A 76-Year-Old Man With Recurrent Clostridium difficile–Associated Diarrhea: Title and subTitle BreakReview of C difficile Infection

DR REYNOLDS: Mr S is a 76-year-old man with hypertension, hypertensive renal disease with cadaveric renal transplantation in 1988 and again in 1998, spinal stenosis, urinary retention due to the spinal stenosis and to benign prostatic hypertrophy (treated with an indwelling Foley catheter), and recurrent Clostridium difficile infection (CDI).

Mr S's current illness began when he was admitted to the hospital with a urinary tract infection. He received ceftriaxone and ciprofloxacin; after discharge, he developed diarrhea. After several weeks of diarrhea, his nephrologist treated him empirically for CDI with a 10-day course of metronidazole; he improved and his stools initially became formed. However, about a month after discharge, he was still having 4 to 5 loose stools daily along with poor appetite, weight loss, fever, chills, and weakness. He was readmitted to the hospital and a stool sample was found to be positive for C difficile toxin. He was placed on contact precautions and treated with oral vancomycin and again improved, but again relapsed and was treated with a long (6-8 weeks) course of oral vancomycin followed by a tapering dose. Four days after stopping the oral vancomycin, Mr S again began experiencing diarrhea. In the ensuing months, he had several hospitalizations for diarrhea; he also had another urinary tract infection and was treated first with cefpodoxime and then with ciprofloxacin. He was prescribed standing oral vancomycin and rifaximin together. This combination controlled his diarrhea at the time of his interview. He is currently having semiformed, soft stools once or twice a day. He is not experiencing any bleeding. His appetite is good. Mr S has had a 10- to 15-lb (4.5- to 6.75-kg) weight loss over the past 6 months and attributes this to his diarrhea.

Mr S also has arthritis of his left shoulder with severe shoulder pain, and he is mostly confined to a wheelchair because of his spinal stenosis and resultant weakness in his legs, although he can walk short distances with a walker. Further evaluation of his urinary problems during his recent hospitalization revealed a neurogenic bladder thought to be due to his spinal stenosis; the Foley catheter was removed and he has learned to use straight catheters. He had been scheduled for surgery for the spinal stenosis but it was postponed when he developed CDI.

Mr S's medications include prednisone, 4 mg daily; cyclosporine, 50 mg daily; and diltiazem, trazodone, atenolol, minoxidil, furosemide, long-acting nitrates, finasteride, and erythropoietin, along with the standing vancomycin and rifaximin.

Mr S is retired and lives with his wife and adult son. He does not smoke and he drinks only 1 alcoholic beverage per month.

On physical examination, Mr S is thin, weighing 135 lb (61 kg), and is 69 in (175 cm) tall (body mass index, 19.9). His blood pressure is 150/70 mm Hg and his heart rate is 68/min. No neurological or spinal examination was performed for the discussion, but his physical examination was otherwise remarkable only for his left lower quadrant renal allograft, which was tender and without bruits, and for 1+ pitting edema in his lower extremities.

Mr S's laboratory test results show a baseline creatinine level of 1.8 mg/dL and a serum urea nitrogen level of 68 mg/dL.

His stool studies were positive for C difficile toxin A during his initial hospitalization for diarrhea; subsequent samples obtained when symptoms have recurred have been negative. Mr S's most recent urine culture showed resolution of a previous Pseudomonas infection.

I would take the regimen of medicines prescribed by the doctor. Three or four days after finishing the regimen, I would get better. Then I would turn right around; I’d get the chills and start shaking, sometimes uncontrollably. I’d have to put the covers on, even in the summertime. I would take [acetaminophen/paracetamol], and it would bring the temperature down and the medicine would control the C difficile.

This illness has been very distressing. Sometimes I would be on the toilet and would think I’m done. I get about halfway back to the bedroom, then I have to go back again. Sometimes this would happen 3 times in a 15-minute period. Right now, it seems the pills are curing me, and I’m getting better. I’m not going as frequently, but I’m still going 4 or 5 times a day. I might wake up in the evening and go once or twice.

My back is giving me a lot of problems. I get a lot of pain when I try to move. I can't walk as well, and I have to use my walker. I’m really slow. I had a procedure done on my neck in November. And I was supposed to have the second procedure done on my spine, but we had to cancel that because of the C difficile.

I’d be kind of frightened if somebody told me I had to go back on antibiotics after this. I wouldn't wish this illness on anybody. C difficile is one of the most terrible things that I’ve been dealing with in my whole life, and I’ve dealt with a lot of things since my kidney transplant. I hope that this latest regimen of pills will cure me—because I’m going out of my mind.

How common is C difficile after an initial episode of antibiotic treatment and how are rates affected by age or immune status? How should the diagnosis be made and is empirical treatment an acceptable strategy? How common is recurrent infection and are these usually relapse or reinfection? What is the evidence for the most effective treatment for first and recurrent episodes? What should physicians do to prevent C difficile in terms of prescribing and infection control practices? What do you recommend for Mr S?

DR KELLY: Mr S is a 76-year-old retiree with a functioning renal transplant, benign prostatic hypertrophy, and spinal stenosis whose case illustrates the challenging problem of CDI. He received both ceftriaxone and ciprofloxacin during inpatient treatment for a urinary tract infection, developed diarrhea, and was treated with oral metronidazole for possible CDI. He subsequently experienced multiple episodes of recurrent CDI leading to hospital admissions and requiring repeated and prolonged treatments with oral vancomycin.

Clostridium difficile is a gram-positive, spore-forming, anaerobic bacillus first isolated in 1935 from the stool of a healthy neonate.¹ The finding that C difficile is an opportunistic pathogen was not made until the late 1970s, when the clinical entity of antibiotic-associated, C difficile–induced diarrhea and colitis was first characterized. Clostridium difficile infection is now by far the most common known cause of infectious diarrhea in hospital patients in North America and Europe, where both the incidence and severity of disease have increased alarmingly since 2000.²

Although C difficile is common in the general environment, only 1% to 4% of healthy adults carry this organism. Quiz Ref IDA normal colonic microflora confers “colonization resistance” against CDI.² Neonates, lacking an established intestinal microflora, frequently have C difficile colonization but, interestingly, seldom experience C difficile toxin–induced diarrhea or colitis, possibly because of a lack of toxin receptor expression in the immature gut.¹ In adults, loss of colonization resistance to CDI is most often the result of antimicrobial therapy. All antimicrobials are not equal in this regard, and the risk of CDI associated with different antimicrobials results from a complex combination of their luminal concentrations, their effects on the normal intestinal microflora, and their activity against C difficile.^{2 - 5} The ability of clindamycin to impair colonization resistance and induce CDI is well known.³ However, in current medical practice, second- and third-generation cephalosporins and fluoroquinolones are more frequently used to treat infections in hospital patients and, as in the case of Mr S, most frequently cause CDI.^{3 - 4} All clinical isolates of C difficile are resistant to cephalosporins, whereas the addition of fluoroquinolones to the short list of agents most likely to cause CDI is recent and parallels the development of high-level fluoroquinolone resistance in C difficile clinical isolates.³ Similarly, a clindamycin-resistant strain of C difficile was implicated in a previous outbreak that was eventually controlled by curtailing clindamycin use.⁵

Most cases of CDI present during or shortly after antimicrobial use.^{3 ,6 - 7} However, it is important to be aware that disease onset can be delayed for 2 or, in rare cases, even 3 months, as these late presentations can lead to misdiagnosis. Quiz Ref IDAlthough the majority of CDI results from antimicrobial use, other events may impair colonization resistance, including bowel preparation for colonoscopy or surgery, cytotoxic chemotherapy (presumably because of its antimicrobial effects), and colitis caused by inflammatory bowel disease where colonization by C difficile is common and may exacerbate the underlying inflammatory bowel disease.^{8 - 10} Furthermore, cases of community-acquired CDI that develop in the absence of any evident antimicrobial exposure or other risk factors are increasingly described.^{7 ,11}

Pathogenic strains of C difficile produce 2 large (approximately 300 kDa) protein exotoxins, toxin A and toxin B.¹² These toxins share amino acid sequence and structural and functional homologies but are antigenically distinct. Hence, enzyme immunoassays (EIAs) directed against one toxin cannot recognize the other. This proved important in outbreaks of CDI caused by strains that lacked toxin A and produced toxin B alone.¹³ Commonly used tests that detected toxin A by EIA yielded false-negative results in a majority of patients in those outbreaks.^{13 - 14} Most, but not all, currently used EIAs detect both toxins. Clostridium difficile infection caused by toxin A–negative, toxin B–positive strains has dispelled the long-held belief that intestinal disease is caused by toxin A alone. In fact, both toxins are highly injurious to the human colon, causing enterocyte cell death, a marked acute inflammatory response, and severe mucosal injury.¹⁵

The risk of developing CDI as a complication of antimicrobial use depends on 3 groups of factors: (1) impairment of colonization resistance, as already discussed; (2) risk of exposure to toxigenic C difficile or its spores; and (3) host health and immune status. Mr S developed CDI after his hospitalization for a urinary tract infection. The risk of developing CDI during or after antimicrobial use in the hospital is far greater than after the use of the same agents in the community.^{7 ,16 - 17} In community-based studies, CDI complicated fewer than 1 in 5000 antimicrobial prescriptions, leading to 7.6 CDI cases per 100 000 person-years.^{16 - 17} In stark contrast, CDI complicates 0.5% to 1% of hospital admissions.^{18 - 19} Quiz Ref IDThe environment of hospitals and other health care facilities is contaminated by C difficile spores.²⁰ These spores are resistant to drying, temperature fluxes, and many antiseptic solutions, yet remain viable for long periods in the environment and cause CDI when ingested, creating a formidable infection control challenge. Host factors also contribute to the concentration of CDI in hospitals. Advancing age as well as acute and comorbid diseases are prominent risk factors for CDI and, as in Mr S's case, the disorder disproportionately affects older and frail patients.^{2 - 4 ,16 ,18 - 24} Conversely, a robust immune response to C difficile and its toxins is associated with less severe disease.^{16 ,25} Many individuals colonized by pathogenic strains of C difficile remain asymptomatic, and this carrier state has been associated with high circulating titers of IgG antitoxin.^{16 ,20} Thus, Mr S exhibited many of the CDI risk factors discussed herein, including antimicrobial use, hospital admission, advancing age, severe underlying disease, and impairment of immune function.

Since 2000, reported cases of CDI in US hospitalized patients have steadily increased.¹⁹ According to the US National Hospital Discharge Survey, 31 CDI-associated hospitalizations were reported per 100 000 population in 1996 compared with 61 per 100 000 in 2003 and 84 per 100 000 in 2005 (the latest year for which data are available). This represents an approximate 25% annual increase each year since 2000. Other countries, notably Canada and European Union members, have experienced similar increases.^{4 ,22 - 24 ,26} Compounding these alarming increases in reported disease incidence is a parallel increase in disease severity and, hence, case-fatality rates.^{22 - 24 ,26} In England and Wales, for example, there were 975 deaths associated with CDI in 1999, 6480 in 2006, and 8324 in 2007, a 750% increase over 7 years.²⁶ In addition to the increased incidence and severity of endemic CDI, there have also been dramatic outbreaks of nosocomial disease. Perhaps the best-characterized outbreak occurred in Quebec in 2003, when CDI cases increased almost 4-fold and deaths associated with CDI increased more than 10-fold compared with the previous year.^{22 - 24} Most isolates (80%) from this outbreak belonged to a single strain type (North American pulsed-field type 1 [NAP-1]), and this same strain has since been associated with outbreaks of severe disease across the United States and in many European countries. The NAP-1 strain has acquired high-level resistance to fluoroquinolones, which may provide a survival advantage in hospital settings.^{23 - 24} The NAP-1 strain also carries deletion mutations in a regulatory gene (tcdC) believed to inhibit toxin production. A resulting increase in toxin transcription and synthesis may contribute to the increased disease severity evident in CDI outbreaks caused by NAP-1.^{23 - 24 ,27 - 28}

Clostridium difficile infection is primarily a clinical diagnosis based on the onset of diarrhea during or after antimicrobial use, often associated with abdominal discomfort, fever, and polymorphonuclear leukocytosis.^{2 ,29} Initially, Mr S was treated empirically without a stool test. Typically, though, confirmation of CDI should be obtained by demonstrating the presence of C difficile or its toxins in a stool sample. Most clinical laboratories use EIA to detect toxin A and toxin B.²⁹ These assays have the advantages of being relatively quick and easy to perform so that same-day results are feasible. However, they are often insensitive (66% sensitivity in 1 study), leading to frequent false-negative results.³⁰ Hence, negative test results do not override a strong clinical suspicion for CDI, as in the case of Mr S, who had several negative stool test results after initial treatment. Testing of multiple samples has been used to try to overcome a lack of sensitivity of EIA, but if 2 samples test negative and there remains a strong clinical indication to confirm or refute CDI, a more sensitive testing method should be used. More accurate stool assays are available, including the tissue culture cytotoxicity assay, often used as a gold standard test, that identifies toxin effects at a cellular level using living cells in tissue culture and can detect toxin concentrations that are 100 to 1000 times lower than those detected by EIA.^{29 - 30} However, these more accurate techniques require greater laboratory resources and personnel training, take longer (24-48 hours) to provide a result, and are more than twice as expensive.^{29 - 31} Sensitive molecular techniques have been used to identify C difficile genetic material in stool samples and are being adapted and simplified for routine clinical laboratory use.³¹ Repeat testing to monitor response is not indicated; a patient's clinical course should supersede the test results.³² Many patients treated successfully for CDI will continue to shed C difficile and toxins A and B in their stool for some weeks despite complete clinical recovery.³² Such symptomless carriage of C difficile does not warrant treatment or barrier precautions but emphasizes the importance of universal infection control measures.^{2 ,16 ,20 ,29 ,32}

Because CDI is a common and severe nosocomial infectious disease, measures to reduce its incidence are essential. Antimicrobial use is the main risk factor for CDI, and prudent use of antimicrobials is the first step in risk reduction.³³ Probiotics have been used by antibiotic recipients in an attempt to conserve colonization resistance and avoid both CDI and simple antibiotic-associated diarrhea.^{34 - 35} Quiz Ref IDProbiotic preparations containing viable lactobacilli or Saccharomyces species taken in conjunction with antimicrobials can halve the incidence of antibiotic-associated diarrhea.^{34 - 35} The ability of probiotics to protect against CDI is less clear. However, a randomized placebo-controlled trial in 135 hospitalized patients, of whom 113 were followed up, found that a lactobacillus-containing probiotic mixture reduced the incidence of nosocomial CDI in antimicrobial recipients from 17% to 0%³⁵ (risk reduction, 17%; 95% confidence interval, 7%-27%), but further studies are needed to confirm or refute this finding. In health care facilities, barrier precautions should be used for patients with CDI and diarrhea and a single room provided if possible. Given that symptomless carriage of C difficile is common in hospital patients, universal precautions including hand-washing before and after patient contact are essential in reducing infection transmision.^{16 ,20} In high-risk situations, hand hygiene using soap and running water is preferred to alcohol-based hand cleansers for removing C difficile spores.^{36 - 37} Environmental cleaning is also important to reduce the burden of spores and, in an outbreak setting, sporicidal agents such as chlorine-containing disinfectants should be used.^{36 - 37} Recent reports of rare cases of CDI in healthy household contacts have emphasized the need for infection control measures after discharge, especially for patients with ongoing diarrhea.⁷ These include hand-washing with soap and water after using the bathroom and before eating, cleaning kitchen work surfaces and bathrooms with a mixture of 1 part household bleach to 10 parts water, and, until diarrhea ceases, using a dedicated bathroom or cleaning the toilet with the bleach-and-water mixture after each use.^{36 - 38}

Treatment of CDI can be stratified according to the severity of illness (Table).^{6 ,39 - 41} Quiz Ref IDMild CDI in an otherwise healthy individual can be managed by discontinuing the antimicrobial agent responsible for causing the infection, requesting stool testing to confirm CDI, and monitoring the patient's progress.⁴³ In the absence of all antimicrobial treatment, approximately 25% of patients have resolution of diarrhea within 48 hours and, importantly, recurrent CDI is unlikely to occur.³⁹ Oral metronidazole, 500 mg every 6 hours for 10 to 14 days, is the treatment of choice for moderately severe CDI, for mild CDI in frail patients as in the case of Mr S's initial presentation, for mild CDI in patients whose inciting antimicrobials cannot be discontinued safely because of intercurrent infection, and for mild CDI that does not respond within 48 hours to conservative management.^{43 - 44} In a large randomized controlled trial, the response rate to metronidazole, as determined by resolution of diarrhea at day 10 of therapy, was 77% in mild and moderate CDI.⁴⁰

Oral vancomycin, 125 mg every 6 hours for 10 to 14 days, is also effective in mild to moderate CDI, with a response rate (82%) similar to metronidazole.⁴⁰ Oral vancomycin is the second-line agent in this situation because of greater cost and lingering concerns regarding encouraging vancomycin resistance, especially in a nosocomial setting.⁴⁴ Mr S developed severe CDI after an initial course of metronidazole gave only temporary improvement. The indications for using oral vancomycin to treat CDI include severe disease, metronidazole contraindications or intolerance, and inadequate response to metronidazole treatment.^{43 - 44} Oral vancomycin has been recommended as the optimal first-line treatment for severe CDI for more than 2 decades. However, evidence from well-designed clinical trials to support this recommendation became available only recently. In 2 randomized controlled trials, individuals with mild to moderate CDI responded similarly to metronidazole compared with vancomycin (as described herein).^{6 ,40} However, in severe CDI, both studies demonstrated a significant advantage of vancomycin over metronidazole, with response rates of 97% vs 76%, respectively, in the first study (P = .02) and 85% vs 65% (P = .04) in the second.^{6 ,40} It is clear, therefore, that oral vancomycin should be the first-line agent in severe CDI. It is less clear how severe CDI is best defined to guide clinicians in their choice of therapy. The variables used to define severe disease in the study by Zar et al⁶ consisted of intensive care unit admission or pseudomembranous colitis at endoscopy or any 2 of the following: age older than 60 years, temperature higher than 101°F (38.4°C), serum albumin less than 2.5 g/dL, and peripheral white blood cell count greater than 15 000/μL. This grading system was established empirically but, nonetheless, was able to prospectively identify patients with CDI at greater risk of failure of metronidazole treatment.⁶

Refractory CDI can be defined as disease that is unresponsive to vancomycin and/or metronidazole.^{43 ,45 - 46} Fulminant CDI can be defined as disease that progresses rapidly to cause systemic manifestations, including marked leukocytosis, hypotension, renal failure, and anasarca.^{47 - 51} In practice, fulminant and refractory disease often overlap, their management is challenging, and their incidence is rising. Fulminant CDI may present with abdominal pain and tenderness, colonic distension, and signs of sepsis including fever, leukocytosis, hypotension, and lactic acidosis.^{47 - 51} Diarrhea may be absent or minimal due to ileus, and this factor may obscure a diagnosis of CDI and instead suggest diagnoses such as ischemic bowel or a perforated viscus.⁴⁹ In this situation, gentle flexible sigmoidoscopy or colonoscopy (without attempting a full bowel preparation and with minimal air insufflation) can be invaluable in immediately identifying pseudomembranous colitis and thereby informing medical and surgical management.^{47 ,49 - 50}

Oral (or intragastric) vancomycin, 500 mg every 6 hours, is the recommended first-line treatment for fulminant or refractory CDI.^{2 ,6 ,40 ,43 - 44 ,48} In patients with an ileus or with megacolon, intravenous metronidazole, 500 mg every 6 hours, should be added. In patients with a complete ileus, for whom oral or intragastric medications are likely to be ineffective, vancomycin can be administered rectally (500 mg in 100 mL of normal saline every 6 hours). Surgical intervention in the form of subtotal colectomy can be lifesaving in fulminant or refractory CDI. However, the optimal timing of surgery is difficult to establish. Patients with severe CDI are typically elderly, critically ill, and at high risk of surgical and anesthetic complications, making them poor surgical candidates. Conversely, delaying surgery until systemic inflammatory response syndrome with organ failure and hypotension have developed almost invariably leads to a fatal outcome with or without colectomy.⁵² Clinical research groups are trying to identify objective parameters that will help identify those patients most likely to benefit from surgery.⁴⁸ For now, the best course is to obtain a surgical consultation as soon as fulminant or refractory disease is suspected and to base a decision to operate on close monitoring of the patient's clinical course. Even with surgery, in-hospital mortality among patients undergoing subtotal colectomy for severe CDI is high (42%).⁵²

Normal pooled intravenous immunoglobulin (IVIG) has been used in this situation in an attempt to avert surgery or death through passive immunotherapy aimed at C difficile toxins.^{50 - 51} It is known that individuals with low concentrations of circulating IgG antitoxin are susceptible to symptomatic and prolonged CDI.^{16 ,25} Conversely, those with high antitoxin titers are more likely to recover quickly from CDI or to be symptomless carriers. Because more than 50% of the population has detectable serum antitoxin A and B IgG, pooled normal IVIG has toxin-neutralizing activity.^{50 ,53} Thus, IVIG (400 mg/kg) may be a worthwhile intervention in fulminant or refractory CDI, despite that it is expensive and in short supply and its efficacy is uncertain, supported only by anecdotal reports and case series.^{50 - 51}

The most common management problem in CDI is disease recurrence, as was the case with Mr S.⁵⁴ Recurrence typically develops 1 to 2 weeks after stopping metronidazole or vancomycin but can be delayed for up to 12 weeks. Recurrence rates following treatment with metronidazole or vancomycin are similar at approximately 20%.⁵⁵ Indeed, recurrent CDI is so common that all patients should be forewarned that this complication can occur after their course of therapy is complete, thereby facilitating prompt diagnosis and retreatment and hopefully avoiding hospital readmission or other complications. Risk factors for recurrent CDI include a history of recurrence, increasing age, use of additional antimicrobials, and an inadequate protective immune response to C difficile toxins.²⁵ Patients with 1 prior episode of recurrent CDI have a greater than 40% risk of an additional recurrence, whereas those with 2 or more previous episodes have a greater than 60% risk.^{41 ,56}

Recurrent CDI is rarely if ever a result of antimicrobial resistance to metronidazole or to vancomycin. Instead, treatment with these agents perpetuates the impairment of colonization resistance that underlies CDI. Continued exposure to C difficile in the hospital or home environment can lead to reinfection when vancomycin and metronidazole concentrations have decreased. Persisting intestinal infection is an alternative mechanism for disease relapse, but strain-typing studies indicate that reinfection with a different strain of the organism is common.⁵⁷

Management of a first recurrence of CDI is identical to a primary episode.^{2 ,43 ,54} Repeat stool testing is useful to confirm a diagnosis. However, it can be difficult to demonstrate the presence of C difficile during recurrence, especially if symptoms and signs of CDI develop shortly after discontinuing metronidazole or vancomycin therapy. If the symptoms, signs, and timing are consistent with recurrent CDI, as they were for Mr S, treatment should be instituted even in the face of negative stool test results. The treatment approach is stratified according to disease severity in an identical fashion to that described for primary therapy (Table).

The optimal management of patients with multiple recurrences of CDI is less clear. One study reported on the outcomes of nonrandomized treatments using a wide variety of physician-selected antibiotic regimens in 163 patients.⁴¹ In that study, the overall subsequent recurrence rate was 45%. The most effective regimens were those using a prolonged tapering or pulse-dosing regimen of vancomycin, in which recurrence rates were 31% (P = .01) and 14% (P = .02), respectively. Based on this study, I use a tapering and pulse-dosing regimen of vancomycin for treatment of a second recurrence of CDI (Table).⁵⁷ Vancomycin, 125 mg every 6 hours orally, is given for 14 days, followed by a tapering dose of vancomycin (125 mg every 12 hours for 7 days, then 125 mg daily for a further 7 days), followed in turn by pulse-dosed vancomycin (125 mg once every 2 days for 4 doses, then once every 3 days for 5 doses).^{41 ,58} Prolonged metronidazole therapy is not advocated because of potential adverse effects, including peripheral neuropathy.

A wide variety of interventions have been reported for management of multiple recurrences of CDI, including probiotics, antibiotics, toxin binders, and immune-based treatments, but none have proven efficacy. One probiotic, Saccharomyces boulardii, was effective in preventing recurrent CDI in patients with prior recurrences.⁵⁶ Saccharomyces boulardii (1 g daily for 4 weeks) or placebo was administered to 60 patients during and after standard treatment with metronidazole or vancomycin. The subsequent recurrence rate in the S boulardii group was 35% compared with 65% with placebo (P = .04). However, a larger follow-up study did not show a significant overall benefit of S boulardii over placebo.⁵⁹ Results with other probiotics have been similarly inconsistent.³⁴ Thus, despite the obvious attractions of using S boulardii or Lactobacilli species to augment colonization resistance and prevent recurrent CDI, their therapeutic efficacy remains uncertain. Two other imaginative probiotic approaches warrant mention. Fecal transplantation uses a small amount of fresh feces (from a healthy donor), suspended in saline, filtered, and administered through a nasogastric tube. A case series in 18 patients found a greater than 95% success rate in preventing recurrent CDI, but no randomized controlled trials have been performed.⁶⁰ Nontoxigenic C difficile is nonpathogenic and may occupy the same microenvironmental niche as toxigenic strains. Studies are under way to identify a suitable nontoxinogenic strain that can be produced for human use as a targeted probiotic to block colonization by pathogenic C difficile.⁶¹

Both rifampin and rifaximin have been used in conjunction with vancomycin to treat recurrent CDI. Both agents have in vitro activity against C difficile, but the pharmacokinetics of rifaximin are preferable because it is not absorbed. In 1 study, rifaximin, 400 to 800 mg daily, was administered for 14 days after vancomycin was discontinued in patients with multiple (4-8) CDI recurrences.⁶² Seven of 8 patients had no further recurrences. Both cholestyramine and colestipol have been used alone or as adjuvants to antibiotics to treat CDI, but their efficacy is poor.⁶³ A novel toxin-binding resin (tolevamer) was tested recently for treatment of CDI but was less effective than either metronidazole or vancomycin.⁴⁰ Individuals with recurrent CDI lack protective immunity to C difficile toxins.²⁵ Several case series support the use of IVIG as passive immunotherapy in patients with multiple recurrences of CDI.^{64 - 65} An active vaccine was successfully used in 1 small case series, but a C difficile vaccine is not currently available for use outside of clinical trials.^{66 - 67}

My current approach to managing recurrent CDI in a patient such as Mr S is to prescribe metronidazole or vancomycin, depending on disease severity, for a standard 14-day course for the first recurrence. For a second recurrence, optimal treatment has not been established definitively, but I use the tapering and pulse-dosing regimen proposed by Tedesco et al (Table).⁵⁸ Adding a probiotic containing either Lactobacillus or S boulardii during the final 4 weeks of vancomycin therapy and for an additional 4 weeks thereafter is also an option.^{34 ,56 ,59} If that too fails, I treat the recurrent symptoms with vancomycin for at least 14 days, discontinue vancomycin, and administer rifaximin, 200 mg twice daily for 14 days.⁶² In my experience, these steps are successful in most patients, but for the few remaining patients with recalcitrant CDI, other options, such as prolonged vancomycin therapy, IVIG, or fecal transplantation, can be considered.

QUESTION: Would you comment on the desirability of a prophylactic immunization against the toxin on all elderly sick people?

DR KELLY: One of my conflict-of-interest statements has to do with my work with Acambis, a company in Cambridge, Massachusetts, that has developed a toxoid-based C difficile vaccine. The vaccine induced strong immune responses in healthy adult volunteers.^{66 - 70} If a safe and effective vaccine can be developed, then those aged 65 or greater and infirm patients would be obvious target populations, as these are the most likely to have CDI. Patients with recurrent CDI may also benefit, although one might imagine that they would be the most difficult people to immunize, because as I’ve shown, they fail to mount a protective immune response during natural infection. Nonetheless, a small number of patients with recurrent disease were successfully vaccinated in an open-label, uncontrolled study.⁶⁷ Passive immunization using IVIG, human antitoxin monoclonal antibodies, or a hyperimmune globulin has also been considered and may be preferable to an active vaccine in some circumstances; eg, when rapid protection is needed or in immunocompromised recipients.^{50 ,64 - 65 ,68 ,70 - 71}

QUESTION: A year or two ago a medical resident became ill with CDI. He had worked at another hospital during a C difficile epidemic; he had not taken any antibiotics. Is nosocomial acquisition of infection possible in those who have not received antibiotics?

DR KELLY: Although most cases of CDI follow antibiotic use, the disease can occur in the absence of antimicrobials. I already mentioned special circumstances such as bowel preparation or inflammatory bowel disease that can lead to CDI. Some recent studies suggest that community-acquired CDI may be more likely to develop in the absence of any evident antimicrobial exposure.^{7 ,11} In December 2005, a Morbidity and Mortality Weekly Report article referred to cases of CDI that occurred in individuals considered to be at low risk for the disease.⁷ This included cases of severe CDI in young, previously healthy individuals, some of whom did not have a history of recent antimicrobial use. As in the case you describe, there was often a history of direct contact with a patient with CDI providing a likely source of infection. So I would say that it's unusual for a healthy health care worker to develop CDI in the absence of antimicrobial use, but it's not impossible.

QUESTION: Does C difficile ever cause traveler's diarrhea in people who have not taken prophylactic antibiotics?

DR KELLY: Clostridium difficile infection can occur in travelers as it can occur in the general population but is not considered to be a common cause of traveler's diarrhea. Of course, travelers may be at increased risk of CDI if they take antimicrobials either as prophylaxis or as treatment for traveler's diarrhea. That said, their substantially decreased risk arises from the facts that they are generally healthy and seldom visit health care institutions.

QUESTION: A lot of patients who are diagnosed with CDI are taking antibiotics, and if they have active ongoing infection at the same time, they may require the antibiotics, either for a short-term course or possibly even for a prolonged course. Would the need to continue antiobiotic therapy for an underlying infection change your treatment strategy for CDI?

DR KELLY: The need for continued antimicrobial treatment for other infections affects treatment for CDI in several ways. First, conservative management of mild disease is not suited to situations in which antimicrobials that exacerbate CDI cannot be discontinued safely. Second, if antimicrobial therapy must be continued, it is important to review the treatment regimen and, whenever possible, adopt antimicrobials less likely to facilitate CDI. Third, the risk of recurrent CDI is increased by additional antimicrobial use during or after treatment of CDI with metronidazole or vancomycin.²⁵

QUESTION: How do you suppose probiotics play a role? Is it simple replacement, or is there some kind of cross-talk between what's in the probiotic and the C difficile bacteria? It seems to me that there's an important possibility there.

DR KELLY: You raise an excellent point, and there have been many studies aimed toward identifying how specific probiotics might work. The probiotic agent that I have studied most is S boulardii, which is very similar to brewer's yeast, Saccharomyces cerevisiae. A number of mechanisms have been shown experimentally for S boulardii, including interference with toxin binding to its receptors, stimulation of an immune response with increased IgA antitoxin production, and alteration of cell signaling in enterocytes with a dampening down of the inflammatory response to C difficile toxins. Other effects such as competition with C difficile for nutrients or blocking binding to receptors needed by C difficile for colonization are also quite possible.^{72 - 76} So, a multifaceted effect is likely, rather than a single distinct mechanism. However, the truth is that very little is understood about the mechanisms of colonization resistance against C difficile and how those protective effects are destroyed by antimicrobial treatments.