Introduction

Carbapenem-resistant Klebsiella pneumoniae (CRKP) has become a major public concern, which was mainly mediated by the production of carbapenemases [1]. As a novel carbapenemase, NDM-1 was initially reported in 2008 from the K. pneumoniae and Escherichia coli isolated from a Sweden patient who had received medical care in India [2]. Since then, carbapenem-resistant Enterobacteriaceae (CRE) bearing bla NDM-1 and its nine minor variants have been identified all over the world [3–5].

Since the first NDM-1-producing K. pneumoniae was detected in Nanchang, China in 2013, they have spread rapidly in mainland China [6]. To date, NDM-1-producing K. pneumoniae have been reported in several regions of China [7, 8]. However, outbreak in neonates remains uncommon in China. In this study, we observed a soaring number of CRKP sourced from neonatal intensive care unit (NICU) and neonatal wards in less than three months.

Materials and methods

Bacterial strains

Twenty-two non-duplicate sequential strains of CRKP were isolated from Shanghai Children’s Hospital to investigate the epidemiological characteristics. Most (86.5 %, 19/22) of the 22 CRKP were isolated from sputum, and 4.5, 4.5, and 4.5 % were isolated from trachea cannula, urine, and pus, respectively. All the strains were identified using the VITEK 2 Compact system (bioMérieux, France). Escherichia coli ATCC 25922, Salmonella ser. Braenderup H9812, and E. coli J53 (sodium azide resistant) were used as the quality control for antimicrobial susceptibility testing, reference marker for pulsed-field gel electrophoresis (PFGE), and recipient strain for conjugation experiment, respectively.

Antimicrobial susceptibility testing and β-lactamase characterization

Antimicrobial susceptibility testing was performed using the agar dilution method. The results were interpreted following the criteria of the Clinical and Laboratory Standards Institute (CLSI; 2014) [9]. Breakpoint minimum inhibitory concentrations (MICs) of tigecycline were determined following the guidelines of the U.S. Food and Drug Administration (MIC ≤2 mg/L denoting susceptibility and ≥8 mg/L denoting resistance). The presence of genes encoding β-lactamase, including CTX-M-type extended-spectrum β-lactamases (ESBLs), plasmid-borne AmpC β-lactamases, and carbapenemases, were investigated using primers previously described [10–14]. Polymerase chain reaction (PCR) amplicons were sequenced and the DNA sequences obtained were compared with those available in the NCBI GenBank database using BLAST searches.

Transfer of carbapenemase resistance, plasmid analysis, and bacterial genotyping

Conjugation experiment was carried out with E. coli J53 as the recipient to determine the transferability of the carbapenemase gene, as described previously [15]. Whole-cell DNA of clinical strains embedded in agarose gel plugs, digested with S1 nuclease, was separated by PFGE. Plasmids obtained by PFGE were transferred to nylon membranes and hybridized with digoxigenin-labeled bla NDM-1-specific probes. Clonal relationships were analyzed using PFGE of XbaI-digested genomic DNA as previously described, and the results were analyzed according to the criteria proposed by Tenover et al. [16, 17]. Multilocus sequence typing (MLST) for these isolates was performed as described previously [18].

Clinical epidemiology

The clinical data were reviewed for each patient. Several parameters were assessed, including demographics, prior use of broad-spectrum antimicrobial agents, particularly carbapenems, and potential risk factors for infection or colonization with CRKP. Infection or colonization with CRKP was defined according to the definition of nosocomial infections from the Centers for Disease Control and Prevention (CDC) [19].

Results

Antimicrobial susceptibility testing and β-lactamase characterization

All 22 isolates were resistant to cephalosporin and β-lactam/β-lactamase inhibitor combinations, while none were resistant to tigecycline or colistin. Overall, 59.1, 77.3, and 100 % of these strains were resistant to imipenem, meropenem, and ertapenem, respectively. And 4.5 % (1/22), 4.5 % (1/22), and 90.9 % (20/22) were resistant to amikacin, ciprofloxacin, and aztreonam, respectively (Table 1). Seventeen isolates were bla NDM-1-positive by PCR and DNA sequencing. Of the 17 bla NDM-1-positive isolates, 47 % (8/17) co-harbored bla CTX-M-15 and 5.9 % (1/17) co-harbored bla DHA-1. No carbapenemase genes were detected among the remaining four isolates; however, 50 % (2/4) of isolates produced bla CTX-M-15. Half (11/22) of the isolates carried class 1 integron, while sequence data showed that none has resistant gene cassettes. No other PCR products were obtained for any of the other genes investigated (Table 1).

Table 1 Antimicrobial susceptibilities and characteristics of the 22 carbapenem-resistant Klebsiella pneumoniae (CRKP) strains and their transconjugants (TC)
Full size table

Transfer of carbapenemase resistance, plasmid analysis, and bacterial genotyping

The 17 bla NDM-1-positive K. pneumoniae isolates were selected for conjugation. The results of conjugation experiments indicated that the plasmids with bla NDM-1 from 15 isolates were successfully transferred from donors to recipient E. coli J53, and the conjugants exhibited high resistance to carbapenems, consistent with the detection of bla NDM-1. The MICs of imipenem, meropenem, and ertapenem for the conjugants ranged from 1 to 8 mg/L, and the antimicrobial susceptibility patterns of the conjugants were similar to their donors (Table 1). Hybridization analysis showed that the plasmids carrying the bla NDM-1 gene were approximately 50–240 kb in size (Fig. 1).

Fig. 1
figure 1

S1 nuclease pulsed-field gel electrophoresis (S1-PFGE) patterns (top) of 15 bla NDM-1-positive Klebsiella pneumoniae and Southern hybridization (bottom). Lane M: marker (Salmonella H9812); lanes 1, 8, 29, 36, 37, 45, 51, 58, 65, 70, 77, 83, 94, 42, and 67: clinical strains

Full size image

Five distinct PFGE patterns (PFGE types A–E) were observed among the 22 CRKP isolates: type A (11/22, 50 %), type C (8/22, 36.4 %), and one isolate each for type B, type D, and type E (Fig. 2). Five distinct MLST sequence types (STs) were observed among the 22 CRKP isolates, including ST76 (n = 8), ST37 (n = 11), ST846 (n = 1), ST11 (n = 1), and ST571 (n = 1). Of the 17 bla NDM-1-positive K. pneumoniae isolates, 16 from inpatients were identified as ST76 (n = 7) and ST37 (n = 9) and one isolate from an outpatient belonged to ST846. One bla KPC-positive K. pneumoniae belonged to ST11 (Fig. 2).

Fig. 2
figure 2

DNA fingerprinting and multilocus sequence typing (MLST) of 22 carbapenem-resistant Klebsiella pneumoniae (CRKP) isolates

Full size image

Clinical epidemiology

The clinical details of the patients are shown in Table 1. The 22 isolates were identified in a diverse population of children. The median age of the 22 children (17 males and 5 females) was 1 month (range, 1 day to 12 years). The most common underlying conditions were neonatal respiratory distress syndrome (27.3 %, 6/22), pneumonia and bronchopneumonia (31.8 %, 7/22), and neonatal asphyxia or perinatal asphyxia (13.6 %, 3/22). Indwelling devices were used in some patients, including endotracheal tube (45.5 %, 10/22), arteriovenous intubation (31.8 %, 7/22), mechanical ventilation (9.1 %, 2/22), urinary catheter (9.1 %, 2/22), and gastric tube (9.1 %, 2/22). Most patients (81.8 %, 18/22) were treated with antimicrobial agents, including ampicillin–sulbactam, cefotaxime, ceftriaxone, imipenem, and meropenem. Most patients were improved or discharged after treatment, except for two deaths (Table 1).

Discussion

Infectious diseases caused by NDM-1-producing isolates were known to be associated with significant morbidity and mortality, which was even worse among pediatric populations due to limited therapeutic options [20]. All the isolates were susceptible to tigecycline and colistin in our study; however, tigecycline is not recommended in children because of the risk of dental staining and, currently, colistin is not available for patients in China. Besides, aminoglycosides and fluoroquinolones are also restricted in children due to nephrotoxicity and ototoxicity. In the absence of effective antibiotic therapy, early monitoring of CRKP infection or colonization on admission may play a more important role for timely control of the spread of CRKP [21].

In our study, K. pneumoniae ST76 (n = 8) and ST37 (n = 11) were predominant epidemic clones. Although both STs did not belong to the most common NDM-1-positive clones (ST14 and ST11), K. pneumoniae ST37 has also been reported in India, the UK, and the USA, while K. pneumoniae ST76 bearing bla NDM-1 was reported for the first time in this study [22]. The results of clinical epidemiology indicated that the risk factors for acquiring CRE isolates included invasive procedures (especially surgical operations), indwelling urinary catheters, change of sickbeds, and previous in-hospital cephalosporin use [23]. Our study indicated that immunodepression, invasive procedures, and prior use of broad-spectrum antibiotics might increase the chance of infection or colonization of CRE isolates. In our study, two pediatric patients died of K. pneumoniae ST37 infection, while no patients died of K. pneumoniae ST76 infection.

It is noteworthy that no carbapenemase resistance genes were detected in 18.2 % (4/22) of the CRKP strains. We suspect that a new mechanism, such as the presence of new metallo-β-lactamases or variants of certain carbapenemases, might contribute to the resistance to carbapenems. Further studies are needed to confirm this point. In this study, we reported a nosocomial outbreak of bla NDM-1-producing K. pneumoniae ST37 and ST76 in neonates. Although two outbreaks involving K. pneumoniae ST17 and ST20 were reported recently, to our knowledge, this is the first report of an NDM-1-producing K. pneumoniae ST37 and ST76 outbreak [24, 25]. The vulnerability to colonization or infection with CRE isolates among pediatric patients highlights the necessity of intervention with strict infection-control measures, including proper hand hygiene, contact precautions, and cohort nursing care, to reduce the cross-infection and avoid the rapid spread or clonal dissemination of carbapenemase-producing Enterobacteriaceae strains in healthcare facilities.