Epidemiology of Clostridium difficile infections in a tertiary-care hospital in Korea

ORIGINAL ARTICLE BACTERIOLOGY Epidemiology of Clostridium difficile infections in a tertiary-care hospital in Korea J. Kim 1, J. O. Kang 2, H. Kim 3, M.-R. Seo 1, T. Y. Choi 2, H. Pai 1, E. J. Kuijper 4, I. Sanders 4 and W. Fawley 5 1) Department of Internal Medicine, College of Medicine, Hanyang University, Seoul, Korea, 2) Department of Laboratory Medicine, College of Medicine, Hanyang University, Seoul, Korea, 3) Department of Laboratory Medicine, College of Medicine, Yonsei University, Seoul, Korea, 4) National Reference Laboratory for Clostridium difficile, Department of Medical Microbiology, Leiden University Medical Centre, Leiden, the Netherlands and 5) Leeds Teaching Hospitals NHS Trust, Leeds, UK Abstract To survey healthcare-associated Clostridium difficile infection (HA-CDI) in a 900-bed tertiary-care hospital, we prospectively investigated the epidemiology of CDI and distribution of PCR-ribotypes. From February 2009 through January 2010, all patients with HA-CDI were enrolled. Epidemiological information and prescription records for antibiotics were collected. The C. difficile isolates were characterized using reference strains and were tested for antibiotic susceptibility. During the survey, incidence of HA-CDI was 71.6 per 100 000 patient-days. In total, 140 C. difficile isolates were obtained from 166 patients with HA-CDI. The PCR-ribotyping yielded 38 distinct ribotypes. The three most frequently found ribotypes made up 56.4% of all isolates; they comprised 37 isolates (26.4%) of PCR-ribotype 018, 22 (15.7%) of toxin A-negative PCR-ribotype 017, and 20 (14.3%) of PCR-ribotype 001. Clostridium difficile PCR-ribotype 018 was present in all departments throughout the hospital during the 11 months, whereas ribotype 017 and ribotype 001 appeared mostly in the pulmonary department. Hypervirulent C. difficile PCR-ribotype 027 was detected in 1 month on two wards. The incidence of CDI in each department showed a seven-fold difference, which correlated significantly with the amount of prescribed clindamycin (R = 0.783, p 0.013) or moxifloxacin (R = 0.733, p 0.025) in the departments. The rates of resistance of the three commonest ribotypes to clindamycin and moxifloxacin were significantly higher than those of other strains (92.1% versus 38.2% and 89.5% versus 27.3%, respectively). CDI is an important nosocomially acquired infection and this study emphasizes the importance of implementing country-wide surveillance to detect and control CDI in Korea. Keywords: Antibiotic usage, Clostridium difficile, epidemiology, hospital-acquired infection, ribotype Original Submission: 23 January 2012; Revised Submission: 3 May 2012; Accepted: 4 May 2012 Editor: S. Cutler Article published online: 21 May 2012 Clin Microbiol Infect 2013; 19: 521 527 10.1111/j.1469-0691.2012.03910.x Corresponding author: Hyunjoo Pai, Division of Infectious Diseases, Department of Internal Medicine, College of Medicine, Hanyang University, 232 Wangsimni-ro, Seongdong-gu, 133-791 Seoul, Korea E-mail: paihj@hanyang.ac.kr Introduction Clostridium difficile infection (CDI) is one of the most important hospital infections, and during the past decade its incidence has increased markedly worldwide [1]. Previous exposure to antimicrobials, which can disrupt the normal flora, is the main risk factor for CDI [1,2]. Clindamycin and extended-spectrum cephalosporins are the antibiotics most frequently implicated in CDI, whereas fluoroquinolones are considered risk factors for CDI caused by the hypervirulent BI/NAP1/027 (ribotype 027) strains that have become resistant to newer fluoroquinolones (moxifloxacin, levofloxacin and gatifloxacin) [2]. Additional risk factors for CDI include age, underlying co-morbidities and hospital admission [3,4]. The incidence of CDI and its molecular epidemiology vary depending on the country and over time. There has been a well-documented increase in the incidence of CDI in North America, largely attributed to the spread of ribotype 027 strains [1,5]. In Europe, the reported mean incidence of CDI was 4.1 per 10 000 patient-days, ranging from 0 to 19.1 in Clinical Microbiology and Infection ª2012 European Society of Clinical Microbiology and Infectious Diseases

522 Clinical Microbiology and Infection, Volume 19 Number 6, June 2013 CMI different countries [6]. Finland had the highest incidence, with two dominant PCR-ribotypes, 002 and 014. There are few studies of the incidence of CDI in Asia. The reported rate in Taiwan was 42.6 per 100 000 patient-days [7], and in our hospital in Korea it was 71.6 [8]. In Japan the predominant ribotype of C. difficile was shown to account for up to 64% of all isolates, but type predominance changed over several years [9,10]. One study reported ribotypes 002 and 014 to be dominant types [10]. Our objective was to examine the epidemiology of healthcare-associated CDI (HA-CDI) in our hospital. To this end we analysed the distribution of the PCR-ribotypes found in isolates from different hospital sites over 1 year, and we studied the correlation between CDI incidence and antibiotic prescriptions. Materials and Methods Patients and study design The study was conducted at the Hanyang University Hospital, a 900-bed tertiary care facility in Seoul, South Korea. All patients with diarrhoea whose stool samples were sent for CDI testing between February 2009 and January 2010 were included in the study. The study was approved by the Institutional Review Board of Hanyang University Hospital (HYUH IRB 2010-R-12). defined the prescription density of antibiotics (defined daily doses per 1000 patient-days). The date of diagnosis, ward and department of admission for each CDI patient were collected as epidemiological data. PCR-ribotyping of C. difficile isolates Stools were cultured after alcohol shock [8] on cycloserine cefoxitin fructose agar containing 0.1% taurocholic acid (CCFA-TA, Oxoid Ltd., Basingstoke, UK) supplemented with 7% horse blood. The PCR-ribotyping was performed as described elsewhere, with minor modifications [15,16]. After electrophoresis of the amplified products, the clustering of banding patterns was checked visually. Each unique pattern was assigned its own ribotype code matched against available reference strains according to international standards and according to our system [17]. A PCR-ribotype code was assigned to all PCR-ribotypes represented by more than three isolates. Ribotype 027 (BI/NAP1/027), ATCC 43598 (ribotype 017) and strains from the ECDC-Brazier collection were used for reference. Antibiotic susceptibility Susceptibility of the isolates to clindamycin and moxifloxacin was measured by the E-test (AB-BIOdisc, Solna, Sweden) as previously described [18]. Breakpoints for each compound were defined by reference to the CLSI [19]. Clostridium difficile ATCC 700057 served as a quality control strain. Definitions and collection of data A diagnosis of CDI was made when C. difficile isolates cultured from stool samples contained toxin genes (tcda, tcdb, cdta or cdtb) confirmed by multiplex PCR [11] or by a positive A & B toxin assay (VIDAS Ò C. difficile Toxin A & B; BioMérieux SA, Marcy l Etoile, France). The HA-CDI was diagnosed in patients who developed diarrhoea at least 72 h after hospitalization or within 2 months of the last discharge provided that they were not residents in a long-term facility, and they tested positive for CDI [12]. A large cluster was defined as more than ten isolates with an identical ribotype during the year, and a smallcluster, as four to nine isolates with identical ribotypes. The diversity index was defined as the number of distinct ribotypes divided by the total isolates, expressed as a percentage [13]. An outbreak was defined when the number of isolates with the same ribotype in a month exceeded double the monthly average. Incidence of CDI was defined as the number of cases per 100 000 patient-days. Prescription records for clindamycin, moxifloxacin, extended-spectrum cephalosporins, b-lactam/b-lactamase inhibitors, fluoroquinolones and total antibiotics were collected and converted to the number of defined daily doses [14]. To compare the use of antibiotics, we Statistical methods SPSS version 18.0 for Windows (SPSS Inc., Chicago, IL, USA) was used for statistical analysis. A Spearman s rank correlation analysis or the chi-square test was performed as appropriate. A p-value <0.05 in a two-tailed test was considered statistically significant. Results Over 1 year, stool samples from patients with diarrhoea were cultured for the presence of C. difficile. Out of 869 stool samples, 194 (22.3%) contained C. difficile with 54 distinct PCR-ribotypes (Fig. 1). Of the 194 isolates, 46 were non-toxigenic and eight were derived from patients with community-onset CDI. In total, 140 isolates of C. difficile were obtained from 166 patients with HA-CDI and included in the study. The incidence of HA-CDI in our hospital was 71.6 per 100 000 patient-days or 64.9 per 10 000 admissions, as previously reported [8]. Table 1 shows the distribution of the identified ribotypes. Among the 140 isolates there were 38 distinct ribotypes, and the diversity index was 27.1%. One

CMI Kim et al. Hospital epidemiology of Clostridium difficile infections 523 FIG. 1. The representative PCR-ribotype electrophoresis patterns from Clostridium difficile isolated from healthcare-associated C. difficile infection. M, 1 kb ladder; lane 1 and 5, ribotypes of isolates 47 and 83 identical with PCR-ribotype 018; lane 2, ribotype of isolate 80 identical with PCRribotype 017; lanes 3 and 4, ribotype of isolates 81 and 82 identical with PCR-ribotype 014; R1, reference strain (PCR-ribotype 017); R2, reference strain (PCR-ribotype 027); N, negative control. TABLE 1. PCR-ribotypes of Clostridium difficile isolates from healthcare-associated C. difficile infections PCR-ribotype Toxin genes Number % Cumulative % 018 A+B+ 37 26.4 26.4 017 A B+ 22 15.7 42.1 001 A+B+ 20 14.3 56.4 014 A+B+ 7 5.0 61.4 112 A+B+ 5 3.6 65.0 002 A+B+ 5 3.6 68.6 130 A+B+CDT+ 4 2.9 71.5 027 A+B+CDT+ 3 2.1 73.6 AB24 a A+B+ 3 2.1 75.7 Others 34 24.3 100 A+B+, toxin A-positive, toxin B-positive strain; A B+, toxin A-negative, toxin B- positive strain; CDT+, binary toxin-positive strain. a Ribotypes of AB24 are under study and nomenclature was adopted from the nomenclature system of Kim et al. [17]. hundred and sixteen of the 140 isolates (82.9%) shared a ribotype with an isolate from at least one other patient. Three large-cluster isolates were identified, comprising 56.4% of all isolates (79/140). Thirty-seven of these isolates (26.4%) belonged to ribotype 018, 22 (15.7%) to ribotype 017 and 20 (14.3%) to ribotype 001. There were four small-cluster isolates, corresponding to 15.0% of all isolates (21/140). The small-cluster isolates were isolated from between four and seven patients, and their ribotypes were 014, 112, 002 and 130. The cumulative percentage of large- and small-cluster isolates was 71.4% (100/140) of the total number of isolates (Table 1). Eleven isolates produced binary toxins (including ribotypes 130, 122 and 027), and three ribotype 027. Another quarter of all ribotypes (34/140 or 24.2%) were found in only one (17.1%) or two (7.1%) isolates. Incidence of CDI and the use of antibiotics The incidence of CDI showed a seven-fold difference between hospital departments, ranging from 26 to 186 per 100 000 patient-days (Table 2). We studied the correlation between incidence and the amount of prescribed antibiotics in each department. Use of clindamycin or moxifloxacin correlated significantly with the incidence of CDI in each department: clindamycin (R = 0.783, p 0.013) and moxifloxacin TABLE 2. The correlation between the incidence of Clostridium difficile infection (CDI) and defined daily dose (DDD) of prescribed antibiotics by departments Department PD GE GS HO NE NR NS PS RM CDI/10 5 patient-days 134.93 33.21 48.83 79.93 185.77 74.61 101.01 26.10 52.11 Spearman s rho p value DDD/1000 patient-days Clindamycin 157.4 2.6 3.6 10.6 7.3 25.2 7.2 1.1 2.7 0.783 0.013 Moxifloxacin 172.1 3.9 7.6 17.2 17.4 19.6 12.1 4.9 2.6 0.733 0.025 Fluoroquinolones a 470.2 110.6 57.6 81.5 176.1 66.5 42.6 34.7 81.2 0.550 0.125 ESC 222.6 132.9 235.1 145.0 165.0 83.3 581.4 275.3 64.5 )0.017 0.966 Cephalosporins b 333.3 160.1 371.1 173.0 213.6 95.3 604.2 939.2 144.4 )0.050 0.898 BLI 45.5 39.4 25.9 79.4 60.5 55.9 31.7 93.1 48.6 0.000 1.000 Total antibiotics c 1135.1 482.9 632.5 495.1 607.4 271.6 784.5 1187.5 366.9 0.050 0.898 ESC, extended-spectrum cephalosporins; BLI, b-lactam/b-lactamase inhibitors; PD, pulmonary department; GE, gastroenterology; GS, general surgery; HO, haemato-oncology; NE, nephrology; NR, neurology; NS, neurosurgery; PS, plastic surgery; RM, rheumatology. a Total fluoroquinolones included moxifloxacin. b Total cephalosporins included extended-spectrum cephalosporins. c Total antibiotics included cephalosporins, fluoroquinolones, b-lactam/b-lactamase inhibitors, carbapenems, clindamycin, colistin, glycopeptides, metronidazole, tetracyclines, trimethoprim/sulphamethoxazole, and linezolid.

524 Clinical Microbiology and Infection, Volume 19 Number 6, June 2013 CMI (R = 0.733, p 0.025) (Table 2). The prescription densities of clindamycin and moxifloxacin in different departments ranged from 1.0 to 157.4 and 2.6 to 172.1 per 1000 patient-days, respectively. Use of fluoroquinolones (including moxifloxacin), extended-spectrum cephalosporins, cephalosporins (including extended-spectrum cephalosporins) or b-lactam/ b-lactamase inhibitors did not show a correlation with the incidence of CDI: fluoroquinolones (R = 0.550, p 0.125), extended-spectrum cephalosporins (R = )0.017, p 0.966), cephalosporins (R = )0.050, p 0.898) or b-lactam/b-lactamase inhibitors (R = 0.000, p 1.000). The total amount of oral or intravenous antibiotics prescribed in each department did not correlate with the incidence of CDI either (R = 0.050, p 0.898), The monthly incidence of CDI varied from 52 to 114 per 100 000 patient-days during the study period [8]. We failed to find correlation between the monthly incidence of CDI and prescription of clindamycin, moxifloxacin, fluoroquinolones, extended-spectrum cephalosporins, cephalosporins, b-lactam/b-lactamase inhibitors and total antibiotics per month: clindamycin (R = 0.049, p 0.880), moxifloxacin (R = )0.042, p 0.897), fluoroquinolones (R = )0.042, p 0.898), extended-spectrum cephalosporin (R = )0.406, p 0.191), b-lactam/b-lactamase inhibitors (R = )0.098, p 0.762) and total antibiotics (R = )0.601, p 0.342). Distribution of PCR-ribotypes over time and by hospital ward Fig. 2(a) shows the monthly distribution of the identified ribotypes. Ribotype 018 isolates appeared continuously throughout the year, with an average of 3.08 per month. An outbreak of CDI associated with ribotype 018 occurred in seven patients in August 2009. Another large cluster strain involved ribotype 017 affecting 11 patients during November and December 2009. Ribotype 001 strains were isolated at a rate of 1.7 per month, and small outbreaks were suspected in June and October 2009 and in January 2010. Fig. 2(b) shows the distribution of ribotypes by hospital ward. Ribotype 018 appeared in all wards, whereas ribotype 017 appeared mainly in wards 16 and 18, and ribotype 001 in ward 16. The ribotypes of ten isolates from patients with healthcare-associated CDI, occurring within 2 months of discharge, were identified as hospital-endemic strains. The mean period from the last discharge to admission for the patients with HA-CDI was 16.6 days (5 39 days). Five of the patients developed CDI for the first time, three experienced their first relapses, and two experienced second or further relapses. A histogram of ribotype distribution by departments is shown in Fig. 2(c). It shows that certain departments were more likely to harbour endemic strains: ribotype 018 in the nephrology and general surgery departments (55.0% and 60.0% of all isolates from each department, respectively), ribotype 017 in the pulmonary, haemato-oncology and neurology departments (29.6%, 30.8% and 50.0%, respectively) and ribotype 001 in the pulmonary department (37.0%). Of the 11 isolates producing binary toxins, eight (72.7%) were detected in the neurosurgery department, and ribotype 130 isolates came only from that department. All PCR-ribotype N 20 15 (a) (b) (c) N 15 N Ribotype 25 018 017 001 014 112 20 002 130 AB24 027 Other 15 10 10 10 5 5 5 2009 2010 Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. Feb. Jan. Period 8 11 12 15 16 18 19 H M S Ward PD GE GS HO ID NE NR NS PS RM Depart. FIG. 2. Ribotype distribution of Clostridium difficile isolates over time and hospital location. (a) Monthly distribution of ribotypes, (b) distribution of ribotypes according to hospital ward, (c) distribution of ribotypes according to admission department; for (a) all the isolates were included, for (b) and (c) hospital ward or department with more than three isolates were included. H, healthcare-associated C. difficile infection (CDI) (community onset CDI within 2 months of discharge from hospital); M, medical intensive-care unit; S, surgical intensive-care unit; PD, pulmonary department; GE, gastroenterology; GS, general surgery; HO, haemato-oncology; ID, infectious diseases; NE, nephrology; NR, neurology; NS, neurosurgery; PS, plastic surgery; RM, rheumatology.

CMI Kim et al. Hospital epidemiology of Clostridium difficile infections 525 027 strains were detected in 1 month, whereas ribotype 130 brought about four episodes in consecutive months in the neurosurgery department. PCR-ribotype 122, a binary toxinproducing strain related to ribotype 027 [9], was isolated from one patient in the neurosurgery department after 13 days of hospitalization. Comparison of antibiotic resistance in the large-cluster isolates with other isolates The cut-off values for resistance to clindamycin and moxifloxacin were defined as MIC 8 mg/l by reference to the CLSI [19]. We categorized the isolates as large-cluster isolates and other isolates, and compared rates of antibiotic resistance in the two groups. Antibiotic susceptibility was tested in 76 out of 79 large-cluster isolates and in 55 out of 61 other isolates. Clindamycin resistance in the large-cluster isolates reached 92.1%, compared with 38.2% in the other isolates (p <0.0001), and the frequency of resistance to moxifloxacin was three-fold higher in the large-cluster isolates than in the other isolates (89.5% and 27.3%, respectively; p <0.0001). There was no significant correlation between prescription of clindamycin or moxifloxacin in each department and the rates of resistance of the relevant isolates to clindamycin or moxifloxacin (R = 0.162, p 0.678 and R = 0.226, p 0.559, respectively). Discussion We studied all CDI cases that occurred in our care facility over 1 year, and characterized the C. difficile isolates. This is the first prospective surveillance study of CDI in Korea. The incidence of CDI per 100 000 patient-days during the study period was 71.6 [8], which was similar to the rate recorded during high-incidence years in a tertiary-care centre in the USA before the B1/NAP1/027 strain outbreak [13]. However, compared with the incidence of 41 per 100 000 patient-days recently reported in a surveillance study in 34 European countries, our rate is high and deserves more local and national attention [6]. Several hospital-endemic C. difficile strains were identified, and these made up a large proportion of all isolates. The diversity index was 27.1% and the predominant ribotype strain accounted for 26.4% of all cases. These findings point to simultaneous occurrence of endemic infections and outbreaks in our hospital. Ribotype 018, the predominant type in our hospital, was also found to be prevalent in Italy, where it was associated with more severe disease [6]. Toxin A-negative ribotype 017 strains were known to be prevalent in East Asia, whereas their prevalence in Europe and North America was as low as 0.2 8% [6,10,20 22]. Studies of the outcomes and clinical manifestations of CDI caused by toxin A-negative toxin B-positive ribotype 017 strains have reported conflicting results: some found no difference from other ribotypes [10,23 25], while others reported more severe disease and significantly worse outcomes [24,26]. Ribotypes 001 and 014 were reported as prevalent in Europe (16% and 10%, respectively), where they were distributed over a large area [6]. The most common type in Germany, ribotype 001 (70%), was sixth in southern Scotland and second in Ireland [18,27,28]. These strains were shown to be resistant to multiple antibiotics, including moxifloxacin, erythromycin, ceftriaxone and clindamycin, which could explain the sudden clonal expansion of ribotype 001 [27]. Ribotype 014 is almost identical to ribotype 020, and some studies represent them as 014/020 [6]. PCR-ribotype 014 was reported as less prevalent than 001, and with less frequent resistance to antibiotics [27]. We identified three C. difficile ribotype 027 isolates and sequencing of tcdc in these isolates revealed mutations that were specific to the NAP1/027 strains. However, as all three isolates were sensitive to moxifloxacin (range of MICs, 1 2 mg/l), they were considered to resemble historical 027 strains [5]. Because they were also susceptible to clindamycin (data not shown), we believe that they did not have a selective advantage over the endemic large-cluster isolates, and therefore did not create a large outbreak or persist in the hospital. There was a seven-fold difference between the lowest and the highest CDI incidence in different departments. The most probable cause of this difference is the amount of antibiotics prescribed, although the spread of C. difficile through environmental contamination or the hands of healthcare personnel could also be a factor. Our results demonstrate a correlation between the incidence of CDI and the use of clindamycin or moxifloxacin in all departments. Previous reports indicated that extended-spectrum cephalosporins and clindamycin were most frequently associated with higher rates of CDI [1,3]. It is possible that despite the hospitalwide prescription of extended-spectrum cephalosporins, b-lactam/b-lactamase inhibitors and fluoroquinolones, the skewed use of clindamycin and moxifloxacin (mostly in the pulmonary department) led to a statistically significant result. Nevertheless, considering the high rates of resistance of large-cluster isolates to clindamycin and moxifloxacin, in addition to the significant correlation between these antibiotics and CDI incidence in this study, it is more likely that clindamycin and moxifloxacin were imposing strong selective pressures on C. difficile, as shown elsewhere [5,26]. A recent report described how a hospital outbreak of resistant C. difficile was successfully controlled by restricting the use of

526 Clinical Microbiology and Infection, Volume 19 Number 6, June 2013 CMI fluoroquinolone [29], showing that it is necessary to implement antibiotic stewardship and infection control practices to reduce the incidence of HA-CDI. There are two potential limitations to our study. Because CDI was diagnosed by stool culture or toxin assay only when requested, the incidence of CDI may be underestimated. Second, although the CDI rate was very high in the nephrology department, that department did not consume significantly larger amounts of clindamycin, moxifloxacin, extended-spectrum cephalosporins or total antibiotics. In addition to antibiotics, other risk factors such as age, co-morbidity and hospital admission could strongly influence the incidence of CDI. Differences of age, co-morbidities and other risk factors between patients in the different departments are to be expected; however, they were not evaluated in this study. This is the first prospective surveillance study of CDI in Korea, and it demonstrates that CDI is an important hospital-acquired infection. Our findings stress the importance of implementing a country-wide CDI surveillance and of introducing infection control measures. Acknowledgements This work was supported by a grant from the National Research Foundation of Korea (KRF-2011-0014685).We thank Thomas V. Riley for matching PCR-ribotyping with their laboratory PCR-ribotype standards. This work was presented at Chicago ICAAC, 2011 Transparency Declaration None to declare References 1. Freeman J, Bauer MP, Baines SD, Corver J, Fawley WN, Goorhuis B, et al. The changing epidemiology of Clostridium difficile infections. Clin Microbiol Rev 2010; 23: 529 549. 2. Razavi B, Apisarnthanarak A, Mundy LM. Clostridium difficile: emergence of hypervirulence and fluoroquinolone resistance. Infection 2007; 35: 300 307. 3. Dubberke ER, Reske KA, Yan Y, Olsen MA, McDonald LC, Fraser VJ. Clostridium difficile-associated disease in a setting of endemicity: identification of novel risk factors. Clin Infect Dis 2007; 45: 1543 1549. 4. Brown E, Talbot GH, Axelrod P, Provencher M, Hoegg C. Risk factors for Clostridium difficile toxin-associated diarrhea. Infect Control Hosp Epidemiol 1990; 11: 283 290. 5. McDonald LC, Killgore GE, Thompson A, Owens RC Jr, Kazakova SV, Sambol SP, et al. An epidemic, toxin gene-variant strain of Clostridium difficile. N Engl J Med 2005; 353: 2433 2441. 6. Bauer MP, Notermans DW, van Benthem BH, Brazier JS, Wilcox MH, Rupnik M, et al. Clostridium difficile infection in Europe: a hospital-based survey. Lancet 2011; 377: 63 73. 7. Chung CH, Wu CJ, Lee HC, Yan JJ, Chang CM, Lee NY, et al. Clostridium difficile infection at a medical center in southern Taiwan: incidence, clinical features and prognosis. J Microbiol Immunol Infect 2010; 43: 119 125. 8. Kim J, Pai H, Seo MR, Kang JO. Epidemiology and clinical characteristics of Clostridium difficile infection in a Korean tertiary hospital. J Korean Med Sci 2011; 26: 1258 1264. 9. Sawabe E, Kato H, Osawa K, Chida T, Tojo N, Arakawa Y, et al. Molecular analysis of Clostridium difficile at a university teaching hospital in Japan: a shift in the predominant type over a five-year period. Eur J Clin Microbiol Infect Dis 2007; 26: 695 703. 10. Iwashima Y, Nakamura A, Kato H, Wakimoto Y, Wakiyama N, Kaji C, et al. A retrospective study of the epidemiology of Clostridium difficile infection at a University Hospital in Japan: genotypic features of the isolates and clinical characteristics of the patients. J Infect Chemother 2010; 16: 329 333. 11. Persson S, Torpdahl M, Olsen KE. New multiplex PCR method for the detection of Clostridium difficile toxin A (tcda) and toxin B (tcdb) and the binary toxin (cdta/cdtb) genes applied to a Danish strain collection. Clin Microbiol Infect 2008; 14: 1057 1064. 12. Mulvey MR, Boyd DA, Gravel D, Hutchinson J, Kelly S, McGeer A, et al. Hypervirulent Clostridium difficile strains in hospitalized patients, Canada. Emerg Infect Dis 2010; 16: 678 681. 13. Belmares J, Johnson S, Parada JP, Olson MM, Clabots CR, Bettin KM, et al. Molecular epidemiology of Clostridium difficile over the course of 10 years in a tertiary care hospital. Clin Infect Dis 2009; 49: 1141 1147. 14. World Health Organization. The anatomical therapeutic chemical (ATC) and defined daily dosing (DDD) system index 2010. 2010; Available from: http://www.whocc.no/ (Accessed on 28 Oct 2010). 15. Bidet P, Barbut F, Lalande V, Burghoffer B, Petit JC. Development of a new PCR-ribotyping method for Clostridium difficile based on ribosomal RNA gene sequencing. FEMS Microbiol Lett 1999; 175: 261 266. 16. Stubbs SL, Brazier JS, O Neill GL, Duerden BI. PCR targeted to the 16S-23S rrna gene intergenic spacer region of Clostridium difficile and construction of a library consisting of 116 different PCR ribotypes. J Clin Microbiol 1999; 37: 461 463. 17. Kim H, Jeong SH, Roh KH, Hong SG, Kim JW, Shin MG, et al. Investigation of toxin gene diversity, molecular epidemiology, and antimicrobial resistance of Clostridium difficile isolated from 12 hospitals in South Korea. Korean J Lab Med 2010; 30: 491 497. 18. Solomon K, Fanning S, McDermott S, Murray S, Scott L, Martin A, et al. PCR ribotype prevalence and molecular basis of macrolide-lincosamide-streptogramin B (MLSB) and fluoroquinolone resistance in Irish clinical Clostridium difficile isolates. J Antimicrob Chemother 2011; 66: 1976 1982. 19. Clinical and Laboratory Standards Institute. Methods for antimicrobial susceptibility testing of Anaerobic bacteria, Approved standard 7th edn. CLSI document M11-A7; Wayne, PA: CLSI, 2007. 20. Cheknis AK, Sambol SP, Davidson DM, Nagaro KJ, Mancini MC, Hidalgo-Arroyo GA, et al. Distribution of Clostridium difficile strains from a North American, European and Australian trial of treatment for C. difficile infections: 2005 2007. Anaerobe 2009; 15: 230 233. 21. Lyerly DM, Neville LM, Evans DT, Fill J, Allen S, Greene W, et al. Multicenter evaluation of the Clostridium difficile TOX A/B TEST. J Clin Microbiol 1998; 36: 184 190. 22. Kim H, Riley TV, Kim M, Kim CK, Yong D, Lee K, et al. Increasing prevalence of toxin A-negative, toxin B-positive isolates of Clostridium difficile in Korea: impact on laboratory diagnosis. J Clin Microbiol 2008; 46: 1116 1117.

CMI Kim et al. Hospital epidemiology of Clostridium difficile infections 527 23. Komatsu M, Kato H, Aihara M, Shimakawa K, Iwasaki M, Nagasaka Y, et al. High frequency of antibiotic-associated diarrhea due to toxin A-negative, toxin B-positive Clostridium difficile in a hospital in Japan and risk factors for infection. Eur J Clin Microbiol Infect Dis 2003; 22: 525 529. 24. Arvand M, Hauri AM, Zaiss NH, Witte W, Bettge-Weller G. Clostridium difficile ribotypes 001, 017, and 027 are associated with lethal C. difficile infection in Hesse, Germany. Euro Surveill 2009; 14: 19403. 25. Alfa MJ, Kabani A, Lyerly D, Moncrief S, Neville LM, Al-Barrak A, et al. Characterization of a toxin A-negative, toxin B-positive strain of Clostridium difficile responsible for a nosocomial outbreak of Clostridium difficile-associated diarrhea. J Clin Microbiol 2000; 38: 2706 2714. 26. Goorhuis A, Debast SB, Dutilh JC, van Kinschot CM, Harmanus C, Cannegieter SC, et al. Type-specific risk factors and outcome in an outbreak with 2 different Clostridium difficile types simultaneously in 1 hospital. Clin Infect Dis 2011; 53: 860 869. 27. Taori SK, Hall V, Poxton IR. Changes in antibiotic susceptibility and ribotypes in Clostridium difficile isolates from southern Scotland, 1979 2004. J Med Microbiol 2010; 59: 338 344. 28. Zaiss NH, Witte W, Nubel U. Fluoroquinolone resistance and Clostridium difficile, Germany. Emerg Infect Dis 2010; 16: 675 677. 29. Aldeyab MA, Devine MJ, Flanagan P, Mannion M, Craig A, Scott MG, et al. Multihospital outbreak of Clostridium difficile ribotype 027 infection: epidemiology and analysis of control measures. Infect Control Hosp Epidemiol 2011; 32: 210 219.