Carbapenem-Resistant Enterobacter cloacae Causing Nosocomial Infections in Southwestern China: Molecular Epidemiology, Risk Factors, and Predictors of Mortality

Introduction

Bacteria belonging to Enterobacterales, such as Escherichia coli, Klebsiella pneumoniae, Enterobacter, Citrobacter, Serratia, Proteus, and Morganella, often cause nosocomial infections.^1,2 During the last decade, the emergence of carbapenem-resistant Enterobacteriaceae (CRE), which led to limited treatment options, has become the main cause of clinical anti-infective treatment failure.^3,4 It is worth noting that carbapenem-resistant Enterobacter cloacae (CR-ECL), which is one of the most common species that has been focused on in studies of a single species of Enterobacterales,⁵ has been reported in many countries such as Korea, the United States, India, and China.^6–8

Resistance to carbapenems is associated with several mechanisms. Among them, carbapenemase production and loss of outer membrane proteins (ompC and ompF) are the main drug resistance mechanisms.⁹ In China, carbapenemase production is attributed mainly to metallo-β-lactamases (MBLs) such as blaVIM-1, blaIMP-4, blaIMP-8, and blaNDM-1. Since the first case of carbapenem-resistant E. cloacae harbouring blaNDM-1 was detected in Chongqing, blaNDM-1-producing E. cloacae strains have emerged in various regions across the country.¹⁰ Notably, genes encoding MBLs are most commonly identified in E. cloacae and can be transmitted frequently through mobile genetic elements, leading to the prevalence of CR-ECL. Previous studies have certified that CR-ECL could increase the mortality rate, especially in vulnerable patients.^11,12 Further understanding of the molecular epidemiology of CR-ECL and investigation of carbapenemase gene-carrying plasmids, which are the most important mechanism of transmission of resistance, are needed to prevent the spread of CR-ECL. In addition, the identification of risk factors for CR-ECL infection would improve the choice and effcacy of empirical therapy.^5,7 Therefore, it is urgent to explore the epidemiological characteristics of CR-ECL infection.

However, the molecular epidemiological characteristics of CR-ECL infection are different from region to region.^13,14 In this study, we attempted to systematically analyse the epidemiological characteristics of CR-ECL infection to provide evidence for effective control of nosocomial infection with CR-ECL.

Materials and Methods

Study Design and Setting

We conducted a retrospective study to investigate the molecular epidemiological characteristics of CR-ECL in Chongqing Renji Affiliated Hospital of the Chinese Academy of Sciences University from January 2014 to December 2017. Carbapenem-resistant E. cloacae was defined as E. cloacae strains resistant to at least one of the carbapenem agents, with the criteria of MICs of ≥ 2 μg/mL for ertapenem, ≥ 4 μg/mL for imipenem, or ≥ 4 μg/mL for meropenem. To explore risk factors for CR-ECL infection, a retrospective 1:2 case-control study was performed. The case group consisted of patients with CR-ECL infection. Patients with CS-ECL infection were defined as the control group.

Identification and Drug Sensitivity of Bacteria

The VITEK 2 Compact system and the VITEK MS system (bioMérieux, Marcy l’Etoile, Lyon, France) were used for isolate identification, and the VITEK 2 Compact AST-GN13 card (bioMérieux) were used to test the antibiotic susceptibilities of all isolates. The MIC values for tested carbapenem agents, tigecycline (TGC) and polymyxin B (PB) were determined by the broth microdilution method. The results were interpreted according to the Clinical and Laboratory Standards Institute (CLSI) criteria.¹⁵ Susceptibilities of tigecycline were determined according to the European Committee on Antimicrobial Susceptibility Testing (EUCAST).¹⁶ In addition, the MIC ranges of carbapenem agents were also analyzed in our study.

The Detection of Resistance Gene

Polymerase chain reaction (PCR) testing was performed for detection of resistance genes using a previously described method.^17,18 Plasmid-mediated AmpC β-lactamase genes were detected using a multiplex PCR assay targeting MIR/ACT gene (closely related to chromosomal EBC family gene) and other plasmid-mediated genes, including blaDHA, blaMOX, blaCMY, blaACC, and blaFOX.¹⁹ The PCR amplicons of carbapenemase and ESBLs resistance genes were sequenced.

Variables and Definitions

The data collected included information regarding demographics, baseline diseases, invasive procedures, and antibiotic exposure were also collected. All variables were analyzed to determine the risk factors for CR-ECL infection. Septic shock was defined as sepsis associated with organ dysfunction and accompanied by persistent hypotension after volume replacement,²⁰ and various infections are determined by the US Centers for Disease Control and Prevention (CDC).²¹

Multilocus Sequence Typing (MLST)

The E. cloacae MLST scheme used internal fragments of the following seven housekeeping genes: dnaA, fusA, gyrB, leuS, pyrG, rplB, rpoB. The PCR conditions were queried, and data were analysed using the E. cloacae MLST website (https://pubmlst.org/ecloacae/). Using its allelic profile, the sequence type (ST) was determined with the E. cloacae sequence definitions database.^22,23 In this study, if six of the seven alleles were homologous, strains would be grouped together.

Sample Size Calculations and Statistical Analysis

Based on the report of the CHINET Antimicrobial Resistance Surveillance Program in 2014,²⁴ we assumed that CR-ECL will comprise 1.5% of the cases and CS-ECL controls will comprise 10.9%. To determine a difference at the 0.05 significance level with 80% power, a 1:2 ratio between cases and controls. We estimated that we would need at least 83 CR-ECL vs 166 CS-ECL control cases (EpiInfo, version 3.3.2).

All statistical analyses were performed with SPSS 23.0 software. For univariate analysis, the results were presented as odds ratios (ORs), 95% confidence intervals (CIs) and P values. To identify the independent risk factors, significant variables with P < 0.10 in the univariate analysis were enrolled into the logistic regression model for multivariate analysis to evaluate risk factors for CR-ECL infection and mortality. For all statistical analyses, P value of < 0.05 was considered statistically significant.

Results

Bacterial Isolates

In this study, a total of 85 consecutive nonduplicate CR-ECL isolates were investigated during the study period. Isolates originated from different anatomical sites: urine (n=24, 28.2%), sputum (n=16, 18.8%), blood (n=13, 15.3%), secretions (n=10, 11.8%), puncture fluid (n=10, 11.8%), bile (n=6, 7.1%), pus (n=4, 4.7%), and cannula (n=2, 2.4%). The majority of patients were in the urinary surgery ward (n=17, 20.0%), followed by patients in the intensive care unit (n=14, 16.5%), hepatobiliary surgical ward (n=11, 12.9%), geriatric ward (n=10, 11.8%), neurology ward (n=10, 11.8%), respiratory ward (n=8, 9.4%), digestive medicine ward (n=7, 8.2%), orthopaedic ward (n=6, 7.1%), and endocrine ward (n=2, 2.4%).

Antibiotic Susceptibility Test

As shown in Table 1, of the 85 CR-ECL isolates, the ETP resistance rate was 100% (85/85), while 51.8% (44/85) and 42.4% (36/85) of the isolates were resistant to IMP and MEM, respectively. In addition, the rate of resistance to cephalosporins was relatively high. Specifically, 98.8%, 97.6%, and 65.9% of the isolates were resistant to CRO, CAZ, and FEP, respectively. However, the resistance rate to AK was relatively low (n=11, 12.9%). Notably, the drug resistance rate of CR-ECL infection group was significantly higher than that of the control group (P < 0.05). In addition, the distribution of carbapenem MIC ranges for CR-ECL with or without carbapenemase is listed in Table 2. Compared to carbapenemase-negative isolates, our results showed that resistance was a significantly greater proportion of CR-ECL isolates that were carbapenemase-positive for carbapenem agents (P < 0.05).

Table 1 The Antimicrobial Susceptibility of CR-ECL and CS-ECL

Table 2 Distribution of Carbapenem MIC Ranges for CR-ECL with or Without Carbapenemase

Detection of Drug Resistance Genes and ST Types

Overall, 50.6% (43/85) of the CR-ECL isolates harboured carbapenemase-encoding genes. Among them, 24 isolates possessed blaNDM-1, three isolates had blaNDM-5, eight isolates contained blaKPC-2, and eight isolates contained blaIMP-8. The blaOXA-48 gene was also detected in two CR-ECL strains in our study, and two CR-ECL isolates were determined to carry the blaNDM-1 and blaIMP-8 genes at the same time. In addition to the production of carbapenemase, the deletion of outer membrane porins such as ompC and/or ompF was been detected in 44 CR-ECL strains. Notably, 83.3% (35/42) of non-carbapenemase-producing isolates had lost at least one porin, which indicated that the deletion of outer membrane porin genes played an important role in carbapenem resistance.The MLST results showed that ST51 was the predominant epidemic type (n=13, 30.2%), followed by ST88 (n=11, 25.6%) and the other types were relatively disperse. In addition, there was no outbreak of carbapenemase-positive CR-ECL isolates in our hospital, but some strains had the same ST type, which indicated that CR-ECL strains were dispersively transmitted in clinical departments (Table 3).

Table 3 Distribution of Resistance Genes and MLST Types of CR-ECL

Risk Factor Analysis of CR-ECL Infection

The comparison of the CR-ECL infection group to controls according to 36 clinical variables is listed in Table 4. The following factors were most frequently associated with the development of CR-ECL infection: ICU admission, mechanical ventilation, drainage tube, central venous catheter, and exposure to carbapenem antibiotics. As shown in Table 5. ICU admission (OR: 3.62, 95% CI: 1.93–6.80, P<0.001), drainage tube (OR: 3.06, 95% CI: 1.50–6.22, P=0.002), central venous catheter (OR: 3.21, 95% CI: 1.43–7.21, P=0.005), and carbapenem exposure (OR: 3.43, 95% CI: 1.53–7.70, P=0.003) were identified as independent risk factors for infection with CR-ECL.

Table 4 Univariate Analysis of Risk Factors for Infection with CR-ECL and CS-ECL

Table 5 Multivariate Analysis of Risk Factors for Infection with CR-ECL and CS-ECL

Clinical Outcome of CR-ECL Infection

At 28 days after infection onset, mortality was significantly higher in patients with CR-ECL infection than in those with CS-ECL infection (23/85 vs.5/170, P<0.001). In the CR-ECL group, 28-day mortality of patients who received carbapenem-including treatment, which was administered to 26 patients, was not significantly different from the mortality of patients without carbapenem treatment (8/26 vs.15/59, P=0.609), which indicated that the use of carbapenem is not associated with mortality in patients with CR-ECL infection. As shown in Table 6, the results showed that solid tumours (OR: 6.06, 95% CI: 1.73–21.21, P=0.005), septic shock (OR: 4.81, 95% CI: 1.30–17.79, P=0.019), and mechanical ventilation (OR: 4.70, 95% CI: 1.20–18.45, P=0.027) were the predictors independently associated with 28-day mortality.

Table 6 Risk Factors Associated with 28-Day Mortality

Discussion

At present, the global spread of CR-ECL strains gravely threatens public health because of the limited treatment options and unfavourable impact on prognosis.^25,26 Previous studies showed that the detection rate of CR-ECL strains in North America, Europe, the Middle East and Southeast Asia was relatively high,⁷ and even outbreaks occurred in some areas.²⁷ In China, the molecular epidemiology of CR-ECL has been reported in Henan, Guangdong and northwestern China, and ST418 and ST78 were the predominant epidemic types of CR-ECL producing blaNDM-1 in Guangdong and northwestern China, respectively.^28,29 However, there have been no reports on the clinical molecular epidemiology of CR-ECL infection in southwestern China, and there is an urgent need to systematically analyse the epidemiological characteristics of CR-ECL infection.

Patients with CR-ECL infection often received more effective empirical treatment due to high clinical suspicion for multidrug-resistant gram-negative bacteria. Previous studies have shown that appropriate antimicrobial treatment can help to improve survival rate.³⁰ Regarding the drug sensitivity of CR-ECL strains, most CR-ECL strains have been shown to be highly multidrug resistant. Actually, tigecycline and polymyxin B are the most effective antibiotics in the treatment of CR-ECL infection.³¹ However, the use of polymyxin B as a monotherapy against CR-ECL infection is debatable. Increasing reports about heterologous resistance to polymyxin B of several gram-negative bacteria indicate that rapid resistance to polymyxins can develop as a result of polymyxin B monotherapy.³² Notably, with the emergence of mcr-1-positive E. cloacae, which may lead to resistance to colistin, the treatment of CR-ECL infection is increasingly diffcult. Therefore, combined antibiotic therapy should be considered as the optimal treatment option for severe infections caused by CR-ECL.^33–35

In this study, we found that most CR-ECL isolates contained multiple drug resistance genes at the same time. The production of ESBLs and AmpC β-lactamases has been demonstrated to play a role in the resistance of E. cloacae to broad-spectrum beta-lactam antibiotics. In addition, blaDHA was reported to be the most common gene in E. cloacae producing AmpC β-lactamases in many studies, which is consistent with our study.^36,37 However, it is worth noting that carbapenemase acquisition and loss of outer membrane proteins are the main causes of bacterial resistance to carbapenem, rather than ESBLs and AmpC β-lactamases.³⁸ Notably, under antibiotic selective pressure, resistance genes can be transmitted frequently through mobile genetic elements, which will ultimately result in the development of extensively drug-resistant (XDR) and even pandrug-resistant (PDR) E. cloacae.^39,40

Most previous studies appeared to be restricted to reports of sporadic cases of E. cloacae isolates harbouring blaNDM-1, with diverse clones from geographic regions such as ST92 in Croatia, ST265 in Australia, and ST78, ST93, ST120, and ST418 in China.^41,42 In the current study, ST51 and ST88 were the predominant epidemic types among the seven different ST types identified, which provides a new model for the spread of blaNDM-1-carrying E. cloacae in China and should be of greatconcern. Interestingly, only two of the seven alleles were different between ST51 (dnaA, fusA, gyrB, leuS, pyrG, rplB, and rpoB: 4, 4, 4, 6, 37, 4, and 6, respectively) and ST88 (dnaA, fusA, gyrB, leuS, pyrG, rplB, and rpoB: 4, 4, 4, 62, 59, 4, and 6, respectively), suggesting that they are closely related to each other. Notably, in contrast to the close connection between ST258/ST11 and blaKPC-2 in K. pneumoniae,⁴³ the blaNDM-1 gene has a high detection rate in the above two ST types (n=20, 83.3%), suggesting that the blaNDM-1 gene was involved in the epidemic dissemination of CR-ECL strains. Diverse clones of blaNDM-1-carrying E. cloacae have been widely geographically distributed.⁴⁴

We also performed a retrospective analysis to assess clinical predictors of CR-ECL infection. Compared to previous studies,^45,46 ICU admission, drainage tube, central venous catheter, and carbapenem exposure were identified as independent risk factors for infection with CR-ECL in our study. As we all know, the immunity of patients in ICU is relatively low, and the possibility of infection of CR-ECL strain will be increased during invasive operations such as drainage tube and central venous catheter. In addition, when the E. cloacae are exposed to carbapenem antibiotics, the sensitive strains are inhibited or killed, and the resistant strains survive and become the dominant strains, even lead to the spread of CR-ECL strains.⁴⁷ Therefore, we should reasonably use carbapenem antibiotics to reduce the production of CR-ECL strain.

Consistent with previous studies,^48,49 our study highlighted the high mortality associated with CR-ECL infection. Although our study showed that exposure to carbapenem antibiotics was strongly correlated with CR-ECL infection, we did not identify an association between the use of carbapenems and mortality in patients with CR-ECL infection, we found that solid tumours, septic shock, and mechanical ventilation were independent risk factors for death caused by CR-ECL infection. Most cancer patients usually have low immunity after radiotherapy or chemotherapy. In addition, the basic condition of these patients is very poor, which increases the risk of infection and even death. Moreover, CR-ECL strains often cause septic shock, which leads to multiple organ failure and death.⁵⁰ Therefore, We should strengthen the monitoring of these patients with a high risk of death to reduce their mortality.

Limitations

This study has several limitations. First, it was a retrospective case-control study and was conducted in a single centre, our sample size was relatively small, which might have led to errors in statistical analysis and the omission of some other risk factors. Second, we were unable to determine all the resistance mechanisms of CR-ECL to carbapenems, such as the mutation of outer membrane porins, overexpression of efflux pumps and so on, owing to the limitations of the research conditions. Third, due to methodological limitations, this retrospective study was not able to evaluate the expected efficacy and outcome of the treatment, and further prospective and multicentre clinical trials are expected to be performed.

Conclusions

Our findings showed that ST51 and ST88, which are closely related, were the predominant epidemic types of CR-ECL producing blaNDM-1 in southwestern China. In addition, we identified important risk factors and predictors of mortality in patients with CR-ECL infection, and provided some recommendations for the diagnosis and treatment of patients infected with CR-ECL strains in southwestern China.