High Prevalence of New Delhi Metallo-β-Lactamase-1 (NDM-1) Producers among Carbapenem-Resistant Enterobacteriaceae in Kuwait

The aim of the study was to determine the prevalence of New Delhi metallo-β lactamase-1 (NDM-1) producing Enterobacteriaceae in Kuwait over a one year period. Consecutive Enterobacteriaceae isolates with reduced susceptibility to carbapenems were collected from four government hospitals in Kuwait from January–December 2014. Their susceptibility to 18 antibiotics was performed by determining the minimum inhibitory concentration. Isolates resistant to carbapenems were tested by PCR for resistant genes. Finger printing of the positive isolates was done by DiversiLab®. Clinical data of patients harboring NDM-1 positive isolates were analyzed. A total of 764 clinically significant Enterobacteriaceae isolates were studied. Of these, 61 (8%) were carbapenem-resistant. Twenty one out of these 61 (34.4%) were NDM-1-producers. All patients positive for NDM-1-carrying bacteria were hospitalized. About half were females (11/21 [52.3%]), average age was 53.3 years and the majority were Kuwaitis (14/21 [66.6%]). Six patients (28.5%) gave a history of travel or healthcare contact in an endemic area. Mortality rate was relatively high (28.6%). The predominant organism was Klebsiella pneumoniae (14 [66.6%]) followed by E. coli (4 [19%]). All NDM-1-positive isolates were resistant to meropenem, ertapenem, cefotaxime, cefoxitin and ampicillin, while 95.2% were resistant to imipenem, cefepime, and piperacillin-tazobactam. They were multidrug resistant including resistance to tigecycline, but 90% remained susceptible to colistin. About two-thirds of isolates (61.9%) co-produced-extended spectrum β-lactamases. During the study period, an outbreak of NDM-1 positive K. pneumoniae occurred in one hospital involving 3 patients confirmed by DiversiLab® analysis. In conclusion, NDM-1-producing Enterobacteriaceae is a growing healthcare problem with increasing prevalence in Kuwait, especially in hospitalized patients, leaving few therapeutic options. A high prevalence of NDM-1 necessitates the implementation of strict infection control to prevent the spread of these organisms.


Introduction
Infections due to Enterobacteriaceae are an important cause of morbidity and mortality worldwide. Use of antimicrobial agents is the main stay of treatment of such infections. However, resistance to these agents has become a serious health concern worldwide [1]. Carbapenems, e.g. imipenem and meropenem, are the last choice to treat multidrug-resistant (MDR) Gramnegative bacterial infections [2]. Infections due to carbapenem-resistant Enterobacteriaceae (CRE) are increasingly reported throughout the world and the Centers for Disease Control and Prevention (CDC), USA considers them as an urgent threat to human health [3]. Carbapenem resistance in Enterobacteriaceae can be mediated by production of carbapenemases which are encoded by different genes [2]. Carbapenemases in Enterobacteriaceae can be divided into Ambler class A βlactamases (e.g. KPC), class B metallo-β-lactamases (MBLs) (e.g. VIMs, IMPs, NDMs) and class D oxacillinases (e.g. OXA-48 like enzymes) [2]. New Delhi metallo-βlactamase-1 (NDM-1), a relatively newly described MBL, can hydrolyze all β-lactams including carbapenems except monobactam. It was first identified in Klebsiella pneumoniae and Escherichia coli isolated from a Swedish patient who was hospitalized in India in 2008 [4]. Since then, it has spread all over the world. To-date, NDM-1-producing-bacteria have been found in over 40 countries creating a significant public health threat [1,5,6]. NDM-1 producing bacteria are particularly important because the gene encoding this enzyme is found on transferable plasmid (of varying size) so that resistance can be easily transferred from one bacterium to another. These bacteria are usually multi-drug resistant making treatment difficult. NDM-1-positive isolates are also colonizers, causing no disease and therefore screening for them can be difficult. Importantly, NDM-1 enzyme is also present in E. coli, the most common cause of urinary tract infection in the community and hospital [7,8].
Kuwait is a Middle-Eastern country of 3.5 million people, the majority of whom are expatriates from numerous countries. This population mix provides ample opportunities for introduction and spread of resistant bacteria. In Kuwait, NDM-1 producing K. pneumoniae was first identified in two hospitalized patients [9]. A subsequent study in a major teaching hospital found that K. pneumoniae was the main species that expresses NDM-1 [10]. However, the magnitude of carbapenem resistance due to NDM-1 in Enterobacteriaceae in Kuwait is unknown. We therefore, conducted a surveillance study to detect carbapenem-resistant Enterobacteriaceae carrying NDM-1 gene in four major hospitals in Kuwait over a one-year period.

Materials and Methods
A surveillance study was conducted in four major teaching/public hospitals in Kuwait (Amiri, Farwaniya, Ibn Sina, and Mubarak Al Kabeer hospitals) responsible for healthcare delivery to a catchment area of 2.5 million people. Consecutive clinically significant (suspected of causing infection or carriage) isolates of the family Enterobacteriaceae were collected from a total of 764 patients during one year, January to December 2014. These were non-duplicate single isolate per patient. The contribution from each hospital was as follows: Amiri hospital, 208; Farwaniya hospital, 123; Ibn Sina hospital, 210; and Mubarak Al Kabeer hospital, 223. The isolates were from different sources e.g. urine, wound and tissue, blood, respiratory secretion, abscess, central venous pressure line tip, peritoneal dialysis fluid, cerebrospinal fluid, pericardial fluid and screening rectal and throat swabs.

Bacterial isolates
A total of 61 non-susceptible isolates to at least one of the carbapenems, with a minimum inhibitory concentration (MIC) of >1mg/L for imipenem or meropenem, or >0.5mg/L for ertapenem, according to the Clinical and Laboratory Standards Institute (CLSI) [11] were included in the study. They were stored at -70°C in 1% protease/peptone broth containing 7% glycerol for further evaluation. Species identification was performed with VITEK II automated system (bioMérieux, Marcy l'Etoile, France).

Molecular finger printing
Clonal relatedness of the K. pneumoniae and E. coli isolates was investigated by semi-automated repetitive sequence-based polymerase chain reaction (rep-PCR) using the DiversiLab 1 strain typing platform (bioMérieux). It was performed in duplicate with the DiversiLab 1 Klebsiella spp. kit and Escherichia spp. kit, according to manufacturer's instructions. The resulting electropherograms were analyzed with DiversiLab software version 3.4 (bioMérieux). Samples were compared by automatically generated dendrograms calculated using the unweighted pairgroup method with arithmetic mean (UPGMA) algorithm. In DiversiLab 1 , the cut-off limits for identity, similarity and dissimilarity for both E. coli and K. pneumoniae are 98%, 95% and < 95% respectively, [14,15].

Ethics statement
Collection of the strains was conducted according to the Declaration of Helsinki and with particular institutional ethical and professional standards. A written informed consent was not obtained from patients or parents of children because the bacterial isolates studied were collected from the routine work of clinical microbiology laboratory for patient care and no additional clinical specimens were collected for the purpose of the study. It is a standard practice not to get written informed consent for use of bacterial isolates unlinked to patient identity from the routine clinical laboratory. Therefore, the waiver for informed consent was granted and the study was approved by the Medical Ethics Committee of Ministry of Health, Kuwait approved the study (permit number 288/MTT).
As shown in Table 2, the most common infections caused by NDM-1 positive isolates were urinary tract infection (UTI) (n = 10; 47.6%) followed by blood stream infection (BSI) (n = 5; 23.8%), wound/tissue infection (n = 3; 14.3%) and central venous pressure tip infection (n = 1; 4.8%). The remaining 2 patients were colonized as demonstrated by positive rectal and throat screening. All Indian, Syrian, Iranian and Egyptian patients had a recent travel history to their native countries within the previous 1-3 months before admission to the hospital, while only 3 of the Kuwaiti patients had a history of travel to Jordan, India and USA within the previous 1-3 months before hospital admission.

Susceptibility testing
The results of susceptibility testing are shown in Table 3. All NDM-1 positive isolates were highly resistant to ampicillin, amoxicillin-clavulanic acid, cephalothin, cefotaxime, cefoxitin, meropenem and ertapenem. All except one isolates were resistant to cefepime, imipenem and piperacillin-tazobactam. Nineteen (90.5%), 19 (90.5%) and 14 (66.6%) isolates were resistant to aztreonam, ciprofloxacin and amikacin, respectively. Approximately 50% of isolates were resistant to tigecycline. Two K. pneumoniae isolates were resistant to colistin with an MIC of 4 and 48mg/L respectively, while the M. morganii and P. stuartii isolates were resistant to colistin.

Clonal relatedness
Analysis of genetic relatedness of 14 K. pneumoniae isolates by Diversilab 1 revealed a similarity of > 95% for 3 isolates (serial numbers 5, 6, 7; i.e. patients 19, 21 and 20) suggesting a close relationship (Fig 1). However, 2 of these isolates (serial numbers 5 and 6) were 98% similar suggesting an identity, thus confirming that the same clone infected two patients within one month period in one hospital. In addition, there was > 95% similarity between 2 K. pneumoniae isolates (serial numbers 1 and 2; i.e. patient 17 and 15). Analysis of genetic relatedness  among all 4 E. coli isolates harboring bla NDM-1 genes revealed an identity of 98% for 2 isolates (serial numbers 1, 2; patients 1 and 16), suggesting clonality of these 2 isolates causing infection in two adjacent wards (Fig 2).

Discussion
The emergence of NDM-1 producing Enterobacteriaceae has become a major public health threat that presents a challenge for treatment of serious infections throughout the world. There are several published reports of NDM-1 carbapenemases worldwide including the Gulf Corporation Council (GCC) countries of Saudi Arabia, Sultanate of Oman, Qatar, United Arab Emirates and Kuwait [16,10], Yemen [17], the West Asian countries of Turkey [18], Lebanon [19] and Israel [20], and North African countries of Egypt [21], Tunisia [22] and Morocco [23]. In our study, 61.9% of the NDM-1 harboring isolates co-expressed ESBLs which is concordant with previous reports [16][17][18][19]23]. The present study reveals the fact that carbapenem resistance among Enterobacteriaceae mediated by NDM-1 MBL production is becoming a significant occurrence in several hospitals in Kuwait. NDM-1 was present in 2.7% of all Enterobacteriaceae in our study and in about 34.4% of all CRE. This is more than double the 1.2% reported by Kumarasamy and colleagues in 2010 in India, Pakistan and the United Kingdom [24] and the 1.1% reported by Tran and colleagues in Vietnam in 2015 [25]. The detection of 34.4% of the CRE in our hospitals as NDM-1-producers is alarming. However, 40 of the CRE isolates were negative for all the tested β-lactamases and carbapenemases genes, suggesting the possibility of occurrence of new genes or permeability problems [26,19].  This report reviews 21 cases of infected or colonized patients caused by NDM-1 positive bacteria in Kuwait. They were recognized over a period of one year. The majority of the isolates were from female patients above the age of 53 years of age. Our data are similar to those reported by Kumarasamy et al. from India [24], where the majority were female patients but different in the mean ages; 53.3 versus 35years. This is in contrast to the isolates reported from the United Kingdom by Jain and colleagues in 2014, where the majority of their patients were males above 60 years of age [27]. About half of our isolates were from UTI (47.6%), a finding similar to the study of Jain et al. [27]. Other frequent sources were blood stream infection (23.8%) and wound/tissue infections (14.3%). History of travel or healthcare contact in endemic areas was given by more than a quarter of our patients. Some patients also travelled to areas regarded as non-endemic as Middle-Eastern countries (Egypt, Syria, Jordan and Iran) and the USA. None-of these patients had a history of travel to the Balkans, which is suggested to be the second epicenter for NDM-positive bacteria [28]. A very high proportion (71.4%) of NDM-1 positive patients had no history of travel to any endemic areas such as India and Pakistan or contact with healthcare facilities which have been reported previously as a source of infection [27,29]. It appears that all our patients acquired the NDM-1 producing strains while they were in the hospital suggesting nosocomial acquisition. A caveat of this assumption is that the patients were not screened for NDM-1 producing bacteria on the day of hospital admission. Infections due to CRE have been associated with high mortality rates ranging from 38-44% [30,31]. The mortality rate in our series is slightly lower (28.6%); all died from septic shock.  In this report, all the CRE were resistant to meropenem and ertapenem while 96% of them were resistant to imipenem. Variable resistance to carbapenems has been reported by other workers [19,32]. This is attributed to alterations in porins leading to variable permeabilities to different carbapenems [33]. Although aztreonam is not inactivated by NDM-1 or other NDM-1 Carbapenemase-Producing Enterobacteriaceae metallo-β-lactamase enzymes, approximately 91% of our CRE were resistant to aztreonam, which is essentially similar to the findings of other studies [24,27] thus reflecting perhaps the presence of other resistance genes such as ESBLs or AmpC enzymes. Among the non-β-lactam antimicrobial agents, ciprofloxacin resistance was noted in 90.5% of the NDM-1 positive isolates. High-level resistance to amikacin and gentamicin was detected in 66.6% which may be related to the production of one or more of the 16S rRNA methyltransferases by some of the CRE isolates [34]. After exclusion of Morganella spp. and Providentia spp. which are intrinsically resistant to colistin, 90% of our isolates were susceptible to this drug, a finding already noted in other previous studies [24,27]. However, after exclusion of Morganella spp. and Providentia spp. which are intrinsically resistant to tigecycline, only 52.6% of CRE were susceptible to tigecycline, a rate which is higher than those reported previously [24,27].
There are multiple clones of NDM-1 producing K. pneumoniae and E. coli strains circulating in Kuwaiti hospitals. However, two K. pneumoniae and two E. coli isolates were clonally identical pointing to the potential of outbreaks due to some clones.
In conclusion, out study has established the prevalence of NDM-1 genes encoding carbapenem resistance amongst Enterobacteriaceae isolates in Kuwait and the high frequency of multidrug resistance among these bacteria. Except for a small outbreak in one hospital, the majority of cases were sporadic. This report highlights the speed at which NDM-1-mediated resistance is spreading in the country which was not the case three years ago, when there was no case of CRE and MBL-positive isolates. This spread appears to be facilitated by different factors such as immigration, poor infection control practices, antibiotic misuse, and environmental spread.
Monitoring and control measures are important to prevent the spread of CRE. The Centers for Disease Control (CDC), Atlanta, GA, USA have published the guidelines for these [35]. The measures include: hand hygiene, contact precautions, healthcare personnel education, minimal use of invasive devices, laboratory reporting of CRE isolation to clinical and infection control staff, identification of CRE patients at admission, antibiotic stewardship, environmental cleaning, patient and staff cohorting, screening contacts of CRE patients, active surveillance testing and chlorhexidine bathing.