Clostridioides difficile infection and One Health: an equine perspective.

Clostridioides (Clostridium) difficile presents a significant health risk to humans and animals. The complexity of the bacterial-host interaction affecting pathogenesis and disease development creates an ongoing challenge for epidemiological studies, control strategies and prevention planning. The recent emergence of human disease caused by strains of C. difficile found in animals adds to mounting evidence that C. difficile infection (CDI) may be a zoonosis. In equine populations, C. difficile is a known cause of diarrhoea and gastrointestinal inflammation, with considerable mortality and morbidity. This has a significant impact on both the well-being of the animal and, in the case of performance and production animals, it may have an adverse economic impact on relevant industries. While C. difficile is regularly isolated from horses, many questions remain regarding the impact of asymptomatic carriage as well as optimization of diagnosis, testing and treatment. This review provides an overview of our understanding of equine CDI while also identifying knowledge gaps and the need for a holistic One Health approach to a complicated issue.

Introduction

First isolated in 1935 from the intestinal flora of human infants, Clostridioides (Clostridium) difficile was initially considered a commensal (Hall and O'Toole, 1935). This perception remained for four decades until C. difficile was finally identified as a causative agent of antimicrobial‐related diarrhoea and life‐threatening pseudomembranous colitis (Bartlett et al., 1978; Larson et al., 1978). Today, C. difficile is recognized as a major cause of gastrointestinal disease affecting both animals and humans, with its ubiquity in the environment becoming increasingly apparent (Lim et al., 2020).

In human populations, C. difficile is the most common cause of infectious healthcare‐associated diarrhoea with the rate of severe cases increasing (McDonald et al., 2018). In the last 20 years, however, focus has turned to the role of C. difficile in animal gastrointestinal disease and the role of animal populations in the amplification and transmission of C. difficile. High prevalence of C. difficile has been consistently reported globally in both swine (mean 43%, range 0%–100%) and cattle (mean 14%, range 0.5%–56.4%), with a 2014 study confirming the relatedness of strains isolated from paired pig and farmer samples (Knetsch et al., 2014; Knight and Riley, 2019). Further genomic studies have provided compelling evidence for a novel zoonotic paradigm for C. difficile infection (CDI) (Knight et al., 2017; Knight et al., 2019). A 2017 study strengthened this animal–human link, identifying a significant association between proximity to livestock farms and the occurrence of community‐acquired CDI case clusters (Anderson et al., 2017). These key aspects have led to a deeper consideration of the impact of C. difficile in a wider range of animal populations.

In horses there are three inter‐related issues pertaining to C. difficile. Primarily there is an animal welfare concern; disease in horses due to C. difficile can cause discomfort in its mildest form and debilitating complications and death at its most severe. Horses are, however, also intertwined with human activity; used as performance, work and companion animals. In 2018/19, the horse racing industry in Australia reportedly contributed AU$9.3 billion to the Australian GDP (Racing Australia, 2020) The socio‐economic impact of CDI on racing, breeding and other equine‐related industries is therefore potentially of great significance. Finally, in the One Health era, which acknowledges the link between human health, animal health and the environment, the role of horses in the dissemination and dispersal of C. difficile to the wider community and environment needs to be considered. The review outlines our current understanding of C. difficile and CDI within equine populations. It reflects on the knowledge gaps and diagnostic shortfalls evident within this emerging field and the importance of adopting a One Health approach to achieve effective infection prevention and control and improved health outcomes for humans and animals alike.

Pathophysiology

The pathophysiology of CDI is similar in horses compared to humans and other animals. CDI refers to the colonization of C. difficile within the host tissue. Disease associated with CDI is toxin‐mediated and exhibits a broad spectrum of signs and symptoms. Mild cases manifest as watery diarrhoea and low‐grade fever. Further infection development may result in a progression to severe CDI, with additional features of haemodynamic instability, pseudomembranous colitis and severe anorexia (Bartlett et al., 1978). In horses, C. difficile is also a known cause of duodenitis‐proximal jejunitis and necrotizing enteritis (Arroyo et al., 2004; Arroyo et al., 2017). Extracolonic manifestations such as bacteremia and organ failure can also develop and extreme cases can result in death (Dallal et al., 2002; Arroyo et al., 2004; Napolitano and Edmiston, 2017).

Transmission of C. difficile occurs through the faecal–oral route. Ingested spores pass to the bowel where bile acids stimulate germination into vegetative cells (Francis et al., 2013). These cells proliferate in the intestinal anaerobic environment, penetrating the mucus layer to attach to the host epithelial cells. Following attachment, toxigenic strains produce toxins that interfere with cell signalling, disrupting the cytoskeleton resulting in cell damage, loss of tight junction integrity and apoptosis (Hecht et al., 1992). This damage induces inflammatory mediator release and fluid secretion which manifests as watery diarrhoea (Pruitt and Lacy, 2012).

Colonization and lesion development associated with host inflammatory response to CDI occurs within the intestinal tract; however, the exact location varies between animal species and stage of life (Keel and Songer, 2006). In neonatal foals (≤1 month old), lesions are predominantly located within the small intestine with extended formation within the large intestine less frequent (Keel and Songer, 2006; Diab et al., 2013b). Conversely, lesion development in older foals and adult horses appears to be restricted to the cecum and ascending colon of the large intestine (Keel and Songer, 2006).

Clinical manifestations may be self‐resolving or chronic. Despite recurrent CDI occurring in 20%–30% of human C. difficile cases, recurrent CDI has not been noted as an ongoing issue in equine populations (Weese et al., 2006; Cornely et al., 2012). A lack of long‐term surveillance of C. difficile and CDI in horses, however, may be impacting this view. Asymptomatic infection can also occur resulting in the shedding of viable spores in the absence of disease, contributing to contamination of the environment (Båverud et al., 2003). This complexity creates difficulty in discerning between states of carriage, colonization and infection. Further investigation into this complexity and the disparity of disease impact within and between species may indeed lead to further understanding of C. difficile pathophysiological nuances (Weese, 2020).

Pathogenicity

The pathogenicity of C. difficile is attributed to the production of potent toxins as well as the ability to form hardy endospores, and these characteristics may appear in human, animal and environmental strains alike. Toxigenicity is influenced by the presence of the Pathogenicity Locus (PaLoc) – a 19.6 kb chromosomal region that encodes toxin A (tcdA) and toxin B (tcdB), as well as positive and negative regulators for toxin expression (tcdR and tcdC respectively) (Braun et al., 1996; Knight et al., 2015b). The presence of an additional binary toxin (C. difficile transferase, CDT), thought to enhance pathogenicity, has also become increasingly significant in the last two decades. CDT appears to be highly prevalent in animal strains (Knight et al., 2013; Gerding et al., 2014; Knight et al., 2015a). The genetic architecture of the C. difficile PaLoc and binary toxin locus (CdtLoc) is shown in Fig. 1. While toxigenic strains of C. difficile are undoubtedly important due to their association with symptomatic disease, it has been demonstrated experimentally that acquisition of the C. difficile PaLoc region by non‐toxigenic strains can occur via horizontal gene transfer (HGT), although the frequency at which this occurs is not known (Brouwer et al., 2013; Elliott et al., 2014; Candel‐Pérez et al., 2019). Recombination and HGT are thought to have played a significant role in the evolution of ‘hyper‐virulent’ C. difficile strains seen today, including PCR ribotype (RT) 027 which caused human CDI epidemics in Canada, the USA and Europe and has been isolated from horses (Songer et al., 2009; He et al., 2013). Genomic studies have determined that approximately 11% of the C. difficile genome is comprised of mobile genetic elements including transposons and plasmids carrying antimicrobial resistance (AMR) genes (Sebaihia et al., 2006).

Fig. 1

Schematic diagram of the C. difficile PaLoc and CdtLoc chromosomal regions which encode for virulence factors, toxin A and toxin B, and binary toxin respectively. Adapted from Francis .

Non‐toxigenic strains of C. difficile are also thought to have a protective function against toxigenic strains (Natarajan et al., 2013). Although this is yet to be investigated in equine populations, protection has been seen experimentally in pigs (Songer et al., 2007; Oliveira Júnior et al., 2019). This attribute could potentially be exploited in the production of preventative medications or vaccines, as seen with the non‐toxigenic C. difficile human strain, NTCD‐M3, which is showing promising results in phase II human trials for the prevention of CDI (Gerding et al., 2015; Zhang et al., 2015; Gerding et al., 2018). For this reason, surveillance of both toxigenic and non‐toxigenic strains in healthy and diseased hosts is critical in understanding the aetiology and epidemiology of CDI, and for the early detection of emerging strains.

Another factor contributing to C. difficile pathogenicity is the ability to form hardy endospores following exposure to stress (Kochan et al., 2018). As an obligate anaerobe, the formation of spores allows survival outside the host and the versatility to persist in diverse environments. Inoculated horse faeces can harbour viable C. difficile for 4 years despite being exposed to the natural environment (Båverud et al., 2003). These C. difficile spores can also withstand extreme temperatures and are impervious to conventional chemicals including alcohol‐based sanitizers commonly used in infection prevention and control (Fawley et al., 2007; Hellickson and Owens, 2008). This highlights the durability of C. difficile in both human and animal settings and is a cause for major concerns for public health, and agricultural and animal husbandry practices. Despite 40 years of investigations, the infectious dose of C. difficile in humans and animals is not known, although murine models suggest that this could be as low as 1 spore cm−2 in healthy mice (Lawley et al., 2010). The process is further complicated by the need for microbiota disruption prior to exposure (Moono et al., 2016). It is therefore important to maximize the detection of even small numbers of C. difficile spores present within samples until further investigation into infectious dose.

Epidemiology

The earliest record of C. difficile in equines was in 1984 (Ehrich et al., 1984); however, the first suggestion of an association with equine enterocolitis was proposed 3 years later following an outbreak in a group of diarrhetic foals (Jones et al., 1987). To date, there have been inconsistencies in the reported prevalence and perceived impact of C. difficile in horses (Diab et al., 2013a). Isolation of C. difficile has long been associated with horses with diarrhoea or acute colitis, with isolation rates ranging from 5% to 90% (Båverud et al., 2003; Frederick et al., 2009; Thean et al., 2011; Morsi et al., 2019).

The proportion of healthy adult horses that carry C. difficile appears to be much lower. Earlier small‐scale investigations of C. difficile in the Northern hemisphere returned relatively low detection rates (0%–4%), (Madewell et al., 1995; Weese et al., 2001; Båverud et al., 2003), while a single preliminary Australian study failed to isolate C. difficile from healthy horses (n = 112) (Thean et al., 2011) This is in contrast to a larger study in Ontario in 2011 which returned an overall faecal prevalence of 7.6% in healthy adult racehorses (n = 540) and, more recently, smaller studies in Minnesota, USA (n = 50) and Italy (n = 24) which recorded a 14% and 25% prevalence of C. difficile respectively (Ossiprandi et al., 2010; Medina‐Torres et al., 2011; Shaughnessy et al., 2018). Table 1 summarizes the prevalence of C. difficile identified in these key studies.

Table 1

Summary of key studies on the prevalence of C. difficile in horses.

Location	Year(s)	Number tested	Health Status	Adult/Foal	No. C. difficile positive (%)	References
Australia	2007–2009	62	Diarrhetic	Unknown	14 (23%)	Thean et al. (2011)
		112	Healthy	Unknown	0 (0%)
Canada (Ontario)	1998–1999	55	Diarrhetic	Adult	7 (12.7%)	Weese et al. (2001)
		255	Healthy	Adult	1 (0.4%)
		31	Diarrhetic	Foal	11 (33.3%)
		47	Healthy	Foal	0
Canada (Ontario)	2006–2008	540	Healthy	Adult	41 (7.59%)	Medina‐Torres et al. (2011)
Italy	2007	24	Healthy	Adult	6 (25%)	Ossiprandi et al. (2010)
		18	Healthy	Foal	8 (44.4%)
Saudi Arabia	2019	30	Diarrhetic	Adult	7 (23.3%)	Morsi et al. (2019)
		286	Healthy	Adult	3 (1.1%)
		49	Diarrhetic	Foal	11 (22.5%)
		42	Healthy	Foal	3 (7.1%)
Sweden	Unknown	227	Enteric disorders/on antimicrobials	Adult	23 (10.1%)	Båverud et al. (2003)
		273	Healthy	Adult	0
		51	Enteric disorders/on antimicrobials	Foal	11 (21.57%)
		226	Healthy	Foal	17 (8%)
USA (California)	1993	10	Diarrhetic	Adult	9 (90%) a	Madewell et al. (1995)
		23	Healthy	Adult	1 (4.3%)
USA (Minnesota)	2011–2013	50	Healthy	Unknown	7 (14%)	Shaughnessy et al. (2018)
USA (Florida)	2003–2008	233	Diarrhetic	Foal	11 (5%)	Frederick et al. (2009)

Possible hospital outbreak.

The prevalence of C. difficile appears higher in foals, with younger animals tending to harbour the bacterium at higher rates, similar to other young animals (Båverud et al., 2003; Morsi et al., 2019). In a 2003 study in Sweden, C. difficile was isolated from 29% of healthy foals under the age of 14 days, and only 0.6% of foals aged greater than 14 days (Båverud et al., 2003). This trend was also identified in a 2019 study in Saudi Arabia where all foals carrying C. difficile (7.1% of healthy and 22.5% of diarrhetic foals) were aged <2 months, with C. difficile not isolated from any foal over this age (Morsi et al., 2019).

It is difficult to draw meaningful conclusions from these variable results given the limited number of investigations, combined with geographical, methodological and temporal differences. Nevertheless, there does appear to be a tendency towards outbreaks and sporadic cases rather than ongoing chronic or recurrent illness (Diab et al., 2013b). Longitudinal studies further revealed the transient nature of horse C. difficile colonization, with an overall prevalence of 5.4% compared to a cumulative prevalence of 40% (Schoster et al., 2012). This was concordant with a recent Swiss study investigating C. difficile in horses with colic, and diarrhetic and healthy horses where the cumulative prevalence (19%) appeared much higher than single‐day testing (10%), questioning the need for multi‐day sampling in at‐risk horses or suspected cases (Schoster et al., 2019).

This ephemeral pattern has been demonstrated in other animals and adds to the complexity of CDI epidemiology and difficulty in comparing studies (Bandelj et al., 2016). Furthermore, the opportunistic nature of C. difficile colonization and the need for both exposure and commensal flora disruption for the establishment of disease creates challenges in determining the significance of an asymptomatic state. Despite these apparent inconsistencies and knowledge gaps, it is evident that C. difficile in horses could potentially act as a reservoir for zoonotic spread and further investigation is needed to clarify this role.

Equine C. difficile strains isolated in studies include both novel strains as well as those identified in other animals, the environment and humans. Notably, in the 2011 Ontario study mentioned 76.5% of C. difficile isolated were strains previously isolated in humans locally, with 57.7% being RTs 001, 027 or 078, which have been implicated internationally in epidemic outbreaks in humans and other animals (Medina‐Torres et al., 2011). Concerningly, equine cases of infection with the highly virulent RT 027 strain have also been identified elsewhere with severe outcomes (Songer et al., 2009).

Predisposing factors for CDI

Risk factors for CDI in horses centre around circumstances that disrupt the host's native intestinal flora, or which create situations of higher exposure. Antimicrobial exposure and hospitalization are the most recognized risks and have long been associated with CDI across human and animal populations alike (Deshpande et al., 2013; Slimings and Riley, 2014).

Antimicrobial use

Antimicrobials contribute to disease by altering the number, diversity and relative composition of the host commensal gut flora, allowing C. difficile to colonize (Robinson and Young, 2010; Reeves et al., 2011). Studies suggest certain antimicrobials may also increase adhesin (Denève et al., 2008) and toxin gene (Drummond et al., 2003) expression in C. difficile, leading to increased pathogenicity.

CDI in horses has been associated with exposure to an array of antimicrobials including β‐lactams (penicillin, ampicillin, cephalosporins), gentamicin, clindamycin, erythromycin, rifampicin and trimethoprim/sulfonamides (Båverud et al., 1997; Båverud et al., 1998; Arroyo et al., 2004; Diab et al., 2013a; Morsi et al., 2019). Of particular note is the association of CDI with the use of ceftiofur, one of the antimicrobials most commonly used in horses (Rodriguez et al., 2014). Ceftiofur is a veterinary third‐generation cephalosporin, the human equivalent of which is also a known risk factor for CDI in humans (Slimings and Riley, 2014). Ceftiofur can significantly disrupt the bacterial flora of the horse hindgut with studies identifying a 75% reduction in lactobacilli and the appearance of C. difficile within 24 h of antimicrobial administration (Harlow et al., 2013). This imbalance can allow opportunists such as C. difficile to colonize. Commensal bacteria in the horse gut are important for understanding C. difficile colonization for several reasons. First, it is believed that these bacteria compete for both nutrients and adhesion sites. Second, studies have also suggested that species such as lactobacilli alter their environment, producing metabolites utilized by certain bacteria and excluding others (Harlow et al., 2013). Commensal bacterial counts remain disrupted for at least 1 week after antimicrobial administration (Harlow et al., 2013). This is an important consideration as longer disruption increases the chance of exposure to pathogens such as C. difficile while in a high‐risk state.

In addition to direct administration, indirect exposure to antimicrobials may also be important. Although not studied in horses, it has been estimated that between 15% and 50% of antimicrobials administered to livestock remain as residue in resulting manure with some thought to persist for over a year (Chee‐Sanford et al., 2009; Kim et al., 2011; Berendsen et al., 2018; Filippitzi et al., 2019). Tetracyclines, macrolides, quinolones and lincosamide appeared to have the longest persistence in manure and the environment (Berendsen et al., 2018). Interestingly, the latter class includes clindamycin which has been linked to a greater risk of community‐associated CDI (CA‐CDI) development in humans (Deshpande et al., 2013). Studies in horses have also shown that mares of macrolide‐treated foals have contracted CDI and hyperacute colitis due to the ingestion of residual antimicrobials, and outbreaks of colitis on horse farms due to feed contamination by tetracyclines have also been documented (Båverud et al., 1998; Keir et al., 1999).

Hospitalization

While hospitalization is generally accepted as a risk factor for C. difficile and CDI in horses, the primary source of exposure remains less clear. Environmental sampling at veterinary hospitals identified rough, hard to clean surfaces (such as concrete and mats), high traffic zones, and areas previously used by individuals with confirmed CDI as high‐risk areas for transmission (Weese et al., 2000). However, in a recent study, nosocomial equine CDI was presented as an increasingly complex and multifaceted issue (Weese et al., 2021). The nature and severity of illness at admission, undefined classification of ‘hospital’ versus ‘community’ acquired cases in an equine setting and extent of contact with treating veterinarians all add to the overall narrative and must be considered in the identification of preventative strategies and infection control protocols.

Other factors

Diet changes, transportation and other causes of stress in animals may act as risk factors for CDI in equine populations (Båverud, 2002). Such influences have been previously identified in cattle and are thought to disrupt the gut flora, providing a window of opportunity for C. difficile to establish; however, the exact mechanisms and the full impact are not known (Bandelj et al., 2016). Despite this extensive list, it should be noted, however, that cases of CDI with no obvious risk factors are common. This is particularly true in foals which may become colonized within days of birth, but also in a proportion of adult equine cases (Båverud et al., 2003). This wide array of potential predisposing factors and uncertainty shows the complexity of CDI and highlight the challenges faced in controlling its impact.

Presentation, detection and diagnosis

Equine CDI can have a rapid onset, with a delay in treatment leading to significant patient deterioration. With reported mortality of up to 83% in confirmed CDI cases (Nomura et al., 2020), a need for timely investigation and diagnosis based on a combination of clinical history, presentation and laboratory testing is apparent.

As with humans, the clinical presentation of CDI in horses can vary in both clinical signs and severity. Horses with CDI may exhibit episodes of watery diarrhoea, abdominal distension, fever, tachypnoea, tachycardia, changes to the mucous membranes and capillary refill times, as well as depression and anorexia (Weese et al., 2006). Intestinal inflammation and lesion development are common in both foals and adult horses, with the region thickened due to oedema, and characterized by haemorrhage, eruption and necrosis of the mucosa, and pseudomembrane formation (Keel and Songer, 2006; Diab et al., 2013b). These clinical signs and symptoms, however, are common to a variety of aetiologies and are insufficient indicators alone for a presumptive diagnosis of CDI. Differential diagnoses cover a diverse selection of infectious agents such as Salmonella species, Equine Coronavirus, Neorickettsia risticii and Clostridium perfringens (Shaw and Stämpfli, 2018). Laboratory identification, therefore, plays an important role in diagnosis, although this is not without its problems.

Standardized testing protocols do not exist across veterinary laboratories (Medina‐Torres et al., 2010). Despite progression in technology and laboratory systems in the last few decades, the most optimal method for detection of C. difficile and subsequent diagnosis of CDI remains a contentious issue across veterinary and human medical fields (Fang et al., 2017). As with all clinical testing, a delicate balance must be struck between sensitivity and specificity as well as efficiency and cost. In the case of C. difficile, however, the complexities of the pathogenesis of CDI, combined with the phenomenon of asymptomatic carriage, create a further obstacle, and the lines between detection and diagnosis begin to blur. There are currently three main laboratory testing methods utilized in the detection of C. difficile and diagnosis of CDI. These include culture, enzyme‐linked immunosorbent assays and PCR. Ongoing research efforts into additional testing options have, however, shown potential.

Culture with cell cytotoxin assay

Techniques involving culturing of C. difficile from faecal samples and testing isolates for toxin production are recognized as the gold standard for laboratory detection (Planche et al., 2013). Due to the supposed difficulty in culturing this bacterium (from which it gained its name), and the low vegetative cell/high spore count in animal faecal samples, various enrichment broths containing antimicrobials are often utilized in addition to direct culture (Knight et al., 2014; Avberšek, 2017). This is followed by sub‐culture onto selective and differential media such as cycloserine‐cefoxitin fructose agar or chromogenic agar (Avberšek, 2017).

Simple culturing C. difficile is insufficient to discriminate between toxin and non‐toxin producing strains. Subsequent tests such as a cell culture cytotoxicity assay are therefore required to determine toxigenicity and, in turn, the capacity to cause disease. The long turnaround time for growth (24–48 h) and toxin assays deem this approach impractical for routine diagnostic use. Furthermore, C. difficile culturing procedures across laboratories are not standardized (Carroll, 2011). Culturing is therefore generally reserved for epidemiological investigations and as a reference method (McDonald et al., 2018).

Enzyme‐linked immunosorbent assay

Enzyme‐linked immunosorbent assays (ELISAs) can detect glutamate dehydrogenase (GHD), a highly conserved enzyme produced by all C. difficile in faecal samples (Carman et al., 2012). This method is quick and inexpensive; however, it lacks specificity to distinguish between toxigenic and non‐toxigenic strains. Several commercial EIA kits have been developed and are utilized in diagnostic laboratories. Assessment of the most commonly used kit for equine C. difficile in North America showed a sensitivity of 86% and specificity of 96%; however, this is likely variable across competing products (Medina‐Torres et al., 2010). ELISA kits aimed at detecting the presence of toxins A and B in faecal samples (with or without GHD) have also been developed. Given the requirement for toxin production for disease development, as well as the ease and availability of ELISA kits, these are now often routinely used as the diagnostic standard despite lower sensitivity and with many lacking formal validation in equine settings (Ramos et al., 2020).

PCR

PCR is being increasingly utilized in commercial laboratories as a quick and very sensitive method for the detection of C. difficile, despite greater expense compared to ELISA (Planche et al., 2013). This method detects the presence of C. difficile genes or its toxin genes within the sample. Caution must be employed for the diagnosis of disease as this method does not identify toxins, just toxin genes, and fails to distinguish between transient carrier and permanent colonization states (Oliver‐Espinosa, 2018). For this reason, CDI overdiagnosis through the reliance on PCR testing alone has become a concern. In human studies, while negative predictive values remain high (96%), CDI positive predictive values can be as low as 46% and are highly dependent on disease prevalence (Lee et al., 2021). This highlights a need to better understand the extent of asymptomatic carriage within a population and the role it plays in CDI development and dissemination in parallel to decisions regarding diagnostic methods.

Future developments in diagnostics

As knowledge of the bacterium and disease progresses, the possibility of additional diagnostic methods increases. For example, a recent study of blood biomarkers in 407 Arabian horses identified increased haptoglobin, serum amyloid A, neopterin and procalcitonin in horses with active C. difficile enterocolitis, as well as evidence of oxidative stress markers (El‐Deeb et al., 2020). Although limited, this investigation into additional C. difficile markers shows the potential for future paths in CDI detection.

To overcome the shortfalls of current testing methodologies, a multistep testing regime may assist to increase the sensitivity and specificity of individual tests. Although specific recommendations in equine testing are yet to be made, the call for a two‐step diagnostic method for animal C. difficile assays is repeatedly echoed throughout the literature. For example, researchers in the evaluation of pig testing recommended the successive use of both real‐time PCR and toxigenic culture to overcome poor performance and inconsistency in EIA kits (Keessen et al., 2011), while Fathy et al. (2021) promote a combination of conventional culture followed by molecular methods to reduce false‐negative results (Fathy et al., 2021).

There is, however, some concern regarding the methodologies currently utilized in equine C. difficile detection and CDI diagnosis. First and foremost is the frequent use of methods developed for human samples, but not yet validated for equine samples (Medina‐Torres et al., 2010). This leaves many questions regarding the appropriateness of use and comparative performance in animal investigations and highlights the need for further analysis and species‐specific testing.

Furthermore, limited understanding of the toxins identified in equine CDI may have an impact on laboratory diagnosis. While it is accepted that C. difficile toxins A and B are associated with cytopathic damage, the implications of the different combinations of toxins are not well known, creating issues in laboratory protocols that may focus only on the detection of a single toxin. In addition, the role of binary toxin in equine disease is not well understood and detection is not usually included in routine testing regimes. On a final note, as with all testing regimes, it is important that decisions on testing and diagnosis do not ignore practical issues such as whether the test outcome will have an impact on clinical decision making and alter the treatment strategy. As suggested by the international Equine Colitis Research Group, to make the most appropriate choice regarding how to test and, indeed, whether or not to test in the first place, the prevalence in healthy populations and the positive predictive value of the test must be known (International Equine Colitis Research Group, 2020). Perhaps more studies in a research setting utilizing the ‘gold standard’ method of toxigenic culture are required.

Treatment and prophylaxis

On initial presentation, equine CDI cases with diarrhoea and endotoxemia associated with the disease can often represent an immediate danger that can lead to dehydration, electrolyte imbalances and haematological abnormalities (Weese et al., 2006; Nomura et al., 2020). Fluid and electrolyte therapy aimed at restoring blood volume and biochemistry is often carried out to stabilize the patient, with nonsteroidal anti‐inflammatory agents administered to minimize deleterious inflammatory responses (Shaw and Stämpfli, 2018). Treatment to avoid complications associated with CDI is also important in the care of equine cases, including hoof cryotherapy to prevent laminitis (Shaw and Stämpfli, 2018).

These initial treatments, however, focus on correcting the effects of the infection rather than controlling the bacteria and toxins, themselves. In equine cases, a combination of antimicrobial and supportive therapies, therefore, remain central in the overall treatment of CDI. Metronidazole is often the first‐line choice in the treatment of CDI in horses, with administration associated with survival (Weese et al., 2006). Concerningly, the existence of metronidazole resistance has been noted in some equine and human studies, highlighting the need for multiple avenues for treatment to be available (Boekhoud et al., 2020). Vancomycin may be utilized in cases where the infecting strain of C. difficile shows resistance to metronidazole; however, this should be avoided where possible due to the heavy reliance on vancomycin in human treatment and the rise of vancomycin resistance (Schoster and Staempfli, 2016). This thinking is being challenged with the increase in C. difficile resistance to metronidazole and a range of other antimicrobials over the last two decades (Peng et al., 2017). The AMR situation has become so dire that in both 2013 and 2019, the United States CDC listed C. difficile in the top five infectious agents posing an urgent threat to the community based on the apparent increase in AMR in circulating strains (Centers for Disease Control and Prevention, 2019).

Adjunctive therapies have also been developed with varying results. Bismuth subsalicylate is thought to prevent attachment of C. difficile to intestinal cells by coating the mucosa as well as providing antimicrobial and anti‐inflammatory activity against C. difficile (Mallicote et al., 2012; Pitz et al., 2015). Despite its common use in diarrhetic horses, its true effectiveness in horse infections of the large intestine has been questioned due to the large volume of contents with little species‐specific testing (McConnico, 2015). Di‐tri‐octahedral smectite also binds and neutralizes C. difficile toxins A and B in vitro, however, while commercial products (such as Bio‐Sponge) are successfully utilized in the general treatment of diarrhetic horses, C. difficile specific in vivo testing is lacking (Weese et al., 2003; Hassel et al., 2009; Oliver‐Espinosa, 2018).

Beyond traditional treatment methods, alternative microbiota restorative therapies are also being developed aimed at re‐establishing commensal microbiota diversity to resemble that of a ‘healthy’ individual. Faecal microbial transplantation (FMT) transfers faecal matter from healthy donors into the gastrointestinal tracts of CDI affected patients (Kelly et al., 2015). This has recently gained popularity in the treatment of recurrent human CDI with cure rates of 87%–90% (Kelly et al., 2015; van Beurden et al., 2017). The concept of FMT is not new in the animal setting. In its most basic form coprophagia, where one individual consumes the faeces of another, is commonplace between foals and their dams as an important process in establishing ‘normal’ gut bacteria during infancy (Quercia et al., 2019). The effects of this on C. difficile and CDI have not been exclusively investigated.

Transfaunation, as FMT is also known in animals, is well established in the treatment of general gastrointestinal ailments in livestock, including horses, although primarily anecdotal data exists for the latter (Feary and Hassel, 2006; Bakken, 2009). In recent years, better studies into the benefits of FMT in horses have emerged with promising results (McKinney et al., 2021). While large‐scale studies in horses and other animals have not been done, isolated cases of treatment in marmosets and dogs have been largely successful (Yamazaki et al., 2017; Sugita et al., 2019). It is clear that much more needs to be done to enable FMT to become a mainstream treatment option in horses. A meeting of the International Equine Colitis Research Group in 2020 cited a lack of robust clinical studies into FMT in horses as a limiting factor in progressing this therapy, advising that many questions remain concerning longevity, screening and best practice protocols in horses (International Equine Colitis Research Group, 2020). The group also highlighted a gap in knowledge regarding the horse microbiome as a whole and a need for targeted investigations into key characteristics of horses affected by infectious agents such as C. difficile.

Preventative therapies may also play an important role in minimizing the effects of CDI in the future; however, to date success has been limited. Probiotics have generated some interest although with varied and inconsistent results (Schoster et al., 2015). Schoster et al. (2014) suggested that this inconsistency may have been a result of strain and dosage selection with some questions surrounding the quality control of commercial products. Despite this, a small number of specific probiotic agents have shown promising results. Lactobacillus reuteri reduces the adhesion of C. difficile to epithelial cells and significantly reduces the number of clostridial cells in the faeces of horses (Dicks et al., 2015). Similarly, Saccharomyces boulardii has also shown potential in the prevention of equine CDI following success in humans (Desrochers et al., 2005; Boyle et al., 2013; Carstensen et al., 2018). This microorganism releases proteases that digest C. difficile toxin A, reduce its ability to bind to host intestinal cells and interfere with host cell signalling to reduce damaging inflammatory responses (Castagliuolo et al., 1996; Chen et al., 2006). Vaccines for animals or humans are yet to be developed although a number have progressed to phases II and III trials (Riley et al., 2019). It is clear that further investigations into prevention and alternative treatments for CDI in horses are required.

A changing landscape and call for a One Health approach

Although traditionally considered a healthcare‐related disease, cases of CA‐CDI are becoming increasingly common, now accounting for up to 50% of all human CDI cases (Ofori et al., 2018). Furthermore, studies have reported that one‐third of patients with CA‐CDI have no apparent exposure to traditional risk factors of hospitalization or antimicrobials (Mooney et al., 2008). The driving factors behind the shift towards CA‐CDI are not well understood, making infection prevention and control, and establishing effective eradication programs challenging.

To date, C. difficile has been detected in a diverse range of sources from compost and lawns to root vegetables and livestock (Moono et al., 2017; Lim et al., 2018a; Lim et al., 2018b). The presence of C. difficile in production animals has been investigated, with evidence of overlap of strains seen in animals and humans (Songer et al., 2009; Medina‐Torres et al., 2011; Knight and Riley, 2013). The emergence of human disease caused by strains previously only seen in animals also adds to mounting evidence that CDI may be zoonotic, highlighting a need for a holistic One Health approach to understand and control this disease (Knight et al., 2015b; Rodriguez et al., 2016).

The ‘One Health’ paradigm recognizes the relationship between human health, animal health and the environment. It highlights the need to review all factors contributing to a health issue in attempts to control, prevent and treat disease. The ubiquitous nature of C. difficile makes a One Health approach vital to public health planning. Interaction between the main constituents influencing C. difficile transmission (seen in Fig. 2) is extensive and complicated. Horses represent an interesting addition in the C. difficile story, with overlapping domains of production animal, companion animal and non‐domesticated populations. The potential for dissemination of C. difficile encompasses transfer through interaction, consumption and indirectly through exposure to horse manure. In countries such as Australia, with an estimated 400 000 feral horses, there is also significant potential for dissemination at the wildlife–livestock–human interface through interaction with other wild and native species, as well as dispersal through shared water sources and possible interaction with free‐range livestock (Csurhes et al., 2016). To date, C. difficile in feral horses and the role they play in dispersal has not been investigated. It was previously noted that commercially available animal manures and compost showed traces of C. difficile. Local horse manure is readily available for use in domestic and market gardens, and on farms is often furrowed back into paddocks allowing the potential for the spread of any existing C. difficile spores into the community. This creates a potential pathway for transfer between horses and humans.

Fig. 2

The proposed interaction of elements in C. difficile dispersion.

Furthermore, the emergence of C. difficile AMR in both human and animal strains and the possibility of bidirectional resistance gene transfer between the two add a further complication to the system (Knetsch et al., 2018). Studies have shown horse derived C. difficile strains show high resistance to commonly used antimicrobials such as ceftiofur and gentamicin, suggesting AMR may be influenced by their use (Rodriguez et al., 2014). Injudicious use of antimicrobials in human healthcare, veterinary practices and farming has come under great scrutiny for its contribution to AMR and the spread of infectious disease. In 2015, the World Health Assembly identified AMR as a critical issue, endorsing the development of a global action plan to tackle this problem. In addition, in 2013 and 2019, the US Centers for Disease Control and Prevention identified C. difficile in the top five microorganisms posing an urgent threat to public health due to its developing AMR (Centers for Disease Control and Prevention, 2019). Given the current global crisis of AMR and the pervasiveness of C. difficile, it is vital that the health and science communities start to look outside their immediate fields for solutions. For this reason, investigations into the aetiology and epidemiology of C. difficile in non‐traditional sources are required.

The way forward

Despite promising developments in the understanding of equine C. difficile, a lack of validation for species‐specific diagnostic testing and treatment, as well as the ongoing threat of AMR, creates challenges in the fight against CDI. The implications of asymptomatic carriage on the dispersal of the bacterium also remains elusive and strain characterization and molecular investigation may prove crucial for a true appreciation of the C. difficile epidemiology to assist in tracing the flow through horse populations and the community as a whole. As the links between C. difficile in animals, humans and the environment become increasingly apparent, a more efficient approach to antimicrobial surveillance, stewardship and AMR investigations is needed for long‐term sustainability. A One Health approach and further appreciation of the possible sources of C. difficile are therefore vital to the development of infection prevention and control strategies, to minimize transmission risk as well as generate protocols for optimal antimicrobial use.

Acknowledgements

This work was supported by a Research and Innovation Strategic Scholarship from Murdoch University awarded to NH‐S and a Fellowship from the National Health and Medical Research Council (APP1138257) awarded to DRK. Open access publishing facilitated by Murdoch University, as part of the Wiley ‐ Murdoch University agreement via the Council of Australian University Librarians. Correction added on 11 April 2022, after first online publication: CAUL funding statement has been added.