Prevention of Endemic Healthcare-Associated Clostridium difficile Infection: Reviewing the Evidence

Transcription

1 CLINICAL AND SYSTEMATIC S nature publishing group 2327 CME Prevention of Endemic Healthcare-Associated Clostridium difficile Infection: Reviewing the Evidence J. Hsu, MD 1, C. Ab a d, M D 1, M. D i n h, M D 1 and Nas i a S afd ar, M D, Ph D 1 OBJECTIVES: METHODS: RESULTS: Clostridium difficile is the most common infectious cause of healthcare-associated diarrhea. Because of the increasing incidence and severity of endemic C. difficile infection (CDI), interventions to prevent healthcare-associated CDI are essential. We undertook a systematic review of interventions to reduce healthcare-associated CDI. We searched multiple computerized databases, and manually searched for relevant articles to determine which interventions are useful in preventing CDI. Studies were required to be controlled in design and to report the incidence of endemic CDI as an outcome. Data on the patient population, intervention, study design, and outcomes were abstracted and reviewed using established criteria. Few randomized controlled trials exist in the area of CDI prevention. The interventions with the greatest evidence for the prevention of CDI include antimicrobial stewardship, glove use, and disposable thermometers. Environmental decontamination also may decrease CDI rates, although the level of evidence is not as strong as for the other proven interventions. Treatment of asymptomatic carriage of C. difficile is not recommended. There is insufficient evidence to make a recommendation for or against the use of probiotics. In cases of known or suspected CDI, hand hygiene with soap and water is preferred over use of waterless alcohol hand rub. Many nonrandomized trials included in our analysis used multiple interventions concurrently, making the independent role of each preventive strategy difficult to determine. We chose to include only studies that focused on endemic CDI because studies of outbreaks have used multiple strategies, making it difficult to measure the relative efficacy of each strategy. Environmental disinfection and probiotics need to be studied further to evaluate their roles in the prevention of CDI. Although there have been no studies assessing the utility of isolation and cohorting for the prevention of endemic CDI specifically, it is a widely used intervention for containment of this and other similar multidrug-resistant pathogens. CONCLUSIONS: Antimicrobial stewardship, glove use, hand hygiene, and disposable thermometers should be routinely used for the prevention of CDI. Environmental disinfection and probiotics should be studied further for their role in reducing CDI. Am J Gastroenterol 2010; 105: ; doi: /ajg ; published online 6 July 2010 INTRODUCTION Clostridium difficile is a major infectious cause of healthcareassociated diarrhea, with as many as 25 % of cases caused by the organism ( 1,2 ). A recent survey showed an overall C. difficile prevalence of 13.1 patients per 1,000 inpatient in US hospitals of which 73 % were healthcare associated ( 3,4 ). In addition, studies have found acquisition rates of C. difficile in 1 13 % of patients hospitalized for less than a week, which increases to 50 % for patients hospitalized for >4 weeks ( 5,6 ). C. difficile infection (CDI) increases length of hospital stay and imposes a significant financial burden on healthcare institutions, with an estimated cost of 1.1 billion per year in the United States ( 7 ). The incidence and severity of CDI seems to be increasing ( 8 10 ). Recently, a fluoroquinolone-resistant strain of C. difficile, referred to as BI / NAP1, toxinotype I, and ribotype 027, with variation in toxin genes caused large outbreaks of CDI in Canada, the United States and Europe, associated with morbidity and considerable mortality ( ). In addition, healthy groups of patients, such as peripartum women, have recently found to be at increased risk (15,16 ). 1 Section of Infectious Diseases, Department of Medicine, University of Wisconsin School of Medicine and Public Health, Madison, Wisconsin, USA. Correspondence: Nasia Safdar, MD, PhD, Section of Infectious Diseases, Department of Medicine, University of Wisconsin School of Medicine and Public Health, Room J5 / 222 CSC, 600 Highland Avenue, Madison, Wisconsin 53792, USA. ns2@medicine.wisc.edu Received 12 January 2010; accepted 24 May 2010

2 2328 Hsu et al. Because of increased antimicrobial exposure and environmental contamination by C. difficile spores, healthcare facilities remain the primary source for acquisition of CDI ( ). In view of the fact that CDI has reemerged as a major healthcare-associated pathogen, we undertook an evidence-based systematic review to identify interventions for the prevention of CDI in healthcare institutions, and to critically evaluate the efficacy of these interventions. We chose to focus on endemic CDI, that is, CDI prevalence in nonoutbreak settings, as studies of outbreaks have generally used multiple interventions making the relative efficacy of each difficult to measure. METHODS Search strategy We searched PUBMED (including MEDLINE) (1966 to 1 May 2009), CINAHL (1982 to 1 May 2009), Cochrane Database of Systematic Reviews, ACP Journal Club (1991 to 1 May 2009), Database of Abstracts of Reviews of Effects, Cochrane Central Register of Controlled Trials, and the NIH clinical trial registry using the following search terms alone and in combination: Clostridium difficile plus the following words gloves, gowns, barrier precautions, antibiotic cycling, antibiotic restriction, antibiotic stewardship, clindamycin, cephalosporins, microbicides, bleach, chlorhexidine, hypochlorite, glutaraldehyde, disinfectant, decontamination, equipment, stethoscope, thermometer, endoscope, cohorting, epidemiology, colonization, handwashing, soap, alcohol-based hand rub, alcohol gel, probiotics, Lactobacillus, and Saccharomyces. References from relevant articles, including published guidelines for CDI prevention and treatment were analyzed to determine appropriateness for inclusion ( ). We also searched abstract proceedings of the Infectious Diseases Society of America ( ), the Interscience Conference on Antimicrobial Agents and Chemotherapy ( ) and the Society for Healthcare Epidemiology of America ( ), and the European Congress of Clinical Microbiology and Infectious Disease ( ). The end date of the search was 1 May No language restrictions were applied to the search. Table 1. Infectious Diseases Society of America US Public Health Service grading system for ranking recommendations in clinical guidelines ( 22 ) Category, grade Definition Strength of recommendation A B C D E Quality of evidence I I Good evidence to support a recommendation for use Moderate evidence to support a recommendation for use Poor evidence to support a recommendation Moderate evidence to support a recommendation against use Good evidence to support a recommendation against use Evidence from 1 properly randomized, controlled trial Evidence from 1 well-designed clinical trial, without randomization; from cohort or case-controlled analytic studies (preferably from > 1 center); from multiple time-series; or from dramatic results from uncontrolled experiments Evidence from opinions of respected authorities, based on clinical experience, descriptive studies, or reports of expert committees Grading of preventive strategies Study quality and recommendations were rated according to defined guidelines by the Infectious Disease Society of America s United States Public Health Service grading system ( Table 1 ) (30 ). We followed the Quality of Reporting of Meta-Analysis (QUORUM) checklist for our systematic review ( 31 ). Analysis We considered a formal meta-analysis inappropriate because of the limited number of studies and heterogeneous study designs within each intervention category. We report results from the individual studies included in our review. Inclusion and exclusion criteria We included studies that (i) assessed the effectiveness of a strategy for the prevention of CDI in the hospital setting, (ii) either provided a risk or rate ratio or provided data to calculate these measures, and (iii) used clinically important outcomes, such as the incidence of CDI, either as a primary or secondary outcome. Because of the paucity of randomized trials in this area of study, we included both randomized and nonrandomized clinical trials for all preventive strategies. However, for trials that studied probiotic use, we included only randomized-controlled trials. We focused on the prevention of endemic CDI in this review. Several recent articles have discussed treatment of CDI, and we have not addressed that in our review ( 28,29 ). Three of the authors undertook data extraction (J.H., C.L.A., and M.D.). Disagreement was resolved based on discussion with the fourth author (N.S.) and arriving at consensus. RESULTS Th e initial search identified 2,578 articles that we reviewed for relevance. Seventy-six articles of relevance were reviewed in detail; the remaining articles were not relevant for the prevention of CDI. Thirty-eight references were excluded because they included multiple interventions, occurred in outbreak settings, or did not provide CDI incidence data. The results of our search are shown in Figure 1. We grouped preventive strategies into the following main categories: hand hygiene, barrier precautions, disposable equipment, environmental disinfection, antimicrobial stewardship, and probiotic use. Tables 2 4 show the studies that assessed these practices. Diagnosis of CDI Most studies included in this review used a positive toxin assay for CDI in patients with healthcare-associated diarrhea (defined The American Journal of GASTROENTEROLOGY VOLUME 105 NOVEMBER

3 Prevention of Clostridium difficile -Associated Disease 2329 Results of initial computerized search, n = 2,578 Articles not relevant to CDI prevention, n = 2,504 Potentially relevant articles retrieved for detailed review, n = 76 Included: Thermometers, n =2 Glove use, n =2 Hand hygiene, n =4 Disinfectants, n =5 Surveillance, n =1 Carriage of C. difficile toxin, n =2 Antibiotic stewardship, n = 17 Probiotics, n =8 Excluded: Thermometers, n =1 Glove use, n =1 Hand hygiene, n =6 Disinfectants, n =4 Surveillance, n =2 Carriage of C. difficile toxin, n =2 Antibiotic stewardship, n = 8 Other (multiple interventions), n =14 Total trials, n = 38 Total trials, n = 38 Figure 1. Database and reference search results. CDI, C. diffi cile infection. as diarrhea developing 2 or 3 after hospitalization). Some defined acquisition of C. difficile as presence of C. difficile toxin in stool with or without diarrhea. Surveillance Unlike other healthcare-associated pathogens such as methicillin-resistant Staphylococcus aureus and vancomycin-resistant Enterococcus, there is a paucity of data on rates of CDI in healthcare institutions. Archibald et al. ( 32 ) reported CDI rates of 5.1 per 10,000 patient in ICUs of teaching hospitals, using National Healthcare-associated Infections Surveillance System data from More recently, Jarvis et al. ( 3 ) published a point prevalence study of CDI in 648 US hospitals that showed 13.1 per 1,000 inpatients had CDI or colonization (94.4 % infection). Gravel et al. ( 33 ) showed a CDI incidence rate of 4.6 cases per 1,000 admissions in Canada. Studies such as these have contributed to improving our understanding of CDI epidemiology. Recently, comprehensive guidelines for surveillance for CDI have been published. The European Centre for Disease Prevention and Control has set forth a proposal for implementing a systematic surveillance program ( 34 ). McDonald et al. (4 ) have developed interim surveillance definitions and recommendations based on existing literature and expert opinions that may help improve CDI prevention efforts. The Society for Healthcare Epidemiology in America and the Infectious Disease Society of America have published joint guidelines regarding the prevention of CDI in acute care hospitals ( 26 ). With these recommendations, surveillance for CDI can be standardized and used to establish baseline rates for comparative purposes. This information is necessary to guide implementation of preventive measures. Hand hygiene Background. The oldest measure advocated to prevent healthcare-associated infection in healthcare institutions is hand hygiene ( ). However, despite universal acknowledgement of handwashing as a cornerstone of healthcare-associated infection control programs, compliance rates > 50 % have been difficult to achieve, and handwashing rates have ranged from 9 % to 50 % in studies of healthcare workers ( 40,41 ). A vigorous 1-min rubbing with a sufficient volume of alcohol to wet the hands completely has been shown to be highly effective at reducing the density of skin flora (42 44 ). Because of their convenience and broad-spectrum germicidal activity, alcohol-based waterless hand rubs (ABHR) are recommended as the first choice for hand decontamination before and after patient care in a guideline released by the CDC in 2002 ( 45 ). However, all have limited efficacy with gross soilage so that visibly soiled hands should always be washed with antiseptic soap and water ( 45 ). ABHR have poor activity against spore-forming organisms such as C. difficile and in the setting of a CDI outbreak, conventional handwashing with soap and water is recommended instead of an alcohol hand rub ( 45 ). Quantification of C. difficile spores on hands of a volunteer group after use of alcohol-based hand rub showed only marginal reduction in spore counts ( logs) ( 46 ). It has been suggested that increased use of ABHR may contribute to transmission of CDI in healthcare institutions. However, a recent retrospective study in a single facility found no increase in CDI despite a progressive increase in the use of ABHR over a 3-year period ( 47 ). Similarly, an interventional time-series analysis showed no correlation between ABHR consumption and CDI incidence (48 ). Summary of the evidence. We identified two trials that determined the incidence of CDI after use of alcohol-based hand rubs (49,50 ). Both were before after nonrandomized studies. Rao et al. ( 49 ) found that the incidence of CDI decreased during a 1-year period after the implementation of a hand hygiene program that promoted use of an alcohol-based hand rub. In contrast, Gordin et al. ( 50 ) reported no change in the baseline incidence of CDI over a 3-year period after alcohol-based hand rub use in a tertiary care medical center. It is possible that the decrease in incidence of CDI in the study by Rao et al. may be due to behavioral change rather than the direct impact of ABHR on CDI, as ABHR does not have sporocidal activity. We concur with the current recommendation from SHEA / IDSA to perform hand hygiene with soap and water, instead of ABHR, after caring for patients with CDI in outbreak setting or in settings with high transmission of C. difficile (B-I). Cost and potential for harm. The cost in this case is related to the expense of soap to be used in lieu of the use of ABHR. The potential for harm with the use of ABHR in the setting of CDI transmission pertains to its lack of reliable sporicidal activity. Barrier precautions Background. There is ample evidence to suggest that C. difficile spreads from the hands, equipment or apparel of healthcare workers (17 19,51 ). Healthcare-associated transmission is well

4 2330 Hsu et al. Table 2. Nonpharmacologic approaches to prevention of CDI Trial Study design Population Results Alcohol-based hand rub (ABHR) Gopal Rao et al. ( 49 ) Gordin et al. ( 50 ) Glove use Johnson et al. ( 53 ) Disposable thermometers Brooks et al. ( 63 ) Jernigan et al. ( 66 ) Hypochlorite disinfectant Mayfi eld et al. ( 69 ) Wilcox et al. ( 67 ) Hydrogen peroxide vapor Boyce et al. ( 74 ) Treatment of carriers Delmee et al. ( 108 ) Johnson et al. ( 107 ) ; nonrandomized historical control ; nonrandomized historical control Before after trial and concurrent ; nonrandomized without control Prospective, randomized cross-over control ; historical Nonrandomized cross-over control Prospective interventional Prospective ; nonrandomized historical control Nonblinded placebo-controlled randomized trial Patients hospital wide (ages not specifi ed) Adult patients hospital wide Adult patients on two medical and two surgical wards Pre-intervention 11.5 Cases per 1,000 admissions 3.24 Cases per 10,000 patient care 7.7 Cases per 1,000 discharges Predominantly elderly 2.7 Cases per 1,000 patient Adult patients on 20 selected wards Adult patients in the bone marrow transplant unit, the neurosurgical intensive care unit, and a general medicine unit Elderly patients on two selected medicine wards Five inpatient units All patients in the hematologic unit CDI, C. difficile infection; CI, confidence interval; HR, hazard ratio. a No decline in control units. b No decline in CDI in other two units. c No decline in CDI in the second unit Cases per 8.6 Cases per 1,000 patient 8.9 Cases per 100 admissions 2.28 Cases per 16.6 % Positive stool cultures Post-intervention 9.5 Cases per 1,000 admissions ( P = 0.02) 3.38 Cases per 10,000 patient care ( P = 0.78) 1.5 Cases per 1,000 discharges 1.76 Cases per ( P < 0.01) 0.16 Cases per 3.3 Cases per 1,000 patient 5.3 Cases per 100 admissions ( P < 0.05) 1.28 Cases 1,000 patient ( P = 0.47) 3.6 % Positive stool cultures > 60 Years of age Number of negative stool cultures after treatment: Vancomycin: 9 / 10 Metronidazole: 4 / 10 ( P = 0.02) Placebo: 2 / 10 ( P = 0.005) Relative risk (95 % CI) 0.82 (Not reported) 1.04 (Not reported) 0.16 ( ) a 0.65 (Not reported) 0.44 ( ) I HR = 0.37 ( ) b Level of evidence 0.59 (Not reported) c 0.56 (Not reported) 0.02 (Not reported) I documented ( 18,52 ). The use of gowns and gloves by healthcare workers when caring for patients with CDI would be expected to reduce healthcare-associated transmission of C. difficile. Summary of the evidence. Glove use: The use of gloves by healthcare workers is recommended to prevent the transmission of hospital-acquired CDI (A-I). In our review, we found a single study that assessed vinyl glove use for the prevention of CDI. Johnson et al. (53 ) performed a prospective, nonrandomized study with concurrent on the universal use of vinyl gloves and found a decrease in healthcare-associated CDI rates from 7.7 per discharges before glove use to 1.5 per 1,000 after implementation of glove use (relative risk = 0.16; 95 % confidence interval = , P = 0.015). The control group, which did not have the interventions of periodic in-services, posters, and a box of gloves placed at bedside but merely had boxes of vinyl gloves in the supply room, also had a small but not statistically significant decrease in CDI rates (5.7 per discharges to 4.2 per 100, P = 0.20). The difference in CDI incidence was lower in the glove ward compared with the control, but this was not statistically significant ( P = 0.14). The American Journal of GASTROENTEROLOGY VOLUME 105 NOVEMBER

5 Prevention of Clostridium difficile -Associated Disease 2331 Table 3. Antimicrobial stewardship for prevention of healthcare-associated CDI Trial Brown et al. ( 81 ) Restricted antimicrobial agent Clindamycin Stewardship mechanism Voluntary restriction with subsequent transition to ID physician approval Study design Intervention, controlled Study duration Results Preintervention Postintervention Level of evidence Comment 12 Months 2.25 % 0.74 % a Contact isolation, empiric therapy emphasized Climo et al. ( 88 ) Multiple b Approval by ID physician or pharmacist 24 Months 11.5 Cases per month 3.33 Cases per month ( P < 0.001) Frank et al. ( 102 ) Multiple b Approval by ID physician or pharmacist Nonrandomized, single group study 12 Months 0.48 Cases per 0.52 Cases per 1,000 patient ( P = 0.70) c McNulty et al. ( 54 ) Cefuroxime Substitution by benzyl penicillin and trimethoprim; antimicrobial use monitoring by pharmacy 7 Months 5.28 Cases per month 2.28 Cases per month ( P = 0.002) Other infection control measures were implemented concurrently Stone et al. ( 93 ) Cephalosporins Limiting use of antibiotics to 7, substitution by ampicillin, trimethoprim, or gentamicin, removal of cephalosporins from ward, and antimicrobial use monitoring by pharmacy 18 Months 3.35 Cases per 100 admissions 1.94 Cases per 100 admission ( P < 0.05) Other infection control measures were implemented concurrently Ludlam et al. ( 91 ) Injectable ceftriaxone, ceftazidime, and cefotaxime Voluntary restriction; other details not provided Intervention with concurrent 24 Months 8.16 Cases per month 3.75 Cases per month ( P < 0.001) Thomas et al. ( 95 ) Ceftriaxone Removal of ceftriaxone from ward stocks, microbiology department approval for ceftriaxone use required Not specifi ed 2.09 Cases per 1,000 discharges 0.87 Cases per 1,000 discharges ( P < ) Khan et al. ( 90 ) Ceftriaxone Policy of early switch from intravenous cephalosporin to noncephalosporin oral antibiotic, subsequent withdrawal of oral cephalosporin for inpatient use, levofl oxacin substituted for ceftriaxone 17 Months 1.46 Cases per patient 0.34 Cases per a Thomas et al. ( 94 ) Ceftriaxone Approval by ID physician or microbiologist required 24 Months 0.61 Cases per 0.28 Cases per (95 % CI = ) c Carling et al. ( 103 ) Aztreonam, imipenem, parenteral fl uoroquinolones, third-generation cephalosporins Pharmacy and infectious disease recommendations for antibiotic use with monitoring and education Quasi-experimental 10 Years 2.2 Cases per 1.4 Cases per 1,000 patient ( P = 0.002) Automatic 7-day stop orders for all antibiotics; limited susceptibility reporting; exclusion of pharmaceutical representative-based detailing of antibiotics Continued on following page

6 2332 Hsu et al. Table 3. Continued Trial O Connor et al. ( 92 ) Wilcox et al. ( 96 ) Fowler et al. ( 97 ) Schulhof et al. ( 101 ) Hansen et al. ( 99 ) Nuila et al. ( 98 ) Weil et al. ( 100 ) Restricted antimicrobial agent Cefuroxime, ceftriaxone Cefotaxime 4.6 Cases per 100 admissions Cephalosporins Amoxicillin / clavulanate Multiple d Quasi-experimental Cephalosporins, extendedspectrum penicillins, fl uoroquinolones Restriction of most parenteral antibiotics Cephalosporins, fl uoroquinolones Stewardship mechanism Policy to discourage use of all second- and thirdgeneration intravenous cephalosporins Voluntary restriction followed by active restriction with routine monitoring; piperacillintazobactam substituted Increase use of narrow spectrum antibiotics Pharmacy and infectious disease recommendations for antibiotic use Implemented antibiotic management program; other details not provided Approval by infectious disease physician or pharmacist Details not provided Study design Quasi-experimental with concurrent Intervention Intervention with concurrent Study duration 4 Months 3.9 Cases per 100 admissions Fixed 3-month interval over 3 years Results Preintervention Postintervention 1.20 Cases per 100 admissions ( P = 0.03) 2.2 Cases per 100 admissions ( P = 0.008) 42 Months 0.35 incidence rate ratio (0.17, 0.73) 37 Months 1.5 Cases per 36 Months 1.4 Cases 12 Months 41.7 Cases per month 36 Months 11.6 Cases per 10,000 patient 0.84 Cases per 22 Cases per month ( P < 0.001) 8 Cases per 10,000 patient Level of evidence Comment Concurrent implementation of hand hygiene education campaign Concurrent implementation of unspecifi ed infection control measures CDI, C. difficile infection; CI, confidence interval. a P value not given. b Amikacin, clarithromycin, azithromycin, ciprofloxacin, ofloxacin, norfloxacin, amoxicillin-clavulanate, ampicillin-sulbactam, ticarcillin-clavulanate, cefaclor, cefixime, cefotaxime, cefoperazone, cefazolin ( > 3 g / day), ceftriaxone ( > 1 g / day), imipenem, aztreonam, oral vancomycin, chloramphenicol, fluconazole, and itraconazole. c Not significant. d Ampicillin-sulbactam, piperacillin-tazobactam, cefepime, ceftriaxone, ciprofloxacin, levofloxacin, and imipenem. Gowns: We did not find any studies that evaluated the use of gowns for the prevention of endemic healthcare-associated CDI. Although several studies of outbreaks have reported that gowns are effective in interrupting healthcare-associated transmission of CDI, in each study, multiple interventions were undertaken concurrently. Thus, the role of gowns in the prevention of healthcareassociated transmission of CDI is unknown ( ). The most recent CDC guidelines on isolation precautions in the hospital recommends hand hygiene and contact precautions when caring for patients with CDI until diarrhea resolves. Contact precautions include the use of a clean, nonsterile gown and nonsterile gloves before entering the patient s environment. There is insufficient evidence for or against the use of gowns for reducing CDI. Cost and potential for harm. In the trial by Johnson et al., an accurate cost comparison was not possible because of the high usage of gloves on the control side. Although latex allergies are well documented, vinyl gloves have not been shown to have allergic potential ( 56 ). The potential for harm from the use of gloves arises from prolonged use by healthcare workers, and outbreaks caused by contaminated gloves have been well described ( 57,58 ). Patient isolation and cohorting Background. Isolation of patients in addition to the use of barrier precautions by healthcare workers is considered essential for the prevention of spread of CDI in healthcare institutions. Cohorting of patients with CDI is also widely practiced as a control measure. Although studies of the efficacy of these approaches exist in the outbreak setting, to our knowledge, no studies have formally examined the efficacy of patient isolation and / or cohorting for containing endemic CDI and we did not identify any studies assessing this approach in our search. The American Journal of GASTROENTEROLOGY VOLUME 105 NOVEMBER

7 Prevention of Clostridium difficile -Associated Disease 2333 Table 4. Randomized controlled trials of probiotics for prevention of CDI Trial Probiotic Duration of probiotic Surawicz et al. a,b ( 118 ) McFarland et al. ( 82 ) Lewis et al. b ( 119 ) Kotowska et al. ( 120 ) Can et al. ( 121 ) Thomas et al. ( 117 ) Plummer et al. ( 116 ) Hickson et al. ( 122 ) S. boulardii Initiated within 48 h of starting antibiotics; continued for 2 weeks after antibiotic completion S. boulardii Initiated within 72 h of starting antibiotics; continued for 3 after antibiotic completion S. boulardii For the duration of antibiotic therapy S. boulardii For the duration of antibiotic therapy S. boulardii Initiated 48 h after starting antibiotics; continued for duration of antibiotic therapy Lactobacillus GG L. acidophilus, B. bifidum L. casei, L. bulgaricus, S. thermophilus Initiated within 24 h of starting antibiotic therapy; continued for 14 CDI outcome measure Patient population (no. of patients) adults (180) adults (193) adults > 65 years of age (69) children 6 months to 14 years of age (246) adults years of age (151) adults (302) Probiotic 22 / 81 (27 % ) Results Placebo 5 / 36 (14 % ) Relative risk (95 % CI) 1.95 ( ) Duration of follow-up 17 Days Not reported by randomization status 7 Weeks 5 / 33 (0.15 % ) 3 / 36 (0.08 % ) 3 (2.5 % ) 10 (7.9 % ) 0 / 73 (0 % ) 2 / 133 (1.5 % ) 20 Days Primary Elderly (150) 2 / 69 (2.9%) Initiated within 48 h of antibiotics and continued 7 after antibiotic completion patients (135) 6 / 57 (12 % ) 2 / 76 (2.6 % ) 3 / 134 (2.2 % ) 5 / 69 (7.2 % ) 19 / 56 (34 % ) 1.81 ( ) 0.3 ( ) 0.20 ( ) 0.67 ( ) 0.40 ( ) 5 10 Days 14 Days 30 Days 21 Days 20 Days Weeks B. bifium, Bifidobacterium bifidum ; CDI, Clostridium difficile infection; CI, confidence interval; L. acidophilus, Lactobacillus acidophilus ; L. casei, Lactobacillus casei ; S. boulardii, Saccharomyces boulardii ; S. thermophilus, Streptococcus thermophilus. a Subgroup analysis in patients from whom 3 stools analyzed for C. difficile toxin. b The outcome in these studies was C. difficile acquisition as determined by the presence of C. difficile toxin in stool. Level of evidence Summary of the evidence. We were not able to identify any studies that have assessed the efficacy of patient isolation and / or cohorting for the prevention of endemic CDI. Given the lack of evidence, no formal recommendation regarding patient isolation and / or cohorting can be made. However, this practice is widely used and has shown benefit in outbreak containment with CDI and other organisms with epidemiology similar to CDI, such as methicillin-resistant S. aureus. Cost and potential for harm. Th e cost of patient isolation and cohorting has not been formally assessed. Studies of patient isolation with methicillin-resistant S. aureus have found that healthcare providers were less likely to document physical examination findings for patients colonized with methicillin-resistant S. aureus and that methicillin-resistant S. aureus patients in isolation were more likely to suffer falls and other adverse effects ( 59 ). We did not find any studies that examined the potential harm posed by isolating patients with CDI. Medical apparatus Background. A large body of literature suggests that contamination of hospital equipment and surfaces with C. difficile spores is extremely common ( 19,51,60 63 ). Although the contribution of environmental contamination to healthcare-associated infection is unclear, use of disposable or dedicated equipment may reduce the risk of CDI ( 64,65 ). Summary of the evidence. We found two studies that addressed the use of disposable rather than electronic rectal thermometers for the prevention of endemic CDI ( 63,66 ). Brooks et al. (63 )

8 2334 Hsu et al. undertook an interventional, nonrandomized trial in which electronic rectal thermometers were substituted with disposable rectal thermometers in an acute care hospital. They found a decrease in the incidence of CDI from 2.71 per to 1.76 per ( P < 0.01) during the 6 months after intervention. Subsequently, Jernigan et al. ( 66 ) reported the results of a randomized-controlled trial comparing the rate of decrease in CDI from the use of disposable thermometers compared to electronic thermometers usage. This trial revealed that there was a significant decrease of CDI with disposable thermometers (0.16 per ) compared with electronic thermometers (0.37 per, relative risk = 0.44; 95 % confidence interval = , P = 0.026). On the basis of these studies, single-use disposable thermometers should be used for the prevention of transmission of CDI (A-I). Cost and potential for harm. B ro o k s et al. ( 63 ) estimated that the cost of using disposable rather than electronic rectal thermometers would cost an additional ~ $ 7,000 per year. Jernigan et al. ( 66 ) also determined that the cost of single-use disposable thermometers was higher compared with electronic thermometers because of both expense of the disposable items as well as requirement for increased nursing time. The investigators estimated the cost to be an additional $ 5,926 per CDI prevented for single-use, disposable compared with electronic thermometers. No harm has been reported from the use of disposable thermometers. Environmental disinfection Background. C. difficile is ubiquitous in the hospital environment and has been cultured from a number of surfaces, such as bedding, toilets, mops, and furniture ( 19,61 63,67 ). A potentially useful approach to reducing healthcare-associated spread of C. difficile is environmental decontamination with agents that have activity against the spore form of C. difficile (18,19,51,68 ). Summary of the evidence. Four studies have assessed environmental decontamination as a means of preventing CDI ( 67,69 ). Three studies used hypochlorite solution as the decontaminating agent and one used hydrogen peroxide vapor (HPV). In a before after intervention study, Mayfield et al. (69 ) used quaternary ammonium solution for 9 months followed by 1:10 hypochlorite solution in the rooms of patients with CDI on the bone marrow transplantation unit, general medicine service or neurosurgical ICU. During the 9 months before the use of hypochlorite solution, CDI rates were 8.6, 3.0, and 1.3 cases per for bone marrow transplantation, neurosurgical ICU, and general medicine patients, respectively. After the initiation of the hypochlorite protocol, CDI rates decreased significantly to 3.3 cases per (hazard ratio = 0.37; 95 % confidence interval = ) for bone marrow transplantation patients but did not change significantly for patients in the other two units. After switching back to quaternary ammonium, the CDI rate increased to 8.1 cases per. Limitations of the study include lack of complete data on antineoplastic and antimicrobial doses and duration, lack of evidence of environmental contamination, and lack of direct observation of housekeeping technique and compliance. Wilcox et al. ( 67 ) showed in a nonrandomized, cross-over control study a decrease in the rate of CDI in only one of the two patient care wards included in the study with the use of a sodium hypochlorite solution compared with liquid detergent. In one ward, rates of CDI decreased from 8.9 to 5.3 cases per 100 admissions ( P < 0.05), whereas in the other ward, rates of CDI increased from 3.5 to 4.7 cases per 100 admissions despite the use of sodium hypochlorite disinfection ( P < 0.05). In 2006, McMullen et al. (70 ) published data from a randomized, controlled trial of education (housekeeping and contact precaution education every 6 months) vs. education plus bleach (education plus change to bleach as routine cleaning agent). The control ward was provided no education and continued to use quaternary ammonium product for routine cleaning. CDI rates on the education plus bleach ward decreased slightly from 3.1 per to 2.7 per ( P = 0.42). They concluded bleach might not impact endemic CDI rates. These conflicting results make it difficult to draw conclusions regarding the efficacy of environmental disinfection with sodium hypochlorite. McMullen et al. ( 71 ) undertook a pre post intervention study in medical and surgical intensive care units during an outbreak, where 1:10 household bleach was used for decontamination purposes instead of the usual quaternary ammonium. The rate of CDI in the MICU decreased from 16.6 cases per to 3.7 cases per (relative risk < 0.22; 95 % confidence interval = ) after the intervention. Similarly, CDI rates in the SICU also decreased from 10.4 per 1,000 patient to 3.9 per. This study was not formally included in this review, as it was done in the setting of an outbreak; however, the downward trend that persisted for up to 2 years after the intervention suggest that use of hypochlorite solution may reduce CDI. At this time, the use of chlorine-containing solutions is recommended for environmental disinfection in areas where ongoing C. difficile transmission is noted (B-) ( 72 ). More recently, interest in hydrogen peroxide mist has arisen. Shapey et al. ( 73 ) showed significant reduction in surface contamination with C. difficile spores after use of a single cycle of dry mist hydrogen peroxide in rooms on an elderly care ward. Boyce et al. ( 74 ) performed a prospective interventional study to evaluate the effectiveness of HPV room decontamination. Five high-incidence wards were manually cleaned, temporarily vacated, and subjected to HPV decontamination. CDI incidence rates fell from 2.28 cases per 1,000 patient to 1.28 cases per post-intervention ( P = 0.047). HPV was then expanded to all rooms vacated by patients with CDI and the hospital-wide incidence fell from 1.89 cases per to 0.88 cases per ( P = 0.047). This is the only study identified in our search relating to clinical application of HPV. Although this is a promising strategy that deserves further investigation, insufficient The American Journal of GASTROENTEROLOGY VOLUME 105 NOVEMBER

9 Prevention of Clostridium difficile -Associated Disease 2335 evidence is available to recommend for or against use of hydrogen vapor decontamination at this time. Cost and potential for harm. Cost comparison of the disinfecting agents was not provided in any of the included studies. In general, the increased cost of hypochlorite-based solutions compared with detergents would be small in contrast to the large cost savings from the prevention of CDI ( 67 ). In considering potential harmful effects of chemical agents, adverse reactions such as asthma exacerbations have been reported related to the use of various cleaning agents such as chlorine gas (by-product of mixing sodium hypochlorite with hydrochloric acid), glutaraldehyde, and quaternary ammonium ( 56,75,76 ). In addition, corrosive effects of disinfectants on surfaces over time have been reported. Pharmacologic approaches to prevention of CDI Antimicrobial stewardship. Background: Antimicrobial use is the major predisposing factor for CDI. Recent systematic reviews of the literature have examined the relationship between antibiotic use and CDI ( 77,78 ). Although certain antimicrobial agents (ampicillin, cephalosporins, and clindamycin) have been associated with a higher risk of CDI, virtually every antimicrobial agent has been implicated (22,79 83 ). Fluoroquinolone use has been associated with an increased risk of CDI in many studies ( ). Antimicrobial stewardship especially restriction of selective antimicrobial agents known to greatly increase the risk of CDI, such as clindamycin, cephalosporins, and fluoroquinolones, may reduce the incidence of healthcare-associated CDI ( 87 ). Summary of the evidence: We identified two interventional nonrandomized trials, using historical ( 81,88 ) that assessed the efficacy of clindamycin restriction for the prevention of CDI. Both found that rates of CDI decreased markedly paralleling decrease in drug usage. Brown et al. (81 ) reported that clindamycin restriction was associated with a decline in CDI rates from 22.5 per 100 discharges to 7.4 per 100 patient discharges over a 1-year period. Climo et al. ( 88 ) reported that requirement of approval of clindamycin by an infectious diseases physician was associated with a 17 % decrease in CDI over a 24-month period. Susceptibility testing of C. difficile isolates was reported in one study, where the predominant strain was found to be resistant to clindamycin ( 88,89 ). In both studies, increased emphasis was also placed on other infection control measures such as education and the use of barrier precautions. We identified eight trials that assessed the utility of cephalosporin restriction for reduction in CDI ( 54,90 96 ). All of these trials found that cephalosporin restriction reduced CDI. All were nonrandomized studies. Cephalosporin stewardship was implemented in a number of ways. One trial required infectious diseases physician s approval in order for the restricted agent to be administered (94 ). Three trials distributed written policies to staff ( 91,93,96 ). Other trials gave few details of restriction methods, but primarily used voluntary restriction ( 54,90,91,95 ). Three trials had a system of routine monitoring of usage of restricted agents performed by the pharmacy ( 54,93,96 ). Two trials reinforced restriction policies with routine reminders via meetings or rounding on the wards ( 92,93 ). The four trials that documented increased usage of alternative agents did not observe any adverse consequences such as an increase in infections caused by other pathogens ( ). Follow-up ranged from 4 months to 7 years. Seven trials restricted more than one antibiotic class and found decreased rates of CDI ( ). Two of the five trials included concomitant change in infection control practices, which may have affected the results. The most common method of restriction is requirement for antibiotic approval by an infectious disease physician or pharmacist. Our review of the use of antimicrobial stewardship shows that there is good evidence supporting antibiotic restriction to prevent healthcare-associated CDI (A-). Trials that implicated clindamycin as a factor for developing CDI with high numbers of clindamycinresistant strains were able to establish that the number of clindamycin-resistant strains decreased with introduction of clindamycin restriction by follow-up susceptibility testing ( 88 ). A study by Pear et al. is a hallmark study in demonstrating the reduction in CDI incidence with restriction of clindamycin use. However, it was not included in the review, as it was in the setting of an outbreak ( 89 ). Periodic susceptibility testing of C. difficile strains in CDI cases may help to identify resistance patterns caused by selection pressure that may respond to antibiotic restriction. Given the correlation of fluoroquinolone usage with the increased incidence of CDI, antibiotic stewardship of fluoroquinolones should be considered as well. Four studies included restriction of fluoroquinolone usage; however, these were implemented in the context of other antibiotic restrictions, making it difficult to discern the impact of fluoroquinolone restriction on the reduction in CDI ( ). Cost and potential for harm: Three studies estimated changes in expenditures after implementation of antimicrobial stewardship. Most of the savings were realized because of alternative antimicrobials to the injectable cephalosporins and clindamycin were more costly. Trials estimated cost savings in the reduction of CDI to range from $114,000 to 212,000 (88,91,94 ). Frank et al. (102 ) and Carling et al. ( 103 ) estimated cost savings from changes in antibiotic usage and not in relation to CDI. Treatment of asymptomatic carriers. Background: Person-toperson healthcare-associated transmission of C. difficile has been well documented ( 5,6,18,52, ). Patients who are asymptomatic carriers of C. difficile may contribute to horizontal spread within the institution. Antimicrobial treatment to eradicate asymptomatic carriage of C. difficile has been studied as a way to interrupt healthcare-associated transmission and prevent CDI. Summary of the evidence: We identified two studies that evaluated the efficacy of treating patients with asymptomatic carriage of C. difficile for the prevention of endemic CDI. In a randomized, placebo-controlled trial, Johnson et al. (107 ) treated 30 patients with asymptomatic C. difficile carriage with metronidazole, vancomycin, or placebo for 10. Metronidazole was not effective at eliminating carriage and although vancomycin initially cleared C. difficile from stools, rate of recolonization was high at follow-up in 2 months. In a quasi-experimental prospective study over 29 months in patients with leukemia, Delmee et al. (108 ) treated all

10 2336 Hsu et al. patients who were carriers of C. difficile (symptomatic and asymptomatic) with vancomycin for 7. The percentage of positive toxin assays fell from 9.9 % to 1.2 % ; however, multiple other interventions were also carried out during the same time frame, including environmental decontamination that may have influenced the results. Our review does not provide sufficient evidence regarding treatment of asymptomatic carriers. This is in agreement with current guidelines, which does not support treatment of asymptomatic carriage of C. difficile (D-I). Cost and potential for harm: The cost of treatment of asymptomatic carriers has not been evaluated. Probiotics. Background: Disruption of normal gastrointestinal flora by antimicrobials may predispose to C. difficile colonization and CDI ( ). Probiotics, such as Lactobacillus, may protect against acquisition of C. difficile and CDI by restoring normal gastrointestinal equilibrium ( 112,113 ). Summary of the evidence: Two meta-analyses of randomizedcontrolled trials have been conducted to determine the efficacy of probiotics for the prevention and treatment of CDI ( 114,115 ). The paucity of trials, small samples sizes, and heterogeneity in type of probiotic used do not allow conclusions regarding the overall efficacy of probiotics for the prevention of CDI. We identified eight randomized-controlled trials that reported C. difficile toxin acquisition and / or CDI as a primary or secondary outcome ( Table 4 ) (82, ). The probiotic used in the trials was either Lactobacillus spp, Saccharomyces boulardii or a combination (82, ). One trial evaluated probiotics in children, whereas the remaining seven trials were conducted with hospitalized adults (120 ). The trials had a small number of cases and short follow-up the longest being 7 weeks by McFarland et al. (82 ). Hickson et al. ( 122 ) showed a statistically significant decrease in CDI with use of a combination probiotic milkshake. No patients in the probiotic group acquired CDI, whereas 9 out of 53 (17 % ) in the placebo group developed CDI ( P = 0.001). None of the remaining trials demonstrated a statistically significant decrease in CDI or C. difficile toxin acquisition with the use of probiotic therapy. However, the trials lacked adequate statistical power to determine the efficacy of probiotics. No trial reported adverse effects from the use of probiotics. There is currently insufficient evidence to recommend routine use of probiotics to prevent CDI. Larger trials with longer followup are needed to better assess the efficacy of probiotics for the prevention of CDI. Harm: Live Lactobacilli are widely consumed in fermented dairy food products because of putative health benefits and are available over the counter in various probiotic formulations. Although generally regarded to be safe, cases of fungemia from S. boulardii and Lactobacillus bacteremia have been reported in immunocompromised patients and those with central venous catheters ( ). Ongoing assessment of the safety of probiotics, especially in immunocompromised patients is necessary. Limitations Our review has several limitations. Most trials that focused on preventive measures were not randomized trials. Confounding and selection bias remain a concern. In many nonrandomized studies included in our review, multiple interventions for prevention of CDI were used concurrently, thus making the individual impact of each difficult to measure. Conclusion A number of strategies have been studied for the prevention of healthcare-associated CDI. Our review shows that although further research is needed on many established as well as novel prevention measures, the following can be recommended for the prevention of CDI: (i) antimicrobial stewardship, (ii) glove use, and (iii) disposable thermometers. Treatment of asymptomatic carriage of C. difficile is not recommended. There is insufficient evidence to recommend for or against the use of probiotics and additional trials are necessary. Environmental disinfection is potentially promising and should be studied further. Table 5 Table 5. Recommendations for prevention of healthcareassociated CDI Preventive strategy Result Grade Recommendation Hand hygiene Soap and water Barrier precautions Benefi t A- The practice of hand hygiene is strongly recommended Benefi t B-I In a setting in which there is an outbreak or an increased CDI rate, instruct visitors and healthcare workers to wash hands with soap (or antimicrobial soap) and water after caring for or contacting patients with CDI Gloves Benefi t A-I Gloves are recommended for contact with all body substances Gowns Disposable thermometers Environmental disinfection Antimicrobial stewardship Treatment of patients with asymptomatic carriage of C. difficile Probiotics Insuffi cient evidence Additional clinical trials are needed to determine the incremental benefi t of gowns for prevention of CDI Benefi t A-I The use of disposable thermometers rather than electronic thermometers is recommended Benefi t B- Chlorine-containing cleaning agents or other sporocidal agents should be useful for disinfection of surfaces in areas of increased CDI transmission. No EPA-registered products are specific for inactivating C. difficile. Additional trials are needed Benefi t A- Antimicrobial usage monitoring is recommended. Stewardship efforts should especially be directed at antimicrobials noted to be associated with high risk of CDI especially cindamycin, cephalosporins, and in more recent years, fl uoroquinolones No benefi t Insufficient evidence D-I Not recommended. Additional trials are needed No recommendation. Additional trials are needed CDI, C. difficile infection; EPA, US Environmental Protection Agency. The American Journal of GASTROENTEROLOGY VOLUME 105 NOVEMBER

11 Prevention of Clostridium difficile -Associated Disease 2337 summarizes our recommendations for the prevention of endemic CDI. Randomized trials to test the efficacy of preventive measures for CDI are needed. CONFLICT OF INTEREST Guarantor of the article: Nasia Safdar, MD, PhD. Specific author contributions: Collected and interpreted initial data: J. Hsu, M. Dinh, and C. Abad; analyzed and finalized data and accepts full responsibility for the conduct of the study: Nasia Safdar. All authors helped with drafting the manuscript, approved the final draft submitted, had full access to all of the data in the study, and take responsibility for the integrity of the data and the accuracy of the data analysis. Financial support: N.S. is supported by institutional training grant 5 K12 AG from the National Institutes of Health to the University of Wisconsin School of Medicine and Public Health. Potential competing interests: None. REFERENCES 1. Kel ly C P, Pot hou l a k is C, L a Mont J T. C lo st r i diu m d i ffi cile colitis. N Engl J Med 1994 ;330 : Mylonakis E, Ryan ET, Calderwood SB. Clostridium difficile associated diarrhea: a review. Arch Intern Med 2001 ;161 : Jar v is W R, S ch l o ss e r J, Jar v is A A et al. National point prevalence of Clostridium difficile in US health care facility inpatients, Am J Infect Control 2009 ;37 : McD ona l d LC, C oi g nard B, D ubb e rke E et al. Recommendations for surveillance of Clostridium difficile-associated disease. Infect Control Hosp Epidemiol 2007 ;28 : Joh ns on S, C l ab ot s C R, L in n F V et al. Healthcare-associated Clostridium difficile colonisation and disease. Lancet 1990 ;336 : S amore M H, D egi rol ami P C, T lu cko A et al. Clostridium difficile colonization and diarrhea at a tertiary care hospital. Clin Infect Dis 1994 ; 18 : Ky ne L, Hamel M B, Pol av ar am R et al. Healthcare costs and mortality associated with healthcare-associated diarrhea due to Clostridium difficile. Clin Infect Dis 2002 ; 34 : McDonald CL, Banerjee S, Jernigan DB. Increasing incidence of Clostridium difficile -associated disease in U.S. acute care hospitals, Proceedings of the 14th Annual Scientific Meeting of the Society for Healthcare Epidemiology of America. Philadelphia, Pennsylvania: The Society for Healthcare Epidemiology of America, 2004: Frost F, Craun GF, Calderon RL. Increasing hospitalization and death possibly due to Clostridium difficile diarrheal disease. Emerg Infect Dis 1998 ;4 : Fro st F, Hurl e y J S, Pete rs e n H V et al. Estimated incidence of Clostridium difficile infection. Emerg Infect Dis 1999 ; 5 : War ny M, Pepi n J, Fang A et al. Toxin production by an emerging strain of Clostridium difficile associated with outbreaks of severe disease in North America and Europe. Lancet 2005 ;366 : Pepi n J, Va l i qu e tte L, A l ar y M E et al. Clostridium difficile-associated diarrhea in a region of Quebec from 1991 to 2003: a changing pattern of disease severity. CMAJ 2004 ;171 : McD ona l d LC, Ki l l gore G E, Thomps on A et al. An epidemic, toxin genevariant strain of Clostridium difficile. N Engl J Med 2005 ; 353 : Bartlett JG. Narrative review: the new epidemic of Clostridium difficileassociated enteric disease. Ann Intern Med 2006 ;145 : R oupha el NG, O D on nel l JA, Bhat nagar J et al. Clostridium difficileassociated diarrhea: an emerging threat to pregnant women. Am J Obstet Gynecol 2008 ;198 : G are y K W, Ji ang Z D, Yad av Y et al. Peripartum Clostridium difficile infection: case series and review of the literature. Am J Obstet Gynecol 2008 ;199 : B ar nett J, Thom li ns on D, Per r y C et al. An audit of the use of manual handling equipment and their microbiological flora implications for infection control. J Hosp Infect 1999 ;43 : McFarl and LV, Mu l li g an M E, Kwok RY et al. Healthcare-associated acquisition of Clostridium difficile infection. N Engl J Med 1989 ; 320 : Feke t y R, Kim K H, Brow n D et al. Epidemiology of antibiotic-associated colitis; isolation of Clostridium difficile from the hospital environment. Am J Med 1981 ;70 : Skoutelis AT, Westenfelder GO, Beckerdite M et al. Hospital carpeting and epidemiology of Clostridium difficile. Am J Infect Control 1994 ;22 : G erd ing DN, Muto C A, O we ns RC. Me asure s to c ont rol and pre vent Clostridium difficile infection. Clin Infect Dis 2008 ; 46 (Suppl 1) : G erd ing DN, Joh ns on S, Pe te rs on L R et al. Clostridium difficile-associated diarrhea and colitis. Infect Control Hosp Epidemiol 1995 ;16 : Simor A E, Bra d l e y SF, St r aus b au g h L J et al. Clostridium difficile in long-term-care facilities for the elderly. Infect Control Hosp Epidemiol 2002 ;23 : Nat i ona l Clostridium difficile Standards Group. Report to the Department of Health, Health Protection Agency, United Kingdom. 25. Association for Professionals in Infection Control and Epidemiology. Guide to the elimination of Clostridium difficile in healthcare settings, Association for Professionals in infection Control and Epidemiology, Washington, D.C. 26. D ubb e rke E R, G erd i ng DN, C l ass e n D et al. Strategies to prevent Clostridium difficile infections in acute care hospitals. Infect Control Hosp Epidemiol 2008 ;29 (Suppl 1) : S Vonb erg R P, Kuijp e r E J, Wi l c ox M H et al. Infection control measures to limit the spread of Clostridium difficile. Clin Microbiol Infect 2008 ;14 (Suppl 5) : Bricker E, Garg R, Nelson R et al. Antibiotic treatment for Clostridium difficile -associated diarrhea in adults. Cochrane Database Syst Rev 2005 ;3 : CD As l am S, Ham i l l R J, Mu s h e r DM. Tre at m e nt of Clostridium difficile - associated disease: old therapies and new strategies. Lancet Infect Dis 2005 ; 5 : Nicolle LE, Bradley S, Colgan R et al. Infectious Diseases Society of America guidelines for the diagnosis and treatment of asymptomatic bacteriuria in adults. Clin Infect Dis 2005 ; 40 : C l arke M. The QUORUM statement. Lancet 2000 ;355 : Archibald LK, Banerjee SN, Jarvis WR. Secular trends in hospital - acquired Clostridium difficile disease in the United States, J Infect Dis 2004 ; 189 : Gravel D, Miller M, Simor A, et al., Canadian Healthcare-associated Infection Surveillance Program. Health care-associated Clostridium difficile infection in adults admitted to acute care hospitals in Canada: a Canadian Healthcare-associated Infection Surveillance Program Study. Clin Infect Dis 2009 ;48 : Kuijp e r E J, C oi g nard B, Tu l l P. E me rge nc e of C l o st r i d iu m d iffi cile-associated disease in North America and Europe. Clin Microbiol Infect 2006 ;12 (Suppl 6) : Sim mons B, Br y ant J, Nei man K et al. The role of handwashing in prevention of endemic intensive care unit infections. Infect Control Hosp Epidemiol 1990 ;11 : Webster J, Faoagali JL, Cartwright D. Elimination of methicillin-resistant Staphylococcus aureus from a neonatal intensive care unit after hand washing with triclosan. J Paediatr Child Health 1994 ; 30 : Pittet D, Boyce JM. Hand hygiene and patient care: pursuing the Semmelweiss legacy. Lancet Infect Dis 2001 ;1 : Ma k i D G. The use of antiseptics for handwashing by medical personnel. J Chemother 1989 ;1 : Jarvis WR. Handwashing the Semmelweis lesson forgotten? (see comments). Lancet 1994 ;344 : Pittet D, Mourouga P, Perneger TV. Compliance with handwashing in a teaching hospital. Infection Control Program. Ann Intern Med 1999 ;130 : Larson E. A causal link between handwashing and risk of infection? Examination of the evidence. Infect Control 1988 ;9 : Rotter ML. Arguments for alcoholic hand disinfection. J Hosp Infect 2001 ;48 (Suppl A) : S Rotter ML, Simpson RA, Koller W. Surgical hand disinfection with alcohols at various concentrations: parallel experiments using the new proposed European standards method. Infect Control Hosp Epidemiol 1998 ; 19 : Rotter ML. Hygienic hand disinfection. Infect Control 1984 ; 5 : Boyce JM, Pittet D. Guideline for Hand Hygiene in Health-Care Settings. Recommendations of the Healthcare Infection Control Practices Advisory Committee and the HICPAC/SHEA/APIC/IDSA Hand Hygiene Task Force. Society for Healthcare Epidemiology of America/Association for Professionals in Infection Control/Infectious Diseases Society of America. MMWR Recomm Rep 2002 ; 51 : 1 45.