Clonal replacement of epidemic KPC-producing Klebsiella pneumoniae in a hospital in China

Background Klebsiella pneumoniae is a frequent nosocomial pathogen causing difficult-to-treat infections worldwide. The prevalence of Klebsiella pneumoniae carbapenemase (KPC)-producing Klebsiella pneumoniae (KPC-KP) is increasing in China. The aim of this study was to investigate the molecular epidemiology of KPC-KP in a nosocomial outbreak. Methods Fifty-four KPC-KP isolates were consecutively collected between November 2013 and August 2014 during a KPC-KP outbreak in a tertiary care hospital in Beijing, China. Antimicrobial susceptibility was determined by agar dilution. Carbapenemase, extended-spectrum β–lactamase, 16S rRNA methylase, AmpC β-lactamase, and plasmid-mediated quinolone resistance determinants were detected by PCR amplification. The genetic relatedness of isolates was analyzed by pulsed-field gel electrophoresis and multi-locus sequence typing. Results All isolates belonged to ST11 except one isolate which was identified as a new sequence type (ST2040). PFGE profile of genomic DNA revealed seven clusters, of which cluster A and C dominated the KPC-KP outbreak and cluster A was replaced by cluster C during the outbreak. PFGE of genomic DNA, S1-PFGE of plasmids, replicon typing, and drug resistant characteristics showed that clonal spread occurred during the outbreak. When compared with isolates within cluster A, all isolates in cluster C harbored rmtB and showed higher level of resistance to cefepime, amikacin, tobramycin, and tigecycline. Conclusion We reported a nosocomial outbreak of KPC-KP with clonal replacement and a new sequence type (ST2040) of KP. High degree of awareness and surveillance of KPC-KP should be given to avoid potential outbreaks, especially in ICU wards.


Patients and bacterial isolates
All Carbapenem-resistant KP strains, which were collected between November 2013 and August 2014 in this study, were isolated from different clinical wards of our tertiary care hospital with 1500 beds. Species were identified by the Vitek 2 system, followed by 16S rDNA sequencing. Crude mortality was defined as death during hospitalization. Infection-related mortality was defined as death only as a direct consequence of KPC-KP infection during hospitalization. A KPC-KP outbreak was defined as two or more laboratory-confirmed patients that were temporally related, epidemiologically linked, and infected by the same KP variant.

Antimicrobial susceptibility testing
Susceptibility tests were performed using the Vitek 2 system and the AST-GN card. MICs of various antimicrobials were determined by agar dilution method and results were interpreted according to the criteria recommended by CLSI (2014) [28].

Pulsed-field gel electrophoresis (PFGE) analysis
The clonal relatedness of KP isolates were analyzed by PFGE as described previously [29]. Prepared genomic DNA was digested using XbaI restriction enzyme on all clinical isolates. The banding patterns were analyzed by the BioNumerics software. The genetic similarity was calculated by dice coefficients and dendrograms were constructed by the unweighted pair group of arithmetic average. The analysis parameters were based on 1.5% tolerance values. Clusters were defined as DNA patterns sharing ≥85% similarity. Plasmid profiling was performed by PFGE of total bacteria genome DNA cut with nuclease S1.

Conjugation experiments
The conjugation experiments were performed by using azide-resistant E.coli J53 as the recipient strain. Cultures of donor and recipient cells in logarithmic phase (0.5 ml of each) were added to 4 ml of fresh Luria-Bertani broth and were incubated overnight without shaking. The transconjugants were selected on Mueller-Hinton agar plates containing 100 μg/ml ampicillin and 300 μg/ml sodium-azide. Antibiotic resistance genes and replicon typing were performed as described previously in this study.

Statistical analysis
All statistical analysis were performed using the SPSS software package version 20.0. Continuous variables (age and Charlson score) were summarized as medians and were compared using non-parametric Mann-Whitney U test. Categorical variables were summarized as percentages and were compared using the pearson chi-square test. A p-value of less than 0.05 was considered statistically significant.

Description of isolates
A total of 54 non-duplicate KPC-KP isolates were included in this study. All patients, where the 54 isolates were isolated from, were identified as clinical infections based on the isolation of KPC-KP from clinical samples and medical diagnosis established by physicians according to the clinical manifestations and the antibacterial effects. The predominant type of samples from which KPC isolates were isolated was sputum (n = 41, 75.9%), followed by bronchoalveolar lavage fluid (n = 5, 9.3%), blood (n = 4, 7.4%), secretions (n = 2, 3.7%), urine (n = 1, 1.9%), and peritoneal fluid (n = 1, 1.9%). Of note, 77.8% (42/54) of isolates were isolated from patients in two ICU wards (ICU-1, respiratory medicine department; ICU-2, department of critical care medicine), and remaining 22.2% (12/54) of isolates were isolated from other seven medical wards. Specifically, the ICU-1 ward (with 14 beds) and ICU-2 ward (with 12 beds) belonged to two departments and were located in two different buildings with a distance of about 100 m. There were no intercommunications of patients, staffs, and medical instruments between the two ICU wards. Patients in the two ICU wards were scattered in patient rooms or beds. Resistance genes encoding class B carbapenemase (bla NDM , bla SIM , bla IMP , bla VIM , and bla SPM ) and AmpC β-lactamase (bla ACC , bla FOX , bla MOX , bla DHA , bla CIT /bla SPM , and bla EBC ) were not detected in any of the isolates.

Distribution of plasmid replicons
PBRT revealed that 55.6% (30/54) and 37.0% (20/54) of KP isolates contained IncFIIAs and IncFII, respectively. Specifically, 96.3% (26/27) of isolates in cluster A contained IncFIIAs, and 100% (20/20) of isolates in cluster C contained IncFII. After sequencing, the IncFIIAs confirmed in this study was identified as a new subtype of IncFII, which was characterized with low nucleotide identity (90%) when compared with the IncFIIAs replicon of Salmonella virulence plasmids. , fewer cases mainly caused by cluster C. Of note, cluster A was replaced by cluster C during this KP outbreak. 85.2% (23/27) of isolates in cluster A contained one plasmid with IncFIIs replicon which harbored resistance gene bla KPC , bla TEM , and bla SHV . According to PFGE of genomic DNA, S1-PFGE of plasmids, replicon typing, and drug resistant characteristics (Fig. 1), two main groups in cluster C were suggested to be derived from different origins, of which one group consisted of isolates from the ICU ward of respiratory department (ICU-1) and the other group consisted of strains from the ICU ward of department of critical care medicine (ICU-2). 75% (9/12) of isolates in the former group mainly contained three plasmids with IncFII replicon which harbored bla KPC-2 , bla TEM-1 , bla SHV-11 (or bla SHV-12 ), bla CTX-M-14 , acc(6′)-Ib-cr, qnrS, and rmtB. However, all strains (n = 7) in the latter group contained two plasmids with FII replicon which harbored bla KPC-2 , bla TEM-1 , bla SHV-11 (or bla SHV-12 ), bla CTX-M-14 , qnrS, and rmtB. Specifically, clonal replacement (from cluster A to cluster C) occurred in ICU wards, especially in ICU-1 (Fig. 1).

Mortality discrepancy between clones
Medical records revealed that there were no significant differences in gender (male percentage, 81.5% vs 70.0%; median age, 75 vs 77; median Charlson scores, 6.0 vs 6.0; patient distributions (from ICU wards or transferred from ICU wards), 92.6% vs 95%) in patients between cluster A (n = 27) and cluster C (n = 20). As for the crude mortality, it was significantly higher in cluster C than those in cluster A (80.0% vs 51.9%, p < 0.05). The infection-related mortality in cluster C was higher than those in cluster A although the difference was statistically nonsignificant (25.0% vs 14.8%, p > 0.05).

Discussion
KPC-KP outbreaks have been previously reported in several provinces of China [4,[7][8][9][10][11][12][13][14][15][16]. Compared with Western countries, the prevalence of KPC-KP in China is relatively low (2.9%), as described in a 4000-bed tertiary care hospital [32]. However, KPC-KP is an increasing cause for great clinical concern due to its high antibiotic resistance and pathogenicity. In this study, a nosocomial outbreak caused by KPC-KP is described. As shown in Fig. 1, the first clonal replacement of epidemic KPC-KP occurred in ICU wards is confirmed by PFGE of genomic DNA, S1-PFGE of plasmids, replicon typing, and drug resistant characteristics. RP-KP also seems to have a higher dissemination advantage over RN-KP. Although this outbreak was severe in the ICU wards, it was terminated by effective measures, such as hand hygiene intervention, patient isolation, and environmental cleaning. This study suggests that continuous surveillance system and strict infection control measures are necessary and urgent to prevent the spread or outbreak of KPC-KP in health care facilities. It is noteworthy that bacterial culture and susceptibility test require 3-5 days and outbreak might occur during this period of time. Thus, rapid identification of the pathogens is crucial in the active surveillance culture program. On the other hand, considering that colonization frequently precedes infection, early identification of colonized cases might also reduce infections and potential outbreaks.
Although KPC-KP isolates have been identified worldwide, the spread has been proved to be caused by the dominant KP clones (ST258 and ST11). In China, ST11 is the dominant clone of KPC-KP [4]. Some new sequence types of KP have also been frequently described [17,33]. Here, we reported a new sequence type (ST2040). Compared with the majority of isolates in cluster C (ICU-2), this clone only contains two plasmids (Fig. 1). During the outbreak period, the new ST2040 was only isolated once and didn't become the epidemic strain. Although resistances are proved to be associated with reduced fitness and virulence of pathogens [34,35], the clinical resistant KP isolates could survive in those high-density antibiotic environments (health care facilities and day care centers). Thus, as a multi-drug resistant clone, the fitness and pathogenicity of ST2040 might need to be investigated further.
The most frequently encountered ESBLs are TEM-, SHV-, and CTX-M-type β-lactamases. The bla CTX-M , of which prevalence is increased in KP, is carried mainly by IncFII-type plasmids [36]. In this study, the bla CTX-M genes in isolates are also carried by IncFII-type plasmids and could be transferred by conjugation. On the contrary, IncFIIAs plasmids (a new replicon subtype), which harbor bla KPC-2 , bla TEM-1 , and bla SHV-11 in isolates of cluster A, seems to be hard to be conjugative.
Isolates producing 16S rRNA methylase frequently exhibit high level of resistance to almost all clinically important aminoglycosides through methylation of the aminoglycoside-binding site [39]. The resistance mechanism caused by 16S rRNA methylase has been reported in KP, and the prevalence of 16S rRNA methylase genes among clinical KP isolates in China is increasing [40,41]. KPC-KP isolates are multi-drug resistant, but are usually susceptible to aminoglycosides [42][43][44]. Aminoglycosides also have a significantly higher microbiologic clearance rate than polymyxin B or tigecycline [45]. Here, 37.0% (20/54) of RP-KP showed strong resistance to amikacin (100%) and tobramycin (100%). Co-production of 16S rRNA methylases in KPC-KP could leave limited therapeutic choices for antibacterial treatment and might be associated with higher mortality in this study.

Conclusions
In summary, we reported a nosocomial outbreak of KPC-KP with clonal replacement and a new sequence type (ST2040) of KP in our hospital. ST11 is the dominant clone. The outbreak mainly occurred in ICU wards. The clonal spread was responsible for this outbreak. Our study also suggested that a high degree of awareness and surveillance of KPC-KP should be given to avoid potential outbreaks, especially in ICU wards.