Background

Since the discovery of Clostridium difficile-associated disease (CDAD) approximately 30 years ago [1, 2], much progress has been made in our understanding of the pathogenesis and management of this infection. In recent years, however, the epidemiology of C. difficile has changed dramatically. Beginning in 2000, widespread regional outbreaks of C. difficile strains involving more severe and refractory disease have occurred, with greater numbers of complications, colectomies, and deaths than previously described [3–6]. Since most C. difficile acquisitions occur within healthcare settings, emphasis should be placed on implementing evidence-based strategies for infection control and prevention, early detection, and effective treatment for severe and relapsing CDAD.

Pathogenesis and epidemiology

Toxin-producing strains of C. difficile, an anaerobic spore-forming bacillus, cause illnesses ranging from mild diarrhea to fulminant colitis and toxic megacolon leading to sepsis and even death. There are two essential requirements for CDAD to develop: exposure to antimicrobials and new acquisition of C. difficile, although the precise timing and order of these events is not well understood. An important third factor, possibly relating to host susceptibility or virulence factors of the bacterial strain, may then determine whether the clinical outcome will be asymptomatic colonization or CDAD [7].

Acquisition of C. difficile occurs by oral ingestion of spores, which resist the acidity of the stomach and germinate into the vegetative form in the small intestine. Disruption of the commensal flora of the colon, typically through exposure to antimicrobials, allows C. difficile to flourish and produce toxins that lead to colitis. The primary toxins produced are toxins A and B, two large exotoxins that cause inflammation and mucosal damage. Both toxins appear to have cytotoxic effects through disruption of the actin cytoskeleton within cells [8]. Although previous evidence suggested that toxin A is the major enterotoxin, C. difficile strains that produce toxin B but not toxin A have recently been isolated from patients with CDAD [9].

Nearly all antimicrobials have been implicated in the development of CDAD [10]. Certain antimicrobial classes that are broad-spectrum and have a propensity for killing colonic bacteria, especially cephalosporins, clindamycin, and most recently, fluoroquinolones, may pose a greater risk for the development of the disease. The other major risk factors for CDAD are advanced age and hospitalization [10]. In addition, severe underlying disease, immunocompromising conditions, chemotherapeutic drugs, gastrointestinal surgery, nasogastric tubes, and gastric acid suppression are predisposing factors [11, 12]. Persons with normal healthy gastrointestinal flora and the ability to mount a brisk immune response are at lower risk for CDAD. Asymptomatic carriers of C. difficile have been found to have high levels of IgG antibodies to toxin A and have a decreased risk of developing CDAD compared to non-carriers [13, 14]. In addition, a vigorous serum antibody response to toxin A during an initial episode of CDAD is associated with protection against recurrent disease [15].

Although community-acquired C. difficile is increasingly recognized, healthcare setting transmission is the primary mode of C. difficile acquisition. Only about 3% of healthy adults in the community are colonized with C. difficile, compared to 20–40% of hospitalized patients [16, 17]. The risk of C. difficile acquisition is also highly correlated with length of hospital stay [18]. The incubation period of C. difficile following acquisition has not been clearly defined. Although one study suggested an incubation period of less than seven days [19], there may be prolonged intervals between onset of diarrhea and hospital discharge or cessation of antimicrobials [20], emphasizing the need for a high level of suspicion for CDAD in patients presenting with diarrhea in the community following a hospital admission.

Changing epidemiology

In the last seven to eight years, the incidence and severity of C. difficile have increased dramatically. CDAD rates in intensive care units increased significantly from 1987 to 2001 in hospitals with greater than 500 beds [21]. In the United States, the number of hospital discharges where CDAD was listed as any diagnosis doubled between 2000 and 2003, with a disproportionate increase for persons aged > 64 years [22] (Figure 1). By 2003, regional reports of CDAD outbreaks from hospitals throughout the US and in Quebec, Canada emerged, describing severe disease associated with greater numbers of complications, including colectomies, treatment failures, and deaths [3–5]. In 2004, the attributable mortality rate of nosocomial CDAD in Quebec hospitals was 6.9% [3], compared to 1.5% among Canadian hospitals in 1997 [23]. In the US, death certificate data suggest mortality rates due to CDAD increased from 5.7 per million population in 1999 to 23.7 per million in 2004 [24]. By contrast, a study in a US medical center in 1998 found no excess mortality attributable to CDAD after adjustment for confounding factors [25].

Figure 1
figure 1

Rates of discharges from US short-stay hospitals of patients with C. difficile-associated disease listed as any diagnosis by age [22].

Full size image

A hypervirulent epidemic strain of C. difficile was found to be associated with the outbreaks in the US and Quebec and subsequently with outbreaks in the United Kingdom and other parts of Europe [3, 4, 26]. The epidemic strain has been characterized as restriction enzyme analysis type BI, North American Pulsed-Field Type 1 (NAP1), and PCR ribotype 027 [4, 27]. Restriction enzyme analysis of the pathogenicity locus containing the toxin and associated regulatory genes also classifies this strain as toxinotype III, a previously uncommon toxinotype among hospital strains [28]. The BI/NAP1/027 strain has been found to produce 16-fold higher concentrations of toxin A and 23-fold higher concentrations of toxin B in vitro [27], a finding that is most likely related to the presence of an early frameshift mutation identified in tcdC within the pathogenicity locus of this epidemic strain [29], since the product of the unaltered gene normally inhibits toxin production [30]. An 18 base pair deletion is also present within this gene in the epidemic strain but is downstream of the frameshift mutation and does not alter the function of the TcdC protein [30]. Another characteristic of the strain is the production of a toxin called binary toxin, the role of which is not yet defined; however, strains that produce the binary toxin may be associated with more severe diarrhea [31]. The cause of the extreme virulence of the BI/NAP1/027 strain may be a combination of increased toxins A and B, binary toxin, or other features particular to toxinotype III strains.

Another feature related to the proliferation of this strain is its universal, high levels of resistance to fluoroquinolones, including the C-8-methoxyfluoroquinolones, moxifloxacin and gatifloxacin [4]. Although BI/NAP1/027 isolates existed previously, historic strains were less resistant to fluoroquinolones, and they were not associated with outbreaks of disease. The emergence of this strain now is likely related to its selective advantage in the presence of widespread increasing use of fluoroquinolones. A similar phenomenon was observed with the clindamycin-resistant 'J strain', which caused outbreaks in the late 1980s and early 1990s [32].

In addition to strain characteristics, host factors also play a major role in CDAD, and the elderly have been particularly affected. In recent studies, CDAD rates, severity, and mortality have been highest in persons > 65 years of age [3, 6, 22, 24]. C. difficile is a significant cause of outbreaks and non-epidemic diarrhea in nursing homes [33, 34]. The association of CDAD with advanced age may be a result of a weaker immune response to C. difficile or other underlying illnesses. However, C. difficile has also recently emerged in populations previously considered to be at low risk, including healthy outpatients, peripartum women, children, and people with no recent antibiotic exposure [35]. More research is needed to determine if these cases are due to strains carrying similar virulence factors to the BI/NAP1/027 strain.

Diagnosis

Early diagnosis is key to preventing complications from severe CDAD and preventing transmission. Rapid diagnosis depends on maintaining a high degree of clinical suspicion for CDAD in patients with diarrhea and recent antimicrobial exposure and hospitalization. The tissue culture cytotoxic assay has long been considered the gold standard diagnostic test but is limited by slow turnaround time (at least 48 hours), work intensity, and cost [8]. The enzyme immunoassay for detection of toxins A and B is the most commonly used test in clinical laboratories because of its ease of use and rapidity. Although the test is highly specific, it has a lower sensitivity (70–87%) than the cytotoxic assay [36–39]. However, testing two or three stool specimens can increase the yield by 10% or more [39]. Assays that test for both toxins A and B are required to detect toxin A-B+ strains.

The detection of C. difficile by culture is rarely performed for diagnostic purposes because it is labor intensive and has a slow turnaround time. However, if culture is combined with toxin testing of the recovered isolate (so-called 'toxigenic culture'), it may be even more sensitive than the tissue cytotoxin assay [40] and offers the additional advantage of providing isolates for strain typing to better understand transmission dynamics in a hospital. One promising approach is the use of the highly sensitive, albeit less specific, C. difficile antigen assay as a screening test with a rapid turnaround time, followed by confirmatory testing of positives using a cytotoxin assay or toxigenic culture [41].

Treatment strategies

When initiating treatment for CDAD, the first basic principle is to stop the offending antimicrobial(s) if possible. In one study, 41% of patients who remained on antibiotics during treatment of CDAD with metronidazole failed, compared to none of those whose antibiotics were discontinued [42]. Therapy should be administered orally if possible, and continued for at least 10 days. Anti-peristaltic agents, including narcotics, should be avoided. For conditions such as toxic megacolon and ileus, alternative routes, such as administration of vancomycin enterally via a nasogastric tube and/or directly into the colon as an enema, should be used [43, 44]. Finally, early surgical consultation may improve survival in selected patients with fulminant CDAD [45–47].

The two primary agents used to treat CDAD are metronidazole and oral vancomycin. Earlier randomized trials showed equivalent response rates of greater than 90% with either drug [48, 49]. Because metronidazole is considerably less expensive than oral vancomycin and possibly less likely to promote the selection of vancomycin-resistant Enterococcus spp., practice guidelines have recommended metronidazole as first-line treatment for CDAD [50, 51]. However, more recent concerns about treatment failure with metronidazole, particularly in cases of severe disease, have been raised [52, 53]. In one randomized, double-blinded trial of vancomycin versus metronidazole for CDAD, cure rates were equivalent for mild CDAD (98% and 90% for vancomycin and metronidazole, respectively). However, for severe disease, the cure rate was significantly higher for vancomycin (97%) than for metronidazole (76%) [54]. Since the majority of the cases in this study occurred before recognition of the hypervirulent BI/NAP1/027 epidemic strain, it is unknown whether these findings can be generalized to the current epidemic. However, in a recent phase 3 study comparing the efficacy of the toxin binder tolevamer to vancomycin or metronidazole for CDAD, vancomycin was found to be superior to metronidazole for severe disease (defined as ≥10 bowel movements/day, white blood cell count ≥20,001/mm3, or severe abdominal pain due to CDAD) with a clinical success rate of 85% for vancomycin versus 65% for metronidazole (p = 0.04) [55]. Therefore, vancomycin may be preferred as initial treatment for patients with risk factors for a complicated outcome or poor response to metronidazole, such as intensive care unit stay, low albumin level, fever, leukocytosis, profuse diarrhea, and elevated creatinine [6, 54–56]. Recommendations for treatment of CDAD based on disease severity are given in Table 1.

Table 1 Antimicrobial Treatment for C. difficile-associated disease based on disease severity
Full size table

Recurrent C. difficile-associated disease

Between 15% and 35% of patients with a first episode of CDAD relapse within two months [57]. Having one recurrence puts patients at high risk for subsequent recurrences [58]. Other risk factors for recurrence include older age and decreased quality of life scores, and women appear to be more affected than men by recurrent disease [58]. An adequate serum immune response to toxin A during a first episode of C. difficile provides protection from recurrence [15]. Reinfections with different strains of C. difficile have been found in a large proportion of recurrences, suggesting that many can be avoided by ensuring adherence to infection control measures [57].

The recommended treatment for a first recurrence is a second course of the initial therapy with either metronidazole or vancomycin. A commonly used treatment strategy for subsequent recurrences is a prolonged, tapering course of oral vancomycin, which may be followed by pulsed dosing [59]. Other strategies include combination antimicrobial therapy and efforts to restore the normal colonic flora by the use of probiotics or stool transplants. Newer antimicrobial agents and other adjunctive therapies for severe or relapsing disease are discussed below.

New and evolving therapies

Several newer antibiotics have been found to have good activity against C. difficile. Nitazoxanide, which is used for other gastrointestinal infections, was as effective as metronidazole in a randomized, double-blind study of hospitalized patients with CDAD [60]. Rifaximin, also used for other gastrointestinal infections, is a rifamycin-based drug that is not absorbed and achieves high fecal drug levels. It has been used in combination with vancomycin or as follow-up therapy after a course of vancomycin for patients with recurrent CDAD [61]. However, the development of rifaximin resistance is a concern, especially when the drug is used as mono-therapy. In addition, resistance to rifampin in C. difficile, which accurately predicts rifaximin resistance, was found to be common among C. difficile isolates belonging to the epidemic BI/NAP1/027 strain in at least one institution, possibly limiting the utility of rifaximin for treating CDAD caused by this epidemic strain [62, 63].

The investigational drug OPT-80 (difimicin) showed promising results in treating patients with CDAD and low rates of recurrence in a phase 2 trial [64] and is currently in phase 3 trials comparing treatment efficacy with oral vancomycin. Tolevamer, a soluble anionic polymer that binds to toxins A and B of C. difficile, is a non-antibiotic therapy under investigation for treatment of CDAD and was found to be noninferior to vancomycin in treating patients with mild to moderate CDAD in a FDA phase 2 study [65]. However, preliminary results suggest it has failed to meet its noninferiority end point in a recent phase 3 trial [55]. Ramoplanin, an agent previously evaluated for eradication of vancomycin-resistant Enterococcus spp. colonization, was equally as effective as vancomycin in in vitro and hamster models of CDAD [66] and compared favorably to vancomycin for CDAD in a phase 2 clinical trial [67].

Probiotics

Adjunctive therapies for refractory disease include efforts to replenish colonic flora with the use of orally administered probiotics, usually Lactobacillus species or Saccharomyces boulardii. A recent systematic review of randomized controlled trials to evaluate the efficacy of probiotic therapies identified only two treatment studies that showed some benefit of S. boulardii, although the benefit was restricted to subgroups of patients with severe or recurrent CDAD [68]. A more recent randomized, controlled study found some benefit of a yogurt containing Lactobacillus spp. and Streptococcus thermophilus in the prevention of antibiotic-associated diarrhea and CDAD in patients over 50 years of age, although the applicability of the study has been questioned due to highly selective exclusion and inclusion criteria [69]. There is a concern over the safety of probiotics in severely ill or immuno-compromised patients with several reports of S. boulardii fungemia [70] and less frequent reports of sepsis due to Lactobacillus spp. [71]. In general, there is insufficient evidence to support the routine use of probiotics to prevent or treat CDAD. Finally, case reports and case series have shown success with administration of donor stool or 'synthetic stool' (bacterial mixtures), either by nasogastric tube or colonoscopy [72–74].

Immunomodulation

Pooled human immunoglobulin contains antitoxin IgG antibodies capable of neutralizing C. difficile, and case reports have described rapid responses to intravenous immunoglobulin in patients with severe CDAD, although randomized controlled trials are needed [75, 76]. A C. difficile toxoid vaccine has been developed and induces high level responses of serum antitoxin A IgG in healthy volunteers [77]. Further studies are needed to determine whether the vaccine responses confer protective immunity against CDAD and whether adequate immune responses are achieved in the elderly or in patients with recurrent C. difficile. Another immune therapy approach, the use of human antitoxin A and B monoclonal antibodies, reduced mortality in a hamster model of CDAD [78].

A detailed review of the current status of investigational therapies for CDAD was recently published by Miller [79].

Prevention

Transmission of C. difficile within hospitals has been observed through time-space clustering of new cases with identical strains and a greater risk of acquisition of C. difficile from exposure to roommates or other patients in close proximity who have positive cultures [17, 80]. C. difficile spores have been found to contaminate the hands of healthcare workers and the hospital environment frequently [17, 81].

Because alcohol-based hand sanitizers do not inactivate the spores of C. difficile, concern over their role in transmission of C. difficile have been raised. However, hospitals using alcohol-based hand rubs as their primary means of hand hygiene have not seen increases in the incidence of CDAD associated with their introduction [82]. Due to the theoretical advantage of hand washing over alcohol-based hand sanitizers, hand washing with a non-antimicrobial soap or antimicrobial soap and water should be considered after removing gloves in the setting of a CDAD outbreak or if ongoing transmission cannot be controlled by other measures [83].

Patients with CDAD should be placed on contact precautions and housed in single rooms with private bathrooms or, if unavailable, cohorted in rooms with other patients with CDAD [84]. Single-use disposable or patient-dedicated noncritical equipment should be used. Wearing gloves is one measure that has been proven to reduce the spread of C. difficile in hospitals [85]. Gowns and gloves should be donned prior to entering the room of a patient with CDAD and removed followed by hand hygiene before leaving the room.

Although all hospital cleaning agents can inhibit the growth of C. difficile in culture, only chlorine-containing agents inactivate C. difficile spores. In the most definitive study evaluating environmental cleaning, the use of a 1:10 dilution of a 6% hypochlorite solution for daily room cleaning of CDAD patients in a bone marrow transplant unit decreased the CDAD rate significantly but had no effect on units with lower baseline CDAD rates [86]. Therefore, the use of hypochlorite might be most effective in units where CDAD is highly endemic. The drawbacks of hypochlorite solutions are that most of them must be prepared fresh daily and they can be caustic and damaging to hospital equipment.

Antimicrobial use restrictions are another potential mechanism of controlling and preventing C. difficile. As with environmental cleaning, the exact role of antimicrobial restrictions is undefined due to the presence of confounding factors in most studies. However, several studies support the use of formulary restrictions promoting the use of narrow-spectrum antibiotics to reduce the incidence of CDAD [87–89]. Formulary substitutions of 8-methoxyfluoroquinolones for levofloxacin have also been proposed to control CDAD outbreaks caused by the BI/NAP1/027 strain. While this appeared to be effective in one study [90], it was ineffective in another, most likely because the overall use of fluoroquinolones in the hospital was not controlled [91].

Since resistance of the BI/NAP1/027 strain to fluoroquinolones is a class effect resulting in higher minimum inhibitory concentrations (MICs) to all fluoroquinolones [4], the incidence of disease caused by such resistant strains is not likely to be reduced without controlling fluoroquinolone use in general.

Conclusion

The increasing incidence and severity of CDAD in North America and Europe present major challenges for control and management of this disease. Continued gathering of data on the epidemiology of C. difficile through disease surveillance both within and outside of healthcare facilities, and on the efficacy of prevention and treatment strategies is essential to reduce the burden of this disease. Meanwhile, all clinicians and especially critical care physicians should maintain awareness of the changing epidemiology of CDAD and undertake measures to reduce the risk of disease in their patients.

Note

This article is part of a review series on Infection, edited by Steven Opal. Other articles in the series can be found online at http://ccforum.com/articles/theme-series.asp?series=cc_Infection