The Environment, Farm Animals and Foods as Sources of Clostridioides difficile Infection in Humans

The recent discovery of the same Clostridioides difficile ribotypes associated with human infection in a broad range of environments, animals and foods, coupled with an ever-increasing rate of community-acquired infections, suggests this pathogen may be foodborne. The objective of this review was to examine the evidence supporting this hypothesis. A review of the literature found that forty-three different ribotypes, including six hypervirulent strains, have been detected in meat and vegetable food products, all of which carry the genes encoding pathogenesis. Of these, nine ribotypes (002, 003, 012, 014, 027, 029, 070, 078 and 126) have been isolated from patients with confirmed community-associated C. difficile infection (CDI). A meta-analysis of this data suggested there is a higher risk of exposure to all ribotypes when consuming shellfish or pork, with the latter being the main foodborne route for ribotypes 027 and 078, the hypervirulent strains that cause most human illnesses. Managing the risk of foodborne CDI is difficult as there are multiple routes of transmission from the farming and processing environment to humans. Moreover, the endospores are resistant to most physical and chemical treatments. The most effective current strategy is, therefore, to limit the use of broad-spectrum antibiotics while advising potentially vulnerable patients to avoid high-risk foods such as shellfish and pork.

The most common risk factor associated with CDI is the use/misuse of broad-spectrum antibiotics. C. difficile is often resistant to a wide range of antibiotics [10], and the administration of antibiotics like clindamycin, cephalosporins, penicillins and fluoroquinolones eliminate competitive bacteria in the colon and promote C. difficile outgrowth [11]. The elderly, infants, other immune compromised, and patients on antibiotic therapies are therefore most at risk [1,2,4], although the incidence of CDI in pregnant women, children and patients with inflammatory bowel disease (IBD) has also increased [12].
The generally accepted route for human CDI is transmission from the healthcare environment [13]. However, in recent years the proportion of community-acquired CDI, where the patient has no association with a healthcare facility, has increased [14]. At the same time, non-human reservoirs, including the natural environment (soil, rivers and lakes) [15] and animals, including domestic pets [16,17], food animals [18][19][20] and wild fauna [21] have been reported. Moreover, food may be contaminated [22,23].
The link between C. difficile and animals has been known for at least 60 years. In 1960, McBee [24] isolated this bacterium from the large intestine of a seal in Antarctica. By 1974 C. difficile had also been detected in animal faeces (donkeys, horses, cows and camels) and in TcdA and TcdB possess the same biological activities, among which is the disruption of the cytoskeleton that leads to cytopathic effects in cultured cells. They also possess proinflammatory activity and can stimulate intestinal epithelial cells and immune cells to produce cytokines and chemokines [66,67]. Even low doses of toxins A and B damage the tight junctions of the gut epithelial barrier, facilitating the translocation of commensal bac- Figure 1. Illustration of the C. difficile Pathogenicity locus (PaLoc). Adapted from [65]. TcdA and TcdB possess the same biological activities, among which is the disruption of the cytoskeleton that leads to cytopathic effects in cultured cells. They also possess proinflammatory activity and can stimulate intestinal epithelial cells and immune cells to produce cytokines and chemokines [66,67]. Even low doses of toxins A and B damage the tight junctions of the gut epithelial barrier, facilitating the translocation of commensal bacteria, inflammation and cell apotheosis [66][67][68]. Sequence variations, deletions, and duplications within the pathogenicity locus account for different toxinotypes of C. difficile, with 27 currently identified. Certain strains can present only one of the toxins genes (A − B + or A + B − ), however, they reportedly still cause severe disease in humans [62]. In addition, the cytotoxicitybetween toxins that belong to different toxinotypes may vary, making the relation between strain type and CDI severity even more complex [59]. Strains lacking toxin A are more frequently reported due to deletions in the receptor-binding repetitive regions of TcdA caused by the recombination between short repetitive sequences highly conserved in this toxin gene [63]. Donta et al. [66] reported TcdB to be 4 to 200-fold more cytotoxic than TcdA in a mouse model. Therefore, strains producing toxin B have a higher severity in humans.
Up to a third of C. difficile isolates also produce the transferase C. difficile binary toxin (CDT) [69,70]. CDT, composed of CDTa (biological activity) and CDTb (binding), inhibits the protein actin, damaging the cytoskeleton of the gastrointestinal tract (GIT) cells [71]. The presence of the full-length CDT locus implies the potential expression of the binary toxin, and although some strains contain portions of the CDT locus, these are predicted as non-binary toxin-producing strains [68,70]. CDT-producing strains have been previously associated with a higher production of toxins A and B, leading to an increased disease severity [71,72]. However, CDT is not always present in severe cases [73,74]. In addition, CDT can also be produced by only B + and non-toxigenic strains (A − B − ) [72]. Although CDT production is commonly associated with higher severity of C. difficile infection, the role of this toxin during infection and its mechanism of secretion is still not well understood.

Ribotypes
There are in excess of 800 C. difficile ribotypes (RT), some of which are associated with increased virulence [6,80,81], including RT027 and RT078 [82,83]. These ribotypes are also more prevalent in human cases. RT027 (toxinotype III) has a mutation in tcdC, resulting in significantly increased production of toxins A and B while also carrying the genes encoding CDT production and fluoroquinolone resistance [84,85]. Although prevalence has decreased in Europe in recent years, RT027 is associated with a higher mortality and morbidity rate than other ribotypes [86]. The fluoroquinolone resistance, which emerged in two genetically distinct epidemiological lineages (FQR1 and FQR2), was a key driver in the rapid emergence of RT027 [57]. Moreover, this is essential to the increased severity of this ribotype, as this strain typically infects elderly hospital patients on fluoroquinolone treatment [5].
RT078 carries a 39 bp deletion in the tcdC gene and therefore overproduces toxins A and B in addition to the binary toxin CDT. In contrast to RT027, which is mostly hospitalacquired, RT078 is more prevalent in younger people and is generally associated with the community [87]. RT078 strains are resistant to fluoroquinolones and erythromycin, which has contributed to their higher prevalence in CDI [88]. Ribotype 126 has the same mutation in its tcdC gene found in RT078, is resistant to moxifloxacin and tetracycline and is also considered hypervirulent [89][90][91]. Other significant ribotypes from a public health perspective include RT017 and RT018. Although the former only produces toxin B, it is resistant to fluoroquinolones and rifampicin and has been associated with numerous outbreaks [92][93][94]. RT018 has high toxin production capacity, increased cell adhesion, is multidrug-resistant (erythromycin, clindamycin and moxifloxacin) and has become endemic in several countries, including Italy, Spain, Austria and Slovenia [95][96][97].

Water
Toxigenic C. difficile has been isolated from a variety of aquatic environments, including drinking water, rivers, sewage effluent and swimming pools [98,99]. Coastal beaches and river sediments are also contaminated [98,99], in some cases by runoff from fields or effluents from wastewater treatment plants [100]. Indeed, C. difficile is often detected in water from treatment plants [101], and contamination of drinking water was the source of at least one C. difficile outbreak in Finland [102]. Thus, C. difficile survives in water and through the effluent treatment process [100].

Farm Environment and Animals
Marcos et al. [110] reported that C. difficile were widespread in soil, water and faeces on beef, sheep and broiler farms, with the prevalence ranging from 7% to 83% and counts from 2.9 to 8.4 log 10 cfu/g or /mL, depending on the animal species and sample type being tested. Other studies also found C. difficile in the faeces of a range of farm animals, including cattle, sheep, poultry and pigs [111][112][113][114][115][116]. Of these, pigs are the most important source of C. difficile [113,116], with the relative prevalence by age being 45%, 3% and 1% in suckling piglets, post-weaning piglets and finishing pigs, respectively [114]. Although these animals may show symptoms (diarrhoea), most are asymptomatic [114]. Other similar studies have reported a prevalence of 37% [115] and 78% [111] in piglets and 4% [115], 62% [117] and 9% [16] in mature pigs. C. difficile are also found in cattle, especially younger animals. Rodriguez et al. [113] reported a prevalence of 11% in calves and 6% in adult cattle. Other studies have found these bacteria in 11% [118], 14% [117] and 22% of calves [111] and 7% of mature animals [16]. Sheep, including lambs, are also potential carriers, with 0.6 to 2% in the former and 7% reported in the latter [16,119].

C. difficile at the Animal Slaughter Stage
Pathogenic bacteria in faeces on the hide/fleece or in the gastrointestinal tract are readily transferred to the carcass during slaughter and dressing [125]. C. difficile was found in 1%, 3% and 28% of porcine gut contents at slaughter in Belgium [126], Austria [18] and the Netherlands [127], respectively. Reported carcass contamination rates include 7% in Belgium [126], 15% in Canada [128] and 23% in Taiwan [129]. The prevalence of bovine carcass contamination ranges from 7-8% [111,126] but may be as high as 34% [130]. Ovine carcass contamination rates of 15% and 25% have been reported in Iran and Turkey, respectively [130,131]. While poultry carcass data is lacking, Candel-Pérez et al. [132] found C. difficile in 28% of gizzard and 6% of liver samples collected in a poultry processing plant in Spain. In Ireland, beef, sheep and broiler carcass contamination rates ranged from 40% to 100%, 40% to 60% and 10% to 40%, respectively, depending on the sampling stage during carcass processing [16].

C. difficile in Retail Foods
C. difficile has been reported in a range of foods at the retail stage. Thus, the consumption of contaminated retail foods, especially ready-to-eat (RTE) foods, is a risk factor for human infection [134]. Marcos et al. tested meat, dairy and vegetable retail foods and detected C. difficile in 9 out of the 240 samples tested [110]. These include corned beef (1), spinach leaves (2), iceberg and little gem lettuce (1 sample each), wild rocket, coleslaw, whole milk yoghurt and cottage cheese (also 1 sample each). Of these samples, direct counts were obtained for the spinach leaves (5.8 log 10 cfu/g), coleslaw (4.3 log 10 cfu/g) and cottage cheese (6.8 log 10 cfu/g).

Meta-Analysis
The data presented in Tables 2 and 3 were analysed using Graphpad Prism version 9.3.1. The odds ratios (OR) (the odds of consuming a contaminated product) were calculated for each food type. Briefly, the OR was calculated as the number of positive over negative samples reported for each study. Turkey (with only two studies) was combined with the chicken data (poultry category), while the single lamb study was omitted. The medians and 95% confidence intervals were obtained and were then used to prepare the forest plots. In these Figures, the vertical line is set at an OR = 1 (50:50 chance of the food being contaminated). When all ribotypes are considered, shellfish and pork present a higher risk to the consumer ( Figure 2). However, when the analysis is repeated, focusing exclusively on ribotypes 027 and 078 (the 2 hypervirulent strains most commonly associated with human infection), the increased risk is only associated with the consumption of pork (Figure 3).

Meta-Analysis
The data presented in Tables 2 and 3 were analysed using Graphpad Prism version 9.3.1. The odds ratios (OR) (the odds of consuming a contaminated product) were calculated for each food type. Briefly, the OR was calculated as the number of positive over negative samples reported for each study. Turkey (with only two studies) was combined with the chicken data (poultry category), while the single lamb study was omitted. The medians and 95% confidence intervals were obtained and were then used to prepare the forest plots. In these Figures, the vertical line is set at an OR = 1 (50:50 chance of the food being contaminated). When all ribotypes are considered, shellfish and pork present a higher risk to the consumer ( Figure 2). However, when the analysis is repeated, focusing exclusively on ribotypes 027 and 078 (the 2 hypervirulent strains most commonly associated with human infection), the increased risk is only associated with the consumption of pork ( Figure 3).

The Epidemiology of Foodborne Infection
In 1978, C. difficile was recognised as the causative agent of pseudomembranous colitis and diarrhoea in patients on antimicrobial therapy and it was a hospital-associated disease [157]. In the 1980s and 1990s, the incidence of CDI increased significantly, driven by the use of broad-spectrum third-generation cephalosporins (to which C. difficile is intrinsically resistant), but the disease was rarely fatal [158,159]. There was a further increase in CDI in the first 10 years of this century driven by the emergence and epidemic spread of the hypervirulent strain, ribotype 027 [160]. The epidemiology of CDI also changed in terms of clinical presentation, response to treatment, and disease outcome. Communityacquired CDI, defined as cases with symptom onset in the community with no history of hospitalisation in the previous 12 weeks or symptom onset within 48 h of hospital admission [161], also emerged. Since then, the incidence of CDI has remained high in developed countries [159,162], and rates of community acquired CDI have increased, accounting for 41%, 30% and 14% of total CDI in the USA, Australia and the EU, respectively [96,159,163]. Furthermore, community acquired CDI patients are generally younger, healthy, often female and lack the traditional risk factors of CDI, including a history of antimicrobial usage [164].
The natural habitat of C. difficile is the mammalian gastrointestinal tract (GIT). These bacteria colonise the neonatal GIT, proliferate and are excreted in the faeces to which other newborn animals are exposed, and the cycle recommences. As mammals develop, other bacterial species colonise the GIT, and the prevalence of C. difficile decreases [165]. The GIT microbiota inhibit germination, vegetative growth and toxin production, thus protecting against C. difficile [48]. However, in the 1990s, this protection was removed when cephalosporins were used in animal husbandry, and food animals became a major reservoir and amplification host for C. difficile [119,166], resulting in the contamination of the environment and a range of foods [100,119,166].
Once the environment is contaminated, there are multiple direct and indirect routes to humans, including via food (as illustrated in Figure 4). It is all but impossible to provide incontrovertible proof of foodborne transmission because of the ubiquitous nature of C. difficile, delayed onset of symptoms, ability to persist for extended periods as an endospore, etc. However, it has been shown that C. difficile endospores in animal waste, wastewater treatment sludge, soil, manure and compost may survive for extended periods of time, facilitating direct contamination of vegetables and fruit or meat via crosscontamination of carcasses during slaughter and processing [108,147]. Water also fre-

The Epidemiology of Foodborne Infection
In 1978, C. difficile was recognised as the causative agent of pseudomembranous colitis and diarrhoea in patients on antimicrobial therapy and it was a hospital-associated disease [157]. In the 1980s and 1990s, the incidence of CDI increased significantly, driven by the use of broad-spectrum third-generation cephalosporins (to which C. difficile is intrinsically resistant), but the disease was rarely fatal [158,159]. There was a further increase in CDI in the first 10 years of this century driven by the emergence and epidemic spread of the hypervirulent strain, ribotype 027 [160]. The epidemiology of CDI also changed in terms of clinical presentation, response to treatment, and disease outcome. Community-acquired CDI, defined as cases with symptom onset in the community with no history of hospitalisation in the previous 12 weeks or symptom onset within 48 h of hospital admission [161], also emerged. Since then, the incidence of CDI has remained high in developed countries [159,162], and rates of community acquired CDI have increased, accounting for 41%, 30% and 14% of total CDI in the USA, Australia and the EU, respectively [96,159,163]. Furthermore, community acquired CDI patients are generally younger, healthy, often female and lack the traditional risk factors of CDI, including a history of antimicrobial usage [164].
The natural habitat of C. difficile is the mammalian gastrointestinal tract (GIT). These bacteria colonise the neonatal GIT, proliferate and are excreted in the faeces to which other newborn animals are exposed, and the cycle recommences. As mammals develop, other bacterial species colonise the GIT, and the prevalence of C. difficile decreases [165]. The GIT microbiota inhibit germination, vegetative growth and toxin production, thus protecting against C. difficile [48]. However, in the 1990s, this protection was removed when cephalosporins were used in animal husbandry, and food animals became a major reservoir and amplification host for C. difficile [119,166], resulting in the contamination of the environment and a range of foods [100,119,166].
Once the environment is contaminated, there are multiple direct and indirect routes to humans, including via food (as illustrated in Figure 4). It is all but impossible to provide incontrovertible proof of foodborne transmission because of the ubiquitous nature of C. difficile, delayed onset of symptoms, ability to persist for extended periods as an endospore, etc. However, it has been shown that C. difficile endospores in animal waste, wastewater treatment sludge, soil, manure and compost may survive for extended periods of time, facilitating direct contamination of vegetables and fruit or meat via cross-contamination of carcasses during slaughter and processing [108,147]. Water also frequently contains C. difficile endospores [99,100,148], and food production may also be contaminated via water used for irrigation or food processing [100,144]. Moreover, the presence of endospores in rivers may contaminate fish and seafood [100,138,139,141]. Transfer from food and wild animals and from domestic pets has also been described [116,149]. quently contains C. difficile endospores [99,100,148], and food production may also taminated via water used for irrigation or food processing [100,144]. Moreover, th ence of endospores in rivers may contaminate fish and seafood [100,138,139,141]. T from food and wild animals and from domestic pets has also been described [116,Figure 4. The cycle of community-associated CDI infections from zoonotic, environmental borne sources. Adapted from [100] using ARASAAC pictograms.
Of particular interest, from the public health perspective, is the detection of C. difficile isolates in farm animals and in humans suffering from CDI, suggesti pathogen may be zoonotic [150]. Whole genome sequencing (WGS) analysis has that ribotypes 078, 126 and 066, commonly found in pigs and/or cattle, are gen identical to those in humans [151][152][153][154][155]. Although ribotype data for sheep is limite types 014, 010 and 045 are common to both humans and ovine sources [119,156,167 human-related ribotypes 001, 014 and 039 are also found in broilers [120][121][122]168].

Control Strategies
CDI can be controlled in hospitals using deep environmental cleaning, appr hand hygiene, stringent infection control and antimicrobial stewardship [169]. Ho the same strategies cannot be used in agriculture and food processing [15]. Reduced of antibiotics in food animal production would reduce C. difficile amplification bu likely as increasing global food demand is driving increased antimicrobial usage in husbandry, which is projected to rise by 67% by 2030 [15,140]. In 2006 the EU ban use of antibiotics as growth promoters, followed by the USA in 2017, but other majo producing countries still allow this practice [170].
Preventing the recycling and dissemination of C. difficile endospores in anim ries applied to land as organic fertilisers would also facilitate reduced environmen tamination and animal carriage. However, research is required to develop effectiv ments [171]. Vaccination of food-producing animals is another possible control st but an effective vaccine has not been developed yet [15]. Controlling C. difficile in dependent on reducing or eliminating the endospores, which are resistant to chille and freezing (−18 °C and −80 °C) temperatures [172,173]. Although the endospo resistant at 80 °C [172][173][174] and will survive the recommended cooking time temp Of particular interest, from the public health perspective, is the detection of similar C. difficile isolates in farm animals and in humans suffering from CDI, suggesting this pathogen may be zoonotic [150]. Whole genome sequencing (WGS) analysis has shown that ribotypes 078, 126 and 066, commonly found in pigs and/or cattle, are genetically identical to those in humans [151][152][153][154][155]. Although ribotype data for sheep is limited, ribotypes 014, 010 and 045 are common to both humans and ovine sources [119,156,167], while human-related ribotypes 001, 014 and 039 are also found in broilers [120][121][122]168].

Control Strategies
CDI can be controlled in hospitals using deep environmental cleaning, appropriate hand hygiene, stringent infection control and antimicrobial stewardship [169]. However, the same strategies cannot be used in agriculture and food processing [15]. Reduced usage of antibiotics in food animal production would reduce C. difficile amplification but is unlikely as increasing global food demand is driving increased antimicrobial usage in animal husbandry, which is projected to rise by 67% by 2030 [15,140]. In 2006 the EU banned the use of antibiotics as growth promoters, followed by the USA in 2017, but other major food-producing countries still allow this practice [170].
Preventing the recycling and dissemination of C. difficile endospores in animal slurries applied to land as organic fertilisers would also facilitate reduced environmental contamination and animal carriage. However, research is required to develop effective treatments [171]. Vaccination of food-producing animals is another possible control strategy, but an effective vaccine has not been developed yet [15]. Controlling C. difficile in food is dependent on reducing or eliminating the endospores, which are resistant to chilled (4 • C) and freezing (−18 • C and −80 • C) temperatures [172,173]. Although the endospores are resistant at 80 • C [172][173][174] and will survive the recommended cooking time temperature combinations recommended for meat [174], they are eliminated at 98 • C for 2 min [175]. The same authors suggested microwave irradiation (800 W/60 s) also achieved complete inactivation by denaturing the outer coat.

Conclusions
Based on the information provided, it was concluded that C. difficile is widespread in the environment and along the food chain. Many food isolates carry the virulence factors required for human infection, and there is no conceivable reason why food is not a source of these pathogens. This conclusion is further supported by the presence of the same ribotypes in food and humans suffering from community-acquired CDI. Based on our analysis, potentially vulnerable consumers should be advised not to handle or consume shellfish or pork.