecal samples
The overall prevalence of C. difficile from canine faecal samples in this study was 18.7% (95% C.I. 14.8–23.2%, n = 61/327). The majority of isolates were nontoxigenic strains (63.4% n = 39/61) while 21/61 were toxigenic strains with the PCR presence of either or both tcdA and tcdB genes. One isolate was not tested for toxin genes. A total of 13 different ribotypes were detected in the canine samples, and 10 of these ribotypes included toxigenic isolates (Fig 1, S2 Table).
Ribotypes of C. difficile from dogs in this study, shown according to whether they were isolated from dogs with or without diarrhoea (A) and whether strains were classed as toxigenic or nontoxigenic (B) based on a PCR positive test for either or both tcdA and tcdB genes.
Detection of toxin activity in fresh faecal samples
A total of 116 faecal samples were tested within 48h of collection for the presence of active toxin as determined by Vero cell rounding. Two samples tested positive, and cytotoxic activity was neutralised by C. sordellii antitoxin. Both faecal samples with active toxin present were from dogs with no clinical evidence of diarrhoea. C. difficile was subsequently cultured from 26 of the 116 faecal samples tested for active toxin (21 dogs without diarrhoea, and 5 dogs with diarrhoea) and 11 of these isolates were toxigenic. This included both samples which tested positive with the cytotoxicity assay. Ribotype analysis revealed these strains to be 020 and 106 types respectively (S2 Table).
Risk factors for carriage of C. difficile by dogs
Results of univariable analysis of risk factors and carriage of C. difficile are shown in Table 1. In a multivariable model, dogs were more likely to carry C. difficile with an increasing length of treatment on antibiotics, for each day of antibiotic treatment (OR = 1.08, 95% C.I. 1.04–1.11), p<0.001 (Table 2). Other explanatory variables including age (months), treatment with multiple antibiotics and the presence of multiple pets in the household were not maintained in the final model.
Testing for an association between C. difficile carriage and diarrhoeal disease in dogs.
Neither carriage of C. difficile (OR = 1.33 95% CI = 0.74–2.38, p = 0.34), or the presence of toxigenic strains of C. difficile in a faecal sample (OR = 0.50, 95% CI = 0.14–1.41, p = 0.20) was associated with diarrhoea in dogs. Toxigenic strains of C. difficile were detected both in dogs with diarrhoea and in dogs with no evidence of gastrointestinal disease (Fig 1).
Comparison of strains detected in dogs with clinical surveillance for C. difficile in humans
Human clinical surveillance of C. difficile strains during the time period of this study are shown in S3 Table. The frequency of the most common 12 ribotypes are shown which represent approximately two thirds of the total number of human isolates detected. Six of these ribotypes, 012, 014, 020, 026, 078, 106 were isolated from dogs in this study. These six ribotypes represent approximately one third of the total number of isolates collected from human CDI surveillance.
This study has found that C. difficile can be frequently isolated from diarrhoeic and non-diarrhoeic canine faecal samples, and carriage of toxigenic strains by dogs is not associated with diarrhoeal disease. As in humans and other species, antibiotic treatment was significantly associated with the carriage of C. difficile by dogs. Several toxigenic ribotypes detected in dogs in this study are among the most frequently reported ribotypes from clinical surveillance of people in the same locality over the period of the study.
The overall prevalence of C. difficile carriage in dogs in this study was 18.7%, similar to previous studies from referral hospitals in other countries which reported rates of 18.4% and 13.7% [21, 27]. Prevalence rates in healthy dogs presenting to primary care veterinary clinics and those living in shelters are reported to be lower, ranging from 0% to 6.1% [20, 31]. Several factors could contribute to these apparent differences, including geographic area, laboratory methods and age and clinical history of dogs recruited to these studies. In our study toxigenic isolates accounted for 34% of positive cultures (n = 21/61); previous studies have reported a prevalence of toxigenic isolates of 36.8% to 69% [21, 23, 25, 28]. The most commonly isolated ribotypes in our study cohort were ribotypes 039 and 010, followed by 020, a toxigenic ribotype that was recovered most frequently from non-diarrhoeic dogs (Fig 1). Isolates from two ribotypes (039 and 012) contained both toxigenic and nontoxigenic strains (Fig 1). This result is unusual as typically 039 isolates are nontoxigenic, while 012 isolates are toxigenic , and will be further investigated using whole genome sequencing. In contrast to studies of production animals and horses where the 078 ribotype often dominates, no dominate strain appears associated with canine colonisation [20, 24, 26, 29, 45, 46]. Other European studies have reported the 010 ribotype as one of the most common strains isolated from dogs [20, 24, 29, 45, 46]. In these studies, ribotypes 014 and 020 were also frequently isolated.
We found risk factors related to antibiotic treatment within the previous three months were significantly associated with carriage of C. difficile by dogs. The length of antibiotic treatment was the only factor supported in a multivariable model, with a seven-day course of treatment predicted to increase the risk of carriage by 1.67 times (95% CI 1.30–2.13, p <0.001). Some previous studies in dogs have found an association between previous antibiotic administration and C. difficile carriage [27, 29, 47] whereas other studies did not [21, 25, 28]. As in humans and horses the mechanism underpinning the positive relationship between antibiotic treatment and C. difficile carriage is likely due to loss of microbiome diversity within the gut . As a result, C. difficile is able to germinate and rapidly multiply in the available niche. Age was also positively associated with carriage of C. difficile on univariable analysis with a slightly increased risk of carriage per year OR = 1.13 (95% CI = 1.02–1.25, p = 0.016), as reported by other studies [21, 28]. This positive relationship could suggest an extended duration of colonisation, though no longitudinal studies of carriage in dogs have been carried out to date. Alternatively, there could be increased host susceptibility with age. We were unable to identify a potential source of C. difficile colonisation of dogs from our questionnaire survey. Previous studies have found that dogs living with an immunocompromised person or contact with a person with diarrhoea can increase the risk of colonisation in dogs, while feeding a dry food diet reduces risk [19, 29, 49, 50]. Although comparison of the results of published studies in dogs is limited by difference in study design and geographic area, a potential explanation for variation in ribotypes detected in dogs among different studies could be that carriage is driven mainly by host susceptibility. If this hypothesis is true, the strains isolated from dogs may be reflective of those which they are exposed to on a daily basis in food and the environment [9, 14].
The significance of C. difficile as cause of disease in dogs is unclear, since toxigenic strains can be isolated from healthy, as well as diarrhoeic dogs [25, 31]. Our study was in agreement with others which did not find an association between carriage of toxigenic strains of C. difficile and diarrhoea [25, 34]. Carriage of toxigenic strains in our study was assessed through culture and PCR to detect the presence of either or both of the toxin genes tcdA and tcdB. Although variability in toxin genes may potentially affect primer binding and ability to detect these genes using PCR (see footnote, S2 Table) , a problem with primer binding is suspected to be an issue in only one 017 isolate for the tcdA gene. Based on comparative sequences from typical isolates of these ribotypes and the fragment amplified by PCR, this isolate would be expected to be positive by PCR. This did not affect classification of this isolate as a toxigenic strain since the tcdB gene was detected. Our finding that active toxin was not detected in the majority of dogs carrying toxigenic strains of C. difficile, also found in another recent study  may suggest one possible reason why carriage does not seem to be commonly associated with diarrhoea. These results are consistent with either toxigenic strains being most frequently carried in the canine gut without active transcription of the toxins, or very small quantities of toxin being produced which were below the detection limits of our assay. In our study only two of eleven canine faecal samples which carried toxigenic strains tested positive for active toxin and both of these samples were from non-diarrhoeic dogs. Some previous studies which indicated a relationship between the presence of active toxin and C. difficile associated disease in dogs may have been affected by low sensitivity and specificity of ELISA’s used to detect toxin in dogs .
We were limited in our study cohort in evaluating associations between diarrhoea and carriage of toxigenic strains of C. difficile by the relatively low numbers of diarrhoeic samples carrying toxigenic strains of C. difficile (n = 4). No pattern in the history, clinical presentation or diagnosis was observed among these cases and endoscopic evaluation of the colon was not carried out as part of clinical investigations in these dogs. Due to time and logistical constraints we were only able to implement the cytotoxicity assay for part of the study period which meant only one of these four cases was tested for active toxin. This was not found to be present, suggesting that C. difficile was not the cause of the diarrhoea. Although C. difficile may still be a potential cause of diarrhoea in dogs, our results suggest that the frequency of disease is likely to be low. The availability of reliable tests for active toxin which are suitable for use in a diagnostic laboratory setting is likely to limit clinical investigations into the significance of toxigenic isolates. Cytotoxicity assays are labour intensive and unlikely to be widely available.
There is evidence that the epidemiology of CDI in humans is changing, with increasing numbers of cases reported from patients residing within the community and attribution of the source of infections in most of these cases is unknown [7, 53]. This study, in agreement with other recent studies shows that ribotypes associated within human clinical disease can be carried asymptomatically with the canine gut. Six of the ten toxigenic ribotypes of C. difficile detected in dogs in this study (012, 014, 020, 026, 078, 106) are also some of the most common isolates detected by human clinical surveillance in Scotland from 2015–2018 (S3 Table) [41–43]. A subset of these ribotypes (014, 020 and 078) are amongst the most prevalent causes of C. difficile-associated diarrhoea in Europe [54, 55]. Results from this and other companion animal studies demonstrating shared ribotypes amongst dogs and humans suggest that dogs could contribute to a reservoir for human infections, either directly or by contaminating the environment. Understanding the potential significance of carriage of toxigenic strains of C. difficile by companion animals to human community CDI will require integrated molecular epidemiology studies of community CDI with investigation of food, environment and potential zoonotic sources.
We have found that C. difficile carriage in dogs presenting to a referral hospital in Scotland is relatively common, and an increasing length of antibiotic therapy is associated with a higher risk of C. difficile carriage. The findings of this study and others suggest that C. difficile is not commonly associated with diarrhoeal disease in dogs. Dogs carried several toxigenic strains associated with human clinical disease and could potentially act as a source of infection for humans, or spore accumulation within the environment.
S1 File. Owner questionnaire for risk factors for C. difficile carriage in dogs and associations with clinical disease.
S1 Table. Primers and amplification conditions for 16S and (tcdA and tcdB) PCR.
* F = Forward, R = Reverse.
S2 Table. C. difficile isolates from dogs in this study; ribotype, results of tcdA and tcdB PCR testing and cytotoxicity testing for active toxin presence.
*Based on comparative sequence of typical isolates of these ribotypes and the fragment amplified by PCR, this isolate would be expected to be positive by PCR. Strains of the ribotype 017 belong to the toxinotype VIII group, and are TcdA negative. However, as failure to produce TcdA is linked to a 1.8kb deletion at the 3’ end of the gene, amplification of the first 400bp should have been feasible and the result here is unexpected.
S3 Table. Frequency of the most common ribotypes collected from human clinical surveillance for Clostridium Difficile Infection (CDI) in Scotland from mild, moderate or severe CDI cases (snapshot surveillance).
Data from Health Protection Scotland Annual Reports 2015–2018 (1–3).
The authors are grateful for support of staff and technicians at the University of Glasgow Small Animal Hospital with the recruitment of animals and collection of samples for the study and for advice on the statistical analysis from Paul Johnson.
- 1. Balsells E, Shi T, Leese C, Lyell I, Burrows J, Wiuff C, et al. Global burden of Clostridium difficile infections: A systematic review and meta-analysis. J Glob Health. 2019;9: 010407. pmid:30603078
- 2. Dumyati G, Stevens V, Hannett GE, Thompson AD, Long C, Maccannell D, et al. Community-associated Clostridium difficile Infections, Monroe County, New York, USA. Emerg Infect Dis. 2012;18: 392–400. pmid:22377231
- 3. Banks A, Brown DJ, Mather H, Coia JE, Wiuff C. Sentinel community Clostridium difficile infection (CDI) surveillance in Scotland, April 2013 to March 2014. Anaerobe. 2016;37: 49–53. pmid:26708405
- 4. Kotila SM, Mentula S, Ollgren J, Virolainen-Julkunen A, Lyytikkainen O. Community- and Healthcare- Associated Clostridium difficile Infections, Finand, 2008–2013. Emering Infect Dis. 2016;22: 1747–1753.
- 5. Cassir N, Fahsi N, Durand G, Lagier JC, Raoult D, Fournier PE. Emergence of Clostridium difficile tcdC variant 078 in Marseille, France. Eur J Clin Microbiol Infect Dis. 2017;36: 1971–1974. pmid:28573471
- 6. Eyre DW, Cule ML, Wilson DJ, Griffiths D, Vaughan A, O’Connor L, et al. Diverse Sources of C. difficile Infection Identified on Whole-Genome Sequencing. N Engl J Med. 2013;369: 1195–1205. pmid:24066741
- 7. Freeman J, Bauer MP, Baines SD, Corver J, Fawley WN, Goorhuis B, et al. The Changing Epidemiology of Clostridium difficile Infections. Clin Microbiol Rev. 2010;23: 529–549. pmid:20610822
- 8. Janezic S, Zidaric V, Pardon B, Indra A, Kokotovic B, Blanco JL, et al. International Clostridium difficile animal strain collection and large diversity of animal associated strains. BMC Microbiol. 2014;14. pmid:24467879
- 9. Janezic S, Potocnik M, Zidaric V, Rupnik M. Highly divergent Clostridium difficile strains isolated from the environment. PLoS One. 2016;11: 1–12. pmid:27880843
- 10. Elliott B, Androga GO, Knight DR, Riley T V. Clostridium difficile infection: Evolution, phylogeny and molecular epidemiology. Infect Genet Evol. 2017;49: 1–11. pmid:28012982
- 11. Keel K, Brazier JS, Post KW, Weese S, Songer JG. Prevalence of PCR ribotypes among Clostridium difficile isolates from Pigs, Calves, and Other Species. J Clin Microbiol. 2007;45: 1963–1964. pmid:17428945
- 12. Keessen EC, Gaastra W, Lipman LJA. Clostridium difficile infection in humans and animals, differences and similarities. Vet Microbiol. 2011;153: 205–217. pmid:21530110
- 13. Songer JG. The emergence of Clostridium difficile as a pathogen of food animals. Anim Heal Res Rev. 2004;5: 321–326. pmid:15984348
- 14. Weese JS. Clostridium difficile in food—innocent bystander or serious threat? Clin Microbiol Infect. 2010;16: 3–10. pmid:20002685
- 15. Casey JA, Kim BF, Larsen J, Price LB, Nachman KE. Industrial Food Animal Production and Community Health. Curr Environ Heal reports. 2015;2: 259–271. pmid:26231503
- 16. Knetsch CW, Connor TR, Mutreja A, van Dorp SM, Sanders IM, Browne HP, et al. Whole genome sequencing reveals potential spread of Clostridium difficile between humans and farm animals in the Netherlands, 2002 to 2011. Eurosurveillance. 2014;19: 1–12. pmid:25411691
- 17. Stoesser N, Eyre DW, Phuong Quan T, Godwin H, Pill G, Mbuvi E, et al. Epidemiology of Clostridium difficile in infants in Oxfordshire, UK: Risk factors for colonization and carriage, and genetic overlap with regional C. difficile infection strains. PLoS One. 2017;12: 1–16. pmid:28813461
- 18. Weese JS, Staempfli HR, Prescott JF, Kruth SA, Greenwood SJ, Weese HE. The roles of Clostridium difficile and enterotoxigenic Clostridium perfringens in diarrhea in dogs. J Vet Intern Med. 2001;15: 374–378. pmid:11467596
- 19. Weese JS, Finley R, Reid-Smith RR, Janecko N, Rousseau J. Evaluation of Clostridium difficile in dogs and the household environment. Epidemiol Infect. 2010;138: 1100–1104. pmid:19951453
- 20. Schneeberg A, Rupnik M, Neubauer H, Seyboldt C. Prevalence and distribution of Clostridium difficile PCR ribotypes in cats and dogs from animal shelters in Thuringia, Germany. Anaerobe. 2012;18: 484–488. pmid:22951303
- 21. Struble AL, Tang YJ, Kass PH, Gumerlock PH, Madewell BR, Silva J. Fecal shedding of Clostridium difficile in dogs: A period prevalence survey in a veterinary medical teaching hospital. J Vet Diagnostic Investig. 1994;6: 342–347. pmid:7948204
- 22. Lefebvre SL, Waltner-Toews D, Peregrine AS, Reid-Smith R, Hodge L, Arroyo LG, et al. Prevalence of zoonotic agents in dogs visiting hospitalized people in Ontario: implications for infection control. J Hosp Infect. 2006;62: 458–466. pmid:16466831
- 23. Usui M, Suzuki K, Oka K, Miyamoto K, Takahashi M, Inamatsu T, et al. Distribution and characterization of Clostridium difficile isolated from dogs in Japan. Anaerobe. 2016;37: 58–61. pmid:26456188
- 24. Orden C, Blanco JL, Álvarez-Pérez S, Garcia ME, Blanco JL, Garcia-Sancho M, et al. Isolation of Clostridium difficile from dogs with digestive disorders, including stable metronidazole-resistant strains. Anaerobe. 2017;43: 78–81. pmid:27965048
- 25. Clooten J, Kruth S, Arroyo L, Weese JS. Prevalence and risk factors for Clostridium difficile colonization in dogs and cats hospitalized in an intensive care unit. Vet Microbiol. 2008;129: 209–214. pmid:18164560
- 26. Álvarez-Pérez S, Blanco JL, Harmanus C, Kuijper EJ, García ME. Prevalence and characteristics of Clostridium perfringens and Clostridium difficile in dogs and cats attended in diverse veterinary clinics from the Madrid region. Anaerobe. 2017;48: 47–55. pmid:28687280
- 27. Hussain I, Sharma RK, Borah P, Rajkhowa S, Hussain I, Barkalita LM, et al. Isolation and characterization of Clostridium difficile from pet dogs in Assam, India. Anaerobe. 2015;36: 9–13. pmid:26393292
- 28. Silva ROS, Santos RLR, Pires PS, Carlos Pereira L, Trindade Pereira S, Carvalho Duarte M, et al. Detection of toxins A/B and isolation of Clostridium difficile and Clostridium perfringens from dogs in Minas Gerais, Brazil. Brazilian J Microbiol. 2013;44: 133–137. pmid:24159295
- 29. Rabold D, Espelage W, Sin MA, Eckmanns T, Schneeberg A, Neubauer H, et al. The zoonotic potential of Clostridium difficile from small companion animals and their owners. PLoS One. 2018;13: 1–12. pmid:29474439
- 30. Voth DE, Ballard JD. Clostridium difficile toxins: Mechanism of action and role in disease. Clin Microbiol Rev. 2005;18: 247–263. pmid:15831824
- 31. Ghavidel M, Salari Sedigh H, Razmyar J. Isolation of Clostridium difficile and molecular detection of binary and A/B toxins in faeces of dogs. Iran J Vet Res. 2016;17: 273–276. pmid:28224013
- 32. Wetterwik K-J, Trowald-Wigh G, Fernström L-L, Krovacek K. Clostridium difficile in faeces from healthy dogs and dogs with diarrhea. Acta Vet Scand. 2013;55: 23. pmid:23497714
- 33. Cave NJ, Marks SL, Kass PH, Melli AC, Brophy MA. Evaluation of a routine diagnostic fecal panel for dogs with diarrhea. J Am Vet Med Assoc. 2002;221: 52–59. pmid:12420824
- 34. Marks SL, Kather EJ, Kass PH, Melli AC. Genotypic and phenotypic characterization of Clostridium perfringens and Clostridium difficile in diarrheic and healthy dogs. J Vet Intern Med. 2002;16: 533–540. pmid:12322702
- 35. Chouicha N, Marks SL. Evaluation of five enzyme immunoassays compared with the cytotoxicity assay for diagnosis of Clostridium difficile-associated diarrhea in dogs. J Vet Diagnostic Investig. 2006;18: 182–188. pmid:16617699
- 36. O’Neill GL, Ogunsola FT, Brazier JS, Duerden BI. Modification of a PCR ribotyping method for application as a routine typing scheme for Clostridium difficile. Anaerobe. 1996;2: 205–209.
- 37. Stubbs SLJ, Brazier JS, O’Neill GL, Duerden BI. PCR targeted to the 16S-23S rRNA gene intergenic spacer region of Clostridium difficile and construction of a library consisting of 116 different PCR ribotypes. J Clin Microbiol. 1999;37: 461–463. pmid:9889244
- 38. Chang TW, Gorbach SL, Bartlett JB. Neutralization of Clostridium difficile toxin by Clostridium sordellii antitoxins. Infect Immun. 1978;22: 418–422. pmid:730363
- 39. Bates D, Maechler M, Bolker B, Walker S. lme4: Linear mixed-effects, models using Eigen and S4. 2019.
- 40. Fox J, Weisberg S. An R Companion to Applied Regression. Third. Thousand Oaks CA: Sage; 2019.
- 41. Health Protection Scotland. Healthcare associated infections, Annual Report 2016. 2016. Available from: https://www.hps.scot.nhs.uk/web-resources-container/healthcare-associated-infection-annual-report-2016/. Accessed 7/10/2020.
- 42. Health Protection Scotland. Healthcare Associated infection, Annual Report 2017. 2017. Available from: https://www.hps.scot.nhs.uk/publications/hps-weekly-report/volume-52/issue-18/hps-publish-2017-hai-annual-report/ Accessed 7/10/2020.
- 43. Health Protection Scotland. Healthcare Associated infection Annual Report 2018. 2018. Available from: https://www.hps.scot.nhs.uk/web-resources-container/healthcare-associated-infection-annual-report-2018/. Accessed 7/10/2020.
- 44. Rupnik M, Brazier JS, Duerden BI, Grabnar M, Stubbs SLJ. Comparison of toxinotyping and PCR ribotyping of Clostridium difficile strains and description of novel toxinotypes. Microbiology. 2001;147: 439–447. pmid:11158361
- 45. Koene MGJ, Mevius D, Wagenaar J a., Harmanus C, Hensgens MPM, Meetsma a. M, et al. Clostridium difficile in Dutch animals: Their presence, characteristics and similarities with human isolates. Clin Microbiol Infect. 2012;18: 778–784. pmid:21919997
- 46. Spigaglia P, Drigo I, Barbanti F, Mastrantonio P, Bano L, Bacchin C, et al. Antibiotic resistance patterns and PCR-ribotyping of Clostridium difficile strains isolated from swine and dogs in Italy. Anaerobe. 2015;31: 42–46. pmid:25316022
- 47. Riley T V., Adams JE, O’neill GL, Bowman RA. Gastrointestinal carriage of Clostridium difficile in cats and dogs attending veterinary clinics. Epidemiol Infect. 1991;107: 659–665. pmid:1752313
- 48. Theriot CM, Young VB. Interactions between the Gastrointestinal Microbiome and Clostridium difficile. Annu Rev Microbiol. 2015;69: 445–461. pmid:26488281
- 49. Lefebvre SL, Reid-Smith RJ, Waltner-Toews D, Scott Weese J. Incidence of acquisition of methicillin-resistant Staphylococcus aureus, Clostridium difficile, and other health-care-associated pathogens by dogs that participate in animal-assisted interventions. J Am Vet Med Assoc. 2009;234: 1404–1417. pmid:19480620
- 50. Hensgens MPM, Keessen EC, Squire MM, Riley T V., Koene MGJ, De Boer E, et al. Clostridium difficile infection in the community: A zoonotic disease? Clin Microbiol Infect. 2012;18: 635–645. pmid:22536816
- 51. Janezic S, Dingle K, Alvin J, Accetto T, Didelot X, Crook DW, et al. Comparative genomics of Clostridioides difficile toxinotypes identifies module-based toxin gene evolution. Microb Genomics. 2020;6: 1–13. pmid:33030421
- 52. Thanissery R, McLaren MR, Rivera A, Reed AD, Betrapally NS, Burdette T, et al. Clostridioides difficile carriage in animals and the associated changes in the host fecal microbiota. Anaerobe. 2020;66: 102279. pmid:33022384
- 53. Turner NA, Smith BA, Lewis SS. Novel and emerging sources of Clostridioides difficile infection. PLoS Pathog. 2019;15: 1–6. pmid:31856240
- 54. Freeman J, Vernon J, Morris K, Nicholson S, Todhunter S, Longshaw C, et al. Pan-European longitudinal surveillance of antibiotic resistance among prevalent Clostridium difficile ribotypes. Clin Microbiol Infect. 2015;21: 248.e9–248.e16. pmid:25701178
- 55. Davies KA, Ashwin H, Longshaw CM, Burns DA, Davis GL, Wilcox MH. Diversity of Clostridium difficile PCR ribotypes in Europe: results from the European, multicentre, prospective, biannual, point-prevalence study of Clostridium difficile infection in hospitalised patients with diarrhoea (EUCLID), 2012 and 2013. Eurosurveillance. 2016;21: 1–11. pmid:27470194