Clostridium Difficile
Stuart Johnson
Dale N. Gerding
Clostridium difficile was identified as the etiologic agent of antibiotic-associated pseudomembranous colitis (PMC) in 1978 (1,2) and is now recognized as the most important identifiable cause of healthcare-associated infectious diarrhea. C. difficile may be the only pathogen sufficiently prevalent to warrant testing on a routine basis during the evaluation of healthcare-associated diarrhea (3,4). It is estimated that each case of healthcare-associated C. difficile infection (CDI) costs between $10,212 and $13,675 and results in 3.0 to 6.4 excess hospital days (5). Readmission for CDI costs $128,200 per hospital per year (6). A conservative annual cost estimate for CDI in the United States is $3.2 billion (5). Our present understanding of the pathogenesis of C. difficile disease and rationale for preventive and interventive measures is supported by (a) observations on antimicrobial use in patients who acquire this pathogen, (b) potential infectious reservoirs, (c) modes of C. difficile transmission, and (d) host risk factors.
PATHOGENESIS
The manifestation of enteric disease due to C. difficile depends on at least three critical events: disruption of the normal colonic microflora, exposure to a toxigenic C. difficile strain, and the presence of one or more host factors. Epidemiologic evidence also supports the order of these events in that, with most cases, exposure to antimicrobials with subsequent compromise of host colonization resistance is followed by C. difficile acquisition from exogenous sources (rather than reactivation from an endogenous source).
The normal colonic flora provides a profound resistance to infection with C. difficile. It appears that the host is susceptible to infection with this pathogen only after disruption of the colonic flora by antimicrobial therapy or by substances that act as antimicrobial surrogates such as antineoplastic agents (7). The next critical event, exposure to toxigenic C. difficile, occurs most often in hospitals and chronic care facilities, which serve as the main reservoirs for this infection. With the recognition that asymptomatic carriage of C. difficile is very common among hospitalized patients, it might seem intuitive that these carriers are at high risk of subsequent CDI. However, a meta-analysis of four prospective studies that included 810 patients followed for 1,348 weeks with weekly surveillance cultures indicated that asymptomatic carriers were, conversely, at decreased risk of subsequent CDI (pooled risk difference -2.3% [95% confidence interval 0.3-4.3], p .021) (8,9).
In our current hypothesis (Fig. 37-1), a patient is admitted to the hospital and, although exposed intermittently to C. difficile, is susceptible to colonization or disease only following antimicrobial therapy. The subsequent clinical outcome is determined shortly after acquisition (within a few days) and, like other enteric and infectious diseases, the majority of patients remain asymptomatic. These asymptomatic carriers are then at decreased risk for CDI when compared with noncolonized patients. Host risk factors comprise the third critical component to the pathogenesis of CDI and are discussed later in this chapter.
C. difficile elaborates two major toxins: toxin B, a potent cytotoxin; and toxin A, a potent enterotoxin that is also cytotoxic (10). Although measurement of cytotoxicity in stool specimens is used to diagnose CDI, the enterotoxic effect of toxin A has been presumed to be critical in the pathogenesis of the disease. The early evidence implicating toxin A includes the observation that disease severity correlates more closely with toxin A production in vivo (11) and that toxin A alone, but not toxin B, given intragastrically reproduces the pathology of C. difficile cecitis in hamsters (12). However, toxin B acts synergistically with toxin A in this model, and clinical data support virulence for variant strains that produce only toxin B, but not toxin A. Genetic knockout experiments indicate that toxin B, not toxin A, is the toxin essential for causing disease in the hamster model (13). Furthermore, a monoclonal antibody directed at toxin A was ineffective at reducing CDI recurrence, whereas two monoclonal antibodies directed at toxins A and B were effective in reducing CDI recurrence in patients (14,15).
Toxin A is a unique enterotoxin unrelated to cholera toxin or the Escherichia coli heat-labile toxin and causes extensive mucosal damage with hemorrhagic fluid response (16). The receptor for toxin A involves a trisaccharide moiety, Gala1-3Galb1-4GlcNAc (17), which is present on antigens within the brush border of human and hamster intestinal epithelium (18). Subsequent toxic cellular events involve internalization of toxin A by receptor-mediated endocytosis and disruption of the cellular cytoskeleton. Toxin A (and B) induces glycosylation of small guanosine triphosphate-binding proteins that are important regulators of actin polymerization (19).
 FIGURE 37-1 Hypothesis of the pathogenesis of C. difficile infection. Although patients are likely exposed to C. difficile throughout their hospitalization, we hypothesize that they are at negligible risk for CDI until exposed to an antimicrobial agent. Following antimicrobial exposure, the patient is now susceptible to infection, and when exposed to C. difficile, one of three outcomes ensues: the patient becomes colonized but remains asymptomatic, the patient develops CDI, or potentially the patient does not develop any detectable infection. Once the patient is established as an asymptomatic carrier, data indicate that the patient is at decreased risk for subsequent CDI (7). (Reproduced from Johnson S, Gerding DN. Clostridium difficile-associated diarrhea. Clin Infect Dis 1998;26:1027-1036, with permission.) |
Variant strains of C. difficile that do not produce toxin A have been recovered from clinical specimens around the world. These toxin A−/B+ strains were initially recovered from asymptomatic children and were not thought to be pathogenic. Recently, a particular toxin A−/B+ variant that has a 1.8-kb deletion in the toxin A gene (referred to as toxinotype VIII, serogroup F, restriction endonuclease analysis [REA] group CF) (20) has been recovered from multiple CDI cases, including a fatal case of PMC (21). In addition, two well-documented hospital outbreaks with the A−/B+ variant (22,23) suggest that toxin B or some virulence determinant other than toxin A in these strains is sufficient to cause C. difficile disease, consistent with the observations in hamsters using toxin A knockout C. difficile strains (13).
In addition to toxins A and B, some strains of C. difficile (but not toxinotype VIII strains) also produce an adenosine diphosphate-ribosyltransferase (referred to as binary toxin) (24). The role of binary toxin in the pathogenesis of CDI has not yet been determined, but during the first decade of the 21st century, there have been multiple outbreaks of a toxin-variant strain of C. difficile that has caused severe CDI in North America, the United Kingdom, and Europe (25,26). It is a toxinotype III strain variously known as type NAP1 by pulse field, type 027 by polymerase chain reaction (PCR) ribotyping, and group BI by REA and is collectively termed BI/NAP1/027 (25). The etiology of the high CDI rates and high severity of illness is not known, but the strains produce large amounts of toxins A and B in vitro, produce binary toxin, and have a deletion in the tcdC gene responsible for downregulation of toxin production (27).
The other important aspect of C. difficile pathogenesis lies within the apparent host resistance of some patients to C. difficile disease. Infants in the first year of life frequently carry C. difficile in high numbers with high levels of both toxins in their stools (28), yet C. difficile disease is rare in this group. Age-dependent expression of the epithelial cell receptor for toxin A may explain this apparent resistance in young infants (29). Additionally, during outbreaks in adult settings, asymptomatic carriage is a more frequent outcome of CDI than is symptomatic infection (8,30). Disease manifestation and severity are not solely strain-specific phenomena (31); although the mechanism of this apparent host resistance in adult asymptomatic fecal excretors is not completely known, host antibody response to toxin A has been one factor implicated (32,33).
CLINICAL DISEASE SPECTRUM
Although the most common clinical manifestation of CDI is diarrhea, the disease spectrum ranges from asymptomatic colonization or fecal excretion to PMC to septic shock with and without toxic megacolon, which may present with signs of an acute abdomen but without diarrhea (34). As with other enteric infections, asymptomatic colonization is two to five times more common than clinical disease associated with C. difficile (8,30). Although colonized patients may carry epidemic strains responsible for illness in other patients, they are not at increased risk of CDI (9), they are not at risk for subclinical protein-losing enteropathy (35), and treatment of these patients with metronidazole or vancomycin is not advised (36). Asymptomatic colonization is also very common in neonates, and it has been difficult to attribute any disease manifestation in neonates to C. difficile. However, the pathogenic role of C. difficile in children cannot be completely ignored, particularly in children over 1 year of age and especially in children with hypogammaglobulinemia (37).
Diarrhea with or without demonstrable pseudomembranes in the colon is the most common manifestation of C. difficile disease. Although C. difficile is the most common recognized cause of antibiotic-associated diarrhea, C. difficile accounts for only 15% to 25% of these cases (38). CDI may occur during antimicrobial administration or several weeks after discontinuation of the antimicrobial. In a prospective study of clindamycin therapy, one-third of the patients developed diarrhea or colitis several days to 3 weeks after completion of clindamycin treatment (39). This marked variability of time between onset of diarrhea and antimicrobial exposure also supports exogenous acquisition as the major source of CDI. The incubation period for diarrhea after acquisition of C. difficile is less than 1 week, with a median onset of 2 days following acquisition (8,30). However, the incidence of CDI in the first 30 days following hospital discharge is very high, suggesting the possibility of either acquisition of the microorganism in the community following discharge or a longer incubation period (40).
The severity and chronicity of diarrhea is also variable. In some cases, symptoms may be mild and respond to simply withdrawing the offending antimicrobial. CDI resolved spontaneously within 48 to 72 hours in 25% of patients in
one series (41). More commonly, the diarrhea becomes chronic and severe if not diagnosed and treated with specific therapy. At presentation, symptoms may consist of only a few loose stools per day or multiple, large-volume, watery stools and signs of dehydration (38). Stools may have mucus or evidence of occult blood but are rarely associated with visible blood (42). Other findings commonly associated with CDI include abdominal pain (22%), ileus (21%), fever (28%), and leukocytosis (50%) (42). CDI should be considered in any hospitalized patient with leukocytosis, particularly those with white blood cell counts >30,000 cells/mm3, even without the presence of diarrhea (43). Complications of severe disease include dehydration, electrolyte imbalance, hypotension, hypoalbuminemia with anasarca, toxic megacolon, colonic perforation, and sepsis, and in 1% to 7% of cases, death can result (26,38). Higher mortality rates have been particularly noted in CDI caused by the epidemic BI/NAP1/027 strain (26,44). A characteristic of CDI is the high rate of clinical recurrence following successful therapy, which may result from either relapse with the same strain or reinfection with a new strain (45,46).
one series (41). More commonly, the diarrhea becomes chronic and severe if not diagnosed and treated with specific therapy. At presentation, symptoms may consist of only a few loose stools per day or multiple, large-volume, watery stools and signs of dehydration (38). Stools may have mucus or evidence of occult blood but are rarely associated with visible blood (42). Other findings commonly associated with CDI include abdominal pain (22%), ileus (21%), fever (28%), and leukocytosis (50%) (42). CDI should be considered in any hospitalized patient with leukocytosis, particularly those with white blood cell counts >30,000 cells/mm3, even without the presence of diarrhea (43). Complications of severe disease include dehydration, electrolyte imbalance, hypotension, hypoalbuminemia with anasarca, toxic megacolon, colonic perforation, and sepsis, and in 1% to 7% of cases, death can result (26,38). Higher mortality rates have been particularly noted in CDI caused by the epidemic BI/NAP1/027 strain (26,44). A characteristic of CDI is the high rate of clinical recurrence following successful therapy, which may result from either relapse with the same strain or reinfection with a new strain (45,46).
In distinction from antibiotic-associated diarrhea in general, C. difficile is responsible for nearly all cases of PMC that have been reported since 1978 (38). The pseudomembranous intestinal lesions associated with C. difficile (which are present in only about half the patients who have diarrhea and C. difficile toxin in stool) have a characteristic gross and histologic appearance (47). Early in the disease course, small (1-2-mm), raised, yellowish white plaques are noted, which may enlarge and coalesce (48). These lesions, which are composed of fibrin, mucus, necrotic epithelial cells, and leukocytes, are restricted to the colon; therefore, this disease should be referred to as PMC rather than enterocolitis. Although PMC can be visualized by the sigmoidoscope in 90% of patients who have PMC, some patients have disease limited to the right colon, and the presentation may mimic appendicitis or Crohn disease (49).
Fulminant C. difficile colitis and toxic megacolon are less common manifestations, but important syndromes to recognize, as they are associated with a high mortality rate and frequently require surgical intervention (50,51). It is ironic that this most severe manifestation of CDI often occurs without diarrhea, and as a result, the diagnosis is frequently missed or delayed (34). Risk factors for severe disease in one study included immunosuppression, prior CDI, and prior surgical procedures (51). A rapidly increasing peripheral white blood cell count with a left shift may be an important clue to impending fulminant disease. A rising white blood cell count during medical treatment that is approaching 50,000/mm3 or a lactic acidosis approaching 5 mmol/L are indication for colectomy in patients failing medical management (52).
 FIGURE 37-2 Time line for definitions of Clostridium difficile infection (CDI) exposures. Case patients with symptom onset during the window of hospitalization marked by an asterisk (*) would be classified as having community-onset, healthcare facility-associated disease (CO-HCFA), if the patient was discharged from a healthcare facility within the previous 4 weeks; would be classified as having indeterminate disease, if the patient was discharged from a healthcare facility between the previous 4 and 12 weeks; or would be classified as having community-associated CDI (CA-CDI), if the patient was not discharged from a healthcare facility in the previous 12 weeks. HO-HCFA, healthcare facility-onset, healthcare facility-associated CDI. (Adapted from McDonald LC, Coignard B, Dubberke E, et al. Recommendations for surveillance of Clostridium difficile-associated disease. Infect Control Hosp Epidemiol 2007; 28:140-145.) |
DIAGNOSIS
The diagnosis of CDI depends first on establishing the presence of clinical diarrhea or other gastrointestinal symptoms compatible with C. difficile disease such as abdominal pain or, rarely, ileus without diarrhea. Definitions for healthcareassociated CDI have been published and are summarized in Figure 37-2 (55,56). Two major avenues are available to establish the diagnosis in a clinically ill patient: endoscopic procedures to detect the presence of PMC and laboratory studies to document the presence of C. difficile or its toxins in the stool. The latter tests include (a) stool culture, antigen tests, and the PCR to detect the microorganism; (b) the cell cytotoxin test for toxin B; and (c) a variety of immunoassay tests for the presence of toxin A or toxins A and B in stool. The best way to establish the diagnosis of CDI remains controversial; however, the availability of commercial PCR tests for the detection of toxigenic C. difficile is likely to provide the best combination of sensitivity, specificity, and rapid results reporting for those labs equipped to do PCR (56,57, 58, 59 and 60). Advantages and disadvantages of each of the diagnostic modalities are discussed below. Sensitivity and specificity are summarized in Table 37-1.
Endoscopic Procedures
Endoscopic procedures indirectly indicate C. difficile disease by demonstrating PMC. The procedure may be performed with a flexible fiberoptic sigmoidoscope or a fiberoptic colonoscope. The latter is the only device that permits visualization of the entire colon, but it is not routinely used for this diagnosis because of the need for extensive colon preparation and the high cost of the procedure. Flexible sigmoidoscopy is the most frequently employed of these procedures and allows visualization of the distal 60 cm of the colon. The diagnosis can be made by direct visualization of the pseudomembranes; biopsy may be required if the lesions are small (48). The major deficiency of endoscopy is that it is highly insensitive. Pseudomembranes were seen in only 51% of patients who had diarrhea, a positive stool cytotoxin assay, and positive stool culture for C. difficile (42). It is probable that PMC is a late manifestation of C. difficile disease that is more likely to occur in the proximal colon first, accounting for the low sensitivity of endoscopy for the diagnosis of CDI (49).
TABLE 37-1 Sensitivity and Specificitya of Diagnostic Tests for Clostridium difficile-Associated Disease | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Clinically, the major advantage of endoscopy is the rapidity with which a diagnosis can be made. This is particularly important in critically ill patients in whom the diagnosis of a surgical abdominal process is being considered. These patients often have symptoms of ileus or obstruction, which can be caused infrequently by C. difficile. Visualization of PMC by endoscopy can avert inappropriate emergency abdominal surgery and permit confident initiation of C. difficile-specific treatment. However, the vast majority of C. difficile-infected patients are not critically ill, and endoscopy has been almost totally supplanted by laboratory tests for the diagnosis of CDI.
Laboratory Diagnosis
Stool Toxin Tests
Cell Cytotoxin Assay The cell cytotoxin assay has been traditionally considered the most specific of stool diagnostic tests for CDI, albeit not the most sensitive when compared to stool culture (61, 62 and 63). A wide variety of cell types may be used for testing, but the test requires that laboratories maintain their own cell lines or purchase them in commercially available kits. For the test to be diagnostic, the cytopathic effect of the specimen (primarily due to the effects of toxin B) must be neutralized by specific C. difficile or C. sordellii antitoxin. Appropriate specimen dilution is critical to achieving test sensitivity and specificity. Sensitivity of the cell cytotoxin assay has ranged from 56% to 100% (depending upon the comparator test) when used to detect C. difficile disease in patients with clinically significant diarrhea (61). Low sensitivity is thought to be due to the inactivation of toxins by proteases or other inactivating substances in stool, an occurrence that is more likely if the specimen remains at room temperature for a long period. Despite its relative insensitivity, stool testing for cytotoxin remains one of the gold standards against which C. difficile laboratory tests are evaluated. A major drawback of this test is the long (48-hour) turnaround time for results in most laboratories.
Immunoassay for Toxins A and B A variety of immunoassays are available that detect C. difficile toxin A, or toxins A and B. Enzyme immunoassays that detect only toxin A have been largely abandoned as they fail to detect clinically significant CDI caused by toxin A−/B+ strains (21,22). These tests have now been widely adopted by clinical laboratories (and are often the only tests offered) because of the rapid turnaround time and decreased workload compared to cytotoxin assays. When compared only to cytotoxin testing, the sensitivity of a battery of immunoassays was 82.8% (range 66.7-91.7%) and the specificity was 95.4% (range 90.8-98.8%) (64). The relative sensitivity of toxin immunoassays compared to culture with confirmation of toxin-producing isolates (also called toxigenic culture) was 75% (range 60-86.4%) and the specificity was 96.1% (range 91.4-99.4%) (64). In an effort to overcome the low sensitivity of toxin immunoassays, clinicians may order multiple stool tests for toxin, but this practice, because of the lack of enzyme immunoassay specificity, leads to more false-positive test results than true positives and this practice should be discouraged by both laboratories and clinicians (56,65). Immunoassays that detect both toxins A
and B are more sensitive than assays that just detect toxin A and they also detect variant A−/B+ strains of C. difficile that are not detected by toxin A immunoassays (66,67). The frequency of toxin A−/B+ variant strains has been low (0.2-3%) in most surveys (67, 68 and 69) but may be considerably higher in A−/B+ hospital outbreaks (22,23). Infections due to these A−/B+ variants are important to recognize clinically because of their propensity to cause clinically severe outbreaks worldwide (21, 22 and 23). Advantages of the toxin immunoassays are the relatively high specificity and rapid turnaround time (if the tests are not batched).
and B are more sensitive than assays that just detect toxin A and they also detect variant A−/B+ strains of C. difficile that are not detected by toxin A immunoassays (66,67). The frequency of toxin A−/B+ variant strains has been low (0.2-3%) in most surveys (67, 68 and 69) but may be considerably higher in A−/B+ hospital outbreaks (22,23). Infections due to these A−/B+ variants are important to recognize clinically because of their propensity to cause clinically severe outbreaks worldwide (21, 22 and 23). Advantages of the toxin immunoassays are the relatively high specificity and rapid turnaround time (if the tests are not batched).
Tests for the Detection of the Microorganism in Stool
Stool Culture Stool culture for C. difficile on selective media (cycloserine-cefoxitin-fructose agar [CCFA]) has proven to be the most sensitive tests for CDI (62,63). Special transport of stool specimens is not required for culture; however, CCFA media must be anaerobically reduced prior to inoculation so that residual oxygen in the media does not kill the microorganism when germinating from spores. Addition of taurocholate or lysozyme to CCFA enhances recovery of spores (70). Clinical laboratories must exercise diligent quality control of commercially purchased media, as recovery rates have been shown to vary widely among manufacturers (71,72). Culture is essential if epidemiologic investigation employing microorganism typing (including molecular techniques) is desired. The major criticism of culture as a diagnostic test for CDI is that it lacks specificity when compared with the cytotoxin assay because nontoxigenic C. difficile may be isolated and because many patients in hospitals may be asymptomatically colonized with toxigenic C. difficile and have diarrhea for reasons unrelated to the microorganism (61,73). When culture and cell cytotoxin assay are both performed for CDI diagnosis, both tests yield positive results in well over half of patients with CDI, but stools are culture-positive and cytotoxin-negative in the remaining patients (61). About 75% of the C. difficile isolates recovered from this latter group produce toxins in vitro. Patients with culture-positive, cytotoxin-negative stool specimens should be considered CDI cases if the isolated microorganism produces toxin in vitro (culture positive for toxigenic C. difficile) (41,74).
Antigen Tests The latex agglutination test is a commercially available stool test that was originally thought to detect toxin A but was subsequently shown to detect a different C. difficile protein, glutamate dehydrogenase (GDH) (75). The test is rapid, simple to use, and relatively inexpensive, but in its latex agglutination, detection format lacked both sensitivity and specificity and did not distinguish toxigenic from nontoxigenic strains of C. difficile (62,63) (Table 37-1). The detection method has been reformulated to use an enzyme-linked immunosorbent test or enzyme immunoassay for GDH, also referred to as “common antigen” (76,77). Because the test is sensitive but not specific due to the lack of distinction between toxigenic and nontoxigenic strains of C. difficile, this test must be used in conjunction with a toxinbased assay or toxin gene detection of C. difficile (77,78). Using this “two-step” method of GDH testing followed by cell cytotoxin testing for all positives, the GDH test can be used as a screening test to rule out CDI in the 80% to 90% of patients whose specimens test negative by GDH; however, up to 50% of the patients who test positive by GDH will not have confirmatory toxin found in the stool and will be ultimately reported as negative for C. difficile toxin (77).
Polymerase Chain Reaction The use of PCR to detect C. difficile in stool is now commercially available and primers for amplification of conserved areas of the toxin B gene are used to detect toxigenic strains of C. difficile with a degree of sensitivity between 86% and 94% that of toxigenic culture (57, 58, 59 and 60). For laboratories that have PCR amplification equipment, the use of PCR for clinical diagnosis will significantly improve sensitivity and specificity of diagnosis of CDI with rapid test results turnaround; however, the cost per test is substantially higher than for other tests. For laboratories that lack PCR capability, the GDH two-step algorithm may be the most effective diagnostic method (56).
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
