Accepted Manuscript

Title: Carriage and acquisition rates of Clostridium difcile in

hospitalized horses, including molecular characterization,
multilocus sequence typing and antimicrobial susceptibility of
bacterial isolates

Author: C. Rodriguez B. Taminiau B. Brevers V. Avesani J.

Van Broeck A.A. Leroux H. Amory M. Delmee G. Daube

PII: S0378-1135(14)00249-1
DOI: http://dx.doi.org/doi:10.1016/j.vetmic.2014.05.013
Reference: VETMIC 6619

To appear in: VETMIC

Received date: 14-2-2014

Revised date: 15-4-2014
Accepted date: 5-5-2014

Please cite this article as: Rodriguez, C., Taminiau, B., Brevers, B.,
Avesani, V., Van Broeck, J., Leroux, A.A., Amory, H., Delmee, M., Daube,
G.,Carriage and acquisition rates of Clostridium difcile in hospitalized
horses, including molecular characterization, multilocus sequence typing and
antimicrobial susceptibility of bacterial isolates, Veterinary Microbiology (2014),
http://dx.doi.org/10.1016/j.vetmic.2014.05.013

This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
1 Carriage and acquisition rates of Clostridium difficile in hospitalized horses, including
2 molecular characterization, multilocus sequence typing and antimicrobial susceptibility
3 of bacterial isolates
4
5 C. Rodrigueza,*, B. Taminiaua, B. Brversa, V. Avesanib, J. Van Broeckb, A. A. Lerouxc, H.
6 Amoryc, M. Delmeb, G. Daubea

t
ip
7
8 a
Food Science Department, FARAH, Faculty of Veterinary Medicine, University of Lige,

cr
9 Lige, Belgium
10 b
Microbiology Unit, Catholic University of Louvain, Brussels, Belgium

us
11 C
Equine Teaching Hospital, Clinical Department of Companion Animals and Equids, Faculty
12 of Veterinary Medicine, University of Lige, Belgium
13
14
15
an
*Corresponding author at: Food Science Department, Faculty of Veterinary Medicine,
University of Lige, B43bis, Sart-Tilman, 4000 Lige, Belgium. Tel. : +32 4366.40.15 ; fax
M
16 +32 4366.40.44. E-mail address: c.rodriguez@ulg.ac.be
17
18
d

19 Highlights
te

20 1. Faecal samples of horses admitted to a veterinary teaching hospital were collected

21 2. Ten isolates were obtained with a total of seven different PCR-ribotypes
22 3. A high resistance to gentamicin, clindamycin and ceftiofur was found
p

23 4. MLST analysis showed that most of the isolates clustered in the same lineage
ce

24 5. C. difficile colonization in horses appears to be is transient by different types

25
26
Ac

27 Abstract
28
29 Clostridium difficile has been identified as a significant agent of diarrhoea and enterocolitis in
30 both foals and adult horses. Hospitalization, antibiotic therapy or changes in diet may
31 contribute to the development of C. difficile infection. Horses admitted to a care unit are
32 therefore at greater risk of being colonized. The aim of this study was to investigate the
33 carriage of C. difficile in hospitalized horses and the possible influence of some risk factors in
34 colonization. During a seven-month period, faecal samples and data relating the clinical
35 history of horses admitted to a veterinary teaching hospital were collected. C. difficile isolates

1
Page 1 of 24
36 were characterized through toxin profiles, cytotoxicity activity, PCR-ribotyping, antimicrobial
37 resistance and multilocus sequence typing (MLST). Ten isolates were obtained with a total of
38 seven different PCR-ribotypes, including PCR-ribotype 014. Five of them were identified as
39 toxinogenic. A high resistance to gentamicin, clindamycin and ceftiofur was found. MLST
40 revealed four different sequencing types (ST), which included ST11, ST26, ST2 and ST15,
41 and phylogenetic analysis showed that most of the isolates clustered in the same lineage.

t
ip
42 Clinical history suggests that horses frequently harbour toxigenic and non-toxigenic C.
43 difficile and that in most cases they are colonized regardless of the reason for hospitalization;

cr
44 the development of diarrhoea is more unusual.
45

us
46 Keywords
47
48 Clostridium difficile, hospitalized horses, multilocus sequence typing
49
50 1. Introduction
an
M
51
52 Most human Clostridium difficile infections (CDI) are acquired in hospitals and nursing
53 homes following antibiotic therapy. It seems that infected patients and contaminated
d

54 environments play an important role in the transmission of this pathogen (Bengualid et al.,
te

55 2011; Weber et al., 2013a). Moreover, colonized new admissions also contribute to C. difficile
56 transmission in hospitals (Clabots et al., 1992). Reported prevalence rates of C. difficile range
p

57 from 5.9% up to 11% in asymptomatic carriers at admission and from 4% up to 21% for
ce

58 hospital acquisition. More than 63% of infected patients remain asymptomatic (Barbut et
59 Petit, 2001).
60
Ac

61 C. difficile is an important agent of diarrhoea and enterocolitis in foals (Uzal et al., 2012) and
62 horses (Arroyo et al., 2006; Weese et al., 2006). Newborn foals can suffer spontaneous C.
63 difficile infection with watery or bloody diarrhoea several days before death (Diab et al.,
64 2013). Additionally, in foals, a possible synergism of C. perfringens type C and C. difficile
65 has been suggested, characterized by the presence of a necrotic mucosa with a superficial
66 pseudomembrane, haemorrhage and vascular thrombosis in the small intestine and colon
67 (Uzal et al., 2012). C. difficile has also been hypothesized to be the aetiology of duodenitis-
68 proximal jejunitis disease in adult horses (Arroyo at al., 2006). As in humans, the major risks
69 factors for the development of nosocomial C. difficile associated disease (CDAD) in horses

2
Page 2 of 24
70 include hospitalization, antibiotic therapy, changes in diet and pre- or post-surgical feed
71 withdrawal. Transmission by oral-faecal route includes contact with other infected horses or a
72 contaminated environment (Diab et al., 2013). An interspecies transmission, including one
73 involving human beings, has also been speculated (Rodriguez-Palacios et al., 2013).
74
75 Carriage of multiple strains of C. difficile in the gastrointestinal tract of healthy horses has

t
ip
76 been reported (Schoster et al., 2012a). As in other young animals (particularly piglets and
77 calves (Rodriguez et al., 2012)), neonatal foals are more likely than adult horses to be carriers

cr
78 of this bacterium (Rodriguez-Palacios et al., 2013). Mare-foal pairs can harbour C. difficile
79 subclinically and therefore potentially serve as reservoirs for cross-colonization (Magdesian

us
80 and Leutenegger, 2011).
81
82 Many articles have reported the prevalence of C. difficile in horses, mares and foals from
83
84
an
different ranches or breeding farms over a specific period of time, but screening for C.
difficile in an equine hospital has rarely been addressed. Only one previous study has
M
85 examined the prevalence of C. difficile in horses with normal faeces admitted on Sundays and
86 Mondays in a large animal clinic (Medina-Torres et al., 2011).
87
d

88 The main objective of this study was to assess the presence of C. difficile in hospitalized
te

89 horses at an equine medical teaching clinic. Faeces of horses with soft or liquid bowel
90 movements were analysed at the moment of the episode of diarrhoea. In addition, horses with
p

91 an extended clinical stay were tracked during their hospitalization. All the isolates were
ce

92 characterized by genotyping, PCR-ribotyping, toxigenic activity and antibiotic resistance.

93 Further characterization was performed by multi-locus sequencing typing (MLST) analysis in
94 order to study clonal relationships of the isolates.
Ac

95
96 2. Materials and methods
97
98 2.1. Sample collection
99
100 A prospective study was conducted over seven months at the Equine Clinic, Department of
101 Companion Animals and Equids, Faculty of Veterinary Medicine, University of Liege,
102 between January - April 2011 and October - December 2011. During this period, a total of

3
Page 3 of 24
103 580 horses were admitted to the clinic (emergencies or consultations), with about 250
104 hospitalizations.
105
106 Between one and three samples were collected from each horse enrolled in the study. Eligible
107 horses were animals with a hospital stay of at least one day. In addition, faeces of horses with
108 soft or liquid bowel movements were analysed at the moment of the episode of diarrhoea.

t
ip
109 Whenever possible, horses enrolled in the study whose hospital admission was prolonged
110 over nine days were monitored every week for the presence of C. difficile. The exclusion

cr
111 criteria were horses with dysphoric mood, intolerable stress or other disease for which
112 sampling by rectal manipulations was not recommended. Horses with a highly contagious

us
113 infectious disease and isolated in quarantine boxes were also excluded from the study. All
114 horses were documented for data relating to clinical history, diagnostic findings and treatment
115 received during hospitalization. In addition, all predisposing factors for developing C. difficile
116
117
an
infection (CDI), such as the prescription of antimicrobials, were carefully recorded. Faecal
sampling was performed directly via the rectum. Samples obtained were scored as normal,
M
118 diarrhoea or bloody diarrhoea faeces. All samples were processed on the same day
119 immediately after transport to the laboratory (with travel at room temperature and for a
120 maximum of 45 min). In the case of a diarrhoea episode during the night or on a non-working
d

121 day, faecal samples were stored at 4C in the hospital for a maximum of 4 days before
te

122 analysis.
123
p

124 2.2. C. difficile culture and characterization

ce

125
126 Fresh faeces (0.1 g) were spread directly on home-made cycloserine cefoxitin fructose
127 taurocholate (CCFT) plates (Delme et al., 1987) and incubated in an anaerobic workstation
Ac

128 (Led Techno, Heusden-Zolder, Belgium) at 37C for three days. At the same time, one gram
129 of faeces was inoculated into 9 ml of CCFT broth as previously described (Rodriguez et al.,
130 2012) and incubated anaerobically for 72 h at 37C. After the enrichment phase,
131 approximately 10 l of the broth was spread on CCFT plates and incubated anaerobically at
132 37C for two days. Initial identification of C. difficile colonies was based on morphological
133 criteria such as yellowish colonies with an appearance of ground glass and a characteristic
134 horse manure odour. One morphologically suspected colony per plate was subcultured onto
135 blood agar (5% Sheep Blood; Biorad, Nazareth, Belgium) and checked using a C. difficile
136 latex agglutination rapid test Kit DR 1107A (Oxoid, Dardilly, France). Identification was

4
Page 4 of 24
137 confirmed by detection of a species-specific internal fragment of tpi and detection of genes
138 for toxin A, B and binary toxin (cdtA) as described previously (Rodriguez et al., 2012). A
139 cytotoxicity assay using confluent monolayer MRC-5 cells was carried out as previously
140 described (Rodriguez et al., 2012).
141
142

t
2.3. Molecular typing of C. difficile isolates

ip
143
144 PCR-ribotyping was performed using the primers and protocol of Bidet et al. (1999).

cr
145 International numbers were used for C. difficile strains that presented a PCR-ribotype profile
146 matching the Cardiff ribotypes from the strain collection available in our laboratory.

us
147 Otherwise, isolates were identified with internal nomenclature.
148
149 All isolates were additionally retested by the Genotype Cdiff system (Hain Lifescience,
150
151
an
Nehren, De), according to the manufacturers instructions, for the presence of the tpi, all toxin
genes (tcdA, tcdB, cdtA and cdtB) and deletions in the regulator gene tcdC and gyrA mutation.
M
152
153 2.4. Antimicrobial susceptibility testing
154
d

155 All of the horse isolates were tested for susceptibilities to a total of ten antimicrobial agents
te

156 by disc diffusion (n=8) and E-test (n=2).

157
p

158 Disc diffusion was performed with standard disc (Becton-Dickinson, Erembodegem,
ce

159 Belgium) of rifampin (25 g), erythromycin (15 g), oxytetracycline (30 g), vancomycin
160 (30 g), penicillin (10 g), clindamycin (2 g), ceftiofur (30 g) and gentamicin (10 g) on
161 Brucella Blood Agar with hemin and vitamin K1 (Becton-Dickinson) according to the French
Ac

162 Society of Microbiology (SFM) (www.sfm-microbiologie.org) protocols. These

163 antimicrobials were chosen because they are widely used in the equine teaching hospital
164 featured in the study, or because they have been associated with C. difficile-associated disease
165 or its treatment. The zone diameters were read after 24 h of anaerobic incubation at 37C. The
166 resistant and susceptible zone diameters were defined as reported by Delme et Avesani
167 (1988): rifampin no zone and >23 mm, erythromycin <13 mm and >20 mm, oxytetracycline
168 <14 mm and >23 mm, clindamycin no zone and >12 mm. The zone diameter breakpoint for
169 vancomycin was 19 mm as proposed by Erikstrup et al. (2012). Susceptibility and resistance
170 to penicillin was defined at the limits of <8 mm and 29 mm established by Cattoir et al.,

5
Page 5 of 24
171 2008. The remaining diameters were based on the only values available for other Gram +
172 bacteria reported by Marie et al. (2000) and the European Committee on Antimicrobial
173 Susceptibility Testing (EUCAST) (www.eucast.org) as follows: ceftiofur <21 mm and 21
174 mm and gentamicin <20 mm and 20 mm.
175
176 Susceptibility to metronidazole and moxifloxacin was determined by the Etest method

t
ip
177 (Lucron ELITechGroup St- Martens-Latem, Belgium) on Schaedler with Vit K1 and 5%
178 sheep blood (Becton-Dickinson) according to the manufacturers instructions. Metronidazole

cr
179 was tested because is the first choice antibiotic for adult horses with diarrhoea in the equine
180 hospital studied (if they are excluded from the human food chain). Resistance to moxifloxacin

us
181 was evaluated because is quite common in human C. difficile isolates (Barbut et al., 2007).
182 Plates were incubated anaerobically at 37C for 48 h. The susceptibility and resistance
183 breakpoints for metronidazole (s 8 g/ml; r 32 g/ml) and moxifloxacin (s2 g/ml; r 8
184
185 an
g/ml) used for interpretation were those recommended by the Clinical and Laboratory
Standard Institute (CLSI, 2010). Bacteroides fragilis ATCL 25285 was tested as a quality
M
186 control.

187 2.5. C. difficile multilocus sequencing

d

188 Seven housekeeping loci (adk, atpA, dxr, glyA, recA, sodA and tpi) were used for the analysis
te

189 of C. difficile isolates by MLST according to the protocol described previously by Griffiths et
190 al. (2010). PCR products were purified with a Wizard SV Gel and PCR Clean-Up System kit
p

191 (Promega, Leiden, The Netherlands). The Sanger sequencing reaction was carried out with
ce

192 the BigDye terminator kit version 3.1 (Applied Biosystems, Life technologies Europe BD,
193 Gent, Belgium) and resolved with a 3730 ABI capillary sequencer (Applied Biosystems) (48
194 capillaries). Results were analysed using the Geneious program (http://www.geneious.com).
Ac

195 The allele number, clade and sequence type (ST) were assigned according to the C. difficile
196 MLST reference database (http://pubmlst.org/cdifficile). A dendrogram was constructed using
197 the Geneious program (Drummond et al., 2013).

198

199 3. Results
200
201 3.1. C. difficile prevalence in hospitalized horses

6
Page 6 of 24
202 During the course of the seven-month study period, 102 faecal samples were collected from a
203 total of 73 hospitalized horses. Eighteen horses were sampled on more than one occasion
204 either because they had a long hospital stay or because they suffered an additional diarrhoea
205 episode after the first collection. Ten out of a total of 73 horses (13.7%) tested positive for C.
206 difficile but only two presented clinical signs of diarrhea associated with CDI. Most of the
207 positive samples (8/10) were detected after three days of enrichment, but two tested positive

t
ip
208 in direct culture without enrichment of the faeces. Because these two samples also tested
209 positive after three enrichment days, a total of 12 strains were obtained from ten horses. None

cr
210 of the horses tested positive for C. difficile on more than one occasion.

us
211 Of the total of 73 animals tested, 41 horses (56.2%) presented gastrointestinal disorders
212 (diarrhea or colic). C. difficile was detected in five (12.2%) of these 41 horses with
213 gastrointestinal problems. Two had diarrhea and were suspected to suffer from CDI (one new

an
214 born foal with bloody diarrhea and septicaemia and one adult horse with diarrhea and
215 anorexia). Three other adult horses tested positive for C. difficile had colic with a diagnosis of
216 impaction and volvulus of the jejunum, nephrosplenic entrapment and incarceration of the
M
217 jejunum respectively. The remaining 36 horses (87.8%) tested negative for C. difficile, with
218 intestinal disorders diagnosed as the following: 18 had diarrhea (50%) and another 18 (50%)
d

219 suffered an episode of colic. The most common causes of colic for these 18 C. difficile
220
te

negative horses included: colon impaction (5/18; 27.8%), nephrosplenic entrapment (4/18;
221 22.2%) and colonic displacement (3/18; 16.7%).
p

222 Of the remaining 32 animals (43.8%) not affected by gastrointestinal problems, C. difficile
ce

223 was isolated from five horses (15.6%). The clinical diagnosis of each positive animal was hip
224 fracture, sinusitis, wound on leg, chorioptic mange and paraphimosis, respectively. In
225 contrast, 84.4% (27/32) of the horses without gastrointestinal problems tested negative for C.
Ac

226 difficile. Their clinical diagnoses were varied; for example mange and other parasites (n=7),
227 wounds (n=6), dislocation and bone fracture (n=2), among others.

228 In relation to clinical interventions and treatments, previous gastrointestinal surgery had been
229 carried out on 11 of the negative horses, but in only one of the 10 C. difficile positive animals.
230 Regarding the antimicrobial therapy, a total of nine horses tested positive for C. difficile had
231 previously received an antibiotic medication. Cefquinome and penicillin were prescribed for
232 six and one positively tested horses respectively. A combined antibiotic treatment composed
233 of two or more different antibiotics was administrated to the two other positively tested

7
Page 7 of 24
234 horses. Among the animals that tested negative for C. difficile, 35 had received antimicrobial
235 therapy. Cefquinome and penicillin was administrated to nine and two negatively tested
236 horses respectively, while 37.1% (13/35) of C. difficile negative horses had received a
237 combined antibiotic treatment (Table 1).

238 Of the ten horses carrying C. difficile in their faeces, two animals had been assessed positive

t
239 less than two days after admission (H7252 and H3113). Another two horses were suspected of

ip
240 being colonized during their hospitalization (H4850 and H0521), as they were both tested

cr
241 negative for C. difficile during initial sampling on admission. For the remaining six positively
242 tested horses, initial sampling could not be performed until five days after admission.

us
243 Monitoring of the hospitalized horses reveals that two C. difficile positive animals tested
244 negative after ten and 20 days of clinical stay (H6647 and H3113). Three horses were
245 euthanized during the study. One of them (H7516) was identified as positive for C. difficile

an
246 only in a second sampling after euthanasia (Table 2).

247 3.2. Toxin gene profiles, toxin activity, PCR-ribotyping and Genotype Cdiff Test
M
248 Seven different PCR-ribotypes were identified. Only one strain has a ribotype profile
249 corresponding to an international collection number (014). The remaining isolates were not
d

250 associated with any reference Cardiff ribotypes (UCL16L, UCL5a, UCL228, UCL9,
te

251 UCL261, UCL16a). Five out of these seven different PCR ribotypes had toxic activity (75%
252 of all isolates). All toxigenic isolates encoded toxin A and B, while only PCR-ribotype
p

253 UCL5a also contained the binary toxin (Table 2).

ce

254 The two animals assessed as C. difficile positive after less than two days of admission carried
255 toxigenic C. difficile PCR ribotypes UCL16L and UCL5a. Isolates from the horses suspected
256 to have been colonized during the hospitalization were also toxigenic and identified as PCR
Ac

257 ribotypes UCL16a and UCL228. Gastrointestinal surgery or diarrhoea was not associated with
258 the presence of a specific PCR-ribotype.

259

260 3.3. Antibiotic resistance

261 Isolates were tested for resistance to a total of 10 antibiotics. All strains were susceptible to
262 vancomycin, metronidazole and rifampicin. In addition, all of the isolates showed full
263 sensibility to moxifloxacin, with the exception of one non-toxigenic isolate (H5197) that had

8
Page 8 of 24
264 intermediate resistance to this drug. These results were correlated with the absence of a
265 mutation in the gyrA gene. Moreover, all the strains were resistant to clindamycin, gentamicin
266 and ceftiofur. For penicillin, only one isolate (H3113) was resistant while all others showed
267 intermediate resistance. Resistance to erythromycin was detected in one non-toxigenic isolate
268 (H5197) and also in one toxigenic strain (H2867) that was also intermediately resistant to
269 tetracycline (Figure 1). The only isolate from a horse without an antibiotic treatment (H7252)

t
ip
270 showed a very similar antimicrobial susceptibility to the other strains with coresistance to
271 clindamycin, gentamicin and ceftiofur and intermediate resistance to penicillin.

cr
272

us
273 3.4. C. difficile MLST analysis

274 In order to determine the allelic diversity between C. difficile strains from the hospitalized

an
275 horses, each isolate was characterized by MLST. The analyses revealed four different
276 sequencing types (ST), which included ST11, ST26, ST2 (two isolates) and ST15 (one
M
277 isolate). Only the two strains included in ST11 and identified as PCR-ribotype UCL 5a were
278 binary toxin positive. The two nontoxigenic PCR-ribotypes UCL9 and UCL261 were related
279 with ST26 and ST15 respectively. All the strains with the same PCR-ribotype (UCL9 and
d

280 UCL16a) clustered in the same lineage. For three strains (H7252, H7516 and H4850) ST
te

281 assignment was not possible as no loci sequence combination matched the allelic profile of
282 the isolates.
p

283
284 Two different clades were assigned. Clade 1 was correlated with PCR ribotypes UCL261
ce

285 (non-toxigenic isolate) and UCL16a. No clade assignation was available for the types UCL9,
286 UCL16L, UCL228 and 014. The two isolates with PCR ribotype UCL5a were included in
Ac

287 clade 5, which appear in a different cluster from all the remaining strains (Figure 1).
288
289 4. Discussion
290
291 Few studies describing the carriage of C. difficile in hospitalized horses exist, and the majority
292 of these focus their investigation on animals with a particular health problem, such as
293 duodenitisproximal jejunitis (Arroyo et al., 2006), post-operative diarrhoea or colic (Niwa et
294 al., 2013), acute haemorrhagic diarrhoea (Uzal et al., 2012), antibiotic associated diarrhoea
295 (Barr et al., 2013), mare-foal pairs infection (Magdesian et al., 2011) or the prevalence

9
Page 9 of 24
296 variation in different intestinal compartments after euthanasia (Schoster et al., 2012b).
297 Overall, we reported an estimated C. difficile colonization rate of 13.7% (10/73) in an equine
298 hospital setting during a total of seven months of study. Only two horses out of 73 (incidence
299 of 2.7%) presented clinical sign of diarrhea associated with CDI. From currently available
300 data, it seems that foals are more likely to suffer the infection, and that antimicrobials increase
301 the risk of developing the disease. Thean et al. (2011) observed a C. difficile prevalence of

t
ip
302 23% (14/62) in horses with diarrhea tested in Australia (ten of them were foals). Baverud et
303 al. (2003) reported a prevalence of 42% (18/48) in horses that developed acute colitis during

cr
304 antibiotic treatment, and 6% (4/72) in diarrhoeic mature horses with no history of antibiotic
305 administration. The results of our study suggest a low endemic CDI incidence in horses

us
306 admitted to the veterinary hospital studied.
307
308 Limited information is available regarding the presence of C. difficile in hospitalised horses in
309
310
an
the absence of clinical signs of C. difficile disease. Medina-Torres et al. (2011) observed a
prevalence of 4.8% (4/82) from patients at a veterinary teaching hospital with normal faeces
M
311 at the time of admission. In the present study, the colonization rate detected in horses without
312 CDI was 11% (8/73) over the seven months. Remarkably, none of the horses tested positive
313 for C. difficile on more than one occasion. These findings suggest that colonisation is transient
d

314 in most cases. This colonisation may be influenced by many stress situations that alter the
te

315 intestinal flora, often difficult to quantify (change of diet, transportation, hospitalization and
316 surgical or medical treatment among others) (Baverud et al., 2004). Transit shedding of C.
p

317 difficile has been previously reported in healthy adult horses (Schoster et al., 2012b) but also
ce

318 in cattle (Rodriguez-Palacios et al., 2011) and humans (Ozaki et al., 2004).
319
320 The mean age of horses enrolled in this study was one year, with the great majority older than
Ac

321 two years; nevertheless an association between an age of more than 24 months and the
322 carriage of C. difficile could not be established. Three foals under 17 months old participated
323 in the study and of these only an eight day-old filly was colonized with C. difficile. As
324 previously reported, foals may develop C. difficile infection in the first days of life (Diab et
325 al., 2013). In this case, the animal presented the classic signs of spontaneous C. difficile
326 infection including depression and bloody diarrhoea (Diab et al., 2013). 48 hours after the
327 hospital admission, the foal was treated with metronidazole, cefquinome and marbofloxacin.
328 An initial sampling of faeces was carried out four days after the start of antibiotic treatment
329 with a negative result for the presence of C. difficile. A second sample was collected 24 hours

10
Page 10 of 24
330 later, directly from the large intestine after euthanasia. The second analysis revealed the
331 carriage of toxigenic C. difficile in the intestinal contents. It has been suggested that even
332 though faecal samples can demonstrate the presence of C. difficile, rectal samples might not
333 absolutely reflect the status of proximal compartments (Schoster et al., 2012b).
334
335 Intestinal flora perturbations and colic may facilitate the proliferation of C. difficile in horses

t
ip
336 (Donaldson et Palmer, 1999). In this study two C. difficile positive horses (H5017 and H6695)
337 presented signs of colic and had received an antibiotic therapy, but only one harboured a

cr
338 toxigenic C. difficile strain (the other was colonized with a non-toxigenic type). Two horses
339 with health problems other than enteric disease and treated with antibiotics tested positive for

us
340 toxigenic C. difficile after nine and 18 days of hospitalization. They had tested negative at the
341 time of admission, and it was therefore suspected they were colonized during the period of
342 hospitalization. However, neither of the horses showed signs of C. difficile infection. Two
343
344
an
other positive horses became negative eight days after stopping cefquinome antibiotic
treatment. This observation correlates with the results of numerous studies reporting antibiotic
M
345 exposure as the most important risk factor for C. difficile colonization in humans, including
346 clindamycin, cephalosporins or fluoroquinolones (Deshpande et al., 2013; Slimings and Riley,
347 2013). Nevertheless, in the present study no relation between C. difficile infections and
d

348 clinical signs or medication could be obtained, since the group of positive horses was too
te

349 small.
350
p

351 A wide variety of types were observed without the predominance of a particular PCR-
ce

352 ribotype. In most instances (six horses out 10), animals did not present any of the clinical
353 signs classically associated with C. difficile infection, such as diarrhoea, abdominal pain,
354 colic, nausea, depression or dehydration (Keel and Songer., 2006; Diab et al., 2013). This
Ac

355 correlates with the findings of previous studies describing horses infected by C. difficile
356 subclinically (Magdesian and Leutenegger, 2011; Schoster et al., 2012a; Schoster et al.,
357 2012b). Moreover, it has been suggested that healthy or sick carriers without any signs of the
358 C. difficile infection may harbour strains that do not produce toxins (Uzal et al., 2012). In the
359 present study, three positive-tested horses were colonized with non-toxigenic C. difficile
360 strains (PCR-ribotypes UCL 9 and UCL 261). Two of these horses showed no sign of enteric
361 disease. Non-toxigenic C. difficile strains have been described in horses (Thean et al., 2011)
362 and other farm animals including pigs and cattle (Pelaez et al., 2013; Rodriguez et al., 2012;

11
Page 11 of 24
363 Zidaric et al., 2012). Furthermore, the non-toxigenic PCR-ribotype UCL 9 has been
364 previously isolated from suckling piglets in Belgium (Rodriguez et al., 2012).
365
366 Resistance to more than one antimicrobial tested was found among the isolates obtained.
367 Resistance of C. difficile to multiple antimicrobials has been previously described in several
368 studies conducted in both humans and animals (Pelaez et al., 2013; Pirs et al., 2013; Zidaric et

t
ip
369 al., 2012; Weber et al., 2013). Furthermore, all of the isolates were resistant to ceftiofur,
370 clindamycin and gentamicin, which is not surprising as these antibiotics are the most

cr
371 commonly used in the equine clinic. Even though gentamicin has been associated with C.
372 difficile diarrhoea in horses (Diab et al., 2013), no relationship between treatment of animals

us
373 with this medication and the appearance of soft or liquid bowel movements could be
374 established. Furthermore, no association between antimicrobial resistance and toxigenic
375 strains was observed, which correlates with the findings of a previous study (Hanna et al.,
376
377
an
2005). Data observed suggested that antibiotics might have an important role in C. difficile
colonization, but that the distribution of other enteric bacteria may be involved in the
M
378 occurrence of diarrhoea. As previously suggested by Diab et al. (2013), metagenomic analysis
379 of the gut microbiome is useful in improving understanding not only of C. difficile associated
380 diarrhoea but also other etiologies of diarrhoea in horses.
d

381
te

Some studies have used MLST to discriminate human C. difficile strains, but only a few have
382 applied this method to animal isolates (Leme et al., 2004; Leme et Pons, 2010; Stabler et
p

383 al., 2012). To date, this study is the first to address genetic analysis of C. difficile by MLST
384 in an equine hospital. We found a clear concordance between some PCR-ribotypes and ST as
ce

385 described in human isolates from hospitals (Dingle et al., 2011; Weber et al., 2013b). Two
386 horse isolates (both of them PCR-ribotype UCL 5a) were included in ST 11. In the literature,
Ac

387 ST 11 is correlated with PCR-ribotypes 078, 126 or 033 among others (Knetsch et al., 2012).
388 Remarkably, all of them are binary toxin positive, as are our two strains PCR-ribotype
389 UCL5a. For three strains, the allelic profile over the seven loci did not match with any
390 number assigned to the unique allelic profiles available on the PubMLST database. Two of
391 these isolates are closely related in the constructed neighbour-joining phylogenetic tree, and
392 the allelic profile is the same. Nonetheless, results obtained do not allow us to establish any
393 relationship between the acquisition of C. difficile in the hospital and a particular circulating
394 strain, or between the presence of diarrhoea and one particular sequence type. The same
395 observation has previously been reported in an epidemiological surveillance of C. difficile

12
Page 12 of 24
396 infection in a tertiary care hospital (Weber et al., 2013b).

397 Our study has several limitations. Firstly, only 18 of 73 horses were monitored for the
398 presence of C. difficile throughout their stay in the clinic because only animals whose
399 admission was prolonged (over nine days) were monitored. In other cases, the animals health
400 status prevented monitoring. Consequently, some horses colonized during hospitalization may

t
401 have been missed, and it was not possible to determine the duration of carriage for some other

ip
402 positive animals. Additionally, the small number of C. difficile positive horses made it

cr
403 impossible to conduct an epidemiological analysis. For this reason, care has to be taken when
404 interpreting the results. For six positive horses, initial sampling could not be performed until

us
405 five days after admission; we were therefore not able to determine if they were colonized
406 during or before admission. Only two positive samples were detected directly without
407 enrichment media, while in eight other cases detection required an enrichment step. This

an
408 finding indicates that in most cases the spore load was low and the colonization may have
409 been transient. Another limitation of the antibiotic resistance test is the need to use different
410 criteria breakpoints when SFM criteria were not available, and with a different culture media
M
411 than usually recommended. Additionally, the lack of sufficient reference strains in our
412 laboratory only allowed us to identified one ribotype profile corresponding to an international
d

413 collection number while the remaining 6 PCR-ribotypes were identified with an internal
414
te

nomenclature.

415 In conclusion, the results of this study show that CDI is very rare in the studied equine clinic
p

416 and that C. difficile colonization is transient by different toxigenic and non-toxigenic types,
ce

417 suggesting that an appropriate infection prevention strategy may reduce the associated
418 disease.
Ac

419 Acknowledgements

420 Our most sincere thanks to Phyllis Smith and Catherine-Painter of the Institut Suprieur des
421 Langues Vivantes (University of Lige) for their support in editing the manuscript.

422 This study was funded by the Federal Public Service of Health, Food Chain Safety and
423 Environment (contract RF09/6226).

424

425 References

13
Page 13 of 24
426 Arroyo, L.G., Henry, R., Stampfli, H.R., Weese, J.S., 2006. Potential role of Clostridium
427 difficile as a cause of duodenitis-proximal jejunitis in horses. J. Med. Microbiol. 55, 605-608.

428 Barbut, F., Petit, J.C., 2001. Epidemiology of Clostridium difficile-associated infections. Clin.
429 Microbiol. Infect. 7, 405410.

t
430 Barbut, F., Mastrantonio, P., Delme, M., Brazier, J., Kuijper, E., Poxton, I., European Study

ip
431 Group of Clostridium difficile (ESGCD), 2007. Prospective study of Clostridium difficile
432 infections in Europe with phenotypic and genotypic characterisation of the isolates. Clin.

cr
433 Microbiol. Infect. 13, 1048-1057.
434

us
435 Barr, B.S., Waldrige, B.M., Morresey, P.R., Reed, S.M., Clark, C., Belgrave, R., Donecker,
436 J.M., Weigel, D.J., 2013. Antimicrobial-associated diarrhoea in three equine referral

an
437 practices. Equine Vet. J. 45, 154-158.
438
439 Baverud V., 2004. Clostridium difficile diarrhea: infection control in horses. Vet. Clin. Equine
M
440 20, 615-30.
441
442 Baverud, V., Gustafsson, A., Franklin, A., Aspan, A., Gunnarsson, A., 2003. Clostridium
d

443 difficile: prevalence in horses and environment, and antimicrobial susceptibility. Equine Vet.
te

444 J. 35, 465-71.

445
p

446 Bengualid, V., Umesh, K.C., Alapati, J., Berger, J., 2011. Clostridium difficile at a
ce

447 community hospital in the Bronx, New York : incidence, prevalence and risk factors from
448 2006 to 2008. Am. J. Infect. Control. 39, 183-187.
Ac

449 Bidet, P., Barbut, F., Lalande, V., Burghoffer, B., Petit, J.C., 1999. Development of a new
450 PCR-ribotyping method based on ribosomal RNA gene sequencing. FEMS Microbiol. Lett.
451 175, 261-266.
452
453 Cattoir, V., Ould-Hocine, Z.F., Legrand, P., 2008. Antimicrobial susceptibility of
454 Clostridium difficile clinical isolates collected from 2001 to 2007 in a French university
455 hospital. Pathol. Biol. 56, 7-8.
456
457 Clabots, C.R., Johnson, S., Olson, M.M., Peterson, L.R., Gerding, D.N., 1992. Acquisition of

14
Page 14 of 24
458 Clostridium difficile by hospitalized patients : evidence for colonized new admissions as a
459 source of infection. J. Infect. Dis. 166, 561-567.

460 CLSI, 2010. Clinical and Laboratory Standards Institute. 20th Informational Supplement, vol.
461 30. No. 1, M100S20. Wayne, PA. USA.
462

t
463 Delme, M., Vandercam, B., Avesani, V., Michaux, J.L., 1987. Epidemiology and prevention

ip
464 of Clostridium difficile in a leukaemia unit. Eur. J. Clin. Microbiol. 6, 623-627.

cr
465
466 Delme, M., Avesani, V., 1988. Correlation between serogroup and susceptibility to

us
467 chloramphenicol, clindamycin, erythromycin, rifampicin and tetracycline among 308
468 isolates of Clostridium difficile. J. Antimicrob. Chemother. 22, 325-331.
469

an
470 Deshpande, A., Pasupuleti, V., Thota, P., Pant, C., Rolston, D.D., Sferra, T.J., Hernandez,
471 A.V., Donskey, C.J., 2013. Community-associated Clostridium difficile infection and
M
472 antibiotics; a meta-analysis. J. Antimicrob. Chemother. 68, 1951-1961.

473 Diab, S.S., Songer, G., Uzal, F.A., 2013. Clostridium difficile infection in horses : a review.
474
d

Vet. Microbiol. 167, 42-49.

te

475 Dingle, K.E., Griffiths, D., Didelot, X., Evans, J., Vaughan, A., Kachrimaidau, M., Stoesser,
476 N., Jolley, K.T., Harding, R.M., Peto, T.E., Fawley, W., Walker, A.S., Wilcox, M., Crook,
p

477 D.W., 2011. Clinical Clostridium difficile: Clonality and Pathogenicity Locus Diversity. Plos
478
ce

One 6, e19993.
479
480 Donaldson, M.M., Palmer, J.E. Prevalence of Clostridium difficile enterotoxin and
Ac

481 Clostridium difficile toxin A in feces of horses with diarrhea and colic. J. Am. Vet. Med.
482 Assoc. 215, 358-361.
483
484 Drummond, A.J., Ahston, B., Buxton, S., Cheung, M., Heled, J., Kearse, M., Moir, R.,
485 Stones-Avas, S., Thierer, T., Wilson, A., 2013. Geneious v4.8, available from
486 http://www.geneious.com.
487
488 Erikstrup, L.T., Danielsen, T.K.L., Hall, V., Olsen, K.E.P., Kristensen, B., Kahlmeter, G.,
489 Fuursted, K., Justesen, U.S., 2012. Antimicrobial susceptibility testing of Clostridium difficile

15
Page 15 of 24
490 using EUCAST epidemiological cut-off values and disk diffusion correlates. Clin. Microbiol.
491 Infect. 18, 266-272.

492 European Committee on Antimicrobial Susceptibility testing. European Comittee on

493 Antimicrobial Susceptibility testing clinical breakpoints.
494 http://www.eucast.org/clinical_breakpoints.

t
ip
495 Griffiths, D., Fawley, W., Kachrimanidou, M., Bowden, R., Crook, D. W., Fung, R.,
496 Golubchik, T., Harding, R.M., Jeffery, M., Jolley, K.A., Kirton, R., Peto, T.E., Rees, G.,

cr
497 Stoesser, N., Vaughan, A., Walker, A.S., Young, B.C., Wilcox, M., Dingle, K.E., 2010.
498 Multilocus sequence typing of Clostridium difficile. J. Clin. Microbiol. 48, 370-378.

us
499

500

501
502 an
Keel, M.K., Songer, J.G., 2006. The comparative pathology of Clostridium difficile-
associated disease. Vet. Pathol. 43, 225-240.
M
503 Knetsch, C.W., Terveer, E.M., Lauber, C., Gorbalenya, A.E., Harmanus, C., Kuijper, Ed. J.,
504 Corver, J., van Leeuwen, H.C., 2012. Comparative analysis of an expanded Clostridium
d

505 difficile reference strain collection. Infect. Genet. Evol. 12, 1577-85.
te

506 Magdesian, K.G., Leutenegger, C.M., 2011. Real-time PCR and typing of Clostridium
507
p

difficile isolates colonizing mare-foal pairs. Vet. J. 190, 119-123.

ce

508 Marie, J., Morvan, H., Berthelot-Hrault, F., Sanders, P., Kempf, I., Gautier-Bouchardon,
509 A.V., Jouy, E., Kobisch, M., 2000. Antimicrobial susceptibility of Streptococcus suis isolated
510 from swine in France and from humans in different countries between 1996 and 2000. J.
Ac

511 Antimicob. Chemother. 50, 201-209.

512 Medina-Torres, C.E., Weese, J.S., Staempfli, H.R., 2011. Prevalence of Clostridium difficile
513 in horses. Vet. Microbiol. 152, 212-215.

514 Niwa, H., Kato, H., Hobo, S., Kinoshita, Y., Ueno, T., Katayama, Y., Hariu, K., Oku, K.,
515 Senoh, M., Kuroda, T., Nakai, K., 2013. Postoperative Clostridium difficile infection with
516 PCR ribotype 078 strain identified at necropsy in five thoroughbred racehorses. Vet. Rec.
517 173, 607.

16
Page 16 of 24
518 Leme, L., Dhalluin, A., Pestel-Caron, M., Lemeland, J.F., Pons, J.L., 2004. Multilocus
519 sequence typing of humans and animal Clostridium difficile isolates of various toxigenic
520 types. J. Clin. Microbiol. 42, 2609-2617.
521
522 Leme, L., Pons J.L., 2010. Multilocus sequence typing for Clostridium difficile. Methods in
523 Mol. Biol. 646, 77-90.

t
ip
524
525 Ozaqui, E., Kato, H., Kita, H., Karasawa, T., Maegawa, T., Koino, Y., Matsumoto, K.,

cr
526 Takada, T., Nomoto, K., Tanaka, R., Nakamura, S., 2004. Clostridium difficile colonization in
527 healthy adults : transient colonization and correlation with enterococcal colonization. J. Med.

us
528 Microbiol. 53, 167-72.

529 Pelaez, T., Alcala, L., Blanco, J.L., Alvarez-Perez, S., Marin, M., Martin-Lopez, A., Catalan,

an
530 P., Reigadas, E., Garcia, M.E., Bouza, E., 2013. Characterization of swine isolates of
531 Clostridium difficile in Spain : A potential source of epidemic multidrug resistant strains ?
M
532 Anaerobe 22, 45-49.

533 Pirs, T., Avbersek, J., Zdovc, I., Krt, B., Andlovic, A., Lejko-Zupanc, T., Rupnik, M.,
d

534 Ocepek, M., 2013. Antimicrobial susceptibility of animal and human isolates of Clostridium
535 difficile by broth microdilution. J. Med. Microbiol. 62, 1478-1485.
te

536 Rodriguez, C., Taminiau, B., Van Broeck, J., Avesani, V., Delme, M., Daube, G., 2012.
p

537 Clostridium difficile in young farm animals and slaughter animals in Belgium. Anaerobe 18,
ce

538 621-625.
539
540 Rodriguez, C., Avesani, V., Van Broeck, J., Taminiau, B., Delme, M., Daube, G., 2013.
Ac

541 Presence of Clostridium difficile pigs and cattle intestinal contents and carcass contamination
542 at slaughterhouse in Belgium. Int. J. Food Microbiol. 166, 256-262.
543
544 Rodriguez-Palacios, A., Pickworth, C., Loerch, S., Lejeune, J.T., 2011. Transient fecal
545 sheeding and limited animal-to-animal transmission of Clostridium difficile by naturally
546 infected finishing feedlot cattle. Appl. Environ. Microbiol. 77, 3391-7.
547
548 Rodriguez-Palacios, A., Borgmann, S., Kline, T.R., Lejeune, J.T., 2013. Clostridium difficile
549 in foods and animals : history and measures to reduce exposure. Anim. Health Res. Rev. 14,

17
Page 17 of 24
550 11-29.

551 Schoster, A., Staempfli, H.R., Arroyo, L.G., Reid-Smith, R.J., Janecko, N., Shewen, P.E.,
552 Weese, J.S., 2012a. Longitudinal study of Clostridium difficile and antimicrobial
553 susceptibility of Escherichia Coli in healthy horses in a community setting. Vet. Microbiol.
554 159, 364-370.

t
ip
555 Schoster, A., Arroyo, L.G., Staempfli, H.R., Shewen, P.E., Weese, J.S., 2012b. Presence and
556 molecular characterization of Clostridium difficile and Clostridium perfringens in intestinal

cr
557 compartments of healthy horses. BMC Vet. Res. 8, 94.

us
558 Slimings, C., Riley, T.V., 2013. Antibiotics and hospital-aquired Clostridium difficile
559 infection: update of systematic review and meta-analysis. J. Antimicrob. Chemother. 8, in
560 press.

561
562 an
Stabler, R.A., Dawson, L.F., Valiente, E., Cairns, M.D., Martin, M.J., Donahue, E.H., Riley,
T.V., Songer, J.G., Kuijper, Ed. J., Dingle, K.E., Wren, B.W., 2012. Macro and micro
M
563 diversity of Clostridium difficile isolates from diverse sources and geographical locations.
564 Plos One 7, e31559.
d

565
566 Thean, S., Elliott, B., Riley, T.V., 2011. Clostridium difficile in horses in Australia- a
te

567 preliminary study. J. Med. Microbiol. 60, 1188-1192.

568
p

569 Uzal, F.A., Diab, S.S., Blanchard, P., Moore, J., Anthenill, L., Shahriar, F., Garcia, J.P.,
ce

570 Songer, J.G., 2012. Clostridium perfringens type C and Clostridium difficile co-infection in
571 foals. Vet. Microbiol. 156, 395-402.
Ac

572 Weber, D.J., Anderson, D., Rutala, W.A., 2013a. The role of the surface environment in
573 healthcare-associated infections. Curr. Opin. Infect. Dis. 26, 338-44.

574 Weber, I., Riera, E., Dniz, C., Prez, J.L., Oliver, A., Mena, A., 2013b. Molecular
575 epidemiology and resistance profiles of Clostridium difficile in a tertiary care hospital in
576 Spain. Int. J. Med. Microbiol. 303, 128-133.
577

578 Weese, J.S., Toxopeux, L., Arroyo, L., 2006. Clostridium difficile associated diarrhoea in

18
Page 18 of 24
579 horses within the community : predictors, clinical presentation and outcome. Equine Vet. J.
580 38, 185-8.

581 Zidaric, V., Pardon, B., Dos Vultos, T., Deprez, P., Brouwer, M.S.M., Roberts, A.P.,
582 Henriques, A.O., Rupnik, M., 2012. Different antibiotic resistance and sporulation properties
583 within multiclonal Clostridium difficile PCR Ribotypes 078, 126 and 033 in a single calf

t
584 farm. Appl. Environ. Microbiol. 78, 8515-8522.

ip
585

cr
us
an
M
d
p te
ce
Ac

19
Page 19 of 24
602
603 Figure 1. Neighbour-joining phylogenetic tree constructed with the MLST results
604 showing the relationships between horse C. difficile strains.
605
606 ST: sequence type
607 tcdC 39bp: Presence of deletions in the regulator gene tcdC
608 gyrA mut: Presence of mutation in the gyrA gene
609 GM-R: gentamicin resistance; XNL-R: ceftiofur resistance; CC-R: clindamycin resistance; MXF-R:

t
610 moxifloxacin resistance; VA-R: vancomycin resistance; LZ-R: metronidazole resistance; RA-R:

ip
611 rifampicin resistance; E-R: erythromycin resistance; P-R: penicillin resistance; TE-R:
612 oxytetracycline resistance.
613 I: intermediate antimicrobial resistance

cr
614
615

us
an
M
d
p te
ce
Ac

22
Page 20 of 24
Table

t
ip
cr
503 Table 1. Clinical characteristics of the 73 horses enrolled in the study and comparison of of C. difficile colonized versus non-colonized
504 horses.
C. difficile positive horses (%) C. difficile negative horses (%)

us
Overall totals (%) 10 of 73 (13.7) 63 of 73 (86.3)
Mean age in years a 9.1 11.8
Sort by gender

an
Mare 4 (40) 29 (46)
Stallion 1 (10) 8 (12.7)
Gelding 5 (50) 26 (41.3)
Sort by size

M
Light horses 8 (80) 53 (84.1)
Heavy horses 2 (20) 4 (6.3)
Ponies 0 (0) 6 (9.5)
Horses with gastrointestinal disorders 41/73 (56.2) 5 of 41 (12.2) 36 of 41 (87.8)

ed
Diarrhea 2 (40) 18 (50)
Colic 3 (60) 18 (50)
Horses without gastrointestinal disorders 32/73 (43.8) 5 of 32 (15.6) 27 of 32 (84.4)
Dislocation/bone fracture 1 (20) 2 (7.4)
pt
Chorioptic mange/other parasites 1 (20) 7 (25.9)
Wounds 1 (20) 6 (22.2)
Others 2 (40) 12 (44.4)
ce
Horses with an antibiotic treatment 44/73 (60.3) 9 of 44 (20.5) 35 of 44 (79.5)
Penicillin 1 (11.1) 2 (5.7)
Penicillinb-Gentamicin 1 (11.1) 7 (20)
Cefquinome 6 (66.7) 9 (25.7)
Ac

Cefquinomec combined treatment 1 (11.1) 6 (17.1)

Others 0 (0) 11 (31.4)
Horses with a surgery intervention 27/73 (37) 1 of 27 (3.7) 26 of 27 (96.3)
Intestinal surgery 1 (100) 11 (42.3)
Other type of surgery 0 (0) 15 (57.7)
505 a Mean age of the hospitalized horses
506 b Antimicrobial treatment of penicillin combined with gentamicin
507 c Association of cefquinome with other antibiotics as gentamicin, metronidazole, penicillin or marbofloxacin

508 c Including single or combined antimicrobial therapy

18
Page 21 of 24
 t
ip
cr
509
510
511 Table 2. Detailed information of hospitalised horses positive for C. difficile and molecular type of the isolates

us
512
Date of Positive animal Age Hospital stay Diagnostic Diarrhea Intestinal Antibiotic Positive culture PCR- Toxin
sampling identification (years) before positive Surgery Treatment Directa 3 daysb ribotype Profile
culture (days)

an
17/01/11 H7252 11 1 Colic - - - - + UCL16L A+B+CDT-
(nephrosplenic
entrapment)
24/01/11 H5017 11 5 Colic (impaction - - CEF - + UCL16a A+B+CDT-

M
and volvulus of
the jejunum)
31/01/11 H4850 3 22 Paraphimosis - - PEN + + UCL228 A+B+CDT-
03/03/11 H5197 16 35 Chorioptic - - CEF - + UCL9 A- B- CDT-

ed
mange
31/03/11 H6695 4 8 Colic - + PEN-GEN + + UCL261 A- B- CDT-
(incarceration of
the jejunum)
08/04/11 H6647 4 17 Wound on leg - - CEF - + UCL9 A- B- CDT-
29/04/11 H7516
pt8 days 7 Septicaemia +c - LZ-MAR-CEF - + 014 A+B+CDT-
14/11/11 H3113 11 2 Anorexia + - CEF - + UCL5a A+B+CDT+
24/11/11 H2867 6 21 Sinusitis - - CEF - + UCL5a A+B+CDT+
ce
16/12/11 H0521 2 16 Hip fracture - - CEF - + UCL16a A+B+CDT-
513 a Positive results detected after direct culture of the feces

514 b Positive results detected after 3 days of enrichment

515 c bloody diarrhea

Ac

516 CEF: cefquinome; PEN: penicillin; GEN: gentamicin; ENR: enrofloxacin; LZ: metronidazole; MAR: marbloxacin

19
Page 22 of 24
Figure

t
Isolate ST Alleleic Profile Clade PCR tcdC GM-R MXF-R VA-R E-R P-R TE-R

ip
Number (adk, atpA, dxr, glyA, recA, sodA, tpi) Ribotype 39bp XNL-R LZ-R
CC-R RA-R

cr
11 5, 8, 5, 11, 9, 11, 8 5 UCL5a + + - - + I I

us
11 5, 8, 5, 11, 9, 11, 8 5 UCL5a + + - - - + -

an
2, 1, 2, 1, 5, 3, 3 014 - + - - - I -

2 1, 1, 2, 1, 5, 3, 1 1 UCL16a - + - - - I -

M
2 1, 1, 2, 1, 5, 3, 1 1 UCL16a - + - - - I -

ed
pt 26 1, 1, 6, 1, 4, 3, 4 UCL9 - + I - + I -

26 1, 1, 6, 1, 4, 3, 4 UCL9 - + - - - I -
ce
15 1, 1, 6, 1, 8, 5, 1 1 UCL261 - + - - - I -
Ac

2, 1, 6, 1, 1, 3, 1 UCL228 - + - - - I -

2, 1, 6, 1, 1, 3, 1 UCL16L - + - - - I -
518
519

20
Page 23 of 24
 t
ip
cr
520 Figure 1. Neighbour-joining phylogenetic tree constructed with the MLST results showing the relationships between horse C. difficile
521 strains.
522

us
523 ST: sequence type
524 tcdC 39bp: Presence of deletions in the regulator gene tcdC
525 gyrA mut: Presence of mutation in the gyrA gene
526

an
GM-R: gentamicin resistance; XNL-R: ceftiofur resistance; CC-R: clindamycin resistance; MXF-R: moxifloxacin resistance; VA-R: vancomycin
527 resistance; LZ-R: metronidazole resistance; RA-R: rifampicin resistance; E-R: erythromycin resistance; P-R: penicillin resistance; TE-R:
528 oxytetracycline resistance.
529 I: intermediate antimicrobial resistance

M
1
530

ed
pt
ce
Ac

21
Page 24 of 24