The primary means of preventing C difficile are to limit the use of antibiotics, particularly those in specific classes believed to carry particularly high risk, and to ad- here to infection prevention measures, such as the use of gloves and gowns and hand hygiene. Additional infection prevention measures include appropriate daily environmental cleaning and disinfecting.

Physicians should participate in antibiotic stewardship programs that are designed in conjunction with microbiologists, infectious disease specialists, pharmacists, infection preventionists, hospital epidemiologists, and hospital administrators. Many strategies have been used, from restricting the use of high-risk drugs to routinely reviewing antibiotic therapy and giving feedback to the treating clinicians. Stewardship protocols can also focus on improving antibiotic use for specific syndromes or conditions, such as urinary tract infections and respiratory infections.

Because of the increasing frequency of community-associated infection, antibiotic use among outpatients may also be an important contributor to C difficile infection and such use may be a target for preventing these infections.

Hospitalized patients with C difficile infection should be assigned to a private room or a room with other similarly infected patients until 48 hours after diarrhea has resolved (6, 32). Patients with C difficile infection who are colonized or infected with another multidrug-resistant organism should not room with patients with C difficile infection who have another multidrug-resistant organism that differs from theirs (6). Routine infection prevention practices include strict hand hygiene before and after every patient encounter and contact precautions that include use of disposable gloves and gowns during care of patients with C difficile infection or when there is the possibility of exposure to their body fluids. Although no published trials have been designed specifically to study the protective effect of gowns, C difficile has been cultured from the uniforms of hospital workers and use of disposable gowns is recommended on that basis (6, 33).

Several studies have shown that conventional handwashing with soap and water is superior to alcohol-based hand sanitizers for removing C difficile spores, and any remaining spores of C difficile are highly resistant to alcohol (35, 36). Because organic matter interferes with alcohol’s ability to inactivate all bacteria, handwashing with soap and water is recommended if there is direct contact with stool or an area with fecal contamination or when hands are visibly soiled. However, studies have failed to show a change in the rate of C difficile infections when comparing alcohol-based hand products with soap and water across patient populations (37, 38). Moreover, even soap and water against C difficile spores cannot achieve the usual 3- to 4-log reductions commonly associated with alcohol against other bacteria (39). Therefore, gloves are the primary method of preventing C difficile transmission. In outbreak settings, the IDSA/SHEA guidelines recommend soap and water over alcohol-based hand products after glove removal (6).

Disposable medical equipment should be used when a patient with C difficile infection is being treated. To the extent possible, medical equipment should be dedicated to the patient’s room, and reusable equipment should be thoroughly cleaned and disinfected after use with a sporicidal disinfectant registered by the U.S. Environmental Protection Agency (6). Daily and terminal cleaning and disinfecting patient rooms, focusing on high-touch surfaces, should also be done (6). In an outbreak setting, interventions that included the use of a sporicidal agent for terminal disinfection have been associated with reductions in C difficile infection. However, in nonoutbreak settings, terminal disinfection with a sporicidal agent has not consistently led to reductions in C difficile infection. Therefore, terminal disinfection with a sporicidal agent is recommended as a supplemental intervention in outbreak settings or if there is concern of environmental transmission (e.g., repeated cases of infection in the same room) (6).

Patients should be asked about the known risk factors for infection (see the Box). Patients commonly develop antibiotic-related diarrhea with C difficile during or shortly after receiving antibiotics, but symptoms can occur up to several months afterward. In patients who have been hospitalized for at least 3 days, C difficile is by far the most common enteric pathogen (40).

C difficile infection should be considered in patients who have diarrhea (≥3 loose stools in 24 hours) with or without abdominal pain, especially if they have a recognized risk factor (including recent antibiotic use, hospitalization, or advanced age) with no obvious alternative diagnosis (including laxative use in the past 48 hours). Nausea, vomiting, and fever are often but not always present. Physical findings vary depending on the length and severity of disease. There may be signs of dehydration, the abdomen may be tender, and in severe cases peritoneal signs may be present. Ileus or toxic megacolon can cause abdominal distention (Table 1).

Several diagnostic options are available, such as cell culture cytotoxicity neutralization assay (CCNA), toxigenic culture, toxin A and B enzyme immunoassays (EIAs), nucleic acid amplification tests (NAATs), and glutamate dehydrogenase (GDH) assay. However, no single test is considered to be the best laboratory testing method.

CCNA has a sensitivity of 94% to 100% and a specificity of 97%; however, it takes 24 to 48 hours to obtain results and the test requires a tissue culture laboratory, which is not present in most hospitals. Toxigenic culture is more sensitive than CCNA but requires at least 48 hours for incubation, plus further testing to confirm that the isolate is a toxigenic strain. (Despite these problems, culture and molecular typing of isolates are important in epidemiologic studies.) Many commercial toxin EIAs are available to diagnose C difficile infection, but they vary widely in sensitivity. Several NAATs are also commercially available and are more sensitive than toxin EIAs, but they may have lower positive predictive values depending on the pretest probability and the assay’s limit of detection (6). Enzyme-linked immunosorbent as-says for GDH are very sensitive but not specific because GDH is present in both toxigenic and nontoxigenic strains.

Some authorities recommend using a multistep algorithm to diagnose C difficile infection (GDH plus toxin EIA, GDH plus toxin EIA with NAAT confirmation if results are discordant, or NAAT plus toxin EIA) (6). In clinical care where testing is limited to appropriate patients (e.g., those with 3 or more unformed stools in 24 hours in the absence of laxatives) and quality measures ensure appropriate implementation, NAAT alone or a multistep algorithm is generally recommended over toxin EIA alone. If there are no prespecified criteria for C difficile testing, toxin EIA as part of a multistep algorithm is recommended over NAAT alone (6) (Table 2).

Another test with diagnostic value is direct inspection via sigmoidoscopy or colonoscopy. Characteristic raised yellow mucosal plaques, or “pseudomembranes,” are highly suggestive of C difficile–associated diarrhea. Flexible sigmoidoscopy without colon preparation is often adequate and has been shown to be effective for establishing the diagnosis in many cases in which toxin tests yield negative results (41). However, sigmoidoscopy may miss cases that could be detected by colonoscopy because of more proximal disease (42). A limitation of direct inspection is that C difficile infection often does not have pseudomembranes, which may be a marker of severe disease; such cases without pseudomembranes can only be detected by toxin testing or other methods.

Other tests that may provide supporting evidence for the diagnosis and are valuable in assessing disease severity and complications include complete blood count, measurement of serum creatinine and lactate levels, and computed tomography of the abdomen (Table 3).

Patients with C difficile infection may have findings suggestive of colitis, such as mucosal thickening, on abdominal imaging (45). Complications, such as toxic megacolon or perforation, may be detected.

Repeated testing (within 7 days) during the same episode of diarrhea should be avoided. If recurrent C difficile infection after successful treatment and cessation of diarrhea is suspected, testing should include toxin detection. Empirical treatment is discouraged. Testing should also not be done in asymptomatic patients. However, for research purposes or as a supplemental intervention in places with high rates of C difficile infection, screening of asymptomatic patients may be reasonable (e.g., to identify asymptomatic patients potentially at risk for shedding so prevention measures can be instituted).

Other infectious and noninfectious causes should be considered in patients with diarrhea and a work-up that is negative for C difficile. Infectious causes include Salmonella; Shigella; Campylobacter; Shiga toxin–producing strains of Escherichia coli; and, in immunocompromised persons, cytomegalovirus, Cryptosporidia, and other opportunistic organisms. However, with the exception of cytomegalovirus and possibly other opportunistic pathogens in immunocompromised persons, these infections are unusual in patients who develop diarrhea in the hospital. Noninfectious causes include intestinal obstruction, ischemic bowel disease, inflammatory bowel disease, gastrointestinal cancer, and drug-associated diarrhea.

The physician should request a consultation or refer the patient if the diagnosis is uncertain. Lower endoscopy by a gastroenterologist may provide a diagnosis when results of stool studies are negative or when a diagnosis is needed more rapidly than available laboratories can provide.

Discontinuing antibiotic therapy should be considered in all patients with C difficile infection if doing so does not jeopardize recovery from other conditions. Patients with mild diarrhea and a normal or nearly normal leukocyte count and normal creatinine levels who are otherwise not at risk for severe disease or complications may be observed for a few days to determine whether discontinuation of antibiotic therapy is sufficient to resolve the condition.

In addition to discontinuing use of antibiotics as soon as possible, fluid and electrolyte imbalances should be corrected. Antiperistaltic agents should be avoided because they may prevent both distribution of the therapeutic antibiotic within the gut and expulsion of the toxin.

The updated IDSA/SHEA guidelines indicate that either vancomycin or fidaxomicin is recommended over metronidazole for an initial episode of C difficile infection in adult patients, regardless of severity (Figure) (6). Severe disease has been variously defined, but the minimum criteria cited in the IDSA/SHEA guideline are a leukocyte count of at least 15 000 cells/mL or a creatinine level greater than 1.5 mg/dL (6). The recommended dosage for nonsevere or severe disease is vancomycin, 125 mg orally 4 times per day, or fidaxomicin, 200 mg 2 times a day for 10 days (Table 4). If either drug is unavailable or contraindicated, oral metronidazole, 500 mg 3 times per day for 10 days, may be used as an alternative for non-severe disease only. Frequent or extended use of metronidazole has been associated with potentially irreversible neurotoxicity.

In patients with fulminant disease (e.g., hypotension or shock, ileus, megacolon), vancomycin should be administered orally or by nasogastric tube at a dose of 500 mg 5 times per day. If ileus is present or other factors prevent adequate distribution through the gut lumen, instillation of vancomycin at a dose of 500 mg in approximately 100 mL of normal saline per rectum every 6 hours should be considered. Intravenous metronidazole at a dose of 500 mg every 8 hours should be added in cases of fulminant infection, especially if ileus is present (6).

Historically, metronidazole was one of the main antibiotic agents used to treat C difficile infection; however, since 2000, a few randomized placebo-controlled trials have found oral vancomycin to be superior. In 1 study of 150 patients with either mild or severe disease, the overall cure rate was 97% with vancomycin versus 84% with metronidazole (P = 0.006); for severe disease, the cure rate was 97% with vancomycin versus 76% with metronidazole (P = 0.02) (47). In another trial that combined results from 2 multinational studies comparing vancomycin, metronidazole, and tolevamer (a nonantibiotic, toxinbinding polymer), clinical success of tolevamer (44%) was inferior to both antibiotics (P < 0.001), and clinical success of metronidazole (77%) was inferior to vancomycin (81%) (P = 0.02) (48).

Fidaxomicin has been compared with oral vancomycin in 2 randomized placebo-controlled trials. In the first study (49), the clinical cure rate with fidaxomicin was noninferior to vancomycin in both the modified intention-to-treat analysis (88% vs. 86%) and the per protocol analysis (92% vs. 90%). Significantly fewer patients treated with fidaxomicin than vancomycin had recurrent infection. In the second study (50), the clinical cure rate with fidaxomicin was also noninferior to vancomycin in both the modified intention-to-treat analysis (88% vs. 87%) and the per protocol analysis (92% vs. 91%).

Teicoplanin, a glycopeptide similar to vancomycin, has been found to be equivalent or perhaps superior to vancomycin but is not available in the United States (51). Nitazoxanide, bacitracin, fusidic acid, tigecycline, rifampin, and rifaximin have also been studied, but data are limited. Limited data have shown that toxin-binding anionexchange resins, such as cholestyramine, colestipol, and tolevamer, are not consistently effective. If these agents are used, administration must be timed to minimize inactivation of concomitant vancomycin.

Patients should be followed for clinical evidence of expected improvement, including, where applicable, resolution of fever, reduction in stool frequency, improvement in stool consistency, normalization of abdominal examination findings, rehydration as indicated by physical examination and laboratory values, and resolution of leukocytosis. Repeated stool testing is not recommended if symptoms have resolved (i.e., no need to test for cure) because it is common for patients to remain positive for C difficile after symptom resolution, and treatment of asymptomatic carriers is not indicated (6). Patients whose symptoms recur after successful treatment and resolution of diarrhea should be retested.

Numerous formulations of probiotics have been proposed to treat C difficile colitis. The premise is that these normally non-pathogenic yeasts and bacteria may repopulate the gastrointestinal tract and limit growth of C difficile; however, data are inconclusive.

Although no recommendations exist for the use of probiotics to prevent C difficile infection in patients receiving antibiotics, some meta-analyses suggest that short-term use might be safe and effective for patients who are not immunocompromised or severely debilitated (53, 54). One study indicated that hospitalized patients at high risk for C difficile infection should at least be informed of the potential benefits and harms of probiotics (53).

Patients who do not improve or who relapse after initial improvement should be reevaluated for an alternative or concurrent diagnosis or another explanation for a lack of improvement (e.g., persistent fever and leukocytosis may be due to a superimposed infection, and persistent diarrhea may be due to supplemental enteral feedings). In patients without an alternative diagnosis who do not improve, the physician must ensure that the dosage and delivery of therapy are maximized. Intravenous metronidazole and rectal instillation of vancomycin should be given if ileus might prevent orally administered vancomycin from reaching the colon. In some patients, especially those treated with metronidazole, duration of therapy may need to be extended to 14 days. In patients with fulminant infection who do not respond to vancomycin and metronidazole, tigecycline or passive immunotherapy with intravenous immunoglobulins has been used, but data are limited(6). Select patients may also benefit from surgery (see below).

Consultation with a gastroenterologist or an infectious disease specialist (or both) should be considered for patients who respond slowly or relapse. Management changes may be guided by endoscopy and may include manipulation of antibiotic therapy. Surgical consultation is essential if there is evidence of perforation and should be strongly considered in patients with toxic megacolon. It may also be valuable for patients with severe illness in whom medical therapy has failed. Indicators for early surgery may include increased leukocyte count (≥ 25 000 cells/mL) and lactate level (≥ 5 mmol/L). Both of these factors have been associated with high mortality (44).

Patients should be hospitalized for severe disease, complications, or other circumstances in which outpatient treatment is not feasible; dehydration and inability to tolerate oral medication; and signs of peritonitis, toxic megacolon, possible sepsis, or other complications. Admission should also be considered for patients who have indicators or risk factors for severe disease, such as elevated creatinine levels, leukemoid reaction, or advanced age.

Admission to intensive care is necessary for patients with severe disease and an unstable clinical condition, such as septic shock, toxic megacolon, peritonitis, or severe dehydration with hypotension or end-organ dysfunction.

Surgery is required for patients with colonic perforation, and those with toxic megacolon, acute abdomen, or septic shock due to C difficile disease (especially those with an elevated lactate level) may also benefit. Surgery may be useful in a patient without toxic megacolon in whom all medical therapies have failed. The currently recommended procedure for C difficile infection is subtotal colectomy. An alternative, less-invasive procedure that preserves the colon is loop ileostomy with antegrade vancomycin lavage (6), but more data are needed.

Recurrent symptoms after an apparent response to initial treatment may either be relapse or infection from a different strain; regardless, diagnosis and management are the same (Figure). The updated IDSA/SHEA guidelines discuss treatment options for the first recurrent episode depending on what was used to treat the incident episode (6). The first recurrence can be treated with a 10-day course of oral vancomycin if metronidazole was used for the incident episode; a tapered, pulseddose regimen of oral vancomycin if a standard 10-day course was used for the incident episode; or a 10-day course of fidaxomicin. Data are insufficient on the utility of extending therapy beyond the recommended duration or restarting C difficile infection therapy prophylactically in patients who require antibiotic treatment for another infection.

Subsequent recurrences may also be treated with oral vancomycin as a tapered, pulsed-dose regimen; a standard course of oral vancomycin followed by rifaximin; or a standard course of fidaxomicin. However, the evidence for these choices is limited.

Fecal microbiota transplantation (FMT) is recommended for patients with at least 2 prior recurrences in whom appropriate therapy has failed. Successful treatment of refractory infection by instillation of stool from healthy donors has been reported in individual cases and small case series (55, 56). From 2013 to 2016, at least 5 randomized controlled trials were published that compared FMT with vancomycin, autologous FMT, frozen versus fresh stool, or administration via colonoscopy versus nasogastric tube (57–61). The reported efficacy of FMT is lower in most randomized studies than in nonrandomized reports, but this may be caused by several factors, including patient selection and prior antibiotic treatment. Current data suggest that FMT is safe in the short term, and most of the associated mild to moderate adverse events are self-limited. Reported infectious complications to date have been rare, but the potential long-term infectious and noninfectious complications of FMT are unknown.

Bezlotoxumab, a human monoclonal antibody that binds to C difficile toxin B, was approved for use in the United States in 2016 to reduce risk for recurrence in high-risk persons aged 18 years or older who are receiving treatment for C difficile infection (62). Bezlotoxumab is given intravenously and should be given only concurrently with antibiotic therapy for C difficile infection.

SHEA and IDSA have published updated clinical practice guidelines for C difficile infection for both adults and children (6). The Healthcare Infection Control Practices Advisory Committee of the Centers for Disease Control and Prevention publishes periodic guidelines on infection control (63).

The Centers for Disease Control and Prevention Web site has information on C difficile for clinicians and patients (www.cdc.gov/hai/organisms/cdiff/cdiff_infect.html).