Molecular Epidemiology, Risk Factors and Clinical Outcomes of Carbapenem-Nonsusceptible Enterobacter cloacae Complex Infections in a Taiwan University Hospital

The genus Enterobacter is a member of the ESKAPE group, which contains the major resistant bacterial pathogens. Enterobacter cloacae complex (ECC) has emerged as a clinically significant cause of a wide variety of nosocomial infections. Carbapenem-nonsusceptible Enterobacter cloacae complex (CnsECC) has become an emerging threat to public health but there is still a lack of comprehensive molecular and clinical epidemiological analysis. A total of 157 CnsECC isolates were recovered during October 2011 to August 2017. hsp60 gene sequencing and pulsed-field gel electrophoresis (PFGE) were applied to discriminate the species, genetic clusters and clonal relatedness. All the isolates were subjected to polymerase chain reaction (PCR) analysis for carbapenemase, AmpC-type β-lactamase, and extended spectrum β-lactamase (ESBL) genes. Clinical data were collected on all patients for comparing clinical risks and outcomes between patients with carbapenemase-producing (CP)-CnsECC compared with non-CP-CnsECC infection. The most commonly identified species was E. hormaechei subsp. hoffmannii (47.1%), followed by E. hormaechei subsp. steigerwaltii (24.8%). Different species of CnsECC isolates showed heterogeneity in resistance patterns to piperacillin/tazobactam, cefepime and levofloxacin. In the present study, we observed that E. hormaechei subsp. hoffmannii was characterized with higher cefepime and levofloxacin resistance rate but lower piperacillin/tazobactam resistance rate relative to other species of CnsECC. CP-CnsECC comprised 41.1% (65 isolates) and all of these isolates carried IMP-8. In this study, 98% of patients had antimicrobial therapy prior to culture, with a total of 57/150 (38%) patients being exposed to carbapenems. Chronic pulmonary disease (OR: 2.51, 95% CI: 1.25–5.06), received ventilator support (OR: 5.54, 95% CI: 2.25–12.03), steroid exposure (OR: 3.88, 95% CI: 1.91–7.88) and carbapenems exposure (OR: 2.17, 95% CI: 1.10–4.25) were considered risk factors associated with CP-CnsECC infection. The results suggest that CP-CnsECC are associated with poorer outcomes including in-hospital mortality, 30-day mortality and 100-day mortality. Our study provides insights into the epidemic potential of IMP-8-producing E. cloacae for healthcare-associated infections and underscores the importance of understanding underlying resistance mechanisms of CnsECC to direct antibiotic treatment decisions.


Introduction
The Enterobacter cloacae complex (ECC) is a group of Enterobacteriaceae widely distributed in nature [1,2]. Currently, ECC has emerged as a clinically significant cause of a wide variety of nosocomial infections, such as pneumonia, urinary tract infections, intraabdominal infection, and bacteremia [3][4][5]. Although ECC is composed of multiple species, including E. cloacae, E. hormaechei, E. asburiae, E. kobei, E. ludwigii, E. nimipressuralis and E. mori [6], the routine identification of bacteria in clinical microbiology laboratories is not able to differentiate the species in the ECC by using commercialized systems [7,8]. Molecular and genomic techniques are needed to identify the ECC species precisely. Sequencing of the heat shock protein 60 (hsp60) gene, which can discriminate of this complex into 13 genetic clusters (C-I to CXIII) appears to be a valuable tool to identify of ECC species [3]. Furthermore, multilocus sequence analysis (MLSA) of housekeeping genes was also strongly validated by whole genome sequence (WGS) for phylogenetic classification of Enterobacter species [9]. Despite the fact that misidentification of ECC species has little impact on antibiotic therapy previously, as the species of the ECC have the similar antibiotic resistance profiles, these molecular approaches still have been used recently for precise identification of the species in the ECC for understanding the epidemiology, pathogenesis, and microbiological features [10].
Because of the extended overuse of antibiotics, multidrug resistant (MDR) ECC strains have emerged and spread globally. The genus Enterobacter is a member of the ESKAPE group, which are described as the major resistant bacterial pathogens and leading cause of nosocomial infections throughout the world [11]. The undesirable antibiotic resistance, especially carbapenem-resistant Enterobacteriaceae (CRE), is a major public health crisis because these agents are regarded as one of the last effective therapies available for treating serious infections caused by extended-spectrumβ-lactamase (ESBL)-or AmpC-producing Enterobacteriaceae [12]. Infections caused by CRE can lead to severe clinical outcomes and high hospitalization cost [13]. Recent surveillance studies have displayed that Enterobacter spp. are often the second or third most common CRE [14][15][16][17]. Therefore, Carbapenemnonsusceptible Enterobacter cloacae complex (CnsECC) may become an emerging threat to public health.
In Taiwan, Lai et al. reported that E. cloacae has the highest rates of carbapeneminsusceptibility among Enterobacteriaceae species from patients admitted to intensive care units (ICUs) in 2016 [18]. The results from Study for Monitoring Antimicrobial Resistance Trends in Taiwan (SMART) between 2016 and 2018 found that isolates of Enterobacter species showed higher rates of nonsusceptibility to ertapenem than E. coli or K. pneumoniae isolates [4]. The production of carbapenemases is the most important causes of carbapenem resistance in CRE [17]. The most common genes encoding carbapenemases are the Klebsiella pneumoniae carbapenemases (KPCs), imipenemase metallo-β-lactamases (IMPs), Verona integron encoded metallo-β-lactamases (VIMs), New Delhi metallo-β-lactamase (NDMs), and oxacillin (OXA) 48-like enzymes [19]. Global surveillance showed diversity of genes encoding carbapenemases in CnsECC [7,17]. In Taiwan, IMP type carbapenemase is highly prevalent in CnsECC [20][21][22][23][24]. IMP type carbapenemases were first discovered in Japan during the 1990s and have up to 18 varieties [19]. IMP-8 was the most common carbapenemase in E. cloacae from Taiwan [23]. However, comprehensive data regarding the species diversity, microbiological features, resistance plasmid characteristics and clinical relevance of CnsECC remains limited. This study was conducted to understand the epidemiology of species, clonal relatedness, carbapenemase-producing genes, and clinical features of CnsECC in Taiwan.

Species Identification of CnsECC Isolates
A total of 157 isolates were collected with initial reports as CnsECC from the Kaohsiung Medical University Hospital from October 2011 to August 2017. To identify species of the ECCs, we used hsp60 gene sequencing to classify the CnsECCs into seven species (clusters) (  (Figure 1). In PFGE analysis, 96 of 150 isolates (64%) can be clustered into 19 pulsotypes (A-S) with >75% similarity in PFGE banding patterns and the two predominant patterns were Pulsotype R (23 isolates) and Pulsotype M (16 isolates). In these two pulsotypes, 22 and 15 isolates belonged to E. hormaechei subsp. hoffmannii, respectively.

Carbapenem MIC Distribution of CnsECC Isolates with and without IMP-8
IMP-8 is the only carbapenemase-encoding gene detected in this study. To evaluate possible effects of IMP-8 on the carbapenems MIC, we conducted antimicrobial susceptibility testing of ertapenem and meropenem. As shown in Table 5, regardless of harboring a carbapenemase-encoding gene, almost all (99.4%) the CnsECC strains were resistant to ertapenem (MIC50 = 4 µg/ mL). In contrast, in 65 IMP-8-positive strains, 35 (53.8%) remained susceptible to meropenem and among the 92 non-carbapenemase producing CnsECC (non-CP-CnsECC) strains, most of them (86 isolates, 93.5%) were susceptible to meropenem.

Discussion
An increasing incidence of CnsECC poses a great threat to public health [7,25]. There, it is urgent to characterize the clinical molecular epidemiology of CnsECC infection. In this study, we demonstrated that the most common species of CnsECC in southern Taiwan was E. hormaechei subsp. hoffmannii (cluster III) (47.1%), which is similar to previous study for ECC isolates co-resistant to carbapenem and colistin in southeast China [3]. The second common species of CnsECC in this study was E. hormaechei subsp. steigerwaltii (cluster VIII) (24.8%), which was the predominant strain in northeast China (38.9%) [17] and Japan (33.3%) [26].
It is well known that ECC has an intrinsic resistance to ampicillin, amoxicillin and first-and second-generation cephalosporins owing to low expression of chromosomal ampC gene encoding an inducible AmpC β-lactamase under a basal condition [7,11,27]. Expectedly, most of the CnsECC isolates in the present study showed resistant to ampicillin, ampicillin/sulbactam, and first-and second-generation cephalosporins. It is noteworthy that the different species of CnsECC isolates showed heterogeneity in resistance patterns to cefepime, levofloxacin and piperacillin/tazobactam. Our data demonstrated E. hormaechei subsp. hoffmannii (cluster III) was characterized with higher cefepime and levofloxacin resistance rate but lower piperacillin/tazobactam resistance rate relative to other species of CnsECC. A similar pattern of results was also revealed in a recent study, which reported different antibiotics resistant pattern between two identical genetic clusters among carbapenem-resistant ECC isolates [17]. These observations indicate multiple drug-resistant mechanisms may participate in CnsECC isolates and warrant further investigation.
In this study, 98% of patients had antimicrobial therapy prior to culture, with a total of 57/150 (38%) patients being exposed to carbapenems. Chronic pulmonary disease and carbapenems exposure were considered risk factors associated with CP-CnsECC. Besides, we observed CP-CnsECC associated with poorer outcomes including in-hospital mortality, 30-day mortality and 100-day mortality. A previous study demonstrated CP-CRE may be more virulent than non-CP-CRE and is associated with poorer outcomes [36]. However, several studies showed there was no difference in clinical outcomes [25,26,37]. The heterogeneity could be related to various carbapenemase-producing genes in CRE were predominant in different species of Enterobacteriaceae or geographic regions and it highlights the need to understand the local epidemiology to tailor prevention efforts in individual regions. This is the first study to make comparison of clinical characteristics and patient outcomes between carbapenemase-producing and non-carbapenemase-producing CnsECC isolates in Taiwan.
This study has several limitations. Due to the lack of available medical information for ascertainment of infections, our data likely overestimated the proportion of lower respiratory and urinary tract infections. Furthermore, we did not have a control group of carbapenem-susceptible ECC. Therefore, we were not able to conduct risk factor analysis on CnsECC. There is also the possibility that novel carbapenemases or rarer enzymes that were not performed. Finally, as in any observational study, information on clinical characteristics and outcomes could not be completely acquired and our analysis of clinical outcome was subject to confounding biases.

Bacterial Isolates
A total of 157 ertapenem-nonsusceptible Enterobacter cloacae complex isolates (minimum inhibitory concentrations [MICs] of ertapenem >1 µg/mL) were collected from Kaohsiung Medical University Hospital during October 2011 to August 2017 and defined as CnsECC. The identification of ECC and testing of ertapenem susceptibility were performed as routine clinical microbiology laboratory procedures. The collection and testing of clinical specimens were approved by the Review Board Committee of KMUH.

Amplification and Sequencing of the hsp60 Gene
Polymerase chain reaction (PCR) analysis for partial sequencing of the hsp60 gene was performed by a protocol described previously [39]. A 341-bp fragment of the hsp60 gene was amplified and sequencing. The sequences of a 272-bp fragment of the hsp60 gene obtained for 157 strains were analyzed by the nucleotide BLAST program searches against the NCBI database and each isolate was assigned to its respective species, subspecies, and cluster according to taxonomic studies published previously [6,39].

Pulsed Field Gel Electrophoresis
Pulsed-field gel electrophoresis (PFGE) XbaI (New England Biolabs, Beverly, MA, USA)-digested genomic DNA was conducted to delineate the genetic relatedness of the isolates using procedures described previously [40]. PFGE patterns were interpreted in accordance with the criteria of Tenover et al. [41]. Restriction fragments were analyzed using GelCompar II software 6.5 (Applied Maths, Austin, TX, USA), and dendrograms of the patterns were constructed using the unweighted pair group method with arithmetic mean (UPGMA) based on the Dice similarity index.

Clinical Data Collection and Statistical Analyses
This was a single-center, retrospective, observational study of patients with positive cultures of CnsECC from October 2011 to August 2017 at KUMH. Patient information was retrospectively retrieved via electronic medical records. The parameters included demographic data, comorbidities, healthcare exposures (such as indwelling devices, hemodialysis, mechanical ventilation, and surgeries), exposure to antimicrobials within 3 months prior to isolation of CnsECC, and the clinical outcomes. Clinical outcomes were assessed by either in-hospital mortality or patient survival. The Student t-test was used for continuous variables. The chi-square test or Fisher exact test was used to compare categorical variables. For categorical variables with more than two categories, we used ordinary logistic regression to perform multiple comparisons. Statistical significance was set at p < 0.05. All statistical analyses were performed using MedCalc statistical software version 20.013 (MedCalc Software Corporation, Ostend, Belgium).

Conclusions
In conclusion, nearly all the CnsECC isolates in the present study were resistant to ampicillin and first-and second-generation cephalosporins but the majority remained susceptible to amikacin and meropenem. Different species of CnsECC isolates showed heterogeneity in resistance patterns to cefepime, levofloxacin and piperacillin/tazobactam. The most common species of CnsECC was E. hormaechei subsp. Hoffmannii, which was characterized with higher cefepime and levofloxacin resistance rate and with higher prevalence of CP-CnsECC. IMP-8 was the only carbapenemases detected among the CnsECC isolates in this study. Analysis on clinical data revealed patient with CP-CnsECC infection had poorer clinical outcomes. This study highlights the need to understand the local molecular epidemiology of carbapenem-nonsusceptible ECC and shows that using novel rapid infectious disease diagnostic tools to identify the type of carbapenemase is important for optimizing antibiotics therapy against CnsECC. Informed Consent Statement: Informed consent was waived because of the retrospective nature of the study and the analysis used anonymous clinical data.