Nosocomial outbreak of KPC-2- and NDM-1-producing Klebsiella pneumoniae in a neonatal ward: a retrospective study

Background The spread of resistance to carbapenems among Enterobacteriaceae has become a major public health problem in recent years. In this study, we describe an outbreak of Klebsiella pneumoniae in the neonatal ward. First, we aimed to study the drug resistance, genetic relatedness, and transmission mechanism of carbapenem-resistant K. pneumoniae; second, we implemented infection control measures to contain the outbreak. Methods We investigated 27 non-repetitive strains isolated from neonates and five strains cultured from around the neonatal ward. Polymerase chain reaction (PCR), the agar dilution method, and multilocus sequence typing (MLST) were used to analyze the resistance gene(s), antimicrobial susceptibility, and homology, respectively. Health-care personnel education, hand hygiene, outer gown changing, and infected patient isolation were strictly enforced. Results Our antimicrobial susceptibility results show that all strains were multidrug-resistant. MLST and PCR results revealed that, in this study, all of the KPC-2-producing strains are Sequence Type (ST) 11 (ST11) (n = 22) and all of the NDM-1-producing strains are ST20 (n = 4) or ST888 (n = 1). The environmental strains were identified as KPC-2-positive K. pneumoniae ST11 (n = 3) and NDM-1-positive K. pneumoniae ST20 (n = 2). The percentages of isolates with the extended-spectrum-β-lactamases CTX-M-15, blaCTX-M-14, blaTEM-1 were 9.4, 84.3, and 68.8 %, respectively. AmpC β-lactamase genes were not detected in our isolates. Conclusions KPC-2-positive K. pneumoniae ST11 and NDM-1-positive K. pneumoniae ST20 were associated with this outbreak. The identification of these isolates in samples from radiant warmers and nurses suggests that hospital cross-transmission played a role in this outbreak. Active infection control measures were effective for controlling this multidrug-resistant K. pneumoniae outbreak. Electronic supplementary material The online version of this article (doi:10.1186/s12879-016-1870-y) contains supplementary material, which is available to authorized users.


Background
Carbapenem-resistant Enterobacteriaceae (CRE) have increasingly been found to cause nosocomial infections during the past decade; they primarily affect critically ill patients [1]. Carbapenem-resistant Klebsiella pneumoniae (CR-KP) is one of the most important types of CRE, and it has spread extensively around the world [2]. Carbapenems are the most effective drug for treating infections caused by K. pneumoniae that produce extended-spectrum beta-lactamases (ESBLs). However, with the increase of carbapenem resistance in K. pneumoniae, options for treatment have become limited [3].
Carbapenemases, including KPC, GES, SME, NMC, IMI NDM, IMP, VIM, and OXA-48, are the main mechanisms of carbapenem resistance in CRE. Other mechanisms of carbapenem resistance in these bacteria include the absence or decreased expression of outer membrane proteins and mutations in the ampC enzyme or ESBLs [4].
Outbreaks of hospital-associated CRE infections are not rare [5,6]. In recent years, outbreaks of CRE infections in neonates have been rising and have attracted increasingly more attention [7,8]. Here, we noted that the frequency of K. pneumoniae isolated from newborns in our neonatal ward had dramatically increased compared with that of previous years. Therefore, we collected CR-KP in 2015 from January to November to study the resistance and transmission mechanisms of carbapenem-resistant strains in the neonatal ward.

Setting of the study
A retrospective study was conducted in an 11months period between January 2015and November 2015 in the neonatal ward of Tongji Hospital (TJH). The neonatal ward of TJH receives approximately 1750 admissions per year and it has six rooms (60 beds) for newborns. The 60 % of patients come from maternity ward and 30 % of neonates were outpatients.

Bacterial isolates collection and identification
Thirty-two carbapenem-resistant K. pneumoniae were included in this study. Twenty-seven isolates were collected from different patients. At the same time, according to previous investigation, hospital infection control staff collected environmental samples from high frequency contact areas in neonatal ward in September 8, 2015, including bed railing, stethoscope, personal digital assistant, suction apparatus, venous catheter, radiant warmer, nurses' hand, thermostat button and ventilator machine. Samplings were accomplished before routine cleaning, 89 environmental samples and 40 samples from nurses' hands were obtained. All isolated strains were identified by conventional biochemical methods, then were confirmed by PCR amplification of 16S rRNA gene and sequencing [9].

Susceptibility testing
Antibiotic susceptibility testing was assayed by agar dilution method to determinate the minimum inhibitory concentration (MICs) of ampicillin, ceftazidime, cefotaxime, cefepime, cefuroxime, cefoxitin, gentamicin, levofloxacin, amikacin, ciprofloxacin, aztreonam, imipenem, meropenem, trimethoprim-sulfamethoxazole, piperacillintazobactam, tigecycline and colistin. All antibiotics, except tigecycline and colistin, were interpreted according to the standard of the Clinical and Laboratory Standards Institute (CLSI) document M100-S22. For tigecycline and colistin, the European Committee on Antimicrobial Susceptibility Testing (EUCAST) breakpoint was used. E. coli ATCC 25922 and P.aeruginosa ATCC27853 were used as quality control.

Infection control interventions
Strict infection control measures were implemented according to the Guidance for Control of CRE in November, 2015 [14]. Healthcare personnel are familiar with proper hand hygiene technique and wash their hands properly before entering and upon leaving patient rooms. Hand hygiene adherence was monitored by video. Patients who are suspected or confirmed with CRE should be placed in a private room in which staff and visitors are required to don gloves and gowns for entry. Extensive cleaning of shared equipment and the use of disposable materials (when feasible) were implemented.

Epidemiological characteristics of K. pneumoniae isolates
The molecular epidemiology of carbapenem-resistant Enterobacteriaceae was studied from January to November 2015. During this period, There were 1519 patients admitted to hospital,including 590 females (38.9 %) and 929 males (61.1 %). The sex ratio of all patients is similar to infection group (61.1 vs 63.0 %). The average age of the CR-KP infection (32.5 ± 22.7) ws older than all patients (11.2 ± 23.4).27 consecutive CRE were isolated from 27 different patients. The characteristics of the CR-KP isolated in this study are shown in Table 1. Twenty seven of the isolates are from newborns in the neonatal unit, including 17 males and 10 females, with a mean age of 32.5 days (range, 1-85 days). The sources of these specimens were blood (n = 9), urine (n = 8), aspiration catheter (n = 5), and sputum (n = 5). Meanwhile, ten suspicious colonies were found and further identified by biochemical method, and finally five CR-KP isolates were cultured from radiant warmers or nurses' hands during the peak of the outbreak. Notably, strains producing NDM-1carbapenemase show different antibacterial susceptibility spectra from strains producing KPC-2 carbapenemase, including gentamicin, amikacin, levofloxacin, ciprofloxacin, aztreonam, and trimethoprim-sulfamethoxazole (see Table 2).

The effect of infection control interventions
We monitored for neonatal infection with K. pneumoniae in future 4 months, and we detected multidrug-resistant K. pneumoniae in five patients, three patients, one patient and one patient, respectively, over these 4 months.

Discussion
K. pneumoniae is an important nosocomial pathogen that can cause pneumonia, sepsis, and meningitis [15]. The emergence of CR-KP has created a challenge for treating K. pneumoniae infections, the mortality caused by CR-KP incection was high [16]. However, one infected neonate has died in our study. Mortality from K. pneumoniae infection is associated with a variety of factors, including underlying disease, nutritional status, and medical treatment, which may explain the discrepancy in mortality among different studies. Underlying disease, low birth weight, antibacterial drug treatment, invasive procedures, and prolonged hospitalization have all been reported as extremely important risk factors of CRE infection [17,18]. Of the 27 patients in this study, the percentages of subjects who had a low birth weight, who were treated with antibacterial drugs, and who underwent invasive procedures were 74.1 % (n = 20), 100 % (n = 27), and 77.8 % (n = 21) (data was not show), respectively. Additionally, we noticed that each nurse needed to take care of several infants and that each newborn was usually cared for by several nurses due to scheduling reasons. Overall, the newborns in our study possessed high risk factors for CRE infection, and this outbreak was likely spread by cross-infection.
As results showed, no significant differences in sex ratio and age between infection patients and non-infection patients. The CR-KP infection probably has no relation with age and gender in our study,which was similar with other study [19]. In addition, the average days for hospitalization in TJH is 11.8 days at present. However, the average hospital stay of the 27 patients including in this study was 29.9 days. Prolonged hospitalization may contribute to spreading of the ST11 and ST20 strains in neonatal ward.  The first KPC-2-producing K. pneumoniae isolate (TJ1) was isolated from a 23-day-old boy that was confirmed as infected with multidrug-resistant K. pneumoniae at another hospital. He was transferred to our hospital for treatment on January 23, 2015, and a sputum culture confirmed his infection with a multidrug-resistant K. pneumoniae. We learned that his mother had suffered from a respiratory tract infection for 2 weeks and 4 days. Through combination therapy with cefoperazone sulbactam, amoxicillin clavulanic acid, fluconazole, and meropenem, the child recovered and was discharged on February 5, 2015. Subsequently, KPC-2-positive strains with the same phenotype appeared in the same ward. All KPC-2-positive strains were detected in patients who had been hospitalized for more than 48 h, so we speculate that the origin of the KPC-2-positive K. pneumoniae ST11 was probably TJ1. Rodloff MD and his colleagues found that carriage of KPC-2-producing K. pneumoniae can last for 3 years [20]. It probably explain continuous outbreak after index case. Regrettably, the origin of the NDM-1-producing isolates in our study is still unknown.
The number of infection with K. pneumoniae decrease sharply in future 4 months. These declining numbers suggest that our efforts to control the continuous outbreak of multidrug-resistant K. pneumoniae were at least partially effective. However, there was one or two CR-KP detected after infection control interventions. It is difficult to eradicate the CR-KP by infection control measures [21]. Infection control measures were still need to improve, including routine screening of hospitalized patients, pathogens culture of medical equipment and screening the colonization status of contemporaneous patients.
This study has several limitations. Specifically, we only performed one environmental sampling and neglect screening of patients with perianal swabs to detect the rate of colonized patients. Additional samples might have provided stronger evidence to support our conclusion that the CR-KP was spread through cross-infection. Our data is also unable to confirm whether or not the CR-KP strains isolated in the later 3 months of our study are the same as the previous epidemic strain.

Conclusions
In summary, we report that KPC-2-positive K. pneumoniae ST11 and NDM-1-positive K. pneumoniae ST20 are the two main clones that caused neonate infection in our hospital during the study period. During the outbreak, we successfully isolated these two main clones from a radiant warmer and from nurses' hands. The radiant warmer and the nurses' hands may have played an important role in harboring the K. pneumoniae producing KPC-2 or NDM-1 carbapenemases that caused persistent infection within the same ward. Following the emergence of CRE on a ward, Screening and infection control measures must be implemented to block nosocomial transmission. Active and effective infection control measures are indispensable for preventing a subsequent outbreak.

Availability of data and materials
All the data on which the conclusions of the manuscript rely is presented in the main paper and additional files 1, 2 and 3.