Prevalence estimates of asymptomatic
C. difficile colonization vary considerably between different patient groups (Table
2). Among healthy adults with no prior risk factors for CDI, asymptomatic
C. difficile colonization prevalence varied between 0 and 15 % [
15,
26‐
33]. The study reporting 15 % was a prospective cohort study carried out on seven groups of healthy individuals representing various occupations in Japan [
32]. The range of asymptomatic
C. difficile colonization prevalence among groups of study subjects was 4 to 15 %; the groups comprised university students, hospital workers, company employees, and defense force personnel. Among healthy newborns and infants, the prevalence of asymptomatic
C. difficile colonization varied between 18 and 90 % [
15,
34].
Table 2
Prevalence of asymptomatic C. difficile colonization in different populations
Healthy neonates and infants | 18–90 % | |
Healthy adults – general population | 0–15 % | |
Elderly in long-term care facilities, chronic care, or nursing homes | 0–51 % | |
Hospital |
Elderly
| 0.6–15 % | |
Inpatients (not specifically elderly)
| 4–29 % | [ 10, 13, 18, 22, 73, 79, 91, 105, 106, 109, 119‐ 121] |
Rehabilitation (spinal)
| 11–50 % | |
HIV
| 4 % | |
Healthcare workers
| 0–13 % | |
Cystic fibrosis
| 18–47 % | |
Hospital surgical patients on antibiotic prophylaxis
| 17 % | |
Intensive care
| 7 % | |
IBD (ulcerative colitis or Crohn’s disease)
| 11 % | |
Hematological malignancies
| 8 % | |
Few studies have examined asymptomatic
C. difficile colonization in acute hospital care settings. In 1982, Gerding and colleagues detected 43/146 (29 %) patients colonized with non-toxigenic
C. difficile strains [
22]. Over the course of 10 years (1982–1991), Belmares and colleagues reported overall colonization with non-toxigenic strains in 10 % of the patients (ranged from 5 % in 1982 to 18 % in 1984) [
35]. Most studies reporting asymptomatic
C. difficile colonization have targeted elderly patients in dedicated long-term care facilities (LTCFs). Prevalence of asymptomatic
C. difficile colonization among elderly residents ranged from 0 to 51 %, possibly because CDI is often endemic in units or institutions with elderly patients [
9,
30,
36,
37].
Among adults, the highest prevalence of asymptomatic
C. difficile colonization has been reported in patients with cystic fibrosis (CF) and in spinal/brain injury rehabilitation. Asymptomatic
C. difficile colonization prevalence ranged from 18 to 47 % in studies among CF patients, substantially higher than other clinical subgroups (e.g. surgical patients) or general hospital inpatients [
38‐
42]. In a case–control study, Bauer and colleagues found 26/55 (47 %) CF patients were asymptomatically colonized [
38]. Yahav and colleagues reported 14 toxin-positive asymptomatic
C. difficile colonized patients without evidence of diarrhea in a study of 30 CF patients compared to no toxin-positive individuals among non-CF patients [
41]. Welkon and colleagues reported asymptomatic
C. difficile colonization in 19/99 CF patients (19 %), with 12 strains being toxigenic [
40]. Another study of CF patients reported asymptomatic
C. difficile colonization in 12/37 (32 %) patients, rising to 43 % if patients were treated with antibiotics [
39]. The heightened vulnerability of CF patients to asymptomatic
C. difficile colonization rather than to disease has been attributed to an electrolyte transport defect in epithelial cells that may offer protection from the effects of clostridial toxins [
41].
Rehabilitation patients also had higher asymptomatic
C. difficile colonization prevalence than other groups. In one study, 11/22 (50 %) spinal cord rehabilitation patients were colonized and remained asymptomatic [
43]. The asymptomatic
C. difficile colonized patients in this study also had a significantly greater length of stay (median 57 days) compared to non-colonized patients (median 6 days). Stevens and colleagues found that for 7-day increments in length of stay, the risk of healthcare-associated CDI increased by 10 % [
44]; this implies that on average, spinal cord rehabilitation asymptomatic
C. difficile colonized patients will be at 52 % increased risk of developing CDI compared to non-colonized
C. difficile patients. Another study of asymptomatic
C. difficile colonization prevalence on admission to a rehabilitation ward reported that 9/54 (17 %) patients without prior colonization became colonized after admission [
45]. Of these nine patients, six showed no symptoms of diarrhea. The increased colonization rate among this group of patients is thought to result from the rehabilitation therapy where group activities and socialization are encouraged, facilitating transmission.
Mechanism of colonization with C. difficile
The first stage in asymptomatic
C. difficile colonization is the ingestion of
C. difficile spores [
46‐
48]. The spores survive the gastric acid and germinate into vegetative cells in the anaerobic environment of the colon.
C. difficile has been isolated from samples of human jejunum, however the primary reservoir is the large intestine [
49]. Vegetative
C. difficile cells penetrate the mucus layer in the large intestine using flagella and enzymatic degradation of the colonic extracellular matrix [
48]. Once the mucosal layer has been breached, in vitro assays have demonstrated that adhesion of
C. difficile cells to intestinal epithelial cells is facilitated by bacterial surface layer proteins [
50].
For colonization with vegetative
C. difficile cells to occur, there must be a disruption of the normal intestinal microbiota which usually provides colonization resistance against
C. difficile [
51,
52]. The inhibitive effect of the natural gut microbiota may occur through competition for space and nutrients or the production of compounds that inhibit
C. difficile proliferation [
53]. The concept of colonization resistance is important to understand the mechanisms that result in the development of disease. Therefore, there is potential to introduce non-pathogenic organisms as probiotic agents or non-toxigenic
C. difficile strains to compete with toxigenic
C. difficile strains as novel prevention and treatment strategies [
54,
55]. However, Brouwer and colleagues have challenged this concept as they found that transconjugation of the pathogenicity locus can occur from toxigenic to non-toxigenic
C. difficile strains [
56].
Toxin production and asymptomatic colonization
Secretion of toxins A and B usually occurs once
C. difficile reaches the stationary phase. The first essential step for these toxins to exert their effects is binding to receptors on gut epithelial cells [
6]. Disease symptoms commence with toxin catalysis in the cytosol. The catalyzed toxin products inactivate guanosine triphosphate binding Rho proteins [
6]. The subsequent depolymerization of the actin cytoskeleton elicits a cellular response that includes neutrophil infiltration, resulting in inflammation, and the subsequent release of cytokines and interferon gamma [
57,
58]. Cell death occurs by apoptosis following disaggregation of the actin cytoskeleton [
59]. Consequently, extensive colonic inflammation and epithelial tissue damage occur, leading to rapid fluid loss into the large intestine, manifesting as acute diarrhea [
6].
The role and importance of toxins A and B in progression to the disease state has been subject to debate. In early studies using hamster models, purified toxin A was shown to elicit symptoms consistent with disease, whereas toxin B would only elicit a response if co-administered with toxin A [
60]. Consequently, it was suggested that toxin B exerted an effect following initial tissue damage by toxin A. The recovery of toxin A-negative, toxin B-positive strains from symptomatic patients has challenged the view that toxin A is the dominant virulence factor in CDI [
61,
62]. Recent work with animal models using antibodies against toxins A and B showed that administration of anti-toxin B antibodies either alone or in combination with anti-toxin A was more effective at preventing the development of gastrointestinal symptoms consistent with CDI [
63]. Lyras and colleagues constructed mutant isogenic strains of
C. difficile capable of producing either toxin A or toxin B. The toxin A producing strains lost their pathogenicity whereas the toxin B producing stains were as pathogenic in animal models as wild type strains [
64]. However, another group using similar gene knockout methods to generate mutant strains produced conflicting findings with a role for both toxins A and B [
65].
Toxigenic strains of
C. difficile are the most prevalent among colonized patients; early studies cultured stool specimens and using enzyme immunoassay (EIA) or cell culture cytotoxicity neutralization assay reported the proportion of toxigenic strains among asymptomatic colonized patients was in excess of 50 % [
31,
39,
40,
66‐
69]. These findings have been corroborated in later studies using real-time polymerase chain reaction (PCR) [
27,
29,
30,
32,
70]. It is important to note that both EIA and PCR methods specifically target toxigenic
C. difficile strains and could therefore bias results reporting a higher prevalence of these strains [
71].
Duration of the colonized state
There is limited information about the duration over which individuals remain asymptomatic after coming in contact with
C. difficile spores or the time taken to revert to a non-colonized state. In a randomized placebo-controlled trial, Johnson and colleagues compared the efficacy of vancomycin and metronidazole for eradication of
C. difficile in asymptomatic colonized patients. Sixty, 80 and 100 % of the patients in the placebo group were negative for
C. difficile after 40, 70 and >90 days follow-up, respectively [
72]. In a prospective study, Samore and colleagues [
73] compared the incidence of colonization in surgical, medical and intensive care wards. Thirty two colonized patients were followed on a weekly basis until they were discharged; 84 % of the colonized patients remained culture positive with median duration of colonization of 8.5 days (range 7–29 days). The study also showed that 3/20 (15 %) of the patients colonized with non-toxigenic strains, none of whom developed diarrhea, were positive for toxigenic strains at follow-up. Longer-term colonization and transmission was investigated among 1234 healthy Japanese volunteers, who included university students, hospital staff, and company employees [
32]. Follow-up was performed on 38 asymptomatic patients between 5 and 7 months later. Of these 38 cases,
C. difficile was re-isolated from 12 (32 %) individuals, half of whom yielded the same PCR ribotypes and pulsed-field gel electrophoresis types as previously. In a subsequent study by the same authors, a 6-month follow-up of 18 colonized healthy students found 10 (56 %) were no longer colonized and 8 (44 %) were colonized more than once, of whom 3 (38 %) harbored the same strain [
27].
These findings suggest that there is marked variation in duration of the colonized state, however the role of repeated exposure from the environment or other colonized individuals was not investigated. Limited longitudinal data available about asymptomatic C. difficile colonization warrants further epidemiological studies to investigate the persistence of colonization and to understand the role of re-exposure to the organism over time.
Transmission from colonized patients
Person-to-person transmission in hospital wards, environmental contamination, and carriage of
C. difficile on the hands of healthcare workers have been described extensively [
74‐
77]. The main modes of transmission are by the fecal-oral route and direct contact with contaminated surfaces and fomites [
78], although transmission between healthy individuals who are asymptomatically colonized has also been reported [
32].
Spores from asymptomatically colonized patients are a potential source of CDI and may contribute to the transmission reservoir [
9] and studies have clearly demonstrated that transmission from asymptomatically colonized patients can occur [
75,
79]. Curry and colleagues investigated transmission potential of asymptomatic
C. difficile colonized patients using multiple-locus variable number tandem repeat analysis. They found that 29 % of isolates from hospital-associated CDI cases were highly related to isolates from asymptomatic
C. difficile colonized patients [
75]. Clabots and colleagues reported that patients admitted from home without prior hospitalization in the previous month had the lowest prevalence of asymptomatic
C. difficile colonization (6 %) but, because they represent the majority of admissions, they contributed the second-highest total number of
C. difficile introductions to hospital, after patients readmitted to hospital within 30 days [
79]. Similarly, the length of stay in hospital can also influence transmission. Fecal excretion of
C. difficile spores occurs for up to 6 weeks following resolution of CDI symptoms [
80,
81]. Furthermore, Riggs and colleagues demonstrated that even colonized patients who did not develop disease during a 6 months follow-up period were shedding spores into the environment [
9]. The current CDI clinical practice guidelines from the Society of Healthcare Epidemiologists of America (SHEA) recommend maintaining contact precautions only until resolution of diarrhea. It has been suggested that contact precautions should be extended until time of discharge for patients recovering from CDI. However, there is no conclusive evidence to support extending contact precautions following CDI while patients remain asymptomatic during their hospital stay [
81].
Asymptomatic
C. difficile colonized patients in hospital have the potential to contaminate the environment and subsequently infect others [
75]; however the transmission potential is lower in asymptomatic
C. difficile colonized patients than in those patients with active disease [
10]. In one prospective study of acquisition rates in an endemic CDI setting, 38/128 (29 %) environmental samples from hospital rooms occupied by asymptomatic
C. difficile colonized patients were contaminated compared to 90/128 (49 %) samples from rooms occupied by patients with disease. This corresponds with findings from another study of LTCF residents in which proportions of positive cultures from skin sites and environmental samples were highest among residents with disease, second highest among asymptomatic
C. difficile colonized patients and lowest among non-colonized residents [
9]. Moreover, Sethi and colleagues found that even 4 weeks after receiving therapy for CDI, the frequency of skin contamination (30/52; 58 %) and environmental shedding (26/52; 50 %) remained high in asymptomatic
C. difficile colonized patients [
81]. Samore and colleagues demonstrated that in an endemic situation carriage of
C. difficile on the hands of healthcare workers was positively correlated with the extent of environmental contamination with
C. difficile [
82].
The spore-forming ability of
C. difficile makes it distinct from other infectious organisms common to healthcare settings and introduces further challenges to reduce transmission. Spores can persist in the environment for long periods and require chlorine- [
83] or peroxide-based [
84] sporicidal agents or ultraviolet radiation devices [
85] for environmental decontamination. Typically, hospital patients colonized with other multidrug-resistant organisms are isolated to prevent transmission, but this appears to be of limited value for asymptomatic
C. difficile colonization. In an epidemiological model, Lanzas and colleagues demonstrated that transmission of
C. difficile within a ward cannot be sustained unless new
C. difficile colonized patients are introduced [
86]. Therefore, the admission of asymptomatic
C. difficile colonized patients plays an important role in sustaining
C. difficile transmission within a ward [
87]. A recent study, has demonstrated that nearly half of the
C. difficile cases were genetically distinct from all previous cases, which suggests genetically diverse sources of infection [
88]. Furthermore, Yakob and colleagues demonstrated, using a stochastic mathematical model, that screening for asymptomatic
C. difficile colonization to segregate colonized patients from non-colonized patients had little impact on infection transmission because patients still in a latent period (exposed but not yet colonized) would not be detected [
89].