Clostridium difficile–Associated Disease: An Emerging Threat to Patient Safety: Insights from the Society of Infectious Diseases Pharmacists

Clostridium difficile–Associated Disease: An Emerging Threat to Patient Safety

Insights from the Society of Infectious Diseases Pharmacists

Robert C. Owens, Jr., Pharm.D.

A formerly infrequently isolated strain of Clostridium difficile known asBI/NAP1 has resulted in geographically diverse outbreaks of C.difficile–associated disease. Such rapid dissemination and distribution of anoutbreak strain of C. difficile are unprecedented, with many regions acrossNorth America, as well as several countries in Europe, being affected, all insuch a short period of time. Also of note is that nontraditional hosts (e.g.,otherwise healthy, noninstitutionalized persons residing in the community,some without antimicrobial exposure) have been reported to have severedisease. Data suggest that certain virulence characteristics may be responsiblefor more severe clinical presentations and poor outcomes. These factors (e.g.,hypertoxin production, hypersporulation, antimicrobial resistance) possessedby a previously uncommon strain of C. difficile, in conjunction with particularhost and environmental factors, may have precipitated the now widespreadestablishment of this pathogen. Antimicrobial intervention has traditionallybeen a mainstay of combating C. difficile–associated disease. Efforts to combatBI/NAP1 should include good antimicrobial stewardship in addition toeffective infection control and environmental intervention.Key Words: Clostridium difficile, BI/NAP1, antimicrobial stewardship, adverseevents.(Pharmacotherapy 2006;26(3):299–311)

The organism first termed Bacillus difficilis inthe mid-1930s, and now known as Clostridiumdifficile, continues to evolve into a formidablepathogen. Clostridium difficile–associated disease(CDAD) has traditionally ranged in presentationfrom self-limiting diarrhea to a fulminant, life-threatening disease. Of considerable interest isthe observation that CDAD appears to be increasingworldwide in both incidence and severity ofdisease.1, 2 Reasons for this are not entirely clear,are most likely very complex and interwoven,

and may be due in part to the emergence of apreviously uncommon strain of C. difficile.3 Tobetter understand the implications of recentfindings, it is fundamental to appreciate the factthat the risks for CDAD are complex and consistof exposure to toxigenic strains of the organism,4

previous use of any antimicrobial agent5 perhapscompounded by the previous use of antimicrobialsthat are specifically inactive or have borderlineactivity against C. difficile,6 exposure to gastricacid suppressants,7 poor host serum immuno-globulin levels,8 advanced age,9 and severity ofunderlying illness of the host.9 The sequence ofthese events is important for acquiring CDAD:first is exposure to antimicrobials, second isexposure to toxigenic strains of C. difficile, andthird is the presence of additional factors justmentioned.9

From the Antimicrobial Stewardship Program,Departments of Pharmacy Services and Infectious Diseases,Maine Medical Center, Portland, Maine; and the College ofMedicine, University of Vermont, Burlington, Vermont.

Address reprint requests to Robert C. Owens, Jr.,Pharm.D., Department of Pharmacy Services, Maine MedicalCenter, 22 Bramhall Street, Portland, ME 04102; e-mail:[email protected].

PHARMACOTHERAPY Volume 26, Number 3, 2006

The organism responsible for the recentlyreported outbreaks in North America and Europehas been labeled BI/NAP1 and is distinguishablefrom previously identified outbreak (J-type)strains from the late 1980s and early 1990s.10

The BI/NAP1 strains of C. difficile producegreater amounts of toxins A (TcdA) and B(TcdB),11 carry an additional clostridial toxintermed a binary toxin,3 belong to toxinotype III,3

have an increased sporulation capacity,12 and arefluoroquinolone resistant.3 These characteristicsdifferentiate BI/NAP1 from traditional C. difficilestrains. It should be noted that whereas thepredominant outbreak strain is BI/NAP1,additional non-BI/NAP1 strains have beenidentified in outbreak areas, to a lesser degree,and contain similar virulence characteristics asthe current outbreak strains.

In addition to identified changes in thepathogen, certain augmenting host andenvironmental factors may be to blame. Theseinclude the fact that traditional hand washingwith soap and water in hospitals has shifted tothe preferential use of alcohol-based handhygiene products in many institutional settingsover the last 5–10 years (alcohol is ineffectiveagainst C. difficile spores).13 In addition,traditional hospital cleaning agents (quaternaryammonium–based products) do not kill C.difficile spores. In fact, quaternary ammonium–based cleaning agents have been shown tosubstantially increase the sporulation capacity ofC. difficile.12 The greater severity of diseaseattributable to BI/NAP1 in combination withthese environmental risk factors may becontributing to a cycle that has led to C. difficileoutbreaks on at least two continents.

Both clinicians and health care administratorsalike are concerned with the changes in the C.difficile landscape. In addition to the notableconsequences of CDAD, this disease is burdeningalready taxed health care resources. Forexample, patients with CDAD have been shownto incur 54% greater costs and remainhospitalized for an average of 3.6 days longerthan patients who do not develop CDAD duringhospitalization.14 Furthermore, because ofchanges in environmental cleaning policiesrequired during outbreak scenarios, additionalexpenditures are required to equip environmentalservices departments to prepare and handle 10%sodium hypochlorite (bleach) solutions or topurchase expensive commercially premixedhypochlorite-based products. Moreover, patientswith CDAD in acute care facilities need to be

isolated or cohorted, which may cause significantfinancial challenges related to bed management.This was exemplified best by ministry officials inQuebec who directed $20 million into hospitalinfection control resources and infrastructurechanges to contain CDAD after more than 100people died over a short period of time.15 Thisreview provides an update regarding what hasbeen learned about the emerging epidemic strainof C. difficile, followed by a practical review of oneinstitution’s challenges related to its presence.

The Changing Epidemiology of Clostridiumdifficile–Associated Disease

Incidence

The National Nosocomial Infections Surveillance(NNIS) system has reported an increasing trendin CDAD rates over the last 2 decades.2 Inaddition, one group of authors reported a 26%relative increase in 2001 over 1998–2000 in theannual proportion of discharges from nonfederalhospitals in the United States, with CDAD listedas any discharge diagnosis.16 Although a recentstudy performed at Prevention EpicenterHospitals suggested no increase in CDAD ratesbetween 2000 and 2003,17 their meanhospitalwide rates of 12.1 cases/10,000 patient-days or 7.4/1000 hospital admissions was 2–6times greater than that reported from the NNISsystem hospitals from 1987–20012 and teachinghospitals during the same period.18 Regardless ofnational rates that appear to be increasing,monitoring rates locally is of utmost importance.

Severity

In 2000, reports from the University ofPittsburgh noted increased CDAD incidence andseverity as indicated by a doubling of diseaserates along with an increased number of colec-tomies and deaths.19–21 In Canada, most notablythe Sherbrooke region in Quebec, a CDADoutbreak attributable to BI/NAP1 resulted inexcess mortality. The authors of a matched case-control study reported that 30-day mortality rateswere 3-fold higher if CDAD complicated apatient’s admission.22 In addition, prolongedhospitalizations, relapses, and readmissionscharacterized infection with C. difficile duringthis outbreak. Recently, reports of severe CDADhave surfaced in previously low-risk populations(otherwise healthy persons living in thecommunity, some without a history of antibioticexposure) in a four-state region.23

300

CLOSTRIDIUM DIFFICILE–ASSOCIATED DISEASE Owens

The Pathogen: BI/NAP1

The strain BI/NAP1 has contributed to outbreaksin several regions of North America, the UnitedKingdom, and the Netherlands. It was identifiedand referred to by polymerase chain reaction asribotype 27 (CD027), by pulsed-field gel electro-phoresis as North American pulse-field type 1(NAP1), by restriction-endonuclease analysis(REA) as group BI, and by toxinotyping astoxinotype III.3 Various publications may refer tothe strain by different terms—BI/NAP1 orNAP1/027—but they are referring to the samestrain. This current outbreak strain is distinctfrom that causing previous outbreaks during thelate 1980s and early 1990s, known as the J strain(REA type J7/9).10 As previously mentioned,notable virulence characteristics associated withthe BI/NAP1 strain include increased toxinproduction (TcdA and TcdB),11 the presence of abinary toxin,3 altered antimicrobial resistancepatterns (fluoroquinolone resistance),3 andincreased sporulation capacity.12

Increased Toxin Production

The TcdA toxin has been termed “the entero-toxin,” as it is responsible for the expression ofdiarrhea and colonic inflammation, and TcdB“the cytotoxin,” which is responsible foractinomorphic changes in tissue culture cells.24

In all but rare cases, both toxins are expressed inpatients with clinical disease; however, strainsthat are TcdA negative but TcdB positive havebeen identified.25 The toxins TcdA (308 kD) andTcdB (270 kD) are among the largest toxins to beharbored by bacteria and are encoded on achromosome within the pathogenicity locus(PaLoc) of the organism. Also located within the

PaLoc are regulatory genes such as tcdC, which isa downstream negative regulatory gene thatcontrols the expression of TcdA and TcdB. Allidentified BI/NAP1 strains contain an 18–basepair tcdC gene deletion3 that is thought to beresponsible for the accelerated kinetics of toxinproduction.11 Traditionally, TcdA and TcdB areproduced most efficiently when the organism isin the stationary growth phase.26 In contrast,BI/NAP1 strains produce 16 times more TcdA and23 times more TcdB, and studies indicate thatmost of this production occurs in the logarithmicgrowth phase (Figure 1).11

Binary Toxin

In addition to the well-characterized TcdA andTcdB toxins, a formerly uncommon binary toxinhas been identified in all of the BI/NAP1 isolates(previously found in approximately 6% of clinicalisolates).3 The structure and function of thistoxin are similar to those of other binary toxins,such as iota toxin found in Clostridium perfringens.Although patients infected with binarytoxin–positive strains of C. difficile trendedtoward having greater disease severity,27 TcdA-and TcdB-negative but binary toxin–positivestrains of C. difficile have been shown to benonpathogenic in classic nonclinical models ofinfection.28 Thus, the pathogenic role of binarytoxin in BI/NAP1 is unknown.

Toxinotyping

Clostridium difficile strains can also beclassified by toxinotyping studies. Subtlesequence variations within the PaLoc accountsfor the various toxinotypes of C. difficile, of

301

Figure 1. Production of toxin A (A) and toxin B (B) by NAP1/027 (BI/NAP1) strains of Clostridium difficile. (From reference 11with permission.)

0

250

500

750

1000

1250

0 24 48 72 Hours

Log phase Stationary phase

NAP1/027 (toxinotype III)

Control strains (toxinotype 0)

Toxi

nA

Con

cent

ratio

n(n

g/m

l)

A

00 24 48 72 Hours

50

100

150

200

250

Log phase Stationary phase

Toxi

nB

Con

cent

ratio

n(n

g/m

l)

BNAP1/027 (toxinotype III)Control strains (toxinotype 0)


which there have been reported to be at least 22different types.26 Toxinotype III, to whichBI/NAP1 belongs, was previously rare,accounting for only 2–3% of clinical isolates.11

Whether toxinotyping can be used to distinguishvirulence potential among strains has yet to bedemonstrated.

Sporulation Capacity

Genotypically distinct strains of C. difficilehave been shown to demonstrate a propensity tohypersporulate and have been reported to beresponsible for previous outbreaks.29 TheBI/NAP1 strain, like other outbreak strains, hasdemonstrated the capacity to hypersporulatecompared with other nonoutbreak strains (Figure2).12 This putative virulence characteristic maybe, at least in part, responsible for its rapidestablishment in many institutions whereoutbreaks have been reported. As with the otherrecently identified characteristics of thisorganism, future studies are required to elucidatethe exact role of hypersporulation in thetransmission or pathology of C. difficile.

The Environment

Experts have pointed out that blaming anti-microbial agents alone is too simplistic and thatpeople often forget about the source of thebacterium.30 Both symptomatic and asympto-matic patients with C. difficile shed vegetativeforms of the organism and spores into theenvironment.4 Clostridium difficile spores may

persist on surfaces for years and remain aproblematic source of environmental contami-nation and infection.31 Although intuitive, aslevels of environmental contamination increase,so does the prevalence of C. difficile found on thehands of health care workers.32 Thus, susceptiblepatients may acquire the organism either directlyfrom the hands of health care workers orcircuitously from the environment.

Disease Severity

Among the factors that may be contributing tooutbreaks of CDAD is the increased severity ofdisease associated with the BI/NAP1 strain.Greater frequency of diarrhea, surgery require-ments, and recurrences would at least theoreticallylead to an advantageous scenario for the estab-lishment of the organism into the environment.In concert with hypersporulation characteristics,BI/NAP1 appears to be ideally suited for wide-spread dissemination and survival.

Alcohol-Based Hand Hygiene Products

The use of alcohol-based hand hygieneproducts in place of hand washing with soap andwater in health care settings has become commonand may be playing a role in the transmission ofC. difficile. Alcohol is not sporicidal and is notefficient in removing C. difficile from the hands.In fact, an average 36% (range 18–60%) of theinitial inoculation of C. difficile spores on acontaminated hand could be transferred byhandshake after using commercially availablealcohol gels.33 In contrast, the mechanical actionof hand washing in a sink with soap and waterfor a specified period of time has proved to beeffective in removing C. difficile from the handsof health care workers.13, 34 To be clear, handwashing is defined as a scrub with chlorhexidinesoap and water in an effort to remove skin oil(which harbors spores), followed by proper handdrying with a disposable paper towel.35 The timeperiod for hand washing is at least 30 seconds,although the exact period of time required foroptimal efficacy of this procedure is unclear. TheCenters for Disease Control and Prevention(CDC) recommends substitution of alcohol-based hand rubs with hand washing with use ofsoap and water during outbreak situations.36

Alcohol-based products do remain highlyeffective against non–spore-forming organisms,are a vital part of increasing hand hygienecompliance in health care settings, and shouldnot be abandoned for these reasons. It cannot be

302

Figure 2. Sporulation capacity of Clostridium difficileoutbreak strains (NAP1/027 [BI/NAP1] and C. difficile UKepidemic strain [CDUKES]) and nonoutbreak strains.(From reference 12 with permission.)

0

5

10

15

20

25

30

35

Nonoutbreak NAP1/027 CDUKES

Per

cent

age

NAP1/027 vs nonoutbreak (p<0.05)CDUKES vs nonoutbreak (p<0.05)


overstated that compliance with proper hand-hygiene practices is one of the most effectivemeasures for preventing the transmission of C.difficile as well as other organisms.

Hospital Cleaning Agents

As common detergents used in hospitals toclean patient rooms are not sporicidal, theintervention of using a 10% sodium hypochloritesolution has resulted in significant reductions inCDAD cases, as well as environmental sporeburden.37, 38 This same solution can be used toclean areas in the patient’s home to kill C. difficilespores and should be used to minimize thepotential for reinfection.

One part of containing the spread of C. difficileduring an outbreak is proper environmentaldisinfection. Replacing standard hospitalcleaning agents, such as quaternary ammonium–based products, with 10% sodium hypochloritesolutions in the affected units (or universally)appears to be an effective intervention.38

Although manually prepared and dilutedstandard 1:10 sodium hypochlorite solutions arerelatively inexpensive, they have the potential tobe more noxious to environmental servicesemployees, nursing staff, and patients, andrequire additional labor resources for dailypreparation. Commercially available premixedhypochlorite-based cleaning agents, such as thedisinfecting agent Dispatch (Caltech Industries,Inc., Midland, MI), are easier to prepare andadminister to surfaces in a less noxious manner,but are significantly more expensive to purchase.In addition, their effect on the sporulation of C.difficile may be in question.12

The Host

A variety of host-related variables influence thedevelopment of symptomatic CDAD, includingimmunity, age, exposure to antimicrobials, andpossibly gastric acid suppression. Like environ-mental risk factors, these variables may provideopportunities for intervention to reduce the riskof C. difficile infection.

Immunity and Age

Patients in whom an immune response(circulation of antitoxin A immunoglobulin[Ig]G antibodies) cannot be properly mountedare more susceptible to recurrent CDAD.39 Infact, patients who develop serum antitoxin A IgGtiters in response to colonization with C. difficile

have been shown to be 48 times less likely todevelop diarrhea than those who do not mount aresponse.24 As a corollary, older age has beenassociated with CDAD and is a plausible riskowing to the senescence of immunity.40 However,another group of authors noted that age wasconfounded by other variables.41 For instance,older patients tended to have more comorbidconditions that led to longer hospitalizations.Because of the strong association between poorimmune response and developing clinicaldisease, immune-based therapies including bothactive and passive immunization strategies areamong the most promising treatment andprevention modalities for the future.24

Antimicrobial Use

All antimicrobials (including certain chemo-therapeutic agents) are viewed as risk factors forCDAD; however, some agents may pose greaterrisks than others. Proposed variables that mayaugment risk include extended durations ofantimicrobial use, use of antimicrobials withantianaerobic activity, as well as exposure toagents lacking in vitro activity against theinfecting strain of C. difficile. Prolongedexposure to antimicrobials further increases therisk for developing CDAD41; thus, shorterdurations of treatment should be used where dataexist (e.g., urinary tract infections, ventilator-associated pneumonia, intraabdominalinfections), and therapy should be stopped whenthe infection has been adequately treated orinfection has been ruled out. Even single-doseadministration of prophylactic antimicrobials hasbeen implicated in inciting CDAD.41 It has beensuggested that antimicrobials with activityagainst anaerobes may result in greater risks forCDAD than agents lacking anaerobic activity42;however, this theory has not been supported byquality evidence.43, 44 For example, antimicrobialswith potent activity against anaerobes such aspiperacillin-tazobactam have been suggested tobe protective against CDAD or at least minimallyoffensive,45 whereas fluoroquinolones lackingactivity against anaerobes (e.g., levofloxacin,ciprofloxacin) have been strongly implicated inwell-conducted studies.20, 41, 46, 47

A case-control study at our institution con-ducted over a 2-year period during the beginningof our outbreak attributed to BI/NAP1 evaluatedrisk factors for CDAD.48 We matched eachconsecutive symptomatic CDAD case patientwith two control patients based on date of

303


hospital admission, admission unit, age, and sex,with all other variables (e.g., days at risk,previous gastrointestinal surgery, exposure togastric acid suppressants, presence of feedingtube, and severity of illness assessments using theCharleson Comorbidity Index and the HornIndex) being controlled for using multivariateanalysis. The following risk factors (relative risk,p value) were identified as independent anti-microbial exposure risk factors for CDAD onmultivariate analysis: cephalosporins (14,0.007), macrolides (5.51, 0.035), vancomycin(4.18, 0.009), and fluoroquinolones (3.19,0.006). Proton pump inhibitor use (2.89, 0.006)and the presence of a feeding tube (10.6, 0.002)were found to be nonantimicrobial risk factorsfor developing CDAD. Of interest, cephalosporinswere associated with the greatest risk (relativerisk 14, p<0.007), clindamycin was not identifiedas a risk factor (similar to the findings of others,47

most of our strains were clindamycin suscep-tible), and differences in risks between fluoro-quinolones based on in vitro potency againstanaerobes or by primary route of elimination(e.g., renal vs hepatobiliary-gastrointestinal)could not be distinguished.

Furthermore, an evaluation of comparativeclinical studies (phases II–IV) of more than21,000 patients, including hospitalized patientsenrolled in the United States and Canada, did notreveal differences in CDAD between the anti-anaerobic fluoroquinolone moxifloxacin andcomparators.49, 50 A recently published double-blind, randomized, controlled trial evaluating thesafety of moxifloxacin versus levofloxacin inhospitalized elderly patients with community-acquired pneumonia also revealed no differencesin CDAD rates between moxifloxacin (1/195patients [0.5%]) and levofloxacin (6/199 patients[3.0%]).51

Finally, and more plausibly, exposure toantimicrobials that lack activity against C. difficilehas been shown to be a risk factor for CDAD.6

This is thought to occur secondary to theselective survival advantage provided by theantimicrobial, resulting in intestinal overgrowthof C. difficile. Based on this, one of the fewproven formulary interventions to curb CDADrates in hospitals was accomplished by restrictingclindamycin use during an outbreak ofclindamycin-resistant C. difficile strains.52

Because clindamycin seems to have a unique,prolonged effect on intestinal flora, resulting inan extended susceptibility window for CDADacquisition, perhaps the success of this formulary

intervention cannot be extrapolated to otherantimicrobials. The implications of BI/NAP1being fluoroquinolone resistant may mean thatthe use of this entire class of antimicrobialsshould be reserved or perhaps restricted duringoutbreak settings.3 Indeed, because BI/NAP1resistance is uniform to all fluoroquinolone classmembers tested, no advantage can be gained byusing one preferentially over another. For manycommon respiratory tract infections, the abolish-ment of fluoroquinolone use seems impractical asit would only lead to the increased use of otherC. difficile precipitants such as the cephalosporins.

Ultimately, evaluating risk factors for CDADhas been inherently difficult due to the numerousand complex relationships among variables. Infact, most studies have failed to adequatelyaddress important confounding variables,resulting in serious threats to the validity of theirfindings.53 Since the published critique of pastCDAD risk factor ascertainment studies,53 severalstudies have been presented or published that arebetter designed to assess risk.20, 41, 48 However,one of the chief difficulties in risk factorascertainment studies for CDAD that remains isthe fact that many patients have received multipleantimicrobials in the 6–8 weeks precedingdevelopment of CDAD, further complicating theability to assign blame to a particular agent.Also, due to the retrospective nature of theirstudy design, it is difficult to obtain a completelyaccurate outpatient antimicrobial history. Andthe more commonly used antimicrobials can bemore commonly implicated in CDAD, as high-lighted by the case of the fluoroquinolones.Thought of previously as low-to-moderate riskantimicrobials, these agents are now used withgreater frequency in many geographic areas andhave been more commonly identified as riskfactors for CDAD.20, 41 Thus, the issues related toassigning risks to antimicrobials for precipitatingCDAD are complex. It is therefore difficult tostrongly support a single unifying theory thatdescribes an antimicrobial’s comparative risk forcausing CDAD (e.g., possessing or lackingantianaerobic activity) based on the currentpublished evidence is difficult. More research isneeded to properly delineate the complexinterrelationship between antimicrobial use andthe other risk factors that contribute to CDAD.

Although practical considerations prevent usfrom knowing susceptibility data for C. difficilestrains to a broad range of antimicrobials at ourinstitutions, and whether or not antianaerobicactivity deserves further attention, it is clear that

304


interventions directed toward optimizingantimicrobial use are beneficial for reducingCDAD rates. This was demonstrated by thefindings from a recent systematic review oninterventions to improve antibiotic prescribingpractices among hospital inpatients.54 Inaccordance with these findings, it is imperativethat antimicrobials be used in a more discrimi-nating manner, as patients may develop life-threatening CDAD as an adverse effect of apractitioner’s treatment decisions.55

Gastric Acid Suppression

Suppression of gastric acid can increase hostsusceptibility to infection and has beendocumented as a risk factor for a variety ofinfections that include travellers’ diarrhea,salmonellosis, cholera, as well as ventilator-associated pneumonia.7 One group of authorsused cohort and case-control study designs todetermine whether or not exposure to protonpump inhibitors was an independent risk factorfor CDAD.7 Multivariate analyses revealedstatistically significant adjusted odds ratios (95%confidence interval) of 2.1 (1.2–3.5) and 2.7(1.4–5.2) in the two studies identifying protonpump inhibitor use as a risk factor for CDAD.

The use of gastric acid suppressants has alsobeen associated with the development ofcommunity-acquired CDAD.56 Others have alsoimplicated the use of proton pump inhibitors asan independent risk factor for CDAD,57, 58

whereas one study did not find an association.41

In our institutional outbreak attributed toBI/NAP1, proton pump inhibitor exposure wasdetermined to be an independent risk factor forCDAD after correcting for confounders.48 Futureprospective studies of this association arewarranted; however, given the current evidence,it seems logical to increase vigilance regardingthe stewardship of proton pump inhibitor use.

Our Experience

The previous sections describe what has beenlearned about the epidemic strain of C. difficileknown as BI/NAP1. In this section, variouschallenges we have faced related to the discoveryand management of BI/NAP1 at our 600-bedcommunity teaching hospital are discussed.Although some of the issues discussed may becommon sense to the erudite reader, we were(and continue to be) surprised by our obser-vations in the clinic.

In 2002 as a part of our Antimicrobial

Stewardship Program’s daily review of anti-microbial therapy, we began to notice anenormous variability in the treatment of CDADalong with the introduction of non–evidence-based practices in the treatment of CDAD at ourinstitution (oral vancomycin plus oral metron-idazole, extended durations of therapy beyond10–14 days, “prophylaxis” with metronidazolefor patients receiving antibiotics, donor stooltransplantation, formulary requests for probiotics,cholestyramine use). Unaware of a simultaneousinvestigation by the CDC at other hospitals in theUnited States, we requested that our laboratoryculture clinical isolates of C. difficile for thepurposes of susceptibility testing by a notableexpert (Dale Gerding, M.D., Hines VA Hospital,Hines, IL) in this field. The chief reason fordoing this was that the previous outbreakscaused by J-type strains had been controlled byrestricting clindamycin, the caveat being that theJ-type strains were uniformly clindamycinresistant. To our disappointment, our isolateswere clindamycin susceptible, thus removing asimple formulary change from our quiver ofcountermeasures; furthermore, the organismstested were of a previously uncommon type.Serendipitously, we found ourselves in the midstof an outbreak due to BI/NAP1.3

Combined efforts among the CDC, Dr.Gerding’s laboratory, and others establishednewly identified characteristics of this previouslyuncommon strain of C. difficile.3 We continuedour antimicrobial stewardship efforts and workedwith our colleagues in infection control, environ-mental services, and hospital administration toaddress environmental issues. Realizing thatCDAD management guidelines had not beenupdated in nearly a decade, we collaborated withthe departments of gastroenterology, surgery,pharmacy, and infectious diseases to establishevidence-based (where possible) treatmentguidelines in order to address our currentsituation in the new era of BI/NAP1. Wedeveloped local guidelines in part because of therash of non–evidence-based strategies being used.We reviewed the literature and consulted withexperts in the field, which resulted in our finalproduct (Figures 3 and 4). Through this process,we have corralled what we termed “the WildWest” approach to CDAD management, whilereintroducing a more evidenced-based mentality.Subsequently, standardized order sets mimickingour treatment guidelines were incorporated intoour computerized physician order entry system.

Realizing that the treatment modalities for

305


refractory CDAD lacked double-blind, randomized,controlled trial data, we were careful not toexclude potentially beneficial options. However,as eluded to, we extinguished a variety of alleged“panaceas” for C. difficile that lacked evidence of

effectiveness and presented potential harm topatients.

First, a regimen consisting of oral vancomycinalong with oral metronidazole for the initialepisodes of CDAD had become almost routine

306

Figure 3. Treatment of first and second episodes of Clostridium difficile–associated disease. aReview drug interaction potentialbefore adding rifampin to a patient’s regimen.

Diarrhea + one of the following: Positive Clostridium difficile toxin test or C. difficile toxin test result pending + clinical suspicion of C. difficile

Start contact precautionsUse soap and water to wash hands

Metronidazole 500 mg p.o. q8h x 10 days C. difficile suspected or documented + ileus or toxic megacolon suspected Gastroenterology/surgical consultation

Metronidazole 500 mg i.v. q6–8h x 10 daysAlso consider: Intracolonic vancomycin 0.5–1.0 g in 1–2 L of normal saline given as retention enema using the following procedure59: 18-inch Foley catheter with a 30-ml balloon inserted into rectum, balloon inflated, vancomycin instilled, catheter clamped for 60 minutes, deflate and remove

Evaluate on days 4–6 of treatment Should not be deemed a treatment failure until an adequate course of therapy has been administered60

Symptoms not resolving or symptoms worsening

Symptoms resolving:Diarrhea should resolve within 2 weeksRecurrence expected in 12–20% after first episode, 50–60% after second episode

Change to vancomycin 125 mg p.o. q6h to complete 10–14-day courseConsider consultations (gastroenterology/surgery/ infectious disease)

Evaluate on day 4 of treatment


Consider adding rifampina

300 mg p.o. q12h to vancomycin 125 mg p.o. q6h to complete 10–14-day courseConsider consultations (gastroenterology/surgery/ infectious disease)


practice. We learned that the primary reason forprescribing both drugs was if clinicians thoughtthe patient was not responding to one drug earlyin therapy (e.g., day 2), the second drug wasadded. Thus, our guidelines now recommendreevaluation of therapy at days 4–6 and note thatdiarrhea should decrease but not necessarilyresolve within this time frame. We state that ifsymptoms do not improve or are worseningduring therapy, then the practitioner shouldswitch from oral metronidazole to oralvancomycin (not add a new agent to an alreadyfailing regimen). This is important because datasuggest that the very drugs used to treat C.

difficile may also predispose patients to futureepisodes of CDAD.5 Thus, if the goal is to allowfor the recovery of the protective commensalorganisms in the gastrointestinal tract, thecombination of the two drugs cannot beconducive to this goal.62

Some authors recently described that patientstreated for CDAD did not respond as well tometronidazole as first-line therapy,63 but qualityevidence to support this is lacking.64, 65

Resistance to metronidazole and vancomycinamong clinical isolates of C. difficile (bothoutbreak and nonoutbreak strains) in the UnitedStates and elsewhere does not seem to be

307

Figure 4. Treatment of third or more episodes of Clostridium difficile–associated disease or severe cases of first and secondepisodes. aReview drug interaction potential before adding rifampin to a patient’s regimen.

Diarrhea + one of the following: Positive Clostridium difficile toxin test or C. difficile toxin test result pending + clinical suspicion of C. difficile

Start contact precautionsUse soap and water to wash hands

Vancomycin 125 mg p.o. q6h x10 days

Evaluate on days 4–6 of treatment Should not be deemed a treatment failure until an adequate course of therapy has been administered60

Symptoms resolving:Diarrhea should resolve within 2 weeksRecurrence expected in 12–20% after first episode, 50–60% after second episode


Evaluate on day 4 of treatment


Consider adding rifampina 300 mg p.o. q12h to vancomycin 125 mg p.o. q6h to complete 10–14-day courseConsider consultations (gastroenterology/ surgery/infectious disease)

Consider high-dose oral vancomycin (500 mg p.o. q6h)61

C. difficile suspected or documented + ileus or toxic megacolon suspected Gastroenterology/surgical consultation

Metronidazole 500 mg i.v. q6–8h x 10 daysAlso consider: Intracolonic vancomycin 0.5–1.0 g in 1–2 L of normal saline given as retention enema using the following procedure59: 18-inch Foley catheter with a 30-ml balloon inserted into rectum, balloon inflated, vancomycin instilled, catheter clamped for 60 minutes, deflate and remove


contributing to clinical failures since resistance isalmost universally nonexistent.66 One possibleexplanation for the observations of poor responseto metronidazole or vancomycin in the clinic isthat patients are experiencing greater diseaseseverity due to hypertoxin production and lastingdisease due to increased sporulation of BI/NAP1.

Second, we received considerable pressure toadd probiotic formulations to our hospitalformulary for the management of CDAD. Thus,we conducted a MEDLINE search for studies thatevaluated probiotic therapy of CDAD. We couldfind little if any quality evidence supporting theirefficacy67 but did find unexpected evidence tosupport their harm, chiefly in the form of numer-ous reports of fungemias due to Saccharomycesboulardii and bacteremias due to Lactobacillus spafter probiotic administration to both immuno-competent and immunocompromised hosts.68–71

A recent review of the literature emphasized thepotential harm associated with S. boulardii.72 Tothis end, probiotics at this juncture are notrecommended at our institution.

Third, we noted the increased use of anion-binding resins or adsorbents (cholestyramine,colestipol) at our institution. Although theo-retically beneficial as binders of select C. difficiletoxins, colestipol has been shown to beequivalent to placebo for its ability to affect thefecal excretion of toxins.73 Current evidencesuggests that these drugs bind to a variety ofother drugs, including vancomycin, in thegastrointestinal tract, thereby rendering knowneffective agents potentially ineffective. Based onthe unproven efficacy of anion-binding resins andtheir potential to neutralize the efficacy of proventherapies, experts have recommended againsttheir use.74, 75 Unfortunately, the literature isreplete with numerous review articles recom-mending the use of anion-binding resins, withoutsupportive clinical efficacy data. Experimentaltoxin-binding polymers such as tolevamer are indevelopment and may be useful if provedeffective.75

Fourth, we found some practitioners using“prophylactic” regimens of either metronidazoleor vancomycin in patients (without diarrhea andwith no suspicion of CDAD) who were receivingantimicrobial therapy for the treatment of anunderlying infection, reportedly in an effort toprevent a future episode of CDAD. This“prophylactic” approach had been gaining inpopularity before the introduction of ourguidelines. We know that exposing patients toantimicrobials (including oral vancomycin and

metronidazole) increases the probability ofinfection with C. difficile, as well as future relapsesin those with documented infection.5 We alsoknow that the treatments (oral vancomycin andmetronidazole) are not effective in eradicatingC. difficile spores.75 Therefore, administering oralvancomycin or metronidazole for “prophylactic”purposes is illogical, non–evidence based, andpotentially harmful in that it may increase thepatient’s risk for CDAD.

Fifth, we noticed that oral vancomycin dosagesvaried widely for initial treatment of first episodesof CDAD and that treatment for extendeddurations beyond 10 days (often weeks) wasbecoming increasingly common. Our guidelinespromote metronidazole as first-line therapyunless multiple episodes are documented,intolerance exists, or the patient is pregnant.When oral vancomycin is used, evidence dictatesthat 125 mg administered every 6 hours isequivalent to 500 mg every 6 hours in terms ofefficacy,76 but the cost is significantly less.Higher doses of vancomycin (e.g., 2 g/day)administered to treat refractory CDAD have beenshown to be effective in descriptive studies.75

For refractory disease, multiple relapses, orpatients with a poorly functioning gastrointestinaltract, we realize that high-quality evidence-basedoptions are lacking. For patients with refractorydisease (not responding at the day-4–6 evaluationpoint), we recommend switching from oralmetronidazole to oral vancomycin (but to not usethem in combination). Rifampin can be added ifthe patient continues not to respond,77 andsurgical consultation is suggested if markers forsevere disease are present (high white blood cellcount, ascites, obstruction, toxic megacolon).We also reserve oral vancomycin at higher doses,such as 2 g/day, for patients with multiplerelapses or those who fail to respond to therapy.Pulse-dosed or tapered vancomycin regimens aresuggested for patients with multiple relapses, aspreviously described.75 In addition, although notsupported by quality data but because there areno better solutions, intravenous metronidazole isrecommended in patients with suspected toxicmegacolon or ileus, with or without the use ofintracolonic vancomycin.

Conclusion

A previously uncommon strain of C. difficiletermed BI/NAP1 has been implicated as a causeof geographically diverse outbreaks of CDAD.Certain virulence characteristics may be responsible

308


for more severe clinical presentations and pooroutcomes. Putative virulence factors (e.g.,hypertoxin production, hypersporulation)possessed by a previously uncommon strain of C.difficile, in conjunction with particular host andenvironmental factors, just may have resulted inthe “perfect storm.” It is too simplistic to assumethat a basic change in an antimicrobial formularyas a single intervention strategy will be successfulin reducing CDAD rates where BI/NAP1 isendemic; policies that incorporate goodantimicrobial stewardship along with infectioncontrol and environmental interventions arelikely to be necessary to combat CDAD. Themisuse of antimicrobials and the failure tocomply with hand washing have increasingly direconsequences for both patients and health careworkers. While awaiting updated nationalguidelines for the management of CDAD, wedeveloped local guidelines to reestablishevidence-based treatment, where possible, andprovide direction in the wake of BI/NAP1.Finally, the reasons for suboptimal infectioncontrol and antimicrobial use in modern timesare similar: derisory resources both human andfinancial, apathy, and ignorance. It is thereforeup to clinicians and administrators alike to workon all three of these elements that stand in theway of curbing the growing epidemic of CDAD.

Acknowledgments

The author would like to express his gratitude toDrs. Dale Gerding, Cliff McDonald, Mark Wilcox,Patricia Stogsdill, and August Valenti, as well as JamesLyden, Charlie Papa, and Gwen Rogers, for theirassistance in the various phases of our investigationand interdiction of BI/NAP1.

References1. Pepin J, Valiquette L, Alary ME, et al. Clostridium difficile

associated diarrhea in a region of Quebec from 1991 to 2003: achanging pattern of disease severity. Can Med Assoc J2004;171;466–72.

2. Archibald LK, Banerjee SH, Jarvis WR. Secular trends inhospital-acquired Clostridium difficile disease in the UnitedStates, 1987–2001. J Infect Dis 2004;189:1585–9.

3. McDonald LC, Killgore GE, Thompson A, et al. An epidemic,toxin gene–variant strain of Clostridium difficile. N Engl J Med2005;353:2433–41.

4. McFarland LV, Mulligan ME, Kwok RY, Stamm WE .Nosocomial acquisition of Clostridium difficile infection. N EnglJ Med 1989;320:204–10.

5. Bignardi GE. Risk factors for Clostridium difficile infection. JHosp Infect 1998;40:1–15.

6. Gerding DN. Clindamycin, cephalosporins, fluoroquinolones,and Clostridium difficile–associated diarrhea: this is anantimicrobial resistance problem. Clin Infect Dis 2004;38:646–8.

7. Dial S, Alrasadi K, Manoukian C, Huang A, Menzies D. Risk

of Clostridium difficile diarrhea among hospital inpatientsprescribed proton pump inhibitors: cohort and case-controlstudies. Can Med Assoc J 2004;171:33–8.

8. Mulligan ME, Miller SD, McFarland LV, Fung HC, Kwok RY.Elevated levels of serum immunoglobulins in asymptomaticcarriers of Clostridium difficile. Clin Infect Dis 1993;16(suppl4):S239–44.

9. Johnson S, Gerding DN. Clostridium difficile–associateddiarrhea. Clin Infect Dis 1998;26:1027–34.

10. Samore M, Killgore G, Johnson S, et al. Multicenter typingcomparison of sporadic and outbreak Clostridium difficileisolates from geographically diverse hospitals. J Infect Dis1997;176:1233–8.

11. Warny M, Pepin J, Fang A, et al. Toxin production by anemerging strain of Clostridium difficile associated withoutbreaks of severe disease in North America and Europe.Lancet 2005;366:1079–84.

12. Underwood S, Stephenson K, Fawley WN, et al. Effects ofhospital cleaning agents on spore formation by N. Americanand UK outbreak Clostridium difficile strains [abstr]. In:Program and abstracts of the 45th interscience conference onantimicrobial agents and chemotherapy. Washington, DC:American Society for Microbiology, 2005:LB-28.

13. King S. Provision of alcohol hand rub at the hospital bedside: acase study. J Hosp Infect 2004;56:S10–12.

14. Wanahita A, Goldsmith EA, Musher DM. Health care costsand mortality associated with nosocomial diarrhea due toClostridium difficile. Clin Infect Dis 2002;34:346–53.

15. Eggerston L. Quebec puts up to 20 million dollars for C.difficile fight. Can Med Assoc J 2005;172:622.

16. McDonald LC, Banerjee S, Jernigan DB. Increasing incidenceof Clostridium difficile–associated disease in U.S. acute carehospitals, 1993–2001 [abstr]. In : Proceedings of 14th annualscientific meeting of the Society for Healthcare Epidemiology ofAmerica. Alexandria, VA: Society for Healthcare Epidemiologyof America; 2004:67.

17. Sohn S, Climo M, Diekema D, et al . Varying rates ofClostridium difficile–associated diarrhea at Prevention Epicenterhospitals. Infect Control Hosp Epidemiol 2005;26:676–9.

18. McDonald LC. Clostridium difficile: responding to a new threatfrom an old enemy. Infect Control Hosp Epidemiol2005;26:672–5.

19. Dallal RM, Harbrecht BG, Boujoukas AJ, et al. FulminantClostridium difficile: an underappreciated and increasing causeof death and complications. Ann Surg 2002;235:363–72.

20. Muto CA, Pokrywka M, Shutt K, et al. A large outbreak ofClostridium difficile–associated disease with an unexpectedproportion of deaths and colectomies at a teaching hospitalfollowing increased fluoroquinolone use. Infect Control HospEpidemiol 2005;26:273–80.

21. McEllistrem MC, Carman RJ, Gerding DN, et al. A hospitaloutbreak of Clostridium difficile disease associated with isolatescarrying binary toxin genes. Clin Infect Dis 2005;40:265–72.

22. Pepin J, Valiquette L, Cossette B. Mortality attributable tonosocomial Clostridium difficile–associated disease during anepidemic caused by a hypervirulent strain in Quebec. Can MedAssoc J 2005;173:1037–42.

23. Centers for Disease Control and Prevention, Chernak E,Johnson CC, Weltman A, et al. Severe Clostridium difficile–associated disease in populations previously at low risk.MMWR Morb Mortal Wkly Rep 2005;54:1201–05.

24. Giannasca PJ, Warny M. Active and passive immunizationagainst Clostridium difficile diarrhea and colitis. Vaccine2004;22:848–56.

25. Alfa MJ, Kabani A, Lyerly D, et al. Characterization of a toxinA-negative, toxin B-positive strain of Clostridium difficileresponsible for a nosocomial outbreak of Clostridiumdifficile–associated diarrhea. J Clin Microbiol 2000;38:2706–14.

26. Voth DE, Ballard JD. Clostridium difficile toxins: mechanismof action and role in disease. Clin Microbiol Rev 2005;18:247–63.

27. McEllistrem MC, Carman RJ, Gerding DN, Genheimer CW,

309


Zheng L. A hospital outbreak of Clostridium difficile diseaseassociated with isolates carrying binary toxin genes. Clin InfectDis 2005;40:265–72.

28. Barbut F, Decre D, Lalande V, et al. Clinical features ofClostridium difficile–associated diarrhea due to binary toxin(actin-specific ADP-ribosyltransferase) producing strains. JMed Microbiol 2005;54:181–5.

29. Wilcox MH, Fawley WN. Hospital disinfectants and sporeformation by Clostridium difficile [letter]. Lancet 2000;356:1324.

30. Wilcox MH. Cleaning up Clostridium difficile infection. Lancet1996;348:767–8.

31. Kim KH, Fekety R, Batts DH, et al. Isolation of Clostridiumdifficile from the environment and contacts of patients withantibiotic-associated colitis. J Infect Dis 1981;143:42–50.

32. Samore MH, Venkataraman L, DeGirolami PC, Arbeit RD,Karchmer AW. Clinical and molecular epidemiology ofsporadic and clustered cases of nosocomial Clostridium difficilediarrhea. Am J Med 1996;100:32–40.

33. Leischner J, Johnson S, Sambol S, Parada J, Gerding DN.Effect of alcohol hand gels and chlorhexidine hand wash inremoving spores of Clostridium difficile from hands [abstr]. In:Program and abstracts of the 45th interscience conference onantimicrobial agents and chemotherapy. Washington, DC:American Society for Microbiology, 2005:LB-29.

34. Bettin K, Clabots C, Mathie P, Willard K, Gerding DN.Effectiveness of liquid soap vs. chlorhexidine gluconate for theremoval of Clostridium difficile from bare hands and glovedhands. Infect Control Hosp Epidemiol 1994;15:697–702.

35. Weir E, Flegel K. Protecting against Clostridium difficile illness.Can Med Assoc J 2005;172:1178.

36. Centers for Disease Control and Prevention. Informationabout a new strain of Clostridium difficile. Available fromwww.cdc.gov/ncidod/dhqp/id_CdiffFAQ_newstrain.html.Accessed November 20, 2005.

37. Wilcox MH, Fawley WN, Wigglesworth N, Parnell P, Verity P,Freeman J. Comparison of the effect of detergent versushypochlorite cleaning on environmental contamination andincidence of Clostridium difficile infection. J Hosp Infect2003;54:109–14.

38. Mayfield JL, Leet T, Miller J, Mundy LM. Environmentalcontrol to reduce transmission of Clostridium difficile. ClinInfect Dis 2000;31:995–1000.

39. Kyne L, Warny M, Qamar A, Kelly CP. Association betweenantibody response to toxin A and protection against recurrentClostridium difficile diarrhoea. Lancet 2001;357:189–93.

40. Ginaldi L, Loreto MF, Corsi MP, Modesti M, De Martinis M.Immunosenescence and infectious diseases. Microbes Infect2001;3:851–7.

41. Pepin J, Saheb N, Coulombe MA, et al. Emergence offluoroquinolones as the predominant risk factor for Clostridiumdifficile–associated diarrhea: a cohort study during an epidemicin Quebec. Clin Infect Dis 2005;41:1254–60.

42. Gaynes R, Rimland D, Killum E, et al. Outbreak of Clostridiumdifficile infection in a long-term care facility: association withgatifloxacin use. Clin Infect Dis 2004;38:640–5.

43. Owens RC Jr, Ambrose PG. Antimicrobial safety: focus onfluoroquinolones. Clin Infect Dis 2005;40:S456–69.

44. Mohr J. Outbreak of Clostridium difficile infection andgatifloxacin use in a long-term care facility. Clin Infect Dis2004;39:875–7.

45. Wilcox MH, Freeman J, Fawley W, et al . Long-termsurveillance of cefotaxime and piperacillin-tazobactamprescribing and incidence of Clostridium difficile diarrhoea. JAntimicrob Chemother 2004;54:168–72.

46. McCusker ME, Harris AD, Perencevich E, Roghmann MC.Fluoroquinolone use and Clostridium difficile–associateddiarrhea. Emerg Infect Dis 2003;9:730–3.

47. Loo VG, Poirier L, Miller MA, et al. A predominantly clonalmulti-institutional outbreak of Clostridium difficile–associateddiarrhea with high morbidity and mortality. N Engl J Med2005;353:2442–9.

48. Owens RC Jr, Lyden J, Prato BS, et al. Risk factors for

Clostridium difficile-associated disease during an outbreakassociated with a toxin gene variant strain. [abstr]. In: Programand abstracts of the 45th interscience conference onantimicrobial agents and chemotherapy. Washington, DC:American Society for Microbiology, 2005:K-1268.

49. Owens RC Jr, Kruesmann F, Reiter C, Korfmann G, ChoudhriS. Comparative frequency of Clostridium difficile–associateddiarrhea in phase II–IV moxifloxacin clinical trials. Presented atthe 7th European congress of chemotherapy and infection,Florence, Italy, October 19–22, 2005.

50. Ball P, Stahlman R, Kubin R, Choudhri S, Owens RC Jr. Safetyprofile of oral and intravenous moxifloxacin: cumulative datafrom clinical trials and postmarketing studies. Clin Ther2004;26:940–50.

51. Anzueto A, Niederman MS, Pearle J, Restrepo MI, Heyder A,Choudhri SH, for the Community-Acquired PneumoniaRecovery in the Elderly Group. Community-acquiredpneumonia recovery in the elderly (CAPRIE): efficacy andsafety of moxifloxacin therapy versus that of levofloxacintherapy. Clin Infect Dis 2006;42:73–81.

52. Climo MW, Israel DS, Wong ES, Williams D, Coudron P,Markowitz SM. Hospital-wide restriction of clindamycin: effecton the incidence of Clostridium difficile–associated diarrhea andcost. Ann Intern Med 1998;128:989–95.

53. Thomas C, Stevenson M, Riley TV. Antibiotics and hospital-acquired Clostridium difficile–associated diarrhoea: a systematicreview. J Antimicrob Chemother 2003;51:1339–50.

54. Davey P, Brown E, Fenelon L, et al. Interventions to improveantibiotic prescribing practices for hospital inpatients.Cochrane Database Syst Rev 2005;4:CD003543.

55. Owens RC Jr, Fraser GL, Stogsdill P . Antimicrobialstewardship programs as a means to optimize antimicrobial use.Pharmacotherapy 2004;24:896–908.

56. Dial S, Delaney JAC, Barkun AN, Suissa S. Use of gastricacid–suppressive agents and the risk of community-acquiredClostridium difficile–associated disease. JAMA 2005;294:2989–95.

57. Cunningham R, Dale B, Undy B, Gaunt N. Proton pumpinhibitors as a risk factor for Clostridium difficile diarrhoea. JHosp Infect 2003;54:243–5.

58. Al-Tureihi FI, Hassoun A, Wolf-Klein G, Isenberg H. Albumin,length of stay, and proton pump inhibitors: key factors inClostridium difficile–associated disease in nursing homepatients. J Am Dir Assoc 2005;6:105–8.

59. Apisarnthanarak A, Razavi B, Mundy LM . Adjunctiveintracolonic vancomycin for severe Clostridium difficile colitis:case report and review of the literature. Clin Infect Dis2002;35:690–6.

60. Gerding DN, Johnson S, Peterson LR, Mulligan ME, Silva J Jr.Clostridium difficile–associated diarrhea and colitis. InfectControl Hosp Epidemiol 1995;16:459–77.

61. McFarland LV, Elmer GW, Surawicz CM. Breaking the cycle:treatment strategies for 163 cases of recurrent Clostridiumdifficile disease. Am J Gastroenterol 2002;97:1769–75.

62. Bingley PG, Harding GM. Clostridium difficile colitis followingtreatment with metronidazole and vancomycin. Postgrad Med J1987;63:993–4.

63. Musher DM, Aslam S, Logan N, et al. Relatively poor outcomeafter treatment of Clostridium difficile colitis withmetronidazole. Clin Infect Dis 2005;40:1586–90.

64. Gerding DN. Metronidazole for Clostridium difficile–associateddisease: is it okay for mom? Clin Infect Dis 2005;40:1598–600.

65. Bricker E, Garg R, Nelson R, Loza A, Novak T, Hansen J.Antibiotic treatment for Clostridium difficile–associated diarrheain adults. Cochrane Database Syst Rev 2005;1:CD004610.

66. Freeman J, Stott J, Baines SD, Fawley WN, Wilcox MH.Surveillance for resistance to metronidazole and vancomycin ingenotypically distinct and UK epidemic Clostridium difficileisolates in a large teaching hospital. J Antimicrob Chemother2005;56:988–9.

67. Dendukuri N, Costa V, McGregor M, Brophy JM. Probiotictherapy for the prevention and treatment of Clostridiumdifficile–associated diarrhea: a systematic review. Can Med

310


Assoc J 2005;173:167–70.68. Lherm T, Monet C, Nougiere B, et al. Seven cases of fungemia

with Saccharomyces boulardii in critically ill patients. IntensiveCare Med 2002;28:797–801.

69. Salminen MK, Rautelin H, Tynkkynen S, et al. Lactobacillusbacteremia, clinical significance, and patient outcome, withspecial focus on probiotic L. rhamnosus GG. Clin Infect Dis2004;38:62–9.

70. Cassone M, Serra P, Mondello F, et al . Outbreak ofSaccharomyces cerevisae subtype boulardii fungemia in patientsneighboring those treated with a probiotic preparation of theorganism. J Clin Microbiol 2003;41:5340–3.

71. Surawicz CM. Probiotics, antibiotic-associated diarrhea andClostridium difficile diarrhea in humans. Best Pract Res ClinGastroenterol 2003;17:775–83.

72. Enache-Angoulvant A, Hennequin C. Invasive Saccharomyces

infection: a comprehensive review. Clin Infect Dis 2005;41:1559–68.

73. Mogg GA, George RH, Youngs D, et al. Randomised controlledtrial of colestipol in antibiotic-associated colitis. Br J Surg1982;69:137–9.

74. Wilcox MH. Treatment of Clostridium difficile infection. JAntimicrob Chemother 1998;41:41–6.

75. McFarland LV. Alternative treatments for Clostridium difficiledisease: what really works? J Med Microbiol 2005;54:101–11.

76. Fekety R, Silva J, Kauffman C, Buggy B, Deery HG. Treatmentof antibiotic-associated Clostridium difficile colitis with oralvancomycin: comparison of two dosage regimens. Am J Med1989;86:15–19.

77. Buggy BP, Fekety R, Silva J Jr. Therapy of relapsing Clostridiumdifficile–associated diarrhea and colitis with the combination ofvancomycin and rifampin. J Clin Gastroenterol 1987;9:155–9.

311

Documents

Clostridium difficile–Associated Disease: An Emerging Threat to Patient Safety: Insights from the Society of Infectious Diseases Pharmacists