Transmission dynamics of ESBL-producing Escherichia coli clones in rehabilitation wards at a tertiary care centre

ORIGINAL ARTICLE EPIDEMIOLOGY Transmission dynamics of ESBL-producing Escherichia coli clones in rehabilitation wards at a tertiary care centre A. Adler 1, M. Gniadkowski 2, A. Baraniak 2, R. Izdebski 2, J. Fiett 2, W. Hryniewicz 2, S. Malhotra-Kumar 3, H. Goossens 3, C. Lammens 3, Y. Lerman 4, M. Kazma 1, T. Kotlovsky 1, Y. Carmeli 1 and the MOSAR WP5 and WP2 study groups* 1) Section of Epidemiology, Tel-Aviv Souraski Medical Center, Tel-Aviv, Israel, 2) National Medicines Institute, Warsaw, Poland, 3) Laboratory of Medical Microbiology, Vaccine and Infectious Disease Institute, University of Antwerp, Antwerp, Belgium and 4) Geriatric Division, Tel-Aviv Sourasky Medical Center, Tel-Aviv, Israel Abstract Increasing resistance due to the production of ESBL in Escherichia coli (ESBL-E. coli) has become a major threat to public health. Our aims were to study the incidence of ESBL-E. coli acquisition during hospitalization and the transmission rates of different ESBL- E. coli clones. This was a prospective case control study, conducted in two geriatric rehabilitation wards in Tel-Aviv. Serial rectal cultures were collected from admission till discharge. All patient-unique ESBL-E. coli isolates were subjected to molecular typing by PFGE, MLST and determination of ESBL genes. An acquisition of ESBL-E. coli was defined as traceable when a patient with the same ST, PFGE type and ESBL gene was hospitalized in the same ward in parallel to the acquisition case. ESBL-E. colis were recovered from 125 patients out of 492 enrolled: 52 were recovered upon admission, 59 acquired ESBL-E. coli during their stay, and there was undetermined status in 14 patients. A low Norton s score was associated with acquisition (O.R. 1.14 for each point, 95% C.I. 1.01 1.29, p < 0.05). ESBL-E. coli infections (n = 9) had occurred only in ESBL-E. coli carriers. The pandemic ST131 clone was the most common (48/125). The majority of the isolates (101/125) produced CTX-M-type ESBL. Of the 59 acquisition cases, 32 were traced to another patient. In-hospital dissemination was highest in the CTX-M-27-producing ST131 and the SHV-5-producing ST372 sub-clones (acquisition/admission ratios of 17/11 and 9/3, respectively), with almost all cases traced to other patients. In conclusion, most ESBL-E. coli acquisition cases were traceable to other patients. The transmission potential varied significantly between ESBL- E. coli clones. Keywords: Clones, E. coli, ESBL, rehabilitation wards, transmission Original Submission: 8 March 2012; Revised Submission: 22 July 2012; Accepted: 23 July 2012 Editor: R. Cantón Article published online: 27 July 2012 Clin Microbiol Infect 2012; 18: E497 E505 10.1111/j.1469-0691.2012.03999.x Corresponding author: A. Adler, Division of Epidemiology, Tel-Aviv Sourasky Medical Center, 6 Weizmann Street, Tel-Aviv, Israel 64239 E-mail: amosa@tasmc.health.gov.il *The MOSAR WP5 and WP2 study groups are listed in the appendix. Introduction Escherichia coli is the most common aetiological agent of urinary tract infection in humans of all age groups. In addition, it is a frequent cause of other types of infections, ranging from intra-abdominal infections to neonatal meningitis [1]. In the last three decades, we have witnessed increasing resistance due to the production of extended-spectrum b-lactamase (ESBL) in E. coli (ESBL- E. coli), in both the community and healthcare settings [2]. Several clones have been identified recently as leading causes of ESBL- E. coli infections globally, most notably the uropathogenic sequence type 131 (ST131), belonging to the phylogroup B2 [2]. Mechanisms of transmission and spread of ESBL- E. coli in healthcare settings are not fully understood, especially outside the boundaries of a defined outbreak. Although person-to-person transmission has been commonly perceived as the most likely mode, this opinion was not supported in studies that sought to Clinical Microbiology and Infection ª2012 European Society of Clinical Microbiology and Infectious Diseases

E498 Clinical Microbiology and Infection, Volume 18 Number 12, December 2012 CMI correlate ESBL- E. coli acquisition cases with previously known carriers [3,4]. However, these studies were conducted in intensive care units (ICUs) [3 5] where direct patient-to-patient contacts are uncommon; therefore, indirect transmission via healthcare providers or equipment may play a more important role. In the present study, we examined the transmission dynamics of ESBL- E. coli in two rehabilitation wards (RWs) in a hospital in Tel-Aviv (Israel). RWs constitute an excellent and so far non-exploited model for such analyses. Patients admitted to RWs are commonly hospitalized for longer periods, and are more ambulant and interactive. Therefore, the potential for direct transmission between patients may be remarkable, as well as the opportunity to track it. Our objectives were to study (i) the prevalence of ESBL- E. coli on admission and the incidence of its acquisition during hospitalization; (ii) different risk factors for ESBL- E. coli acquisition; (iii) the transmission rate of different ESBL- E. coli clones identified by molecular typing; and (iv) the role of patient-to-patient transmission in the acquisition of ESBL- E. coli. Methods Settings This study was a part of the project MOSAR (Mastering Hospital Antimicrobial Resistance and its Spread into the Community), a trans-disciplinary network funded by the European Commission, and devoted to combating and controlling resistance in bacteria. The project is focused on endemic and epidemic nosocomial pathogens in high-risk medical units, including ICUs, RWs and surgery wards in different European countries and Israel. The study was conducted in two geriatric RWs (50 beds combined) at the Tel-Aviv Sourasky Medical Center (TASMC), and included elderly patients (>65 years of age) who were admitted to the wards between October 2008 and May 2009 because of orthopaedic or neurological disorders. The patients are mostly admitted requiring heavy assistance in daily activities. Some are incontinent and most are wheelchair bound. As they improve, they are able to be partially independent, and towards discharge some will be independent or almost independent and the rest will still need moderate help. Most of the patients will be ambulated on discharge but some will need close observation while walking. During rehabilitation, the patients stay in a two-bed room. Patients spend most of the day in a common room, where they conduct various activities including occupational and recreational therapy, socialize with each other and dine. The patients are transferred also to specialized areas such as physical therapy and the pool. The study was approved by the ethics committee of the TASMC. Design and data collection This was a prospective study, aimed at examining risk factors for ESBL- E. coli carriage among RW patients. The study had two parts: (i) a case control study, where ESBL- E. coli carriers on admission were compared with a control group, which included patients without ESBL- E. coli on admission; (ii) a cohort study, where patients not carrying ESBL-Ent on admission were followed for the acquisition of ESBL- E. coli (see definition ). Surveillance rectal cultures were collected from all patients at admission, 2 weeks later, then once monthly, and at discharge. According to the local infection control policy, contact isolation was not carried out for ESBL- E. coli carriers. The following data were recorded: demographic (patient s age and sex), medical history including underlying conditions and co-morbidities, prior hospital or long-term care facility (LTCF) stay and its duration, Norton score [6], antibiotic treatment during the last month prior to admission, the presence of medical devices, history of surgery or other invasive procedures, the incidence of infection due to ESBL- E. coli, and the discharge destination. Molecular typing and identification of ESBL genes were performed on all patient-unique isolates as described below. Detection of ESBL-EC isolates and their phenotypic characterization Rectal swabs were streaked onto BrillianceÔ ESBL Agar (Oxoid, Basingstoke, UK). Putative ESBL-producing Enterobacteriaceae (ESBL-Ent) colonies were identified according to the manufacturer s instructions. Pure cultures were frozen at )80 C and shipped to the MOSAR ESBL central laboratory (National Medicines Institute in Warsaw, Poland) for definite identification and further analysis. Species identification was carried out using the Vitek 2 system (biomérieux, Marcy l Etoile, France). ESBL production was verified using the doubledisk synergy test with disks containing cefotaxime, ceftazidime, cefepime and amoxicillin with clavulanate on Mueller Hinton agar plates (Oxoid) that were not supplemented and supplemented with 250 mg/l cloxacillin as previously described [7]. Molecular typing of ESBL-EC isolates Pulsed-field gel electrophoresis (PFGE) was performed according to Struelens et al. [8]. PFGE types and subtypes were discerned by visual analysis using the criteria of Tenover et al. [9]. In order to verify the results, electrophoretic patterns were compared also with the BioNumerics Fingerprinting software (Version 6.01, Applied Maths, Sint-Martens- Latem, Belgium), using the Dice coefficient and clustering by

CMI Adler et al. Transmission dynamics of E. coli clones E499 UPGMA (unweighted pair group method with arithmetic mean) with 1% tolerance in band position differences. Representative E. coli isolates of all PFGE types were subjected to multilocus sequence typing (MLST) as described previously [10]. A database available at http://mlst.ucc.ie was used for assigning sequence types (STs) and clonal complexes. Identification of ESBLs ESBL- E. coli isolates were subjected to b-lactamase profiling by isoelectric focusing as described previously [11] by using a Model 111 Mini IEF Cell (Bio-Rad, Hercules, CA, USA). Identification of the ESBL bla CTX-M-1 -, bla CTX-M-2 -, bla CTX-M-9 -, bla CTX-M-25 - and bla SHV -like genes was carried out by PCR as previously described [12]. Sequencing of the genes was performed for representative isolates as reported [12,13], using sets of consecutive primers specific for each gene type. Definitions and data analysis ESBL- E. coli carriers were divided into admission and acquisition groups, according to the ESBL- E. coli identification time: before and after 72 h from admission, respectively. In cases when the first rectal culture was collected more than 72 h from admission, the acquisition status was not determined. An acquisition of ESBL- E. coli was defined as traceable when a patient with E. coli of the same ST, PFGE type and ESBL gene was hospitalized in the same ward in parallel to the acquisition case; in these cases, patient-topatient transmission was assumed. Risk factors were analysed by comparing the ESBL- E. coli carriers of the admission and acquisition groups with an ESBL- Ent-negative group. The control group was selected randomly from the overall ESBL-Ent-negative patients, using random study numbers. Equivalent numbers of control and case patients were selected. Data were analysed using univariate analysis: continuous variables were compared between the groups using an unpaired t-test within each group. Categorical parameters were compared by using the Pearson v 2 test. p-values of 0.05 were considered as a significant difference between the groups. Multivariate analysis using binary logistic regression prediction models was constructed using forward stepwise selection. All data were analysed using the SPSS software package version 15.0 (SPSS, Chicago, IL, USA). Results Demographic and clinical characteristics of ESBL- E. coli carriers at TASMC During the study period, 492 patients were enrolled in the two RWs; ESBL- E. coli isolates were recovered from 125 patients. Fifty-two patients were colonized upon admission (admission group), 59 patients acquired ESBL- E. coli during their stay (acquisition group), and the carriage origin status was undetermined in 14 patients. Of the 367 ESBL-Ent patients, 52 were randomly selected as controls. The demographic and clinical characteristics according to the ESBL- E. coli carriage status are presented in Table 1. The mean (range) age was 83 years (65 105), 62 were male (35%) and the mean (range) length of stay prior to admission was 14 days (1 119). Longer length of stay at an acute-care facility prior to rehabilitation, incontinence, low Norton score and cardiovascular disease (CVD) were more common in the acquisition group when compared with the negative group. In the multivariate analysis, a low Norton score remained a significant risk factor for the ESBL- E. coli acquisition (OR, 1.14 for each point; 95% CI, 1.01 1.29; p < 0.05). No significant differences were found between the groups in the overall use of antimicrobial therapy in the preceding month and in the use of penicillins, b-lactamase+b-lactamase inhibitors, cephalosporins, carbapenems, quinolones, macrolides and aminoglycosides. Transfer to an LCTF or acute-care facility was significantly more frequent in both the admission and acquisition groups when compared with the negative group. Two patients in the admission group passed away. Infections caused by ESBL- E. coli, including three bloodstream infections, occurred in nine patients; four patients were in the acquisition and admission groups each, and one was in the group with the undetermined carriage status. No ESBL- E. coli infections occurred among all the 367 ESBL- E. coli -negative patients hospitalized in the RWs during the study. In addition, we compared the molecular characteristics of the E. coli strains (STs and ESBL types) between the groups according to their defined carriage status. We found that the rate of strains producing SHV-type ESBLs (SHV-5, - 12) was significantly higher in the acquisition than in the admission group: 16/59 vs. 6/52 (p < 0.05). Molecular characteristics of ESBL- E. coli The results of the molecular analysis of the 125 ESBL- E. coli isolates are shown in Table 2. The isolates were classified into 26 STs (clones) and 49 PFGE types. Thirteen clones were represented by more than one isolate and these were mainly the clones that have been disseminated globally and/ or belonged to widespread clonal complexes, for example ST131 (n = 48; 38.5%), ST398 (n = 9; 7%), ST38 (n = 8; 6.5%), ST405 (n = 8; 6.5%), ST69 (n = 6; 5%), ST648 (n =6; 5%), ST10 (n = 3; 2.5%), ST410 (n = 3; 2.5%) or ST354 (n = 2; 1.5%). The second most prevalent clone was ST372 (n = 13; 10.5%). Most of these clones were differentiated into several PFGE types; larger clusters belonged to ST131

E500 Clinical Microbiology and Infection, Volume 18 Number 12, December 2012 CMI TABLE 1. Demographic and clinical characteristics of patients according to ESBL-producing E. coli (ESBL-E. coli) acquisition status ESBL-E. coli acquisition status Variable Negative (n = 52) Admission (n = 52) Admission vs. Negative Acquisition vs. Negative OR (95% CI) b p-value Acquisition (n = 59) OR (95% CI) b p-value Male, n (%) 18 (35) 18 (35) 1 (0.44 2.44) NS a 22 (37) 1.12 (0.51 2.44) NS Mean age, years (95% CI) 81 (79 84) 84 (82 86) 1.04 (0.98 1.09) NS 83 (81 85) 1.02 (0.97 1.08) NS LTCF stay in last 6 months, n (%) 2 (4) 2 (4) 1 (0.13 7.38) NS 5 (8) 2.31 (0.43 12.47) NS Mean duration of hospitalization prior to 11 (10 13) 13 (10 16) 1.02 (0.97 1.06) NS 19 (12 25) 1.03 (1 1.07) <0.05 rehabilitation, days (95% CI) Full continence on admission, n (%) 33 (63) 32 (61) 0.92 (0.41 2.03) NS 25 (43) 0.42 (0.2 0.9) <0.05 Mean Norton scale (95% CI) 14.9 (14 15.9) 14.31 (13.2 15.5) 0.95 (0.85 1.05) NS 13.3 (12.3 14.3) 0.87 (0.77 0.98) <0.05 Active infection on admission, n (%) 4 (8) 4 (8) 1 (0.23 4.23) NS 7 (12) 1.61 (0.44 5.86) NS Cardiovascular disease 36 (69) 43 (83) 2.12 (0.83 5.37) NS 51 (86) 2.83 (1.09 7.32) <0.05 (including hypertension), n (%) Congestive heart failure, n (%) 3 (6) 4 (8) 1.36 (0.28 6.4) NS 4 (7) 1.18 (0.25 5.57) NS History of a stroke, n (%) 9 (17) 10 (19) 1.13 (0.42 3.07) NS 14 (24) 1.48 (0.58 3.78) NS Chronic lung disease, n (%) 5 (10) 12 (23) 2.23 (0.7 7.07) NS 4 (7) 0.68 (0.17 2.69) NS Renal impairment, n (%) 9 (17) 10 (19) 1.13 (0.42 3.07) NS 9 (15) 0.86 (0.31 2.36) NS Diabetes, n (%) 20 (38) 12 (23) 0.48 (0.20 1.12) NS 14 (24) 0.49 (0.21 1.13) NS History of malignancy, n (%) 8 (15) 8 (15) 1 (0.34 2.9) NS 14 (24) 1.71 (0.65 4.48) NS Urinary catheter/other invasive device, n (%) 47 (90) 43 (83) 0.5 (0.15 1.63) NS 47 (80) 0.41 (0.13 1.27) NS Surgery/invasive procedure in last year, n (%) 41 (79) 35 (67) 0.55 (0.22 1.33) NS 39 (66) 0.52 (0.22 1.23) NS Antibiotic in the past month, n (%) 40 (77) 37 (71) 0.74 (0.3 1.78) NS 49 (83) 1.47 (0.57 3.75) NS Antibiotic Rx on admission, n (%) 10 (19) 5 (10) 0.44 (0.14 1.41) NS 8 (14) 0.65 (0.23 1.81) NS Mean duration of hospitalization in 29 (24 33) 24 (20 29) 0.67 (0.31 1.47) NS 35 (29 41) 1.15 (0.54 2.45) NS rehabilitation ward, days (95% CI) Transfer to acute care/ltcf, n (%) 1 (2) 7 (14) 8.5 (1 71) <0.05 9 (15) 9.18 (1.12 75) <0.05 LTCF, long-term care facility. a NS, p-value > 0.05. b OR and 95% CI for continuous variables are calculated per 1 unit. ST (n) ESBL genes (n) PFGE a types Acquisition types (AD, AQ, ND) 131 (48) CTX-M-15 (14) 4 AD-8, AQ-5, ND-1 2 CTX-M-27 (31) 1 AD-11, AQ-17, ND-3 16 CTX-M-14,-39,-55 (3) 2 AD-2, AQ-1 1 372 (13) SHV-5 (12) 1 AD-3, AQ-9 8 CTX-M-15 (1) 1 AQ-1 0 398 (9) CTX-M-39 (8) 1 AD-4, AQ-4 2 SHV-5 (1) 1 AQ-1 0 38 (8) CTX-M-9, -14, -15, -27 7 AD-4, AQ-4 0 405 (8) CTX-M-9, -15; SHV-12 6 AD-2, AQ-3, ND-3 1 69 (6) CTX-M-14 (5) 1 AD-2, AQ-1, ND-2 0 CTX-M-15 (1) 1 AQ-1 0 648 (6) CTX-M-14 (5) 2 AD-2, AQ-2, ND-1 1 CTX-M-15 (1) 1 AD-1 NA 10 (3) CTX-M-14; SHV-5, -12 3 AD-2, AQ-1 0 410 (3) SHV-12 2 AD-1, AQ-2 1 216 (2) SHV-12 2 AQ-1, ND-1 0 354 (2) CTX-M-2,-15 1 AD-1, AQ-1 0 1196 (2) CTX-M-2 1 AD-2 NA 1598 (2) CTX-M-15 1 AD-1, ND-1 NA Ms c (13) CTX-M-2 (1),-14 (2),-15 (6),-55 (1); SHV-5 (1),-12 (2) 13 AD-6, AQ-5, ND-2 0 Total (125) AD-52, AQ-59, ND-14 32 Acquisition traced b TABLE 2. Clonal structure, resistance genes and acquisition status of ESBL-producing E. coli isolates at TASMC ST, sequence type; PFGE, pulse-field gel electrophoresis types; AD, isolated on admission to rehabilitation ward; AQ, acquisition at the rehabilitation ward; ND, not determined; NA, not applied. a Partially presented in Fig. 2. b Acquisition source traced to another patient. c Ms, miscellaneous ST (one each): 48, 59, 62, 95, 348, 449, 469, 641, 746, 929, 940, 1596, 1597. (31 isolates of type I), ST372 (12 isolates of type Q), ST398 (all nine isolates of type AS) and ST69 (five isolates of type B). The majority of the isolates (n = 101; 81%) produced CTX-M-type ESBLs, namely CTX-M-15 (n = 39; 31%), CTX- M-27 (n = 32; 25.5%), CTX-M-14 (n = 12; 9.5%), CTX-M-2 (n = 4; 3%), CTX-M-9 (n = 3; 2.5%) and CTX-M-55 (n =2; 1.5%). The other types were SHV-5 (n = 15; 12%) and SHV- 12 (n = 9; 7%). A single PFGE type usually produced a single

CMI Adler et al. Transmission dynamics of E. coli clones E501 ESBL type (e. g. the ST131 PFGE type I had CTX-M-27 and the ST372 type Q had SHV-5). Only 14 ST131 isolates produced CTX-M-15, and these belonged to four PFGE types. Clonal transmission of ESBL- E. coli strains at TASMC The carriage status of each patient was analysed against the clonality (ST) and ESBL data (Table 2). The overall acquisition to admission ratio was 1.1 (59/52). Of the major ST and ESBL combinations (>5 isolates each), the highest ratio of 3.0 was observed for the SHV-5-producing ST372 sub-clone (9/3, p = 0.06), while it was 1.5 (17/11) for the CTX-M-27-producing ST131, 1.0 (4/4) for the CTX-M-39-producing ST398, and 0.62 (5/8) for the CTX-M-15-producing ST131. In order to understand the transmission dynamics of the various ESBL- E. coli clones, we compared in parallel the hospitalization periods of the patients carrying the same ST, PFGE and ESBL types (Table 2). The results obtained for the main sub-clones are presented in Figs 1 and S1. Overall, the transmission could be traced in 32 out of the 59 acquisition cases (54%). The rate of traceable cases was significantly higher for the SHV-5-producing ST372 (8/9, p < 0.05) and the CTX-M-27-producing ST131 subclones (16/17, p < 0.001) than for the remaining subclones. Discussion In the present study, we analysed the incidence and transmission of ESBL- E. coli clones in a rehabilitation centre. We have analysed all cases of colonization with ESBL- E. coli, which allowed us to assess the factual complexity of the epidemiological issues and trace the dissemination of ESBL- E. coli more precisely. The incidence of ESBL- E. coli acquisition was common (59 out of 426 ESBL- E. coli -negative patients, 14%), exceeding the number of patients positive at admission (n = 52/492, 10.5%). The rate of ESBL- E. coli carriage on admission was similar to previous studies in Israel [14,15]. The acquisition incidence was lower compared with a previous study in an acute care hospital in Israel [14] but was higher compared with the incidence reported from France [16]. This might be explained by the differences in the prevalence of ESBL- E. coli and the overall quality of infection control practices between the institutions. In particular, the lack of implementation of contact precautions might have contributed to the high acquisition rate. This policy was implemented due to the high rate of ESBL-Ent carriage upon admission. The high acquisition rate found in our study may serve as a warning and guide other institutions that do not have such a high baseline carriage rate to implement contact isolation of these patients, even in the set-up of LCTF. Despite similar clinical characteristics upon admission, ESBL- E. coli carriers had a worse outcome in terms of discharge destination (higher rate of transfer to another healthcare facility vs. home), as well as developing infections caused by ESBL- E. coli. This study is also the first large molecular analysis of ESBL- E. coli in Israel. The pandemic ST131 clone spread both in hospitals and in the community, was highly predominant, accounting for 38.5% of the isolates, similar to other countries [2]. Other globally spread E. coli clones, including ST10, ST38, ST69, ST354, ST398, ST405, ST410 and ST648 [17 23], were observed as well; together with ST131 all these clones comprised around 74.5% of the isolates (93/ 125). Of the relatively less common clones, the SHV-5-producing ST372 [22] was remarkably prevalent, and had the highest acquisition to admission ratio (3.0) when compared with the overall value (1.1). Therefore, it contributed to the role of SHV-producing ESBL- E. coli as a significant factor associated with acquisition, which is congruent with the general view of SHV producers being associated more with hospital environments than the community [24,25]. Surprisingly, of the ST131 sub-clones, the one with CTX-M-27 had a higher acquisition to admission ratio (1.5), than those producing CTX-M-15 (0.62), the enzyme so often found in ST131 [2]. There are several possible explanations for this finding. First, incidental events might have led to an outbreak of the CTX-M-27-producing ST131 sub-clone. Second, while CTX-M-15-producing ST131 spread readily in hospitals but also in the community [2], the CTX-M-27-producing subclone may fit better to in-hospital spread. Third, as most molecular epidemiology studies, unlike the current work, focused mainly on clinical rather than surveillance-culture isolates [2], the relative rate of colonization to infection of the particular clones may be different. Fourth, it is possible that the CTX-M-27-producing ST131sub-clone (as is the SHV-5-producing ST372 sub-clone) is currently emerging and may become more important in Israel in the coming years. Of note, the CTX-M-27-producing ST131 sub-clone was remarkably homogeneous (corresponding to a single predominant PFGE type among all seven ST131 types). In contrast, CTX-M-15 was produced by ST131 isolates that were classified into four PFGE types. All these findings should be analysed with caution as this study was limited to a single facility, and unidentified differences between the individual patients or particular circumstances may have played a role in facilitating transmission of particular sub-clones. In this study, traceable patient-to-patient transmission was identified in 32/59 cases (54%) and was highest in the SHV- 5-producing ST372 (8/9) and the CTX-M-27-producing ST131 (16/17) sub-clones. Several studies have investigated the transmission of different ESBL-Ent clones in ICUs [3 5].

E502 Clinical Microbiology and Infection, Volume 18 Number 12, December 2012 CMI (a) (b) (c) (d) FIG. 1. Transmission dynamics of ESBL-producing E. coli sub-clones at TASMC. (a) ST131-CTX-M-27, ward A; (b) ST131-CTX-M-27, ward B; (c) ST131-CTX-M-15, ward B; (d) ST372-SHV-5, ward A. Lines correspond with hospitalization s timing for each patient. The X on each line marks the date of first positive culture; the colours on Panel c represent different PFGE types.

CMI Adler et al. Transmission dynamics of E. coli clones E503 FIG. 2. Hypothetical clonal-independent bla CTX-M-15 gene transmission between different E. coli clones in ward B. The X on each line marks the date of first positive culture; the different colours represent different E. coli STs and PFGE types. In the study by Harris et al., using similar definitions, the traceable acquisition rate was only 3/23 (13%) [3], leading to the conclusion that patient-to-patient transmission does not play an important role in ESBL-Ent acquisition. Compared with that analysis, our study had more opportunity to study the natural course of ESBL- E. coli clonal transmission for several reasons: first, unlike in ICUs where direct patient-topatient contact is minimal, patients in RWs are ambulant and share many common facilities inside the ward, a condition that may facilitate direct transmission; second, the number of ESBL- E. coli carriers identified (125/492) was high enough to allow the analysis of the transmission potential of different clones; third, the longer length of patients stay in our study allowed for the serial survey of ESBL- E. coli carriage and thus better detection of acquisition events. Several reasons may account for the 27 untraceable cases. First, ESBL-E. coli might have been transmitted from clinically-infected patients who were not analysed in our study. However, this is rather unlikely as all ESBL- E. coli infections occurred in colonized patients. Second, infection might be transmitted from other sources, such as colonized staff or food [26]. Indeed, 7% of staff members were found to be colonized by ESBL- E. coli in two surveys conducted during the study (data not shown). Third, previously colonized patients may have been misclassified as new acquisitions due to false-negative results of the admission culture. And last, it is possible that transmission of ESBL genes via mobile genetic elements [23,27] might account for some of the acquisition cases. In order to explore this hypothetical possibility, we analysed the patient-to-patient transmission in the 25 clonally-diverse (n = 16), bla CTX-M-15 -carrying ESBL- E. coli isolates in ward B (Fig. 2). We were able to trace only two of the nine acquisition cases by clonal transmission, whereas the mobile genetic element transmission hypothesis would allow us to trace all the seven additional cases. Also, it is possible that such transfer might occur from non-e. coli species. Further analyses of plasmids and the genetic environment surrounding the ESBL genes are required in order to examine this hypothesis. An important limitation of our study relates to the tracing of the sources of ESBL-E. coli acquisition, which was limited to other patients only. However, other sources such as staff members or the environment might have contributed to the transmission in some of the acquisition cases and therefore individual patient-to-patient transmission cannot be proven with certainty. This study highlights the importance of patient-to-patient transmission in the acquisition of ESBL- E. coli during hospitalization in rehabilitation centres and the varying dissemination potential of different clones. Consequently, we believe that infection control practices should be adapted and implemented in these institutions. Further studies are required to explore the dynamics of other Enterobacteriaceae species in other healthcare settings. Acknowledgements This work was part of the activities of the MOSAR integrated project (LSHP-CT-2007-037941) supported by the European Commission under the Life Science Health priority of the 6th Framework Programme (WP5 and WP2 Study Teams). MG, AB, RI, JF and WH were financed also by the MOSAR-complementary grant No. 934/6 PR UE/ 2009/7 from the Polish Ministry of Science and Higher Education.

E504 Clinical Microbiology and Infection, Volume 18 Number 12, December 2012 CMI Transparency Declaration Nothing to declare. Supporting Information Additional Supporting Information may be found in the online version of this article: Figure S1. PFGE patterns of selected ESBL-producing E. coli sub-clones at TASMC. Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article. Appendix *The MOSAR WP5 and WP2 study groups are: Shiri Navon- Venezia, Mitchell J. Schwaber, Maya Shklyar, Lilach Keren, Rivi Glick, Shiri Klarfeld-Lidji, Meirav Hochman, Anat Klein, Eti Mordechai, Shimrit Cohen, Ruth Fachima, Yelena Zdonevsky and Iris Pfeffer, Section of Epidemiology, Tel-Aviv Sourasky Medical Center, Tel Aviv, Israel; Wasef Naamnih and Bacanda Suonov, Division of Geriatrics, Tel-Aviv Sourasky Medical Center, Tel Aviv, Israel; Anna Grabowska and Małgorzata Herda, National Medicines Institute, Warsaw, Poland. References 1. Mandell GL, Bennett JEDR. Mandell, Douglas, and Bennett s Principles and Practice of Infectious Diseases, 7th edn. Philadelphia, PA: US Elsevier Health Bookshop. 2010. 2. Rogers BA, Sidjabat HE, Paterson DL. Escherichia coli O25b-ST131: a pandemic, multiresistant, community-associated strain. J Antimicrob Chemother 2011; 66: 1 14. 3. Harris AD, Kotetishvili M, Shurland S et al. How important is patientto-patient transmission in extended-spectrum beta-lactamase Escherichia coli acquisition. Am J Infect Control 2007; 35: 97 101. 4. Thouverez M, Talon D, Bertrand X. Control of Enterobacteriaceae producing extended-spectrum beta-lactamase in intensive care units: rectal screening may not be needed in non-epidemic situations. Infect Control Hosp Epidemiol 2004; 25: 838 841. 5. Wu T-L, Chia J-H, Su L-H, Kuo A-J, Chu C, Chiu C-H. Dissemination of extended-spectrum -lactamase-producing Enterobacteriaceae in pediatric intensive care units. J Clin Microbiol 2003; 41: 4836 4838. 6. Justo D, Guy N, Halperin E, Lerman Y. Admission norton scale scores are associated with long-term mortality following rehabilitation in older adults. J Rehabil Med 2012; 44: 172 175. 7. Drieux L, Brossier F, Sougakoff W, Jarlier V. Phenotypic detection of extended-spectrum beta-lactamase production in Enterobacteriaceae: review and bench guide. Clin Microbiol Infect 2008; 14 (Suppl 1): 90 103. 8. Struelens MJ, Rost F, Deplano A et al. Pseudomonas aeruginosa and Enterobacteriaceae bacteremia after biliary endoscopy: an outbreak investigation using DNA macrorestriction analysis. Am J Med 1993; 95: 489 498. 9. Tenover FC, Arbeit RD, Goering RV et al. Interpreting chromosomal DNA restriction patterns produced by pulsed-field gel electrophoresis: criteria for bacterial strain typing. J Clin Microbiol 1995; 33: 2233 2239. 10. Wirth T, Falush D, Lan R et al. Sex and virulence in Escherichia coli: an evolutionary perspective. Mol Microbiol 2006; 60: 1136 1151. 11. Bauernfeind A, Grimm H, Schweighart S. A new plasmidic cefotaximase in a clinical isolate of Escherichia coli. Infection 1990; 18: 294 298. 12. Empel J, Baraniak A, Literacka E et al. Molecular survey of beta-lactamases conferring resistance to newer beta-lactams in Enterobacteriaceae isolates from Polish hospitals. Antimicrob Agents Chemother 2008; 52: 2449 2454. 13. Woodford N, Fagan EJ, Ellington MJ. Multiplex PCR for rapid detection of genes encoding CTX-M extended-spectrum beta-lactamases. J Antimicrob Chemother 2006; 57: 154 155. 14. Friedmann R, Raveh D, Zartzer E et al. Prospective evaluation of colonization with extended-spectrum beta-lactamase (ESBL)-producing enterobacteriaceae among patients at hospital admission and of subsequent colonization with ESBL-producing enterobacteriaceae among patients during hospitalization. Infect Control Hosp Epidemiol 2009; 30: 534 542. 15. Ben-Ami R, Schwaber MJ, Navon-Venezia S et al. Influx of extendedspectrum beta-lactamase-producing enterobacteriaceae into the hospital. Clin Infect Dis 2006; 42: 925 934. 16. Buke C, Armand-Lefevre L, Lolom I et al. Epidemiology of multidrugresistant bacteria in patients with long hospital stays. Infect Control Hosp Epidemiol 2007; 28: 1255 1260. 17. Gibreel TM, Dodgson AR, Cheesbrough J, Fox AJ, Bolton FJ, Upton M. Population structure, virulence potential and antibiotic susceptibility of uropathogenic Escherichia coli from Northwest England. J Antimicrob Chemother 2012; 67: 346 356. 18. Peirano G, van derbijak, Gregson DB, Pitout JDD. Molecular epidemiology over an eleven-year period (2000-10) of Extended-spectrum b-lactamase-producing Escherichia coli causing bacteraemia in a centralized Canadian region. J Clin Microbiol 2011; Available at: http:// jcm.asm.org/content/early/2011/12/01/jcm.06025-11.full.pdf+html. 19. Croxall G, Hale J, Weston V et al. Molecular epidemiology of extraintestinal pathogenic Escherichia coli isolates from a regional cohort of elderly patients highlights the prevalence of ST131 strains with increased antimicrobial resistance in both community and hospital care settings. J Antimicrob Chemother 2011; 66: 2501 2508. 20. van der Bij AK, Peirano G, Goessens WHF, van der Vorm ER, van Westreenen M, Pitout JDD. Clinical and molecular characteristics of extended-spectrum-beta-lactamase-producing Escherichia coli causing bacteremia in the Rotterdam Area, Netherlands. Antimicrob Agents Chemother 2011; 55: 3576 3578. 21. Suzuki S, Shibata N, Yamane K, Wachino J-ichi, Ito K, Arakawa Y. Change in the prevalence of extended-spectrum-beta-lactamase-producing Escherichia coli in Japan by clonal spread. J Antimicrob Chemother 2009; 63: 72 79. 22. Lau SH, Reddy S, Cheesbrough J et al. Major uropathogenic Escherichia coli strain isolated in the northwest of England identified by multilocus sequence typing. J Clin Microbiol 2008; 46: 1076 1080. 23. Woerther P-L, Angebault C, Jacquier H et al. Massive increase, spread, and exchange of extended spectrum b-lactamase-encoding genes among intestinal Enterobacteriaceae in hospitalized children with severe acute malnutrition in Niger. Clin Infect Dis 2011; 53: 677 685.

CMI Adler et al. Transmission dynamics of E. coli clones E505 24. Warren RE, Harvey G, Carr R, Ward D, Doroshenko A. Control of infections due to extended-spectrum beta-lactamase-producing organisms in hospitals and the community. Clin Microbiol Infect 2008; 14 (Suppl 1): 124 133. 25. Cantón R, Novais A, Valverde A et al. Prevalence and spread of extended-spectrum beta-lactamase-producing Enterobacteriaceae in Europe. Clin Microbiol Infect 2008; 14 (Suppl 1): 144 153. 26. Calbo E, Freixas N, Xercavins M et al. Foodborne nosocomial outbreak of SHV1 and CTX-M-15-producing Klebsiella pneumoniae: epidemiology and control. Clin Infect Dis 2011; 52: 743 749. 27. Mathers AJ, Cox HL, Kitchel B et al. Molecular dissection of an outbreak of carbapenem-resistant Enterobacteriaceae reveals intergenus KPC carbapenemase transmission through a promiscuous plasmid. MBio 2011; 2: e00204 e00211.