Dominance of CTX-M-Type Extended-Spectrum β-Lactamase (ESBL)-Producing Escherichia coli Isolated from Patients with Community-Onset and Hospital-Onset Infection in China

Objective To investigate CTX-M genotypes among extended-spectrum β-lactamase-producing Escherichia coli (ESBL-EC) isolated from patients with community-onset and hospital-onset infections in China, their clonality and the distribution of CTX-M variants in different specimens of community-onset and hospital-onset infections. Methods ESBL-EC isolates were collected from general hospitals from 2011 to 2012 in China. Broth microdilution method antimicrobial susceptibility testing of 16 antibiotics was performed. Clinical data from community-onset and hospital-onset infections due to ESBL-EC were analyzed. ESBL-encoding genes were amplified by PCR and sequenced, and multilocus sequence typing (MLST) was performed for a random selection of predominant CTX-M type strains identified. Results A total of 1,168 ESBL-EC isolates were obtained from various clinical specimens, 41.7% of which were responsible for causing community-onset infections. The presence of urinary calculi was higher in community-onset infections, whereas malignancy, cardiovascular and cerebrovascular diseases, dementia, chronic renal disease, diabetes mellitus and surgical treatment were found to have higher proportions in hospital-onset infections. There was no significant difference in trauma between community-onset and hospital-onset infections. 96.2% of the isolates were detected to harbor bla CTX-M genes. bla CTX-M-1 group and bla CTX-M-9 group were detected at 40.7% and 48.7% respectively, and both positive group accounted for 10.6%. bla CTX-M-55 (24.8%) and bla CTX-M-15 (18.2%) were the major genotypes in bla CTX-M-1 group while bla CTX-M-14 (46.8%) was predominant in bla CTX-M-9 group. A comparison of bla CTX-M distribution in different specimens between ESBL-EC causing community-onset and hospital-onset infection showed no significant difference. A total of 229 isolates were tested for MLST. ST131 (14%) was the predominant type. ST648, ST405 and ST1193 were also detected. Conclusions Community-onset ESBL-EC has emerged as a common pathogen in China. CTX-M-14 is the most commonly encountered, CTX-M-55 and CTX-M-15 have spread rapidly. ST131 is the predominant clonal group, and the great diversity of CTX-M-producing isolates of E. coli has emerged in China.


Introduction
The production of extended-spectrum b-lactamases (ESBLs) is one of the primary mechanisms conferring resistance to broadspectrum b-lactam antibiotics. A large surveillance of antimicrobial resistance has shown that the detection rate of ESBLs in Escherichia coli has increased significantly over that in Klebsiella pneumoniae and exceeded 70% in mainland China [1]. A genotypic epidemiological study on ESBLs indicated that bla CTX-M type genes have replaced bla SHV and bla TEM as the most common ESBL gene [2]. In addition, the prevalence of ESBL producing Enterobacteriaceae is increasing in community settings globally [3]. According to a recent study in the US, 36.8% of ESBLproducing E. coli (ESBL-EC) isolates were responsible for community-onset bloodstream infections (BSI) and urinary tract infections (UTI) [4]. Results from the Study for Monitoring Antimicrobial Resistance Trends (SMART) surveillance in European countries revealed that the prevalence of ESBL-EC causing community-onset intra-abdominal infections rose from 4% in 2002 to 7.4% in 2007 [5]. Surprisingly, a high incidence of ESBL-EC was also found in developing countries. A survey in rural Thai communities showed that faecal carriage of CTX-M-type ESBL-producing Enterobacteriaceae among asymptomatic individuals was up to 69.3% in 2010 [6]. The global spread of the E. coli ST131 clone was presumed to be partial responsible for the current situation [7].
Genotypic epidemiology data of ESBL-EC causing communityonset infections in China was insufficient. A multicenter study of the Ministry of Health National Antimicrobial Resistance Surveillance Net (MOHNARIN) in China during January 2007 to March 2008 revealed that 42.8% of urinary tract infections caused by ESBL-EC were community-onset infections [8]. However, there were few studies that compared the genotypic epidemiology of ESBL-EC isolated from community-and hospital-onset infections.
Previous regional surveys showed that CTX-M-14 and CTX-M-15 were the most common ESBLs produced by Enterobacteriaceae in China [2,9], and some epidemic plasmids played a major role in the dissemination of bla CTX-M- 14 and bla CTX-M-15 among E. coli and K. pneumoniae isolates [10,11]. Nevertheless, the current genotypic epidemiology of ESBL-EC in China remains largely unknown. This study was proceeded to provide information on genotypic profile of CTX-M-type-carrying ESBL-EC in community-and hospital-onset infections and their clonality in mainland China.

Ethics Statement
This study protocol was approved by the Ethics Committee of The First Affiliated Hospital of Guangzhou Medical University. All subjects signed written informed consent prior to the study. Patient information was anonymized and de-identified prior to analysis.

Bacterial Isolates and Data Collection
This was a national laboratory-based multicenter study of ESBL-EC genotypic epidemiology in 30 general hospitals of 24 provinces in mainland China. The hospitals were distributed in five regions across China: four in the northeast (Heilongjiang, Jilin, Liaoning and Inner Mongolia); five in the north (Beijing, Tianjin, Hebei, Shandong and Shanxi); five in the mid-east (Henan, Hubei, Zhejiang, Jiangxi and Anhui); five in the south (Sichuan, Yunnan, Guangxi, Guangdong and Hainan); and five in the northwest (Xinjiang, Gansu, Ningxia, Qinghai and Shaanxi). From July 2011 to July 2012, the first 60 non-repetitive ESBL-EC isolates (or all non-repetitive ESBL-EC isolates if the total number of nonrepetitive ESBL-EC was less than 60) were sent to the Department of Clinical Laboratory in Peking Union Medical College Hospital, where they were stored at 280uC until further study. These bacterial isolates were isolated from different specimen types, including urine, blood, secretions, pus, bile, abdominal drainage fluid, bronchoalveolar lavage fluid and others.
Demographic and clinical information of the patients from whom ESBL-EC were isolated was retrieved retrospectively, which included age, gender, history of underlying diseases, specimen type, the date of patient admission and sample collection and so on. Community-onset infection was defined as an infection diagnosed within the first 48 h of hospitalization or outpatient service and no history of hospitalization within 30 days [12]. Hospital-onset infection was defined as an infection occurring more than 48 h after admission to the hospital. Health care associated infection as defined by Friedman [13] was included under hospital-onset infection.

Antimicrobial Susceptibility Testing
ESBL-EC isolates were initially identified by routine methods in each participating hospital according to local operation procedures. When these ''presumptive'' ESBL-EC isolates were sent to the Department of Clinical Laboratory in Peking Union Medical College Hospital, production of ESBL was confirmed by broth microdilution method per Clinical and Laboratory Standards Institute (CLSI) guidelines [14]. The minimum inhibitory concentrations (MICs) of the following 16 drugs were determined with the broth microdilution method according to the CLSI performance standards [14]; imipenem (IPM), meropenem (MEM), ertapenem (EPM), tigecycline (TGC), colistin (CST), cefoperazone/sulbactam (CPS), piperacillin/tazobactam (TZP), cefepime (FEP), cefotaxime (CTX), ceftazidime (CAZ), ceftriaxone (CRO), aztreonam (ATM), amikacin (AMK), levofloxacin (LVX), ciprofloxacin (CIP), and minocycline (MIN). The results were interpreted by using the criteria of the CLSI for broth dilution [14]. E. coli ATCC 25922 was run in parallel for quality control, The interpretation breakpoint of cefoperazone/sulbactam (Pfizer Inc., USA) corresponded to that of cefoperazone, and that of tigecycline was based on the criteria proposed by the Food and Drug Administration (susceptibility at #2 mg/L).

Molecular Detection of ESBL
All the isolates were screened by PCR with specific primers listed in Table S1 in File S1 for bla CTX-M detection and group assignment, as described previously [15]. Detection of bla TEM and bla SHV was performed on both bla CTX-M -positive and bla CTX-Mnegative samples. Purified PCR products were directly sequenced from both ends or cloned in pMD18-T and then sequenced. The DNA sequences and deduced amino acid sequences were analyzed using the NCBI BLAST program (http://www.ncbi.nlm.nih.gov/ ), and the subtypes of b-lactamase genes were confirmed by referring to the Lahey system (www.lahey.org/studies/).

Multilocus Sequence Typing
Multilocus sequence typing (MLST) was performed for a random selection of predominant CTX-M type (including CTX-M-15, CTX-M-55 and CTX-M-14) strains identified in this study. MLST was performed with the seven conserved housekeeping genes (including adk, fumC, gyrB, icd, mdh, purA and recA). A detailed protocol of the MLST procedure, including allelic type and sequence type (ST) assignment methods, available at the University College Cork MLST website (http://mlst.ucc.ie/ mlst/dbs/Ecoli), was used in this study.

Statistical Analysis
The data were analyzed using SPSS 13.0 software package (SPSS, Chicago IL, USA). Continuous variables were compared using the Mann-Whitney U test. Qualitative variables were compared using the chi-square test or the Fisher's exact test. P, 0.05 was considered statistically significant.
A comparison of clinical data in community-onset versus hospital-onset infection groups was shown in Table 2. In the two groups, 78.2% and 38.5% of patients had no history of underlying diseases when admitted to hospital. The presence of urinary calculi was higher in community-onset infections, whereas common underlying diseases including malignancy, cardiovascular and cerebrovascular diseases, dementia, chronic renal disease, diabetes mellitus and surgical treatment were found to have higher proportions in hospital-onset infections. There was no statistically significant difference in trauma between community-onset infections and hospital-onset infections.
In comparison of the drug susceptibility of ESBL-EC from hospital-and community-onset infections, it revealed that there were similar resistant profiles, and no statistically significant difference was found for any drug (P.0.05) (see Table S2 in File S1).   In the bla CTX-M -positive isolates, TEM-type ESBLs were not detected. TEM PCR amplification was positive in 1011 strains, in which 998 isolates were confirmed to be TEM-1, 8 TEM-105, 5 TEM-like by sequence analysis. Ten strains harbored an SHVtype ESBLs in the bla CTX-M -positive isolates: SHV-12 (n = 7) and SHV-5 (n = 3). Among 44 bla CTX-M -negative isolates, 11 of them were detected to harbor bla TEM-1 (n = 11), and also produce bla SHV-12 (n = 9) and bla SHV-5 (n = 2). bla TEM and bla SHV type ESBLs were not detected in the remaining 33 isolates.

Molecular Detection of ESBL
A comparison of bla CTX-M distribution in different specimens between ESBL-EC causing community-onset and hospital-onset infections showed that no significant difference was noted with regard to the proportions of different types of ESBLs between the two groups ( Table 3).
The five study regions showed certain differences in the bla CTX-  Table  S3 in File S1.

Discussion
To our knowledge, this work represents the first national largescale study on the clinical information, antimicrobial resistance and genotype epidemiology of ESBL-EC in China.
Greater than 50% ESBL-EC isolates in this study were resistant to all the third-generation cephalosporins, cefepime, and fluoroquinolones. It has been suggested that cephalosporin and fluoroquinolone are not considered effective choices for treatment of patients with ESBL-producing Enterobacteriaceae infection because of relatively high mortality associated with inappropriate therapy [16]. With regard to b-lactam/b-lactamase inhibitor combinations, ESBL-EC had low resistance rates to piperacillin/ tazobactam and cefoperazone/sulbactam (,20%) in this study, which suggest they might be appropriate drugs against non-BSI caused by ESBL, though there is controversy surrounding whether b-lactam/b-lactamase-inhibitor combinations are a good option [16,17]. As expected, the ESBL-EC had high susceptibility rates to imipenem, meropenem, and ertapenem (94.8-99%), which suggest that carbapenems are considered as last resort antibiotics for the treatment of ESBL-producing E. coli infections in China.
The prevalence of ESBL-EC isolates in community-onset infections has been increasing worldwide. Similar results were found in the present study. Of community-onset UTI, BSI and IAI infections, ESBL-EC took up a notable proportion (19.4%-33.7%), which was in accordance with those reported in Europe and America [3,4,12]. Many risk factors have been associated with the prevalence of ESBL-EC isolates in the community. For example, recurrent UTI and dependent functional status were associated with isolation of CTX-M E. coli responsible for community-onset UTI [3], which correlated with the results of this study, where 11.2% patients in the community-onset UTI group had a history of urinary tract calculus, a risk factor for and consequence of recurrent UTI [18]. Previous exposure to second or third-generation cephalosporins has been considered as an important risk factor for the occurrence of ESBL-producing Enterobacteriaceae in community-onset infection by many researchers [3,19,20]. Although the number and distribution of community health-care settings, e.g. nursing homes, in China are not comparable to developed countries, the frequency and proportion of prescribed antibiotics in the rural areas of China are higher compared with the developed countries, and the number of antibiotics per 100 prescriptions was 54.62, and personal use of broad-spectrum penicillins or cephalosporins and fluoroquinolones is common [21]. This potentially may have contributed to the high incidence of ESBL-EC in communityonset infections.
A 2010-2011 survey in the United States showed that CTX-M type accounted for 85.4% of ESBL-producing E. coli strains [3]. In China, CTX-M-type ESBLs accounted for more than 70% ESBL-producing E. coli strains over the past 10 years; CTX-M-14 remains the most abundant genotype, although the detection rate of CTX-M-15 has shown a continuously increasing trend in recent years [9][10][11]. Similar findings were obtained in the present study, in that the proportion of CTX-M-producing E. coli had increased to 96.2%, and CTX-M-14 remained the most common genotype of ESBLs, accounting for 46.8% of isolates. CTX-M-55, a variant of CTX-M-15 by one amino acid substitution of Ala-77-Val [22], was scarcely found in clinical isolates previously [22,23]. However, the prevalence of CTX-M-55 (24.8%) has increased significantly and become the secondary genotype of ESBLs, following CTX-M-14 in this study. Interestingly, CTX-M-55 was first reported to be highly prevalent (25%) in ESBL-producing E. coli isolated from the intestinal samples of pets in China in 2010 [24]. Subsequent surveillance indicated that CTX-M-55 was predominant amongst ESBL-EC from animal sources [25][26][27]. In addition, CTX-M-27, CTX-M-65 and CTX-M-24 commonly identified from animal derived isolates accounted for a notable proportion in clinical isolates, as discovered in the present and from previous studies [25][26][27]. The distribution of the above ESBLs was generally the same between hospitals and communities and across different regions in China. The results may suggest that the injudicious use of antimicrobial agents amongst livestock and hospitals may result in mutation and subsequent epidemiological change of major ESBL genotypes circulated (e.g. and human-derived E. coli via plasmids needs to be investigated. The lateral transfer of resistant genes between animals and humans may also contribute to the rising prevalence of ESBL-EC in community, and E. coli from healthy food animals can be important reservoirs of bla CTX-M genes and may contribute to the dissemination and transfer of these b-lactamase genes throughout China. In the present study, the bla CTX-M-64 , bla CTX-M-123 , and bla CTX-M-132 genes, new mutant genotypes containing at least three hybrid sites of CTX-M-14 and CTX-M-15 [28], were first detected in clinical isolates in China; the three genotypes were first found in animal isolates in China [29]. Although the detail mechanisms of resistance remain unknown, it is possible that this may be due transmission of isolates and their plasmids between animals and humans, and remains an area for future studies. The majority of CTX-M-producing E. coli particularly CTX-M-15-producing E. coli belonged to a specific clone defined as MLST profile ST131 in many countries [30,31], which was presumed to be some reason for the growing problem of antibiotic resistance. Same as previous research, ST131 (14%) was the predominant clonal group established among these strains. It was mostly detected in the CTX-M-14 strains, but also accounted for a significant proportion in CTX-M-15 and CTX-M-55 strains. What's more, the majority of ST131 strains were obtained from urine (50.0%), which may be connected with its variable virulence potential [32]. In addition, the finding of non-ST131 profile clones like ST405, ST167, ST 3177, ST1193, and ST648 and so on showed the great diversity of CTX-M-producing isolates of E. coli had emerged in China.
There were some limitations in this study. Firstly, some of isolates were not susceptible to carbapenems, it would be important to know if some of these isolates produce carbapenemases. However, the related experiment is ongoing by one of our collaborative researching group and they plan to report our results in a separate paper. Secondly, some clinical information such as previous antibiotic treatment, medical exposure (e.g., received renal dialysis, nasogastric tube use), and outcome of the patients, were not acquired. The lack of this data made it impossible to analyze the risk factors for ESBL-EC infections in the same manner as preformed previously. However, the acquired data makes it possible to classify the all the cases into community-onset and hospital-onset infection, and the large sample size provide some clues of the risk factors for ESBL-EC infections in China. It would be worthwhile in future investigations to collect more detailed clinical data to learn in a more comprehensive manner the risk factors for different CTX-M variants in China. In addition, it will be valuable to discuss zoonotic transmission routes and possible transmission of bla CTX-M positive E. coli between human and animals in China. It is regrettable that we could not obtain some data now. We hope to examine ESBL-EC isolated from both human and animals in parallel study in the follow-up research.
In summary, this study examined the drug resistance and genotypic epidemiology of 1,168 clinical isolates of ESBLproducing E. coli. We found that the CTX-M type is still the primary genotype of ESBL in China and that its proportion (96.2%) has increased. The specific genotype of CTX-M has undergone great changes, and it is worth noting that main resistant genotypes, represented by CTX-M-55 and CTX-M-15, have spread rapidly. ST131 (14%) was the predominant clonal group, and at the same time the great diversity of CTX-M-producing isolates of E. coli has emerged in China. Further studies should continue to monitor ESBL-producing bacteria and explore the mechanisms of spread of CTX-M genotypes in these bacteria to provide reference data to enable relevant infection control.

Supporting Information
File S1 Contains the files: Table S1. Primers used for polymerase chain reaction amplification of bla genes in this study. Notes: Degenerate bases, K for G or T; R for A or G; S for C or G; Y for C or T. Table S2. Susceptibility of extended-spectrum b-lactamase-producing Escherichia coli (ESBL-EC) isolates against common antimicrobial agents and comparison of antimicrobial susceptibility of Extended-Spectrum b-Lactamase-producing