Carbapenemase-Producing Klebsiella pneumoniae in Romania: A Six-Month Survey

This study presents the first characterization of carbapenem-non-susceptible Klebsiella pneumoniae isolates by means of a structured six-month survey performed in Romania as part of an Europe-wide investigation. Klebsiella pneumoniae clinical isolates from different anatomical sites were tested for antibiotic susceptibility by phenotypic methods and confirmed by PCR for the presence of four carbapenemase genes. Genome macrorestriction fingerprinting with XbaI was used to analyze the relatedness of carbapenemase-producing Klebsiella pneumoniae isolates collected from eight hospitals. Among 75 non-susceptible isolates, 65 were carbapenemase producers. The most frequently identified genotype was OXA-48 (n = 51 isolates), eight isolates were positive for bla NDM-1 gene, four had the bla KPC-2 gene, whereas two were positive for bla VIM-1. The analysis of PFGE profiles of OXA-48 and NDM-1 producing K. pneumoniae suggests inter-hospitals and regional transmission of epidemic clones. This study presents the first description of K. pneumoniae strains harbouring bla KPC-2 and bla VIM-1 genes in Romania. The results of this study highlight the urgent need for the strengthening of hospital infection control measures in Romania in order to curb the further spread of the antibiotic resistance.


Introduction
Carbapenemase-producing Enterobacteriaceae (CPE), notably Klebsiella pneumoniae, produce serious infection (urinary tract infections, septicemia, pneumonia, and intra-abdominal infections) in debilitated and immunocompromised patients, in association with prolonged hospitalization and increased fatality ranging from 24% to 70%, depending on the study population [1]. CPE are spreading globally as multidrug-resistant pathogens for which there are only few treatment options available [1].
There are no national statistics in Romania regarding CPE or a national surveillance system for CPE, but a national sentinel surveillance system for nosocomial infections including invasive Klebsiella pneumoniae infections is operating. Also, as member of EARSNet, Romania is reporting yearly the number of invasive K. pneumoniae isolates which are resistant to carbapenems, but not the mechanisms of resistance, including carbapenemase producing capacity or type of enzyme. Clinical microbiology laboratories are currently performing the phenotypic diagnostic of carbapenem-non-susceptible isolates. Voluntary some of them which are investigating the carbapenemase-production capacity of K. pneumoniae isolates, are sending the strains for confirmation and further characterization to National Expert Laboratory (NEL) from Cantacuzino National Institute of Research-Development for Microbiology and Immunology (Cantacuzino NIRDMI).
In Romania, little is known about the distribution and spread of carbapenemase-producing Klebsiella pneumoniae, and the type of carbapenemases produced [2,3,4]. Only two studies have been published about CPE strains in Romania [3,4]. First study is describing the local distribution of carbapenemase encoding genes (bla NDM-1 , bla OXA-48 and bla OXA-181 ) in nine clinical isolates collected in an emergency university hospital located in the center of Romania, whereas the second study presents the distribution of CPE in two hospitals in Bucharest during one year (2011-2012) [3,4].
The present study was performed as part of the European Survey on Carbapenemase-Producing Enterobacteriaceae (EuSCAPE) [2] aiming to provide a six-months snapshot about the occurrence of the carbapenemase-producing K. pneumoniae isolated from clinical specimens of individual patients in Romanian hospitals. Moreover, this study identified the type of carbapenemase and determined the genomic diversity of carbapenemase-producing K. pneumoniae using Pulsed Field Gel Electrophoresis (PFGE). The results of this study are expected to contribute to a better understanding of the resistance genotype and geographical spread of carbapenemase-producing K. pneumoniae strains in Romania, which could inform more targeted infection control measures.

Study population
According to EuSCAPE protocol each participating country should enroll a number of hospitals in a geo-demographically representative manner, proportional with population size of the country. Romania is divided in eight developmental zones: north-western, north-eastern, western, center, southwest, southern, southern-east and Bucharest.
A number of twelve hospitals from every developmental zone of the country, having laboratory capacity to detect carbapenem resistance were invited in September 2013 to participate to the study. Six hospitals from Bucharest and two hospitals from Iasi and Bacau counties located in the north-eastern part of the country contributed with isolates. The population catchment of hospitals from capital city are comprising people living in Bucharest city but also in counties from southern and southern-east zones of Romania.
All hospitals participated on a voluntary basis to the project. During the six-month survey, starting in November 2013 ending in April 2014 (according to the EuSCAPE protocol), each hospital collected and dispatched the first 10 carbapenem-non-susceptible K. pneumoniae isolates from single successive patients to the NEL at the Cantacuzino NIRDMI for species confirmation, and further phenotypic and genotypic characterization. Each hospital also sent to NEL the information about the total number of K. pneumoniae isolates (both first successive nonsusceptible and susceptible to carbapenem) enrolled in the study as well as the total number of isolates collected during the survey period (Table 1).
According to the protocol, the participating hospitals collected information about their bed size, catchment populations and some epidemiological data, such as age and gender of the patients, patient location within the hospitals (outpatient, normal ward, ICU), clinical relevance of the isolate (colonization versus infection), hospital acquisition or community onset, hospital admissions and history of travel during the previous six months.

Bacterial strains
Bacterial isolates from hospitalized patients were identified in the hospital diagnostic laboratories using routine identification procedures [5], and antimicrobial susceptibility testing was performed by the disk diffusion method.

Confirmation of carbapenem susceptibility
K. pneumoniae isolates were sent to the National Expert Laboratory and confirmed as carbapenem-non-susceptible by disk diffusion method using imipenem, meropenem and ertapenem disks. Screening for carbapenemase production was performed by a combination of Kirby-Bauer disk-diffusion methods according to the EUCAST guideline [6], MastDisc ID inhibitor combination disks (MDI) (Mast Group Ltd, UK) and the biochemical Carba NP II tests based on manufacturer's instructions [7]. Moreover, a temocillin disk (30μg) was used for presumptive detection of OXA-48-like enzyme producers. MastDisc ID and Carba NP II were used for phenotypic confirmation of carbapenemase, whereas PCR assays were used for their genetic confirmation.

Molecular detection and genetic characterization of carbapenemase encoding genes
Screening for the presence of most clinically relevant K. pneumoniae carbapenemases according to the EuSCAPE protocol was performed by amplification of the genes bla OXA-48-like , bla NDM , bla KPC and bla VIM using PCR primers described in Table 2. Total genomic DNA from 75 carbapenem-non-susceptible K. pneumoniae strains used in all PCR assays was obtained by thermal lysis (boiling a fresh cultured colony in 200 μl distilled water and subsequent centrifugation at 14,000 rpm).
For further characterization of genes encoding NDM, VIM and KPC carbapenemases a second PCR was performed with specific primers ( Table 2). For OXA-48-like carbapenemase the same primers used for PCR screening were used for sequencing. Obtained amplified products were sequenced using BigDye Terminator v 3.1 and 3100 Avant Genetic Analyser (Applied Biosystems). The sequences were edited and aligned using the BioEdit version 7.0.5.3. software package.
The genetic relatedness among carbapenem-non-susceptible isolates was evaluated using PFGE analysis according to the standard protocol designed by the Centre for Disease Control and Prevention Atlanta (http://www.cdc.gov/pulsenet/pathogens/index.html), and recommended within the ECDC study project for molecular typing of bacteria involved in food and water diseases (FWD) [14]. PFGE of XbaI-digested total DNA was performed as described for E. coli. PFGE profiles were compared using BioNumerics v 5.1. We defined a PFGE type based on a similarity cut-off of 80% (Dice coefficient, represented by UPGMA, 1.5% optimization and 1.5% tolerance). Different PFGE types were given in alphabetical order. Every band difference within a PFGE type resulted in a subtype which was given a numerical order.

Ethics Statement
Informed written consent was obtained from all participants in this study after explanation of the procedure and the purpose of the study. The study was approved by the review boards of the Research Ethics Committee, Cantacuzino National Institute of Research-Development for Microbiology and Immunology.

Bacterial isolates
All the 75 isolates preliminarily identified in the hospital laboratories participating in the EuSCAPE project were confirmed as K. pneumoniae by the National Expertise Laboratory in Cantacuzino NIRDMI. They originated from different anatomical sites: urine (n = 39), lower respiratory tract (n = 17), blood (n = 13), wound (n = 4), puncture sites (n = 1) and peritoneum (n = 1). About 51% of carbapenemase-producing isolates were from patients with hospital-acquired infections, 26% had a community onset, whereas for 23% the epidemiological context was unknown. The history of previous hospitalization and travel remained unknown for all patients.
Phenotypic tests for carbapenemase detection showed that 65 out of 75 carbapenem-nonsusceptible isolates contained one of three Ambler classes of carbapenemase, A (KPC-type), B (metallo-beta-lactamase-MBL) or D (OXA-type), whereas for 10 isolates no carbapenemase activity was detected.
For 51 isolates, testing for temocillin susceptibility showed absence of inhibition zones in the disk diffusion test (disk load 30μg) indicative of OXA-48-like carbapenemase. MastDisc ID and Carba NP II identified 10 isolates positive for class B carbapenemase, 4 isolates as class A carbapenemase producers, whereas the remaining 51 carbapenemase-producing isolates could be suspected as class D carbapenemase producers.

Genetic confirmation of carbapenemase-producing isolates
Distribution of genes encoding carbapenemases among carbapenemase-producing isolates. The PCR results showed that 51 strains harbored the bla OXA-48-like gene, 8 strains had the bla NDM gene, whereas the bla KPC and bla VIM genes were present in 4 and 2 strains, respectively.
bla OXA-48-like gene positive K. pneumoniae isolates were identified in all hospitals, while bla NDM was detected in half and bla KPC and bla VIM were identified in only three and one of the participating hospitals, respectively (Table 3). Among the remaining 10 carbapenem-non-susceptible isolates none of the four carbapenemase target genes were detected.
The sub-type of carbapenemase genes among carbapenemase-producing isolates. DNA sequencing allowed for an identification of the sub-type of the respective carbapenemase genes in all carbapenemase-producing K. pneumoniae isolates. Thus, all sequenced KPC producers' harboured bla KPC-2 genes, whereas all OXA-48-like producers had a bla OXA-48 gene. All NDM producers had a bla NDM-1 gene, whereas all VIM producers carried the bla VIM-1 gene.

Discussion
This is the first systematic multi-centre study describing the occurrence of carbapenem-nonsusceptible K. pneumoniae isolates and their genetic resistance determinants responsible for carbapenemase production in Romania.
Our study shows that OXA-48 carbapenemase has been the most prevalent carbapenemase during the study period, being detected in 79% of carbapenemase-producing K. pneumoniae isolates in Romanian hospitals. Importantly, strains that produce this carbapenemase were spreading among all participating hospitals in this study. OXA-48 had first been identified from a clinical K. pneumoniae isolate recovered in Istanbul, Turkey, in 2001 [15]. For several years, OXA-48 appeared to be confined to Turkey, as almost all OXA-48 beta-lactamase producers were reported either among patients hospitalized in Turkey or with a link to that country [16,17]. Since 2008, this carbapenemase has been identified in many other countries, mainly in K. pneumoniae [15,17]. In addition to sporadic cases, an increasing number of outbreaks due to OXA-48-producing K. pneumoniae are currently reported, not only in Turkey but also in Belgium, France, Greece, the Netherlands and Spain [15,[17][18][19][20]. The successful spread of OXA-48 producers is now considered an epidemic threat as it represents an important source of multidrug resistance in K. pneumoniae in Europe [21,22].
In Romania the presence of OXA-48 producing K. pneumoniae isolates has only occasionally been reported. The first study published was conducted between January 2010 and September 2012 in an emergency university hospital from the central part of Romania, where three K. pneumoniae ST101 isolates with indistinguishable PFGE profiles were reported [3]. The second study was performed between October 2011 and November 2012 in two hospitals in Bucharest, which identified 31 patients with K. pneumoniae isolates harbouring bla OXA-48 gene [4]. No typing results were reported.
In the present study PFGE analysis of the 51 OXA-48 producing K. pneumoniae strains showed that 47 out of 51 (92%) strains isolated from all hospitals belonged to the same epidemic clone, whereas the remaining four isolates had different PFGE profiles. These results suggest an epidemic expansion of a major OXA-48 positive clone which evolves hospitals in different regions of the country.
We also identified eight K. pneumoniae isolates harbouring the bla NDM-1 gene. This gene may be carried by different plasmid types, most of them co-harboring multiple but variable resistance determinants [23]. A few plasmids carrying bla NDM-1 also encode other carbapenemases, including OXA-181-and VIM-types [24]. The first New Delhi metallo-beta-lactamase (NDM-1) producing strains detected in Europe were associated with importation from the Indian subcontinent. More recently, countries in the Balkan region (Serbia, Montenegro, Bosnia and Herzegovina) also reported patients infected and colonized with NDM-1 producers [23,[25][26]. In Romania, the presence of this enzyme among K. pneumoniae was recognized previously [3,4]. In addition, anecdotal reports suggested the introduction of NDM-1 producing strains by travelers into Romania [27]. Their PFGE profiles exhibited a 80% similarity, indicative of a single clone, expanding within and between four hospitals from different Romanian regions. According to an international staging system [28] Romania thus would assume an epidemiological stage 4 for both OXA-48 and NDM-1 producing K. pneumoniae. We also found isolates producing VIM-1 as well as KPC-2 carbapenemase. Strains with these types had previously not been identified in Romania. The PFGE profiles of two VIM-1 positive strains collected from patients hospitalized in the same hospital (H7) were seemingly different, showing a similarity < 80%. The PFGE profiles of four K. pneumoniae isolates harbouring bla KPC-2 gene were however rather similar, with similarity ranging from 90% to 98%, being part of the epidemic clone, showing both intra-and inter-hospitals (H3, H4 and H5 hospitals) transmission.
The PFGE results of this study are testimony of a clonal dissemination of K. pneumoniae harbouring all of the four carbapenemase genes across the sample of Romanian hospitals which have been enrolled and bears the hallmark of a much larger i.e. national epidemic of carbapenemase-producing Enterobacteriaceae in Romanian hospitals.
Hospitalized patients from whom the carbapenemase-producing K. pneumoniae isolates were collected where kept in contact isolation according to guideline regarding prevention of hospital infection transmission.
The results of PCR tests performed on 10 carbapenem resistant strains with no carbapenemase type detected by phenotypic tests showed that in these strains none of the carbapenemase targeted genes were detected, suggesting that other resistance mechanisms to carbapenems (ESBL S or AmpC β-lactamases combined with the loss of outer membrane porins OmpK35 and/or OmpK36) could have been involved [29,30]. It could also be possible that carbapenemases other than those included in this study may have caused resistance in these isolates. But this seems unlikely given the negative results of our phenotypic confirmation tests and the scarcity with which other carbapenemases are reported in Europe. These results demonstrate the importance of molecular tests in the confirmation of carbapenemase-producing Klebsiella pneumoniae.

Limitations of the study
This study has several limitations regarding the rate of response of 67% of invited hospitals participating in the survey as well as the time of only six months. Although hospitals were invited in a geo-demographically representative manner the selection bias could occur due to the fact that participating hospitals in the survey were those having a catchment population for counties from southern, southern-east, north-eastern and Bucharest zones of Romania and obtained data could be extrapolated only for those zones. Clinical data of the isolates as well as data on antibiotic consumption are not presented in this study.

Conclusions
This study shows the first detection of K. pneumoniae strains harbouring bla KPC-2 and bla VIM-1 genes in Romania. The carbapenemase most frequently detected is OXA-48, representing 79% of carbapenemase-producing K. pneumoniae strains. This carbapenemase was detected in isolates collected from patients hospitalized in all of the participating hospitals, and the PFGE results suggest the spread of a single clone in all sampled hospitals with outbreaks in two of them. Likewise, NDM-1 positive strains have also been identified in four hospitals and PFGE results indicate a similar epidemic dynamic as for OXA-48. Based on these findings Romania would assume a epidemiological stage 4 for OXA-48 and NDM-1 carbapenemases according to an international staging system. Before starting this survey by self-assessment realized by the national expert in March 2013, the epidemiological stage 1 was established for carbapenemase-producing isolates from Romania. The present study allowed updating and upgrading the epidemiological stages for all investigated carbapenemases.
The results of this study show that for efficient monitoring of the rapidly evolving spread of carbapenemase-producing K. pneumoniae the development of a national management system for CPE through a dedicated surveillance program, the provision of references services, the obligation to notify cases to health authorities and a national plan for containment of/or preparedness to contain CPE are necessary. In addition European Council Recommendation of 15 November 2001 [31] on the prudent use of antimicrobial agents in human medicine should strongly be considered including Strategies towards sufficient surveillance systems on usage of antibiotics and antimicrobial resistance, control and preventive measures, promoting education and training of health care workforce and inform the general public.