The Resistance Phenotype and Molecular Epidemiology of Klebsiella pneumoniae in Bloodstream Infections in Shanghai, China, 2012–2015

Klebsiella pneumoniae (K.pneumoniae) is a common nosocomial pathogen causing bloodstream infections. Antibiotic susceptibility surveillance and molecular characterization will facilitate prevention and management of K. pneumoniae bloodstream infections. K. pneumoniae isolates causing bloodstream infections were consecutively collected between January 2012 and December 2015 in Shanghai. Eighty isolates (20 per year) were randomly selected and enrolled in this study. Drug susceptibility were determined by the disk diffusion method. Polymerase chain reaction (PCR) was employed to detect extended-spectrum β-lactamases (ESBLs), carbapenemases, and seven housekeeping genes of K. pneumoniae. eBURST was used for multi-locus sequence typing (MLST). More than 50% isolates were resistant to cefuroxime, ampicillin-sulbactam, and piperacillin, while carbapenems had lower resistant rates than other antibiotics. Of the 80 isolates, 22 produced ESBLs, and 14 were carbapenemase producers. In the ESBL-producing K. pneumoniae isolates, the most common ESBL genes were blaSHV and blaCTX−M. Thirteen carbapenemase producers harbored blaKPC−2 and one other carried blaNDM−5. ST11 (14/80) was the most frequent sequence type (ST), followed by ST15 (7/80) and ST29 (4/80). Our data revealed high prevalence of antibiotic resistant K. pneumoniae isolates from bloodstream infections but their genetic diversity suggested no clonal dissemination in the region. Also, one K. pneumoniae isolate harbored blaNDM−5 in this study, which was firstly reported in Shanghai.


INTRODUCTION
Klebsiella pneumoniae (K. pneumoniae) can cause ventilatorassociated pneumonia, urinary tract infection, sepsis, catheterrelated infection, and bacterial meningitis (Tumbarello et al., 2006;Du et al., 2014), and is also the most common causative gram-negative bacterium in nosocomial and communityacquired infections (Hu et al., 2015). Unfortunately, extended and overuse of antibiotics may potentiate antibiotic resistance of K. pneumoniae strains to cephalosporins, aminoglycosides, fluoroquinolones, and even carbapenems (Guana et al., 2014), which will be difficult and costly to control.
Extensive use of invasive procedures and glucocorticoids has increased the incidence of bloodstream infections and caused high mortality in patients (27.2-40.8%;Marra et al., 2006;Yang et al., 2010;Lv et al., 2014;Wang et al., 2016). Of the gram negative bacilli implicated in nosocomial bloodstream infections, K. pneumoniae was one of the most common pathogens, second only to Escherichia coli (Ghadiri et al., 2012;Lv et al., 2014). The prevalence of Klebsiella spp. bloodstream infections was 7.6% in the United States, Canada, South America, and Europe according to SENTRY (Biedenbach et al., 2004). On the basis of a largescale survey from China, K. pneumoniae caused 11.3% of all bloodstream infections in 2011-2012 .
The wide dissemination of drug resistant pathogens leads to their increasing prevalence in bloodstream infections. Data from China showed that 53.3% K. pneumoniae isolates were multidrug-resistant . In Europe, Latin America, and North America, 21.7, 42.7, and 5.8% of Klebsiella spp. had the extended-spectrum β-lactamase (ESBL) phenotype (Biedenbach et al., 2004), while as high as 60.6% K. pneumoniae isolates in Greece were carbapenemase producers (Daikos et al., 2014). And the most prevalent group of carbapenemases was K. pneumonia carbapenemase (KPC). Studies have shown that the drugresistant organisms bloodstream infection was one important risk factor for mortality and negatively impacted the treatment outcome of patients (Kim et al., 2002;Tumbarello et al., 2006). Limited data on susceptibility and molecular epidemiology of K. pneumoniae causing bloodstream infections were available in Shanghai. In the study, we have monitored resistance phenotype, the prevalent resistant genes and sequence types (STs) of K. pneumoniae isolates from bloodstream infections between 2012 and 2015 in the region.

Patients and Bacterial Isolates
This retrospective study was conducted in Ruijin Hospital, an 1800-bed general university-affiliated hospital located in Shanghai, with ∼1,15,000 patient visits per year. Patients with at least one positive blood culture of K. pneumoniae from January 2012 through December 2015 were enrolled in the study. A total of 254 episodes of K. pneumoniae bloodstream infections (66 in 2012, 62 in 2013, 64 in 2014, and 62 in 2015) were identified during this period. Only the first positive blood culture was reviewed and recorded. Eighty isolates were enrolled: twenty isolates were selected from each year using the random number generation function in Microsoft Office Excel 2010 (Microsoft Corporation, Redmond, WA, USA). Isolates identification was carried out using matrix-assisted laser desorption ionizationtime of flight mass spectrometer (bioMérieux, Marcy-l'Étoile, France).
This study was approved by Ruijin Hospital Ethics Committee (Shanghai Jiao Tong University School of Medicine). The Review Board exempted request for informed consent because this retrospective study only focused on the bacteria and did not have impact on the patients.

ESBL-Producing and Carbapenemase-Producing Isolates Screening and Confirmation
According to the CLSI criteria [Clinical and Laboratory Standards Institute (CLSI), 2015], ceftazidime and cefotaxime were used as screening tests for ESBLs. Ceftazidime, ceftazidimeclavulanate, cefotaxime, and cefotaxime-clavulanate were used as confirmatory test. Imipenem and meropenem were used as screening for carbapenemase production. Detection of carbapenemase genes was performed to confirm production of carbapenemases.

Multilocus Sequence Typing
Multilocus sequence typing (MLST) was carried out as described previously (http://bigsdb.web.pasteur.fr/klebsiella/primers_used. html; Diancourt et al., 2005). Briefly, Seven hosekeeping genes (gapA, infB, mdh, pgi, phoE, rpoB, and tonB) for K. pneumoniae were amplified, sequenced, and analyzed. Alleles and STs were determined according to the database (http://bigsdb. web.pasteur.fr/perl/bigsdb/bigsdb.pl?db=pubmlst_klebsiella_ seqdef_public&page=profiles). STs that could not be found in the database were submitted to the curator of the database (klebsiellaMLST@pasteur.fr). eBURST version 3.0 software was used to analyze the clustering of related STs. In this study, isolates were grouped together if six of the seven alleles were homologs.

Statistical Analysis
Data in this study were analyzed by SAS 8.2 (SAS Institute Inc., Cary, NC, USA). Continuous variables were presented as the mean ± SD or median and interquartile range. For categorical variables, the chi-square test was used to compare the disparity between different groups. P < 0.05 was considered to be statistically significant.

Patient Data
From January 2012 to December 2015, there were more male patients (182/254) than females (72/254). The age of patients ranged from 1 to 91 years. Most of the cases (40/254) were derived from the surgery. Fifty-eight males and 22 females were enrolled and their median age was 61 years (range: 12-91 years). Most patients were from the Department of General Surgery (17/80), Transplantation (11/80), and Infectious Disease (9/80).

DISCUSSION
Local epidemiologic data on prevalence of specific clones of K. pneumoniae bloodstream infection were indispensible to develop clinical treatment regimen and evaluate outcomes of different therapeutic strategy (Pai et al., 2004;Marra et al., 2006;Neuner et al., 2011;Harris et al., 2015). However, there were limited data on antibiotic resistance, resistant genes and STs of K. pneumoniae bloodstream isolates in Shanghai. In our pilot study conducted in 2012, we found that K. pneumoniae was the second frequent gram negative bacillus from blood cultures in our hospital, representing 14.3% of all the isolates (Zhao et al., 2014). In current study, we have extended our research through 2015 to acquire more comprehensive molecular epidemiologic data. K. pneumoniae bloodstream isolates in the study showed threatened resistance to the most of routine antibiotics and only meropenem, imipenem, amikacin, and piperacillin-tazobactam had relative low resistant rates (<30%) which made them better candidates for empiric therapy of K. pneumoniae bloodstream infections. However, prudent and rational uses of antibiotics based on the results of antimicrobial susceptibility tests are still essential to the success of treatment. Notably, the resistant rates to meropenem, imipenem, amikacin, cefepime, and ciprofloxacin determined in this study were much higher than what were found in the surveillance of 2011-2012 in China . This dynamic change of the pattern of drug resistance warrant active ongoing surveillance on antibiotic resistance and consistent prevention and control of K. pneumoniae.
The proportion of ESBL-producing K. pneumoniae (27.5%) in this study was lower than that in Italy (32.6%), America (51.8%), Korea (52.9%), and Russia (60.8%) (Kim et al., 2002;Edelstein et al., 2003;Marra et al., 2006;Tumbarello et al., 2006), but the multidrug-resistant phenotype, and the dominate ESBL enzyme (CTX-M) of ESBL producers were similar (Edelstein et al., 2003). Typing of ESBL producers revealed a high level of genetic diversity, with ST15 (22.7%, 5/22) and ST11 (13.6%, 3/22) as the most common STs. ST15 K. pneumoniae has been identified in both animals and humans in several countries. Although it was not yet a dominant clone in our study, ST15 K. pneumoniae had achieved a highly successful clonal spread in some countries, such as Bulgaria, Portugal, and Thailand (Netikul et al., 2014;Rodrigues et al., 2014;Markovska et al., 2015), and we should be cautious about its high potential of becoming a major clone associated with ESBL producers in Shanghai in future. Isolates belonged to ST11 were isolated from different departments, in different years with different antibiotic resistance profiles. This suggested no clonal dissemination in the region.
Different from previous studies where no carbapenemaseproducing E. coli bloodstream isolate was found (Zhao et al., 2015;Wang et al., 2016), the carbapenemase-producing K. pneumoniae accounted for 17.5% in our hospital, and they, as reported elsewhere (David et al., 2013;Du et al., 2014), also showed extremely high resistant rates to major antibiotics except aminoglycosides and trimethoprim-sulfamethoxazole. Molecular analysis suggested 92.9% of the carbapenemase producers were harbored with bla KPC−2 , and belonged to ST11, confirming that ST11 was associated with KPC. Unlike the widespread of KPC-producing ST258 K. pneumoniae in Europe, the dominant ST of KPC-producing K. pneumoniae was ST11 in China (Dautzenberg et al., 2016). Also, One ST1 carbapenemase producers carried bla NDM−5 was found in this study. NDM-5, which mostly found in E. coli, was firstly reported in K. pneumoniae in Shanghai. Taken together, our data indicated the KPC-producing ST11 K. pneumoniae isolate was a highrisk clone in our hospital, and should be taken as the major consideration when developing strategy to control resistant isolates dissemination. NDM, along with KPC, VIM, and OXA-48, were four major carbapenemases detected in K. pneumoniae.     FIGURE 1 | Clonal groupings among K. pneumoniae bloodstream isolates. There were 35 singletons, four groups (group 1: ST37, ST1779; group 2: ST11, ST340; group 3: ST347, ST1023; group 4: ST485, ST45) and one clonal complexes (ST218, ST23, ST2248) with ST23 as the putative founder. One blob represented one ST and the size reflected the number of isolates belonging to it.
The most common type of NDM found was NDM-1. NDM-5, an emerging carbapenemase in K. pneumoniae, should attract our attention.
Our study described the phenotypic and molecular properties of K. pneumoniae bloodstream isolates in Shanghai for the first time. This study also suggested ST11 K. pneumoniae harbored bla KPC−2 had absolute predominance in carbapenemase producers, and NDM-5 was an emerging carbapenemase. Although our conclusion based on a single hospital cannot be directly extrapolated to the whole area, it provides the stepstone for the future expanded research associated with multicenter and further resistant mechanism surveillance to prevent further possible dissemination in this region.

AUTHOR CONTRIBUTIONS
Conceived and designed the experiments: LH, SX, SW, and SZ. Performed the experiments: SW and SX. Analyzed the data: SX and SW. Contributed reagents/materials/analysis tools: LH, YN, and XG. Wrote the paper: SX, SW, LH, WW, FG, and JQ.

FUNDING
This work was supported by Special Fund for Health-scientific Research in the Public Interest: Research & application for the prevention & control of nosocomial infections caused by multi-drug resistant bacteria (201002021) and The Shanghai 3-Year Plan of the Key Subjects Construction in Public Health-Infectious Diseases and Pathogenic Microorganism (15GWZK0102). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.