Chapter 188. Clostridium Difficile–Associated Disease (CDAD)

1. Which factors are associated with an increased risk of primary or recurrent Clostridium difficile–associated disease (CDAD)?

2. What factors affect the accuracy of diagnostic testing of CDAD?

3. What are the most common complications of CDAD?

4. What therapies are most efficacious for primary, recurrent, or refractory CDAD?

5. What system improvements can reduce the incidence of CDAD?

Clostridium difficile was first described in 1935 by pediatricians studying the acquisition of bowel flora in newborns and was termed “difficult” because of its resistance to growth on conventional media. It was not until 1978 that it was found to be associated with toxins in the stools of patients with colitis. Clostridium difficile–associated disease (CDAD) has now become a major nosocomial problem, affecting 3 million inpatients yearly in the United States. Acquisition of C difficile occurs in up to 30% of hospitalized patients, although only 30–60% have symptoms. Patients with symptomatic disease stay an average of 3.6 days longer in hospital than expected, at a cost of over $1 billion annually in the United States.

Normal gut flora confers resistance to C difficile colonization; only 1–4% of healthy adults carry C difficile in their bowel flora. Loss of normal colonic flora, usually from antibiotics, allows C difficile proliferation (clindamycin, cephalosporins, and fluoroquinolones are common offenders). C difficile colitis is also associated with other events that disturb gut flora or host immunity, such as bowel preparations, cytotoxic agents, and colonic inflammation due to inflammatory bowel disease.

CDAD was first described in frail or elderly hospitalized patients after antibiotic exposure, but it is now occurring in younger, healthy individuals with no apparent exposure to health care environments or antibiotics. Up to 20% of CDAD cases are now community acquired. C difficile and its spore form are easily transmitted from person-to-person or by contact with objects in the hospital environment, such as stethoscopes, clothing, and toilets. The spore form is ubiquitous in the hospital environment, resistant to alcohol-based treatment and other disinfectants, and can survive for months. Even after routine cleaning, 10–50% of room surfaces still contain spores.

C difficile is an anaerobic spore-forming bacterium that produces two exotoxins (A and B) that adhere to intestinal epithelial cells, causing cell death. These toxins are transcribed from a locus composed of five genes (two toxin genes and three regulatory genes). The current epidemic strain of C difficile (BI/NAP1/027) has several factors that increase its virulence. A mutation in one of the regulatory genes results in a massive increase in amount of toxin production. This strain also shows high-level fluoroquinolone resistance, giving it a competitive advantage in environments where fluoroquinolone use is widespread. This strain also produces an additional toxin (binary toxin), the clinical significance of which is not clear.

Patient risk of acquiring CDAD and severity of clinical illness are variable (Table 188-1). The clinical spectrum of CDAD ranges from minor diarrhea to fulminant colitis with toxic megacolon. Most patients present with some combination of acute, nonbloody, watery diarrhea (although some present with obstipation), lower abdominal pain, and leukocytosis. Fever is relatively uncommon (present in less than one-third of patients). Mimickers of CDAD include bowel obstruction, fecal impaction, or toxic megacolon.

The differential diagnosis of acute diarrhea in hospitalized patients is extensive, and less than a third of cases of nosocomial diarrhea are due to CDAD. Non-CDAD causes of diarrhea should be considered in all patients, especially those for whom C difficile testing is negative, or those who do not respond to appropriate treatment. Common benign conditions causing nosocomial diarrhea include medications (stool softeners or laxatives as part of standardized order sets), foods (including tube feedings), oral contrast (from computed tomographic scans), and antibiotic-induced alterations of bowel flora. Less common microbial causes of diarrhea include Klebsiella oxytoca, Klebsiella pneumoniae, and Staphylococcus aureus. For a full discussion of the approach to nosocomial diarrhea, see Chapter 80.

The diagnosis should be made on the combination of symptoms and microbiologic data. Testing for C difficile is imperfect (Table 188-2). The cytotoxic assay is the gold standard, with excellent sensitivity (94–100%) and specificity (99%), but it is expensive and time consuming, hence most medical centers no longer routinely perform it. More commonly used are enzyme-linked immunosorbent assay (ELISA) tests for toxins. Most detect both toxin A and B, but some only detect toxin A. ELISA tests have the advantage of being quicker and cheaper but have a sensitivity of only 70–90%. Sensitivity is reduced in some strains that do not produce toxin A (or have mutant toxin A), as well as in patients that have already received antibiotic treatment for CDAD. False-positive tests are rare, and any symptomatic patient with a positive test should be treated. Culture of the organism is not routinely available, and is best reserved for outbreak investigations.

Severe colitis with pseudomembranes is virtually pathognomonic for C difficile (Figure 188-1). It is diagnosed endoscopically by the appearance of raised yellow or white plaques scattered over the colonic mucosa.

In summary, 20–30% of hospitalized patients are colonized with C difficile, but some strains fail to produce toxins, and some patients do not develop symptoms from the toxin when it is present. Most patients without diarrhea should not be tested for C difficile. Some patients with symptomatic C difficile will have a negative test (20–30%) and should be treated if clinical suspicion is high. Symptomatic patients with a positive test should be treated.

Patients with mild CDAD usually do not require hospital admission, but those with moderate to severe CDAD usually do (Table 188-3). Proper infection control measures should be instituted to protect other patients. Contact precautions should be instituted immediately for patients with proven or suspected C difficile infection. This includes gowns and gloves for each patient contact, and dedicated equipment in each room, such as blood pressure cuffs and stethoscopes. Strict handwashing before and after all patient contact is required as C difficile spores are resistant to alcohol-based treatment.

Treatments depend on the severity of the disease at presentation and the response to initial therapy (Table 188-3). The offending antibiotic should be stopped whenever feasible. Symptomatic improvement is likely in mildly ill patients after antibiotics are ceased. It is also recommended to discontinue all antiperistaltic agents (although data for their causal link in worse outcomes are retrospective only). These include loperamide, diphenoxylate with atropine (Lomotil), bismuth subsalicylate (Pepto-Bismol), and tincture of opium.

First-line treatment with either oral vancomycin or metronidazole is usually successful, although only vancomycin is approved for this indication. A recent Cochrane review concluded that no single antibiotic was superior in symptomatic or bacteriologic outcomes for unselected patients (Table 188-4). However, a more recent randomized, controlled trial of 172 patients found that vancomycin and metronidazole had equivalent cure rates in low-risk patients, but that vancomycin had higher cure rates in high-risk patients (Table 188-4). If intravenous therapy is required, metronidazole is more effective than vancomycin. Treatment should be continued for at least 14 days (or at least 10 days after the symptoms have abated).

There are few antibiotic therapies available other than vancomycin and metronidazole. In a recent Cochrane review, vancomycin and metronidazole were compared to several other agents, including teicoplanin, fusidic acid, bacitracin, rifaximin, and nitazoxanide. None of the comparisons showed a significant advantage or disadvantage, other than bacitracin appeared inferior to vancomycin for initial bacteriologic response, and teicoplanin, which is not currently available in the United States, appeared superior to vancomycin for initial bacteriologic response and cure (Table 188-5).

There is some evidence that adding probiotics to standard antibiotic regimens may enhance treatment of CDAD. One study found that adding the probiotic Saccharomyces boulardii to a standard antibiotic regimen improved symptoms and reduced the risk of recurrence compared to placebo (Table 188-4). However, these over-the-counter preparations are not regulated, and there have been case reports of fungemia in elderly and debilitated patients. Therefore, they cannot be recommended for routine use in all CDAD cases but may be appropriate for immunocompetent patients at high risk of recurrence.

Few data exist to support other therapies. The toxin binder tolevamer appeared to be equivalent to vancomycin in phase 2 studies but failed to reach noninferiority to vancomycin or metronidazole in a phase 3 study. Therapies studied in uncontrolled settings or case series include intravenous immunoglobulin, stool transplantation, vancomycin enemas, and anion binding resins, none of which can be recommended for routine use. Active and passive immunization is being studied but is not ready for clinical use.

In patients with refractory or fulminant disease unresponsive to metronidazole or vancomycin, second-line therapies should be considered, including vancomycin enemas (especially for patients with ileus), intravenous immunoglobulin, tigecycline, or subtotal colectomy.

Recurrence of CDAD occurs in about 20% of patients. Relapses probably result from persistent colonization with C difficile spores, which revert to toxin-producing forms after antibiotics are stopped, and not antibiotic resistance. Recurrences usually present within 5 days of cessation of treatment but can be delayed up to a month later. Risk factors include the number of prior episodes, resumption or continuation of antibiotics, poor antibody response, female gender, and diverticular disease. If the patient has more than one recurrence, the chance of another is more than 60%. Noncompliance and other causes of diarrhea should also be considered as possible explanations for continued symptoms.

The first recurrence should receive the same treatment as the first occurrence. Multiple recurrences usually require additional treatment. Tapering or pulsed antibiotics are frequently utilized. One small study used a tapering vancomycin regimen over the course of 6 weeks. The regimen started with vancomycin four times daily for a week, then twice daily for a week, then daily for a week, then every second day for a week, then every third day for 2 weeks. Of the 22 patients treated in this study, all were disease free at 6 months. Oral rifaximin (for 14 days following traditional treatment) has also been effective in reducing subsequent recurrences in uncontrolled trials.

Preventing primary or recurrent CDAD is of paramount importance, but there is little evidence that lactobacilli or other agents are effective in primary or secondary prevention. There is also no evidence that routine use of vancomycin or metronidazole in patients starting antibiotics for other conditions reduces the incidence of CDAD.

Refractory or fulminant CDAD develops in some patients, leading to systemic inflammatory response (characterized by leukocytosis, hypotension, anasarca, and renal failure). The most devastating complication of CDAD is toxic megacolon. In this condition the large bowel becomes massively dilated (> 7 cm) with characteristic thumbprinting due to submucosal edema. These patients are at high risk of perforation and need surgical consultation to be assessed for emergent colectomy. Mortality is high, approximately 50%. Unfortunately, atypical presentations of toxic megacolon are not unusual. In one study of patients requiring colectomy, 20% had a negative C difficile antigen test, and almost 40% had no diarrhea. Therefore, CDAD should be entertained in any patient with colonic dilation.

Patients can be discharged as soon as they are clinically stable (able to eat, walk, and toilet). Transitional care provisions may be needed in those unable to perform these elements of self-care. Continued infection control practices (contact isolation and hand hygiene) need to remain in effect wherever the patient is being discharged; it can be presumed that patients remain infectious (toxin-shedding) for the duration of the diarrhea. Discharge planning should ensure the completion of the full course of antibiotics prescribed, with contingency planning if the diarrhea recurs or worsens after discharge.

Four major areas of system improvement can decrease the hospital prevalence of C difficile.

1. Antibiotic stewardship. Several studies have shown that antibiotic restrictions can decrease rates of C difficile infection. This should be part of a systemwide attempt to reduce unnecessary antibiotics by the use of automatic stop days and restricted formularies.

2. Early diagnosis and isolation. Early diagnosis, treatment, and immediate contact isolation of patients are essential in curbing the nosocomial acquisition of C difficile. Rapid turnaround time of diagnostic testing is essential, and immediate provider notification of positive tests should be built within the laboratory notification system. Systematic, reproducible protocols should exist for initiating contact precautions as soon as a C difficile test is ordered.

3. Hand hygiene. Hand washing is essential; C difficile spores are environmentally hardy and resistant to alcohol-based hand rubs. A hand hygiene program with active surveillance and provider feedback should be part of a systemwide attempt to reduce all hospital-acquired infections, including C difficile.

4. Environment. Appropriate environmental cleaning and environmental surveillance are crucial, to ensure patient and workplace areas are not contaminated.